THE HISTORY OF A HISTORIAN:

PERSPECTIVES ON THE AUTHORIAL ROLES OF SIMA QIAN

Volume 1

Esther Sunkyung Klein

A DISSERTATION

PRESENTED TO THE FACULTY

OF PRINCETON UNIVERSITY

IN CANDIDACY FOR THE DEGREE

OF DOCTOR OF PHILOSOPHY

RECOMMENDED FOR ACCEPTANCE

BY THE DEPARTMENT OF EAST ASIAN STUDIES

Adviser: Martin Kern

September 2010
 UMI Number: 3414152

All rights reserved

INFORMATION TO ALL USERS

The quality of this reproduction is dependent upon the quality of the copy submitted.

In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.

UMI 3414152
Copyright 2010 by ProQuest LLC.
All rights reserved. This edition of the work is protected against
unauthorized copying under Title 17, United States Code.

ProQuest LLC
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, MI 48106-1346
© Copyright by Esther Sunkyung Klein, 2010. All rights reserved.
 Abstract

In the first century BCE, Sima Qian compiled the Shiji, a history of China from

mythological beginnings to his own day. Soon after, readers of the Shiji began producing

both continuations and extensive comments. This dissertation examines the changing

ways in which pre-modern Chinese readers understood authorship through an analysis of

their perspectives on Sima Qian. I argue that we cannot know the „real‟ Sima Qian: the

tragic authorial figure of Sima Qian is a construction by later readers. I trace the

development of this authorial construction from the Han through Song dynasties,

examining readers‟ comments within their historical contexts.

The dissertation has three parts. The first outlines Sima Qian‟s fortunes in the

textual world of traditional China, exploring how his authorial role was seen in relation to

the Classics and to historical texts (in chapter 1), as well as to literary theory and

composition (in chapter 2). Chapter 3 discusses formal aspects of the Shiji and how they

were considered an aspect of Sima Qian‟s creative authorship.

The second part juxtaposes two competing interpretations of the Shiji. Chapters 4

and 5 analyze how Sima Qian‟s personal tragedy was thought to relate to his work on the

Shiji. Initially such motivations were viewed in a primarily negative light. It was not

until the Song that the autobiographical connection came to be valorized. In chapter 6, I

consider an alternative position, that the Shiji was a „true record,‟ and how the meaning

of that term changed over time. I show how this aspect of Shiji interpretation reflected

and influenced traditional Chinese attitudes toward history.

The third part explores textual issues. In chapter 7, I consider three problems

related to Shiji authorship that go beyond Sima Qian: the question of Sima Tan, the work

i
of Chu Shaosun, and the idea of a damaged Shiji text. In chapter 8, I discuss issues

related to the authenticity of the “Letter in Reply to Ren An,” which is often read as Sima

Qian‟s finest autobiographical statement and a crucial interpretive key to the Shiji.

ii
To my parents, who always told me

I could do anything I set my mind to.

iii
 TABLE OF CONTENTS

Abstract………………………………………………………………….…………. i
Dedication………………………………………………………………………….. iii
Table of Contents..…………………………………………………………………. iv
Acknowledgements………………………………………………………………… vi

INTRODUCTION…………………………………………………………………. 1
Shiji Studies in the West……………………………………………………. 4
A Note on Translation………………………………………………………. 30
Structure of the Current Study……………………………………………… 34

PART I
CONTEXTUALIZATON: SIMA QIAN‟S PLACE IN THE TEXTUAL WORLD 41
Chapter 1: Original Conceptions and the New Historical Tradition……….. 44
Original Conceptions……………………………………………….. 45
Early Views of the Shiji…………………………………………….. 58
The New Historical Tradition………………………………………. 72
Chapter 2: Sima Qian in the Realm of Literary Prose……………………… 95
Ancient-style Prose of the Tang…………………………………….. 96
Ancient-style Prose of the Northern Song………………………….. 107
Chapter 3: The Shiji‟s Form in Context…………………………………….. 124
The Overall Form of the Shiji………………………………………. 125
Intention and Invention in the Shiji‟s Five Sections………………... 152

PART II
AUTOBIOGRAPHY OR TRUE RECORD? ……………………………………… 178
Chapter 4: Autobiographical Readings and the Prehistory of the Sima Qian
Romance……………………………………………….…………………… 181
Sources for Sima Qian‟s Biography………………………………… 183
Early Autobiographical Readings…………………………………... 201
Six Dynasties Developments……………………………………….. 221

iv
 Autobiographical Readings in the Tang…..………………………… 235
Chapter 5: Personal or Political: Autobiographical Readings in the Song
Dynasty……………………………………………………………………... 255
Northern Song………………………………………………………. 257
Southern Song………………………………………………………. 288
Chapter 6: Shiji as „True Record‟: The Rhetoric of Reliability…………….. 319
On the Term „True Record‟………………………………………… 322
Seeking a Standard…………………………………………………. 338
Against „Defamatory Text‟ Readings………………………………. 353
On the Dangers of the Straight Brush………………………………. 360
Song Dynasty Developments……………………………………….. 368

PART III
TEXTUAL PROBLEMS AND SHIJI AUTHORSHIP……………………………. 384
Chapter 7: Multiple Authors, Damaged Text………………………………. 386
The Sima Tan Problem……………………………………………... 386
The Role of Chu Shaosun…………………………………………... 403
A Drastically Damaged Shiji……………………………………….. 424
Chapter 8: Qian‟s “Letter in Reply to Ren An” and the Idea of Authenticity 446
On the Notion of Authenticity……………………………………… 446
The Parallel Passage Problem………………………………………. 452
Debates on the Dating of the Letter..……………………………….. 471
Debates over the Letter‟s Intended Purpose………………………… 479

CONCLUSION ...…………….………...…………………….…………………….. 482

BIBLIOGRAPHY…………………………………………………………………... 497

v
 Acknowledgements

This dissertation owes a great deal to the help and encouragement I have received

from my professors, classmates, and friends. Stephen Durrant, my adviser at the

University of Oregon, first awakened my interest in early Chinese historical narrative,

and has always been an invaluable source of inspiration and encouragement. Thanks also

to Michael Fishlen and Maram Epstein, who oversaw my first steps into the field of

Chinese literature, and Carl Robertson, my first Chinese teacher.

My adviser Martin Kern has been tremendously patient and helpful, and has

helped me grow and mature as a scholar. Robert Bagley, Benjamin Elman, Willard

Peterson, and Andrew Plaks have been generous mentors, and set a high standard of rigor

to which I continue to aspire. I also owe much to my conversations with my Princeton

classmates, especially Ian Chapman, Scott Gregory, Mick Hunter, Lin Hsueh-Yi, Mark

Muelenbeld, Brigid Vance, and Zuo Ya.

I was supported in part by a generous grant from Fulbright IIE which enabled me

to spend 2006-07 doing research in Beijing. Yang Haizheng, my teacher and mentor at

Beida, was unfailingly generous with her time, resources, and vast knowledge of the

history of Shiji studies. My classmates at Beida, especially Pillar (Wang Dongliang) and

Ben Hammer, contributed greatly both to my morale and my classical Chinese abilities.

While completing my dissertation, I have been grateful to Edward Shaughnessy

for his kind encouragement and for including me in the East Asian Studies community at

the University of Chicago. Particular thanks also to Daniel Morgan, Jeffrey Tharsen, and

Stephen Walker, all of whom kept asking when this would be done so they could read it.

vi
Special thanks is due to Franklin Perkins for his friendship, support, and classical

Chinese reading groups.

Finally, gratitude beyond words to my husband Colin Klein, who lent me his

philosopher‟s clarity of mind, his strength, and his thinking cap. His patience has verged

on the superhuman.

vii
 INTRODUCTION

Generations of scholars have approached the Shiji 史記 [Archivist‟s Records] as a

great and enduring monument of Chinese civilization. Their scholarship has been for the

most part focused on three main areas: research that builds on traditional beliefs about the

historicity and construction of the text; research that attempts to ascertain its ideological

underpinnings and hidden agenda; and research that approaches the Shiji through the

methods of textual criticism. Viewed from an overall perspective, the results of this

research are often contradictory and inconclusive. Theories may hold for some parts of

the Shiji but not for others, or they may depend on types of evidence that lack the power

to be wholly convincing. The Shiji is such a massive and heterogeneous text that there

may in fact be no possibility of fully and accurately characterizing it. Any theory we

create depends on a selection process which is in turn guided by our starting assumptions

about the text, and most particularly, about the intentions and motivations of the man held

to be its author, Sima Qian 司馬遷 (b.145 BCE).

I approach the study of the Shiji using a radically different methodology. I do not

try to prove or disprove traditional assumptions. Instead, I contextualize and historicize

these assumptions, tracing how they arose and how they changed the view of the Shiji

text for later readers. The most successful of traditional readings have become attached

to the Shiji, and have played a key role in forming our ideas about it. Without a rigorous

process of historical contextualization and critical attention, it is easy to be heavily

influenced by these authoritative readers. We tend to accept not only their views, but

also their underlying assumptions. In doing so, we fail to reflect upon what these

assumptions and views were responding to in the readers‟ own times. I argue that what

1
Shiji studies needs is not the addition of yet another story to its massive scholarly edifice:

rather, it is a re-examination of that building‟s foundations.

My purpose is two-fold. First, I want to create some distance between the Shiji

itself and later interpretations of it, showing how they are informed by their own

historical circumstances and should not be treated as privileged, authoritative statements

on the Shiji‟s „true nature.‟ Second, my work shows the process by which readers create

an author, in this case, Sima Qian. Literary theorists have argued that the reading process

is fundamental to how authorship functions, and yet in most cases it can only be

reconstructed imaginatively or described subjectively (with the critic as his or her own

case study).1 For the Shiji, however, an unusually large and rich body of reader responses

has been produced and preserved over the centuries. It is this heterogeneous and varied

group of texts that forms the primary material for my study.

Sima Qian‟s current identity—as author of a foundational text of Chinese

civilization—did not spring fully formed from a single decade in the Western Han

dynasty. As I trace a history of the Shiji and its readers from the first century BCE up

through the Song dynasty and beyond, I show that the development of “Sima Qian” as a

concept depends as much on the contingent circumstances of its most influential

interpreters as it does on facts about a single man who lived 2100 years ago. My thinking
1
Wolfgang Iser, whose work has been one of the methodological inspirations for this study, discussed both
the potential and limitations for the study of readers‟ responses for European literature:
The real reader is invoked mainly in studies of the history of responses, i.e., when attention is
focused on the way in which a literary work has been received by a specific reading public. Now
whatever judgments may have been passed on the work will also reflect various attitudes and
norms of that public, so that literature can be said to mirror the cultural code which conditions
these judgments. This is also true when the readers quoted belong to different historical ages, for,
whatever period they may have belonged to, their judgment of the work in question will still
reveal their own norms, thereby offering a substantial clue as to the norms and tastes of their
respective societies. Reconstruction of the real reader naturally depends on the survival of
contemporary documents, but the further back in time we go, beyond the eighteenth century, the
more sparse the documentation becomes. (Act of Reading, 28)
The field of Shiji reception studies suffers no such sparseness (see below).

2
on this issue is partly inspired by Foucault‟s notion of an “author-function” (as distinct

from the historical person who wrote the text). This author-function, wrote Foucault, “is

not formed spontaneously through the simple attribution of a discourse to an individual. It

results from a complex operation whose purpose is to construct the rational entity we call

an author.” In the process of this construction, any characteristics that we as readers

assign to the constructed author reflect ourselves as much as they do the author, or as

Foucault put it, “these aspects of an individual, which we designate as an author (or

which comprise an individual as an author), are projections, in terms always more or less

psychological, of our way of handling texts.” 2

While it is important to recognize that all interpretations of the Shiji are

contingent on the interpreters‟ own circumstances, these interpretations also cannot be

ignored. Traditional Chinese literary culture has left behind an abundant record of its

reactions to the Shiji. These records allow us to understand the ways in which the Sima

Qian author-function was constructed by the readers of successive eras. If we skip over

these generations of well-informed, well-educated, traditional readers and try to create

our own Sima Qians, we simply substitute our constructions for theirs, glossing over the

extent to which our constructions are dependent on theirs (as well as being in tension

with them). By investigating the history of the Sima Qian author-function, we can not

only better understand our own ideas about Sima Qian: we can also get a much better

sense of what authorship is, in traditional China and beyond.

2
Michel Foucault, “What is an Author?” 124.

3
SHIJI STUDIES IN THE WEST

Western scholarship on Sima Qian and Shiji, though more extensive than for any

of the other dynastic histories, explores only a fraction of the subject‟s potential.3

Considerable scholarly energy has gone, and continues to go, toward translation.

Édouard Chavannes (1865-1918) produced the thoroughly annotated multi-volume series

Les Mémoires historiques de Se-ma Ts’ien, which covers the first 52 chapters.4 Burton

Watson has produced a number of highly readable though largely un-annotated

translations, most recently a two-volume set of Han dynasty chapters and a volume for

the Qin dynasty.5 William H. Nienhauser Jr. is also supervising an ongoing collaborative

project to translate the entire Shiji in a literal manner with scholarly annotations. To date,

five volumes of this series have been published.6

As some continue the project of translating Shiji, and many use it unreflectively as

a source for the history of ancient China, other scholars have turned to studying the Shiji

3
See William Nienhauser‟s “A Century (1895-1995) of Shih chi Studies in the West,” 1-51, for a general
introduction to the state of the field as of 1995. In my discussion, I attempt to at least mention all the major
works but focus primarily on aspects and issues most closely related to my focus.
4
Chavannes, in collaboration with an unknown Chinese scholar during the years 1889-1893 supposedly
completed a draft translation of the entire Shiji, but only published five volumes (up to chapter 47) of his
translation during his lifetime. A sixth volume was published posthumously in 1969.
5
Records of the Grand Historian: Han Dynasty, 2 vols., and Records of the Grand Historian: Qin dynasty.
Taking into account these and his earlier work, Records of the Historian: Chapters from the Shih chi of
Ssu-ma Ch’ien, Watson has translated all but chapters 1-4, 14, 21-27, 31-47, 60, 62-65, 67, 69-70, 74-78,
81, 83-84, 105, and 128. Despite the lack of annotations, Watson‟s interpretations are based largely upon
the fine tradition of Japanese Shiji scholarship, and are very often solid and reliable.
6
The series, of which volumes I, II, V pt.1, VII, and VIII have come out, is published under the title The
Grand Scribe’s Records. Scholarly reviews of the Nienhauser translations have been somewhat mixed.
Derk Bodde reviewed the first volume rather favorably (1995) while Grant Hardy, reviewing the first two
volumes (1996), cited an excessive number of errors (“His Honor,” 147) and complained about passages
that the translators had failed to make sense of (Ibid., 148).

4
itself: in the last fifteen years, the West has seen the publication of two full-length

monographs7 and a number of substantial articles.

The first large-scale overview of Shiji by a Western scholar, however, was

actually Chavannes‟ lengthy introduction to the first volume of Les Mémoires historiques

(i-ccxlix). It is divided into five chapters, which discuss the work‟s authorship, historical

background, sources, form and methodology, and later fortunes. Appendices provide

French translations of Sima Qian‟s “Letter in Response to Ren An” (HS ch.62) and Ban

Biao‟s evaluation of Sima Qian (HHS ch.40). Although some sections are out of date,8

Chavannes gave a thorough introduction with a level of care and detail that more recent

works often do not attain.

Chavannes was one of the first Westerners to encounter the Shiji as a whole, and

his changing attitude towards Sima Qian as an author is of particular interest to my

project. As Henri Cordier revealed in his obituary for Chavannes, the latter began the

project of translating the “Basic Annals” of the Shiji as a fallback when his initial

ambition of translating the Yili 儀禮 proved too problematic.9 Soon afterward, however,

Chavannes became distracted from his interest in China‟s early dynastic record and

became fascinated instead with the author‟s story. He wrote, “I continue to read Sima

Qian; but I see better now what I want to do with him. I intend to write a book about

Sima Qian himself, which recounts his life and traces his character, determines which

7
I.e., Stephen Durrant‟s The Cloudy Mirror: Tension and Conflict in the Writings of Sima Qian and Grant
Hardy‟s Worlds of Bronze and Bamboo.
8
Nienhauser, for example, singled out the material on editions and the title taishigong 太史公 (“Century,”
4). For a discussion of the latter, see below.
9
As stated in a letter of July 12, 1889, reproduced by Henri Cordier, “Obituary,” 115. Nienhauser also
translated this letter and the next in “Century,” 3.

5
parts of the Shiji are not by him, and finally shows the plan and historical value of his

work.”10

Both aspects of the project materialize in the first volume of Les Mémoires

historiques, but by the time of its publication in 1895 Chavannes had become curiously

ambivalent about Sima Qian and his authorship of Shiji. He devoted a substantial section

to trying to sort out the unrewarding question of division of labor between the father and

the son,11 and another to Chu Shaosun, the ten lost chapters, and interpolators generally.12

This should not surprise us, as it was part of his project as stated above. But what does

seem odd is that Chavannes‟ attitude toward Sima Qian seemed to have soured.

Discussing the patchwork tendency of Chinese historical writing, Chavannes wrote of

Sima Qian‟s method, “Rather clumsily does he fit stones into the vast mosaic which he

spreads before our eyes.”13 Chavannes did concede that this very awkwardness might be

useful in the “scientific” study of history, but the negative connotation remained.

Perhaps more serious than impugning Sima Qian‟s editorial style, Chavannes also began

to doubt altogether the extent to which Qian had a meaningful authorial role:

At the very least, one can say that the documents are linked together, brought
forth, and commented upon by a narrative; are these narrations not the work of
Sima Qian himself, or should we again search for sources? A categorical
response will never be found for this question. It is undeniable that Sima Qian
himself must have written certain pages of his work; but they are less numerous,
perhaps, than one might think, and in many places there is nothing original in
them. To put it more precisely, it is not his style that one notices in his narratives,
but rather that of the Han dynasty. This style has its own qualities: it is simple,

10
“Je continue à lire Sse-ma Ts‟ien; mais je vois mieux maintenant ce que j‟en veux faire; j‟ai l‟intention
de faire un livre sur Sse-ma Ts‟ien lui-même, de raconteur sa vie et de retrace son caractère, de fixer quels
sont les livres qui ne sont pas de lui dans le Che Ki, enfin de montrer le plan et la valeur historique de cet
ouvrage” (qtd. Cordier, “Obituary,” 15, emphasis added). The translation is my own.
11
Mémoires historiques, I.xlvii-lxi. See discussion in chapter 7 below.
12
Ibid., I.cxcv-ccx.
13
“Il encastre assez maladroitement les pierres dans la mosaïque immense qu‟il étale à nos yeux”
(Mémoires historiques, I.ccxxiii).

6
 concise, clear, and vigorous; but on the other hand, it is singularly cold and
impassive…14

Far from exciting the passionate sympathy evinced by some traditional Chinese

commentators, the “Father of Chinese History” seems to have left Chavannes cold. In his

“Introduction”, Chavannes now distanced himself from a theory of Sima Qian as a strong

author. I think this change in attitude is clearly expressed, in the passage quoted above,

by the shift from “his style” to “that of the Han dynasty.”

“It is hardly possible to be enthusiastic about Sima Qian,” Chavannes wrote in his

conclusion, “patient collector of old documents, he amazes us with his erudition rather

than seducing us with his genius; but his work has become great through the greatness of

his subject…”15 Perhaps Chavannes‟ eventual ambivalence is understandable, either as a

reaction to the unfamiliar values of Chinese historiography (in contrast to the Western

tradition Chavannes clearly knew well) or because the parts of Shiji he worked on in

detail are (with notable exceptions) the driest and most unpromising from a literary point

of view: the basic annals, charts, treatises, and early hereditary houses. I think the

problem goes beyond that, however. In fact, there is a perceptible disjunction between

the “Sima Qian romance” (the tragic life-story, the enormously influential theory of

artistic creation, the claims consequently generated about the inestimable value of Shiji)

and the Shiji itself, an overwhelmingly massive patchwork of earlier sources and quoted

documents; dry chronology; obviously fictionalized scenes; missing, interpolated, or

14
“Il est indéniable que Se-ma Ts‟ien a dû écrire lui-même certaines pages de son oeuvre; mais elles sont
moins nombreuses peut-être qu‟on ne pense, et, en maint endroit, il n‟est aucunement original. A parler
exactement, ce n‟est pas son style qu‟on remarque dans ses récits, c‟e st celui de l‟époque des Han: ce style
a ses qualités propres: il est simple, concis, clair et vigoureux ; en revanche il est singulièrement froid et
impassible” (Mémoires historiques, I.clxiii). Burton Watson refers to this passage (Ssu-ma Ch’ien, 179-
180) but I consider that he exaggerates and distorts Chavannes‟ implication.
15
“Il n‟est guère possible de s‟enthousiasmer pour Se-ma Ts‟ien : collectionneur patient de vieux
documents, il nous étonne par son érudition plus qu‟il ne nous séduit par son génie; mais son oeuvre est
devenue grande par la grandeur de son sujet” (Mémoires historiques, I.ccxxv).

7
questionable chapters; and tantalizing but ultimately sparse evaluative comments. The

high estimate most traditional readers have had for the text and its author depends upon

not a broad perspective but a narrow one: it is necessary to blithely ignore all but a

carefully culled selection of “Shiji‟s greatest hits.”16

The pioneering study of Shiji in English17 was Burton Watson‟s 1958 monograph,

Ssu-ma Ch’ien, Grand Historian of China. Intended primarily for the general reader, it

nonetheless set a high standard for the field. Like Chavannes, Watson included sections

on historical background, the author‟s life, and the form of Shiji. In lieu of writing about

sources or later reception of the text (projects he explicitly declined in his “Introduction,”

x), he included a section on “The Beginnings of Chinese Historiography,” and one on

“The Thought of Ssu-ma Ch‟ien.” Two appendices give translations of select passages

from the Shiji (his selection of “Shiji‟s greatest hits”) 18 and a discussion of the dating of

the letter to Ren An.19

One interesting aspect of Watson‟s study is the way he chose to approach the

chapter on Sima Qian‟s life. Unlike Chavannes, who paraphrased and expanded upon the

Han dynasty sources, Watson presents them (i.e., the presumably autobiographical post-

face of Shiji [ch.130] and the additions by Ban Gu when he recopied the post-face into

his chapter on Sima Qian, HS ch.62) in full translation. He gives little discussion of the
16
David Rolston is unusual in expressing a less than glowing opinion of the Shiji, and his remark is
revealing in this connection. He wrote, “By itself, the text of the Shiji is somewhat cold and lifeless. The
traditional way to enliven it has been to construct an image of the author, Sima Qian, in between the lines
of the text” (Traditional Chinese Fiction and Fiction Commentary, 153-154).
17
Linguistic limitations prevent me from doing justice here to Shiji scholarship in German or Russian.
18
The sections Watson translates include: the taishigong comments from the “Annals of the Five Emperors”
ch.1, the prefaces to the “Chronological Table of the Three Dynasties” ch.13 and the “Chronological Table
of the Six States” ch.15, the taishigong comments from the “Hereditary House of Wei” ch.44, the full text
of the “Traditions of Bo Yi and Shu Qi” ch.61, and the taishigong comments from the following: the
“Traditions of Wu Zixu” ch.66, the “Traditions of the Disciples of Zhongni” ch.67, the “Traditions of Su
Qin” ch.69, the “Traditions of the Assassins” ch.86, the “Traditions of Meng Tian” ch.88, the “Traditions
of Ji Bu and Luan Bu” ch.100, and the “Traditions of Da Yuan” ch.123.
19
For further discussion of this issue, see chapter 8 below.

8
text, but the translation is far more copiously annotated than is his usual wont. Timoteus

Pokora,20 in his 1963 review of the book, took Watson to task for this way of presenting

the author‟s life. Pokora wrote:

It would have been appropriate to examine whether all the information contained
in Ssu-ma Ch‟ien‟s autobiography is absolutely reliable, as suggested already
more than thirty years ago by P.Pelliot: “Je ne suis pas sûr que cette
autobiographie, ou plutot cette post-face au Che-ki, ait la valeur absolue qu‟on est
accoutumé de lui accorder.”21

Watson, for his part, defended his choice when he wrote that Ban Gu‟s chapter 62

“constitutes for all practical purposes the final word on the life of the historian.

Contemporary Han sources give us no more than passing references to the man; later

writers have been able to do little more than mull over, and occasionally confuse, the

evidence.”22 By giving Ban Gu‟s biography to us as a text among texts, Watson could

have avoided overt commitment to its factuality and authenticity. Of course, he has no

such desire, and also wrote of his translation, “This is what Ssu-ma Ch‟ien has chosen to

tell us of himself, and it seems only fair to let him speak in his own words.”23

In marked contrast to Chavannes, Burton Watson maintained throughout his book

a concept of Sima Qian as author of Shiji in a strong sense. For example in his discussion

of form, Watson held that each of Shiji‟s 130 chapters is “a significant formal unit whose

20
Pokora, a Shiji scholar in his own right, perhaps deserves more mention than I give him here, though his
scholarship will be discussed again in Part III. Articles that make up his contribution to Shiji scholarship
include: “The First Interpolation in the Shih chi”; “Ironical Critics at Ancient Chinese Courts (Shih chi
126)”; “The Less Eminent Followers of Kao-tsu: Shih-chi 98”; “Ch‟u Shao-sun—The Narrator of Stories in
the Shih chi”; “The Chronological Tables in the Shih-chi and Their Interpolations”; “Shih chi 127, the
Symbiosis of Two Historians.”
21
“Review,” 297. “I am not certain that this autobiography, or rather, this post-face to the Shiji, had the
absolute worth the one is accustomed to ascribe to it.” As Pokora noted, this comment appears in Pelliot‟s
“Review of Arthur Hummel, trans., Autobiography of a Chinese Historian,” 132. For further discussion,
see chapter 8 below.
22
Ssu-ma Ch’ien, 41.
23
Ibid.

9
contents have been selected and disposed with care and intention.”24 Watson even

ascribed a kind of literary function to the many omens appearing in Shiji: “Ssu-ma

Ch‟ien‟s concept of history was essentially poetic and these omens and wonders were

often his symbols.”25 And although Watson denies that anyone can derive from “Ssu-ma

Ch‟ien‟s scattered utterances a consistent system of thought,”26 he does ascribe to him

certain interests or aims. Foremost of these, to judge by the frequency with which

Watson returns to it, is “the individual and the influence of the individual in history.”27

From Chavannes‟ picture of Sima Qian as a mere compiler barely detectible in his

text, to Watson‟s picture of him as a nostalgic hero-worshipper and champion of the

individual, there is already a considerable distance. Joseph Allen went even farther than

Watson in his 1981 article, “An Introductory Study of Narrative Structure in Shiji.”

Following Watson‟s contention that the Shiji is a work of literature as well as of history,28

Allen provided a structural analysis of two Shiji chapters (66 and 109). He employed

categories derived from Western narratologists, Robert Scholes and Robert Kellogg, in

their book The Nature of Narrative. The categories—at least as Allen defined them—are

somewhat foreign to the ancient Chinese context: does it make sense to talk about

“character,” “plot,” and “point of view” in a culture that had not yet separated out a genre

for “fiction” (and therefore had not yet developed a discourse for analyzing it apart from

24
Ibid., 95.
25
Ibid., 100.
26
Ibid., 144.
27
Ibid., 7. Other references to Sima Qian‟s elevation of the individual are found on Ssu-ma Ch’ien, 8, 122,
126, and 128. The one on page 122 is particularly interesting because it contrasts the emphasis on “trends”
in the historiography of Watson‟s own time with Sima Qian‟s supposed emphasis on the individual, much
to the former‟s disadvantage.
28
See for example sections on “The Shih chi as Literature” (159-174) and “The Style of the Shih chi” (174-
182) with which Watson chooses to conclude his book. He writes, in the former section, “It is undoubtedly
partly because Ssu-ma Ch‟ien possessed…literary sense in a degree seldom equaled by historians of his
country that his work has been so much read and so deeply reverenced in China, Korea, and Japan” (159).
For a discussion of the gradual elevation of the Shiji as a literary text, see Part I below.

10
its factual and moral content)? It is not that these categories are not present, but that they

would not have made sense to the author or early readers. As for the remaining category,

“meaning,” Allen defined it not in terms of “moral” but in terms of an “intellectual

connection between the fictional and actual world,”29 again a problematic approach when

the very existence of a separate “fictional world” was not necessarily part of the available

conceptual vocabulary of the time.

Despite Allen‟s appeal to similar studies of Zuo Zhuan by John C.Y. Wang and

Ronald Egan, and his use of Jaroslav Průšek and Andrew Plaks to “sinify” his discussion

somewhat, the Shiji is somewhat less amenable than Zuo Zhuan to this kind of

narratological analysis. The Zuo Zhuan is a text with no known author. Its sources,

intended audience, and even date of compilation are similarly impossible to determine.

Narratological analysis of Zuo Zhuan provides a way in to what is fundamentally a

mysterious text.30 But the Shiji is a different work entirely, and one of the key

differences is the issue of authorship: not only is Shiji presented as a text with an author,

but authorship is clearly an important concern within it. Concern with the issue of

authorship not only runs through the Shiji‟s most frequently read chapters, but also plays

a major role in interpretations by traditional readers. Perhaps part of the reason that

Allen‟s work remains not wholly satisfying is that he explicitly declined to discuss the

issue of authorship: “This is a study of Shiji as literature,” Allen wrote, “and in most

29
“Narrative Structure,” 34.
30
For a literary treatment of the Zuozhuan, see David Schaberg‟s A Patterned Past. For other recent
scholarship on Zuozhuan, see Yuri Pines‟ Foundations of Confucian Thought and Wai-yee Li‟s The
Readability of the Past in Early Chinese Historiography.

11
cases the identity of the author is not of critical importance.”31 Is it or is it not? This is

one of the most important questions the current study seeks to explore.

Stephen Durrant has devoted a significant portion of his career to studying Sima

Qian and the Shiji,32 and has published a major study, The Cloudy Mirror: Tension and

Conflict in the Writings of Sima Qian. Like Watson and Allen, Durrant tends to approach

Sima Qian as a literary author. His work does not claim to be a comprehensive study of

the Shiji, but instead “focuses attention upon only a limited number of thematic

issues…critical to understanding and appreciating Sima Qian‟s literary and

historiographical art.”33 His chapters describe the anxiety and frustration with which

Sima Qian receives and carries out his task (ch.1); Sima Qian‟s view of Confucius (ch.2)

and of the Confucian creative/editorial task (ch.3); Sima Qian‟s relationship with his

material, both in terms of sources and in terms of his own psychological predispositions

(ch.4-5); and an argument that Sima Qian‟s ideology exists in tension with his narrative

art (ch.6).

In his introduction, Durrant imaginatively describes Sima Qian‟s early life, filled

out with resonances from other sources. He also introduces two figures who loom large

in the Shiji, Confucius and the First Qin Emperor, and briefly discusses their significance.

Finally, he outlines the formal structure of the Shiji and brings up questions of authorship

and authenticity. Ultimately, he side-steps the issue by appealing to the limited scope of

his study: “My arguments do not rest on those sections of the text most often held in
31
Ibid., 31 n.1.
32
Durrant has produced a number of articles on Sima Qian, including “Self as the Intersection of
Traditions: The Autobiographical Writings of Sima Qian”; “Ssu-ma Ch‟ien‟s Conception of Tso chuan”;
and “Ssu-ma Ch‟ien‟s portrayal of the first Ch‟in emperor.” This last, a thoughtful discussion of the Shiji
portrayal of the Qin, is also particularly interesting in that it makes use of traditional readers‟ opinions as
gleaned from the Shiji pinglin 史記評林 [Forest of Comments on the Shiji], a compendium of traditional
commentators‟ words on the Shiji.
33
Cloudy Mirror, xviii.

12
doubt.”34 Durrant explicitly was not undertaking a study of the entire Shiji, and it seems

reasonable to ask which chapters he relies on most. Certainly the most heavily cited

chapter is the “Self-Narration” (SJ ch.130).35 Like Watson, he admits that it is a

document that cannot be cross-checked: “What we do know of Sima Qian derives almost

exclusively from his own hand; he creates himself.”36 Durrant treats the chapter as being

explicitly autobiographical, and in fact in an earlier article37 provides a thorough review

of theoretical literature on autobiography as well as a thoughtful discussion of this

“autobiography” in particular. While one might not always agree with what Durrant tries

to do with chapter 130 and its complement, the “Letter in Reply to Ren An,”38 he has

certainly addressed many of the issues involved.39

Durrant‟s close readings and his methodology in examining the Shiji‟s use of

sources go some way toward resolving the anxieties raised by Chavannes—and earlier by

Wang Chong40—that Sima Qian was perhaps just a cut-and-paste compiler. Durrant‟s

answer: he was that, of course, but his influence went beyond mere selection. He made

his sources say what he wanted them to say—a conclusion that may delight the literary

scholar while throwing the historian into consternation. Of course it remains open to

34
Ibid., 1.
35
This text forms the basis for Durrant‟s discussion in chapter 1 and the first part of chapter 6, and is also
frequently referred to throughout the rest of the book.
36
Cloudy Mirror, 1.
37
“Self as the Intersection of Traditions,” noted above.
38
However, Durrant‟s attempt at a “psychological reading” has been criticized, for example by Hans van
Ess (“Recent Studies on Sima Qian,” 521) and most adamantly by Michael Puett (“Review of The Cloudy
Mirror,” 290-301). It seems worth noting, however, that William Nienhauser and Paul Kroll both
expressed enthusiastic approval of Durrant‟s approach and methodology, so perhaps reactions fall divided
along disciplinary lines.
39
Other major chapters in Durrant‟s Shiji include the “Traditions of Bo Yi and Shu Qi” ch.61 (discussed in
Durrant 20-26), the “Hereditary Household of Confucius” ch.47 (29-46), the “Traditions of Wu Zixu”
ch.66 (74-98), the “Traditions of Assassins” ch.86 (105-110), the “Traditions of Lu Zhonglian and Zou
Yang” ch.83 (110-116, 120-121), the “Traditions of Wei Gongzi” ch.77 (116-120), and the “Annals of
Xiang Yu” ch.7 and of “Gaozu” ch.8 (129-143).
40
See LH 39.607-608.

13
question in most cases whether all these small changes were actually made by Sima Qian,

or whether they were not instead made by later (or earlier) editors. One of Durrant‟s

most important assumptions, which he shares with the majority of post-Song dynasty

Shiji readers, is that it is possible and permissible to form a reading of Sima Qian‟s

psychology (based on the “Self-Narration” and the “Letter”) and then project that

psychology onto the Shiji, reading it almost as a roman à clef that employs historical

rather than fictional characters. The resonances that Durrant discovers seem convincing:

a preoccupation with the decision about whether or not to commit suicide,41 a deep

sympathy with those who appreciate others but are not themselves appreciated, and a

preoccupation with ancestors, both literal (Sima Qian‟s father Sima Tan) and

moral/intellectual (Confucius).42 Chinese readers of the Shiji would recognize Durrant‟s

Sima Qian as having much in common with the traditional and much-beloved figure.

Durrant does not debunk that view of Sima Qian, but takes it farther, analyzing the well-

known story in Western psychological and literary terms.

Another study, even more strikingly traditional, is Wai-yee Li‟s, “The Idea of

Authority in the Shih chi” (1994).43 Li begins her article by tracing what she considers a

transformation from the “magical authority” of earlier times to a Han dynasty “moral

authority.” She writes that this moral authority was subject to an “anti-historical

tendency,” as evinced, for example, in the Gongyang and Guliang Commentaries to the

Spring and Autumn Annals and in the thought of Dong Zhongshu. She goes on to argue

that it was in opposition to this “anti-historical tendency” that Sima Qian created Shiji,

41
For a preoccupation with the decision about whether or not to commit suicide, see for example Durrant,
Cloudy Mirror, 18-19, 109-110, 114.
42
This theme is outlined in Cloudy Mirror, chapters 1-2.
43
Although this article was published slightly before Cloudy Mirror, the two works seem to have been
completed independently, and do not cite one another.

14
his “personal statement.” Sima Qian‟s identification with Confucius, according to Li, is

linked to “a critical attitude toward those in power…. This critical spirit and moral

authority Ssu-ma Ch‟ien readily arrogates to himself.”44 Li‟s discussion of how Sima

Qian used his sources is rather similar to Durrant‟s—they cite the same chapters and

sometimes even the same passages.45 Li goes on to discuss Sima Qian‟s specific

rhetorical devices, from the structure of the history overall to the use of evaluative

comments. A separate section analyzes devices that express uncertainty, skepticism, or

irony. Finally, in a section entitled “Pattern and Meaning in History,” Li explores Sima

Qian‟s use of “mutual illumination” (hujianfa 互見法) and how it might express Sima

Qian‟s larger concept of history. She ends with the conclusion that the historian‟s

authority is based on history-writing‟s ability to reconcile “two levels of causality—

historical destiny and individual endeavor.” 46

Wai-yee Li‟s Sima Qian, whose „authority‟ is the raison d‟être of her article, is an

extremely strong author. The Shiji as a text is also trimmed and pressed into service of

Li‟s thesis. It may be therefore understandable that she gives short shrift to problems

with authorship or the state of the text as it has come down to us.47 One footnote

acknowledges Sima Tan‟s possible authorship of the “Traditions of the Assassins”

chapter 86, a chapter which plays a significant role in her argument. Li concludes that

“the sympathies expressed in [ch.86] are more characteristic of Ssu-ma Ch‟ien than of his

44
“Idea of Authority” 360.
45
A few examples: like Durrant, Li discusses the relationship between the “Annals of Xiang Yu” ch.7 and
the “Annals of Gaozu” ch.8 (“Idea of Authority,” 397-399, 401). Also in common with Durrant, Li makes
extensive use of the “Traditions of Bo Yi and Shu Qi” ch.61 (ibid., 380-383) and the “Traditions of
Assassins” ch.86 (ibid., 371-377).
46
“Idea of Authority,” 405.
47
Li completely neglects to mention the issue of possible interpolations to the Shiji.

15
Taoist father.”48 Leaving aside the question of whether it is valid to call Sima Tan—but

not Sima Qian—a „Daoist‟ (whatever that meant in the Western Han): how does Li know

what is „characteristic‟ of Sima Qian? Which chapter‟s authorship is so secure that we

may call the Sima Qian of this chapter the characteristic Sima Qian? This characteristic

Sima Qian, whether we admit it or not, is a construction. And as for the (now) much-

beloved “Traditions of the Assassins” (SJ ch.86), which is such an integral part of „Sima

Qian‟s spirit‟ as we conceive it: would we want a Sima Qian who is not its author, or a

Shiji in which the “Traditions of the Assassins” was written by someone else? Before

saying too much about the characteristic Sima Qian, I propose that we must examine our

construction of him.

I have mentioned the overlap between Durrant‟s Shiji and Li‟s—both a rather

small subset of chapters. In some ways, Li‟s Shiji seems more extensive, but only until

one realizes that she has focused her attention mainly on the Taishigongyue 太史公曰

comments at the end of the chapters. Perhaps in defiance of Watson‟s statement that “It

is impossible, as those who have tried will know” to get a “consistent system of thought”

from these comments,49 Li takes up the challenge and at least claims success: “In the

Records a unified voice emerges from a wide spectrum of attitudes, ranging from ironic

detachment to sympathetic identification, from verification to skepticism.”50 One

wonders, in the face of such a range, where the supposed unity comes from, if not from a

preconceived idea of “Sima Qian.” Li‟s Sima Qian is a valiant rebel, resisting monolithic

conceptions of history, the anti-historical tendencies of Dong Zhongshu 董仲舒 (ca.179-

48
“Idea of Authority,” 372, n.49.
49
Ssu-ma Ch’ien, 144.
50
“Idea of Authority,” 388.

16
ca.104 BCE), and the didacticism of other historians. He is portrayed as clever,

empathic, ironic, critical, sympathetic to the individual, able to appreciate plurality,51 and

fully in control of the entire text. If there were not such a figure in Han China, it would

be necessary to invent one, and perhaps that is what Li has done. Her reading is lively

and creative, yet that does not necessarily make her Sima Qian more true than the many

other possible Sima Qians that could be derived from the vast text of the Shiji.

Grant Hardy‟s 1999 monograph, Worlds of Bronze and Bamboo, is the most

ambitious treatment of Sima Qian and Shiji since Watson‟s, and its disciplinary

perspective is explicitly as much historical as literary.52 Hardy proposes and attempts to

defend a highly complex strategy for understanding Shiji. His basic contention is that

Shiji is a microcosm, “the world in miniature”53: “When we hold the Shiji in our hands,

we are holding a model of the past itself, which intentionally replicates, though to a lesser

degree, the confusing inconsistencies, the lack of interpretive closure, and the

bewildering details of raw historical data.”54 Hardy admits that in his evaluation of Shiji

narratives, “historical interpretation is intimately connected to literary criticism.”55 He

produces a list of “literary techniques” through which “Sima Qian is able to suggest

interpretation.”56 However, he declares himself “wary of overemphasizing Sima [Qian]‟s

literary ambitions and achievements.”57 Locating himself thus on the opposite end of the

spectrum from Joseph Allen, closer to Chavannes, Hardy prefers his Sima Qian firmly on

51
“Idea of Authority,” 399.
52
Prior to the publication of Worlds of Bronze and Bamboo, Hardy also published several articles on the
Shiji: “Can an ancient Chinese historian contribute to modern Western theory?”; “Form and Narrative in
Ssu-ma Ch‟ien‟s Shih chi”; and “The Interpretive Function of Shih chi 14, „The Table by Years of the
Twelve Feudal Lords‟.”
53
Bronze and Bamboo, xiv.
54
Ibid., 48.
55
Ibid., 87.
56
Ibid., 96-97.
57
Ibid., 124.

17
the historian‟s side of the disciplinary divide. Approaching a text as massive and

heterogeneous as Shiji, while trying to maintain a picture of Sima Qian as a strong author,

is of course problematic. Hardy suggests that Sima Qian must have “deliberately held

back from complete control of his material.”58 He lingers over and returns to his

explanation of Shiji‟s contradictions and inconsistencies to such an extent that William

Nienhauser commented:

This theme [of contradictions and inconsistencies] has been repeated in each of
the chapters, pigeonholed or apparently resolved, only to emerge again, as hale as
ever. If „the image of Confucius haunts the pages of the Shi ji‟ ([Bronze and
Bamboo] 116), it is these contradictions and inconsistencies that haunt Hardy.59

Hardy admits that “the Shiji seems to function somewhat independently of its author”60 or

more bluntly writes, “the Shiji almost presents itself as a book without an author.”61

I would argue that Hardy has it backwards: the Shiji very much presents itself as a

book with an author, but as Chavannes found out to his great disappointment, the claim

seems difficult to support. Hardy credits Sima Qian with almost superhuman subtlety

and sophistication: he assumes that Sima Qian intentionally conceived his work exactly

as it appears today in order to serve as a pedagogical tool for readers. There is room for

doubt on this point. Perhaps the Shiji gradually became what it is today as it passed

through the hands of readers who used it for a wide variety of purposes, pedagogy being

only one. Is today‟s Shiji a product of (Sima Qian‟s) intelligent design, or did it evolve?

Hardy acknowledges debates over the problematic nature of Sima Qian‟s

authorship. One of his longer discussions of the topic occurs in chapter 2, where he

58
Rather than “losing control” as Durrant more plausibly suggests (Cloudy Mirror, 129), though in either
case one might question whether Sima Qian even had standards of “control” that matched ours.
59
“Sima Qian‟s Conquest of History,” 161.
60
Bronze and Bamboo, 215.
61
Ibid., xii.

18
presents a number of possible explanations for the recalcitrant nature of the text,

including that Sima Qian was merely a compiler, that he had “contradictory inclinations”

(an opinion ascribed to Durrant), that he ran out of time (Nienhauser seems to espouse

this idea in his published comments on the subject, see below), or that he was just

“overwhelmed.”62 Hardy‟s own answer is that “Sima Qian‟s notions of accuracy,

consistent, evidence, and rationality were similar to ours” (not specifying who “we” are)

and that Sima Qian was a “very active editor” who had “a coherent conception of

history” and “brought his project to a successful conclusion.”63 This conclusion, too,

continues a Western scholarly tendency to idiosyncratically paint Sima Qian in each

scholar‟s own image. In this case, Sima Qian, like Hardy, is a sort of postmodern

historian with a lot of material on his hands. He struggles in the shadow of monolithic

predecessors and is ultimately engaged in the attempt to overthrow them.64

Despite the many criticisms I have raised, aspects of Hardy‟s theory and his work

are exemplary. Although he does tend to use for his analysis the same few narratives as

other scholars,65 in several cases he expands his discussion to include parallel accounts in

other chapters as well. For example, in discussing the conflict between Xiang Yu and

62
Ibid., 47.
63
Ibid.
64
Hardy explicitly denies the charge: “Sima Qian is by no means a postmodernist who views all truth as
relative and conditioned. There is Heaven after all, and Confucius!” (217). One might reply that both
Heaven and Confucius, as portrayed in Shiji overall, provide only relative and conditioned answers. They
are merely guiding lights one may still question but never discount—analogous perhaps to the roles played
for the postmodernists by Foucault and Derrida.
65
He makes extensive use of the “Self-Narration” ch.130 of course, as well as the “Annals of Xiang Yu”
ch.7 (Bronze and Bamboo, 87-113) and of “Gaozu” ch.8 (ibid., 75-83, 87-113), the “Hereditary Household
of Confucius” ch.47 (ibid., 153-168), the “Traditions of Bo Yi and Shu Qi” ch.61 (ibid., 125-127), the
“Traditions of Wu Zixu” ch.66 (ibid., 143-150), the “Traditions of Assassins” ch.86 (ibid., 150-152). This
last so clearly revisits a very well-trodden corner of the field that Hardy politely footnotes both Durrant and
Li, confessing that “our interpretations overlap at several points” (ibid., 251, n.33). Was Hardy so pressed
for examples that he was unable to come up with a fresher one? If so, what does that say about his theory?
In addition to these, Hardy does use a few other, slightly less familiar pieces, such as the “Hereditary
Household of Duke Zhou of Lu” ch.33 (ibid., 62-65), and the “Traditions of Wei Bao and Peng Yue” ch.90
(ibid., 75-83).

19
Liu Bang, Hardy does not confine himself to the two relevant “Annals” chapters but

points out that there are twenty different chapters in the Shiji that contain information

about their struggle.66 Unfortunately, for his (questionable) interpretation of the Qin as

portrayed in Shiji, he gets his information almost exclusively from the “Annals of the

First Emperor of Qin” 秦始皇本紀 (SJ ch.6), for the most part ignoring parallel accounts

or supplementary passages that appear in more than ten other chapters. Of course, in

discussing the Qin-Han transition, Hardy‟s goal is to show the Shiji‟s multi-faceted use of

“mutual illumination” 互見法, which benefits from a thorough search of all the evidence.

In discussing the Qin, however, Hardy is bent on cultivating his notions of a demonized

First Emperor and a symbolic “bronze world”—notions which only thrive in an evidence-

poor environment. The Shiji portrayal of the Qin is far more complex than Hardy is

willing to admit.67

A major part of Hardy‟s thesis is the idea that the Shiji is a microcosm, and

perhaps it is. But is that necessarily because a single author/editor made it that way?

Han scholars, including Sima Qian, believed that it was through the agency of Kongzi

that the Classics became the sacred texts they were for nearly all traditional Chinese

scholars. The Shiji, like the Classics, eventually became canonical. John Henderson has

shown, in Scripture, Canon, and Commentary (1991), that commentators play a key role

in developing an ever-grander conception of canonized works. Henderson writes that

66
Bronze and Bamboo, 88.
67
Another weak point of Hardy‟s selection is the fact that he ignores the Han chapters almost completely,
despite the fact that he states, “As Sima‟s history approached his own day, he did more and more of the
writing” (Bronze and Bamboo, 236, n.25). It is almost as if, like Chavannes, he prefers to see Sima Qian as
a “patient collector of old documents,” or as a “very active editor,” but not as an actual writer. Whether
Hardy‟s assumption about the Han chapters is true or not (and there are those who might question it) given
that he does hold the assumption, he should have tried to at least occasionally make use of the Han
chapters.

20
 The most universal and widely expressed commentarial assumption regarding the
character of almost any canon is that it is comprehensive and all-encompassing,
that it contains all significant learning and truth…. This „obsession with
exegetical totalization‟ reflects a profound desire to see a wholeness or totality
somewhere in the world.68

In other words, Hardy may be a late example of the commentarial tendency that makes

the chosen text into a model of the world—just the kind of universal history embodying

eternal truths that Hardy ascribes to Sima Qian. Generations of readers and

commentators, acknowledged or not, have contributed to this picture.

It seems appropriate here to say a few words about the Shiji scholarship of

William H. Nienhauser, mentioned above as the head of the Shiji translation team at the

University of Wisconsin-Madison. Nienhauser has also produced several articles on the

Shiji and several more articles which one might call „tertiary‟—state of the field pieces

for both China and the West, investigations of famous figures associated with Shiji

studies, and so forth.69 Pieces in this last category provide a useful starting point for Shiji

research and can sometimes turn up information new even to the specialist. However, I

feel that and his articles on the Shiji itself are problematic.70

If there is any truth to my theory that Western scholars tend to create Sima Qian

in their own image, then Nienhauser is perhaps the best evidence for it. Unlike other

Shiji translators (Chavannes or Watson), he has not yet produced a large-scale study of

the Shiji that takes advantage of familiarity gained through the process of producing a

68
Scripture, Canon, and Commentary, 89.
69
In addition to “A Century (1895-1995) of Shih chi Studies in the West,” mentioned above, some of his
other articles in this vein include: “Travels with Édouard—V.M. Alekseev‟s Account of the Chavannes
Mission of 1907 as a Biographical Source”; “The Study of the Shih-chi (The Grand Scribe‟s Records) in
the People‟s Republic of China”; and “Historians of China” (which is primarily an examination into the
role of Gu Jiegang in the production of the Zhonghua shuju Shiji edition).
70
These include “A Reexamination of „The Biographies of the Reasonable Officials‟ in the Records of the
Grand Historian”; “A Note on a Textual Problem in the Shih Chi and Some Speculations Concerning the
Compilation of the Hereditary Houses”; and “Tales of the Chancellor(s): the Grand Scribe‟s Unfinished
Business.”

21
translation. One can deduce from scattered remarks, however, that his picture of Sima

Qian is a sort of harassed academic whose official duties (committee work on such issues

as the imperial calendar?) encroached on his private research. Nienhauser‟s Sima Qian

was apparently troubled by issues such as filing systems, research assistants, and not

having a spacious enough office.71 He sometimes put his work together “almost entirely

from existing texts,” and his occasional egregious carelessness and failure to catch

mistakes results from the fact that he was “working under extreme pressures of time.”72

He also collaborated with a group of unnamed assistants who “prepared draft texts of

various pre-Han chapters” and helped “in copying out the final two copies of his text.”73

The group apparently filed their texts with the aid of a Han dynasty equivalent of sticky-

notes, whose contents occasionally find their way by mistake into the received text of

Shiji.74

While I do not find Nienhauser‟s concept of Shiji‟s authorship wholly convincing,

it has certain merits. One is the recognition of material conditions and their possible

effect on the text. Many of the Nienhauser‟s examples for “Textual Problems” are

undoubtedly scribal errors of some sort, though it seems a bit of a mystery why

carelessness should be limited to the moment of compilation and not extend to the more

than a thousand of years of manuscript transmission thereafter. Some of Nienhauser‟s

examples do look like labels, but on the other hand, anyone preparing an edition of the

Shiji could have labeled parts of the text and then gotten the labels accidentally copied

in—especially if commentarial passages were also being added. As for the idea that

71
“Textual Problem,” 55-56.
72
“Reexamination,” 232.
73
“Textual Problem,” 56-57.
74
Ibid., 58.

22
Sima Qian was under intense time-pressure, I do not see that there is any way of proving

it. Perhaps no modern person could have accomplished what Sima Qian supposedly did,

but then Sima Qian did not suffer the addictive distractions of television, movies, or the

internet. He may have been, and probably was, an extraordinarily intelligent person,

perhaps even a genius to whom ordinary standards would not necessarily apply. He had a

command of his language and sources that no modern person can imagine. Or “he”, as

Nienhauser suggests, may actually represent a series of different hands that operated on

the text under his name—this series could just as well be diachronic and extend far

beyond Sima Qian‟s own death. In short, it seems pointless to judge “Sima Qian” by our

own intuitions of what is possible, especially when considering that he lived deeply

immersed in a scribal culture we know so little about.

Nienhauser is to be congratulated for expressing interest in overlooked portions of

the Shiji. His casualness in dealing with the often-complicated textual history of these

lesser-known chapters is more problematic, as for example in his 1991 article, “A

Reexamination of „The Biographies of the Reasonable Officials‟ in the Records of the

Grand Historian.” Admittedly this piece was written nearly twenty years ago. It seems

relevant, however, to discuss briefly the problems with “Reexamination,” as it is directly

relevant to my own methodology and project.

Very much like the Sima Qian he hypothesizes, Nienhauser goes to the archives

and gathers up a lot of material. This includes a page-long quotation from Cui Shi (1852-

1924)—seven points against the authenticity of the chapter75—one paragraph of which

reappears later76 but in a different translation and without acknowledgement of the

75
“Reexamination,” 209-210.
76
Ibid., 225.

23
repetition. Nienhauser also cites a variety of opinions by other traditional Chinese

scholars on the nature and authenticity of the chapter (the nearly-universal tendency is to

doubt that it came from Sima Qian‟s brush). He translates and discusses both the chapter

in question and large parts of its companion chapter, “Traditions of Harsh Officials.” The

evidence is all there, and the one good point of this article is that it illustrates how much

the richness and creativity of traditional scholarship can add to our picture of a Shiji

chapter. Yet Nienhauser‟s use of the resources at hand is disappointing. He does not

explicitly disagree with his sources—a different matter entirely—but rather quotes them

as if they support his point when in fact they do not.

In the end he puts forth his own theory as a “more logical explanation.”77 This

involves Sima Qian ignoring information readily available in sources he was known to

have used, and instead resorting to wholesale uncritical use of prearranged archival

materials with dubious historical value. The proposed motivation is an unproven

intuition that Sima Qian was suffering under extreme time pressure. The theory is not a

bad one, if instead of Sima Qian one casts an Eastern Han forger or interpolator into this

role, but Nienhauser anticipates that. His argument against it is that a forger would never

have made all those mistakes in chronology, for “a forger could have checked such

matters at his leisure.”78 Never mind that a forger, by the very fact of his writing under

another‟s name, has already demonstrated a certain disregard for historical accuracy.

Nienhauser ignores many other arguments by traditional critics that are clearly relevant,

or counters them only weakly.79 Opinions of educated elite scholars—writing hundreds

77
Ibid., 231.
78
Ibid., 232.
79
Nienhauser concludes his article by opining that “a thorough linguistic comparison of the language of
this chapter to that of other sections of the Shih chi would strengthen” his theory (which has in the

24
of years closer to the time of Sima Qian and living deeply immersed in a cultural milieu

much more like his than ours is—should be treated with more respect than this. Their

standards of evidence may have been different from that of scholars today, but that does

not necessarily imply that theirs were less convincing.

The final work I will consider in this section is Michael Nylan‟s “Sima Qian, A

True Historian?” (1988-1999). Nylan begins her study with a fairly devastating, and in

some cases accurate, critique of the state of the field. She divides Shiji scholars (in China

as well as in the West) into two camps, the “social scientific” (stressing that Sima Qian

was a good and reliable historian even by our standards) and the “lyric/romantic”

(stressing Sima Qian‟s personal motivations). That the two readings are not wholly

incompatible with each other, she herself proves in advancing her own “religious

reading.”

Nylan‟s interpretation aims to transcend the problems with each of the two types

of readings previously delineated. However, elements of the lyrical reading do begin to

creep into her discussion: she says Sima Qian wrote to “honor his father‟s dying wish”

and from a desire “to raise his father‟s menial status posthumously”80—which most

lyricalists would claim as part of Sima Qian‟s tragic story. (It plays a highly significant

role in Durrant‟s reading, for example, as described above, and Nylan even includes it in

her own description of the lyric/romantic position .81) When Nylan describes Sima

Qian‟s “preoccupation with the concepts of bao „obligatory requital‟ and zhi

„commitment‟,” one also wonders how much this differs from the reflections of Sima

meantime become the “only means logically to resolve” the issue [“Reexamination,” 233]). In short, he
admits that he has not done such a comparison, and really offers no evidence as to how it would support his
speculations.
80
“True Historian,” 211.
81
Ibid., 205.

25
Qian‟s “own individual preoccupation with certain highly emotional themes”82 that she

accuses the lyricalists of reading into Shiji? Perhaps only in the use of the word

“individual”—or not even that, given that, in Nylan‟s account, Sima Qian operates under

the belief that “his singular dedication to the past will be recognized and rewarded by

likeminded persons.”83

It may be (as Nylan complains) that Sima Qian was not trying to “establish his

own „name‟” (which the lyricists supposedly claim84). However, when Nylan writes of

Sima Qian‟s supposed hope to “gain for himself and his line a kind of immortality,”85 she

is merely substituting “family name” for “own name.” Nylan defines the lyrical reading

narrowly in hopes of escaping it—but she fails, waxing lyrical even about Sima Qian‟s

castration and how by writing his history Sima Qian can “erase some part of the shame

and physical loss that [he] had inflicted on his forebears in the unhappy Li Ling affair.”86

In short, though Nylan highlights “religious vocabulary and religious impulses,” she does

not succeed entirely in distracting us from the fundamentally lyrical/romantic framework

of her ideas. When she describes Shiji as an example of (Sima Qian‟s?) “willingness to

lose the self in rapt devotion to an entity perceived as infinitely finer,”87 it is a more

lyrical reading than most of the so-called lyricalists would venture.

I come finally to Nylan‟s most devastating and accurate critique of the

lyrical/romantic reading, namely, its reliance on a “mere handful” of Shiji chapters,

which I alluded to above. Because Nylan explicitly sets aside the project of explication

82
Ibid.
83
Ibid., 212.
84
Ibid.
85
Ibid., 213.
86
Ibid., 212.
87
Ibid., 224.

26
du texte,88 it is not entirely clear what chapters she is thinking of in constructing her own

reading. A survey of her own footnotes, however, reveals that in approximately 250

references to the Shiji, more than one-fifth are to a small set of very familiar chapters.89

She also refers to the Hanshu biography of Sima Qian (which contains the “Letter”) 35

times, and revisits the same tired Nie Zheng passage in chapter 86.90

The degree to which Nylan slips into the “social scientific” view she decries is

admittedly less (though I might stress that the formation of her very concept of

“religion” is probably deeply influenced by various social science disciplines). One

notable instance, however, is where she writes, “By my reading, Sima Qian would wish a

faithful recreation of the past for religious reasons. Distortion, let alone outright

invention of the past was not the way to accrue merit and charismatic power. Pious

offspring were to make the dead appear exactly as they had lived.”91 First of all, certain

types of distortion were allowed and even famously approved by Confucius in the

Zuozhuan.92 Second, if distortion and invention were off-limits and “piety” supposedly

required an exact mimetic representation, does not that put Nylan in her own social

scientific camp, implying that Sima Qian had standards of exactitude similar to ours?93

88
Ibid., 208.
89
Chapters 6-9 (i.e., the “Basic Annals” of the First Qin Emperor, Xiang Yu, Gaozu, and Empress Lü); 47
(the “Hereditary Household of Kongzi”); 61, 66, and 86 (the “Arrayed Traditions” of Bo Yi, Wu Zixu, and
the Assassins); and 130 (the “Self-Narration”).
90
She also relies very heavily on SJ ch.129, referring to it 19 times in footnotes. That is to say, she relies
on only ten chapters for approximately 30% of her Shiji citations. This is not a significantly better ratio
than most of the other scholars I have discussed above.
91
“True Historian,” 230.
92
See the story in Zuozhuan Xuan 2, where the Senior Archivist Dong Hu 董狐 wrote that “Zhao Dun
assassinated his lord” [趙盾試其君] merely because Zhao Dun had fled but not fully left the state before
the ruler in question was killed by Zhao Dun‟s brother Zhao Chuan 趙穿. Kongzi affirmed that Dong Hu
was “a good scribe of old” [古之良史] (CQZZ zhu 2.662-3).
93
I would note that Nylan‟s placing Watson in this camp (“True Historian,” 204 n.4) is also unfair: though
Watson did fall into creating the Sima Qian he would like to see, he explicitly admitted that Shiji “does not
meet the test of straightforwardness and true objectivity we demand of historical writing today” (Ssu-ma

27
 A more subtle manifestation of Nylan‟s belief that Shiji is (as tradition and her

“social scientific” scholars claim) a “true record,” is her treatment of the authorship

question. She begins well by claiming that her religious reading “sees the Shiji more as a

reflection of its intellectual milieu than as the product of Sima Qian‟s tortured, solitary

genius, incidentally relieving readers of the impossible burden of trying to establish

conclusively which parts of the Shiji may be ascribed safely to Sima Qian and which to

other authors.”94 A footnote gives a fair summary of Shiji authorship problems, including

Sima Tan‟s contribution, Chu Shaosun and other possible interpolators, the ten missing

chapters and their reconstructions, and a short list of textual studies.95 But having

relieved her readers of this burden, she relieves herself of it as well, and for the rest of the

article refers to Sima Qian as the author of Shiji, though with the occasional inclusion of

Sima Tan for rhetorical effect. She does not content herself with evidence about Sima

Qian‟s intellectual milieu but, as seen above, includes very specific details from his

biography. In short, the possibility that today‟s Shiji might be in part the product of other

hands does not resurface in her arguments.

Some of Nylan‟s better points have been made before. She argues, for example,

that “the Shiji is no work of ordinary recordkeeping”96 but is “one of the first texts to

propose that the Central States history is lengthy, continuous, and thus sacred.”97 That

may be the case, but Nylan is certainly not the first scholar to say so. In 1994, Willard J.

Peterson made a brief but very cogent argument that Sima Qian was writing cultural

Ch’ien, 136). He also anticipated Nylan‟s “religious reading” by calling the writing of Shiji “an act
resembling religious salvation” (Ssu-ma Ch’ien, 128).
94
“True Historian,” 210.
95
Ibid., 210-211 n.27.
96
Ibid., 216.
97
Ibid., 217.

28
history.98 Peterson did not use the word “sacred,” but this may have been a wiser choice:

words such as “sacred” and “religious” bear strong connotations in our own cultural

context. It is not entirely clear that these connotations are necessary or appropriate to an

understanding of ancient China. (Nylan herself admits that her intent is largely

polemical.99) In short, the basic import of her “religious reading” is extremely similar to

Peterson‟s argument that Sima Qian was trying to save from oblivion the culture he

considered to be transmitted to him by such figures as Kongzi and the Duke of Zhou. It

is unclear, however, to what extent the Sima Qian of the Shiji as we know it would have

agreed with Nylan‟s claim that the Shiji should be viewed as a sort of latter day sacrifice

to the spirits of the dead100: the “Self-Narration” (SJ ch.130) explicitly declares that the

intended audience is men of the future, not of the past.

In my discussion of these major studies on Shiji, I have shown that there are as

many Sima Qians as there are Western scholars to write about him. This is all the more

striking given that for the most part each one a) takes for granted the authorial status of

Sima Qian; b) considers the text of Shiji essentially unproblematic, at least for the

purposes of his/her study; and c) relies heavily on essentially the same group of Shiji

chapters. I think the only reasonable conclusion is that the Shiji‟s relationship to its

author is extremely underdetermined. We simply cannot know, and the attempt to better

understand the Shiji by trying to squeeze out increasingly speculative new details about

98
In “Ssu-ma Ch‟ien as Cultural Historian,” 70-79. Note that Nylan does acknowledge Peterson‟s work
(“True Historian,” 212 n.33), but the reference is superficial, and she does not seem to recognize the
relationship that actually exists between her “religious reading” and the “cultural history” hypothesis.
99
Ibid., 208.
100
She argues, for example, that one of his motives was the “pious hope that these particularly potent spirits
among the civilized Chinese dead would choose in return to confer benefits on Sima Qian and his family as
long as the Shiji continued to be read” (212).

29
Sima Qian has ceased to be a productive endeavor. Grant Hardy could be speaking for us

all when he admits that “the Shiji is so vast and suggestive that…a historian of ancient

China arguing for any hypothesis could draw from Sima Qian‟s opus at least a few items

of pertinent evidence.”101 But perhaps the most serious and justified criticism comes

from Michael Loewe, who writes (in his review of Worlds of Bronze and Bamboo): “It

remains open to question how far it is justifiable to handle the Shiji as a single well

thought out work of history deliberately shaped to bring the author‟s ideas to the attention

of readers.”102

A NOTE ON TRANSLATION

Before going any further, I should say a few words about my translations of two

key terms central to my study, namely, Shiji 史記 and taishigong 太史公. The really

difficult character is, of course, shi 史. It is a character attested in the most ancient

sources,103 and its meaning has undergone a process of change tied to developments in

Chinese textual culture. Noting the Shuowen‟s 說文 [1st c. CE] gloss on the word as “one

who records events” [記事者], Ji Xusheng suggests that the ancient meaning was broader

in scope, citing the Shijing poem, “Bin zhi chu yan” 賓之初筵 (Mao #220): “On every

occasion of drinking,/ Some get drunk and some do not./ An inspector is appointed,/ With

101
Bronze and Bamboo, 212.
102
“Review of Worlds of Bronze and Bamboo,” 223. I have said little, in my discussion thus far, about the
branch of Shiji studies studying the „authenticity‟ of the Shiji text. Given the involved and technical nature
of both these scholars‟ arguments and their conclusions, I include them instead in my discussion of textual
issues surrounding Shiji authorship, found in Part III below.
103
For Shang and Zhou dynasty oracle bone forms of this character, see Ji Xusheng 季旭昇, Shuowen
xinzheng 說文新證 [New evidence regarding the Shuowen], 1.199.

30
a recorder to assist him” [凡此飲酒。或醉或否。既立之監。或佐之史].104 It is not

events but, apparently, names that the shi (“recorder” or “scribe”) is charged with writing

down, and that at the direction of an inspector who presumably specified what should be

written.105

The shi of the Zuozhuan, at least those referred to as taishi 太史, appear to have

had more authority, and a job description more closely corresponding to the Shuowen

gloss. As noted above, Kongzi praised the taishi Dong Hu for his choice of words in

recording the death of his ruler.106 Furthermore, the taishi of Qi and his two younger

brothers were executed one after another for their insistence on recording that Cui Zhu 崔

杼 had assassinated his ruler.107 Such officials were most likely the ones responsible for

composing announcements made in the ancestral temple. These announcements were

recorded and archived, and are the probable origin of the Chunqiu 春秋 [Spring and

Autumn Annals].108 Sima Qian himself refers to such texts as shiji 史記 [archivists‟

records].109

In translating for shi, some Western Shiji scholars have opted for „astrologer.‟110

The motivation for this is the fact that the title taishigong was clearly derived from name

of an official position held successively by both Sima Tan and Sima Qian, taishiling 太史

104
Trans. Legge, She-king, 399.
105
Nienhauser et al. accordingly use “Grand Scribe” to translate taishigong (Grand Scribe’s Records etc.).
106
CQZZ zhu 2.662-3.
107
Ibid., 3.1099.
108
See discussion on Pines, “Chinese History Writing,” 318-323.
109
See, for example, SJ 14.509. With Michael Nylan (“True Historian,” 203 ff.), I have opted for
“archivist” as my translation of shi. My reasons will be explained below. See Wai-yee Li, “Authority,”
345 ff. for a discussion of this sense of shi.
110
For example, Chavannes‟ “Duc grand astrologue” (Mémoires historiques, I.ix) and Stephen Durrant‟s
“Gentleman Grand Astrologer” (Cloudy Mirror, 1). Note that both of these phrases are translating
taishigong 太史公. For Shiji, both scholars choose words related to “history” (“Mémoires
historiques”/“Records of the Historian”).

31
令. Hans Bielenstein, who translates this title as “Prefect Grand Astrologer,” has

described the duties of this office as follows:

[The Prefect Grand Astrologer] was in charge of drawing up the annual calendar
and memorializing it shortly before each New Year‟s day. One first of each
month, he also informed the emperor about the calendar of that month. For state
rituals, he identified days which were auspicious and days to be avoided. He kept
a record of portents and auspicious omens…. In addition, he supervised two tests
which had to be passed by those aspiring to appointment as Masters of
Documents.111

Bielenstein‟s sources are all Eastern Han texts, so it is difficult to say if this is exactly the

job as the Simas understood it. However, it does seem clear that their official duties were

indeed connected with the calendar.112 The problem with using this understanding of shi

in the translation of taishigong113 is that whatever Sima Qian‟s understanding of

taishigong may have been, it was not merely an honorific synonym for taishiling.114

Instead taishigong seems to have been used in connection with the compilation of the

Shiji (original title: Taishigong shu 太史公書 [Writings of the Taishigong]), which

scholars almost universally agree was a privately-conceived project, not part of the Simas‟

official duties.

This line of reasoning has resulted in the most common translation choice for shi,

namely, „historian.‟115 There are two possible objections to translating shi in this way.

The first is that traditional China lacked anything that could measure up to the rather

exalted status the word “history” has enjoyed in certain modern Western academic

111
Bureaucracy of Han Times, 19.
112
See HS 21A.974-5, 52.2406.
113
Even leaving aside the problematic connotation the word carries in our culture.
114
See Hucker, Official Titles, 482; Hucker draws a slightly weaker distinction than I am attempting to do,
but makes the two terms distinct nonetheless.
115
This is Watson‟s choice (Ssu-ma Ch’ien, Grand Historian of China etc.). Wai-yee Li also accepts this
translation (see “Authority,” 345), as does Willard Peterson (see “Cultural Historian,” 71; my discussion
has benefitted from the useful references included there). Furthermore, as noted above, both Chavannes
and Durrant follow this understanding for “Shiji,” if not for “taishigong.”

32
contexts.116 This objection can probably be dismissed out of hand. The classical roots of

the word (Latin historia: a narrative of past events; Greek : an account of one‟s

inquiries) can comfortably accommodate Sima Qian‟s project. Nor are the systematic,

scientific connotations of “history” necessarily present in all its possible uses even today.

Finally to deny Sima Qian the status of historian is tantamount to denying such a status to

anyone in the Chinese tradition, a Eurocentric and excessively fastidious understanding

of both the word and the endeavor as a whole.

A second objection, however, is more subtle and more closely tied to my project

in this study. When we call something a „work of history,‟ we do tend to mean

something more than a mere chronicle or story. Hayden White, in his discussion of how

„history‟ differs from „chronicle,‟ ties the historian‟s function to the notion of authority,

arguing that:

In order for an account of events to be considered a historical account… it is not

enough that they be recorded in the order of their original occurrence. It is the
fact that they can be recorded otherwise, in an order of narrative, that makes them
at once questionable as to their authenticity and susceptible to being considered
tokens of reality. In order to qualify as “historical,” an event must be susceptible
to at least two narrations of its occurrence…. [Otherwise] there is no reason for
the historian to take upon himself the authority of giving the true account of what
really happened.117

Whether or not one agrees exactly with White‟s criterion, the important point is that the

English word „history‟ is entangled—visibly or invisibly—with the issue of authorship.

In the Chinese context, as I will argue, the influence the Shiji exercised over its

readers played a pivotal role in forming the very notion of authorship. That being so, it

seems anachronistic to use the word „history‟ to refer to the Shiji from its very inception.

116
Thus, the Oxford English Dictionary registers a tendency to elevate the historian as “one who produces a
work of history in the higher sense,” and to distinguish such a person from the annalist, chronicler, or
“mere compiler of a historical narrative” (emphasis added).
117
“Value of Narrativity,” 19.

33
If by „historian,‟ we mean someone whose authorial function extends beyond the mere

recording of events and into higher realms of textual endeavor, then it is my contention

that those who read and wrote about the Shiji played a necessary role in making Sima

Qian into a historian. This process did take place: at some point Sima Qian became a

historian, and the Shiji a history. Yet to uniformly translate shi in this way implies that

the transformation occurred the moment that Sima Qian laid down his brush (or before).

It is an important point in my argument that this is not, in fact, the case.

To translate shi as „archivist‟ is, in some sense, to under-translate the term. I do

not mean to imply that, like the disillusioned Chavannes, I see Sima Qian‟s primary

endeavor as the reshuffling of pre-existing documents. The Shiji we have today begins

with that endeavor but ends somewhere far beyond it. By using the term „archivist,‟ I am

simply avoiding the assumption that the Shiji always was as it is today.

STRUCTURE OF THE CURRENT STUDY

My study is divided into three major parts. Part I addresses different ways in

which Sima Qian and his work on the Shiji have been contextualized vis-à-vis other

authors and texts. Part II addresses what I consider to be the heart of the interpretive

problems with the Shiji: the choice between reading it is an autobiographically motivated

text and reading it as a “true record”—a reliable history. I end in Part III with an

exploration of textual issues related to the Shiji that question or go beyond the notion of

Sima Qian‟s authorship.

Part I, Chapter 1, considers the overall contextualization of Sima Qian and the

Shiji, first from within the text itself, and then its later development. During the Han and

34
Six Dynasties periods, comparison and contrast among different writers was a

predominant mode of literary thought. Although in later times different ways of thinking

also developed, the significance of juxtaposition never really faded. To name a modern

example, the fact that Lu Xun once famously described the Shiji as “the ultimate song of

a historian, a Li Sao without rhyme” [史家之絕唱, 無韻之《離騷》] has now

permanently established a deep association in the modern Chinese mind between Sima

Qian and the beloved “patriotic” poet Qu Yuan 屈原 (ca. 340 BCE-ca.278 BCE)—much

to the benefit of the former.

Prior to the appearance of the Hanshu 漢書 [History of the Former Han], the Shiji

was often compared to its predecessor and frequent source history, the Zuozhuan 左傳

[Tradition of Master Zuo], to the Classics, and to the nearly contemporary philosophical

text, the eclectic Huainanzi 淮南子 [Master(s) of Huainan]. Then in the early Eastern

Han, Ban Gu 班固 (32-92) compiled the Hanshu, in which he appropriated and

occasionally rewrote the Han dynasty portions of the Shiji, continued the record up to the

time of Wang Mang, and issued a sharp reprimand of Sima Qian‟s moral judgments and

treatment of events. This forever changed the fate of the Shiji, pulling it into the orbit of

the weighty dynastic history tradition which continues up to this day. For most of

Chinese history, Ban Gu remained the most frequent juxtapositional partner of Sima Qian,

and comparison and contrast between the two spawned an entire sub-field of historical

studies known in Chinese as Ban Ma yitong 班馬異同 [Comparisons between Ban Gu

and Sima Qian]. I also discuss the Shiji‟s place in the development of competing views

about how or what a history ought to be.

35
 Under the shadow of Ban Gu‟s condemnation, Sima Qian remained a slightly

unsatisfactory exemplar of the dynastic history tradition. At the same time, there was

something leftover in Sima Qian, something not exhausted by the description “historian.”

One strand of reactions to the Shiji, which I examine in chapter 2, placed emphasis on

Sima Qian as a genius of literary style. In this context he appeared together with Qu

Yuan, Sima Xiangru 司馬相如 (ca.179 BCE-127 BCE), Liu Xiang 劉向 (79-8 BCE), and

Yang Xiong 揚雄 (53 BCE-18 CE).118 Later, due to the Ancient-style Prose (Guwen 古

文) movements in the Tang and Song dynasty, others were added to the list as well, most

frequently Han Yu 韓愈 (768-824). Another unexpected comparison that was

surprisingly prevalent during the Song was between Sima Qian and the poet Du Fu 杜甫

(712-770), who in his Xin Tang shu 新唐書 (New Tang history) biography is called “the

poet-historian” 詩史.119 It would be a mistake to consider these comparisons to be

merely on the level of style. It was a central Ancient-style Prose tenet that moral values,

individual character, and literary style were and should be inextricably linked. A number

of Song writers, such as Qin Guan 秦觀 (1049-1100) and Chao Gongwu 晁公武

(ca.1105-1180), attempted to refute Ban Gu‟s old accusations against Sima Qian‟s moral

judgement, in some sense clearing the path for Sima Qian to take Ban Gu‟s place, not just

as a model of prose style, but as a moral exemplar as well.

In chapter 3, I switch tracks and consider the issue of the Shiji‟s contextualization

from the perspective of the text‟s formal structure. A long-accepted fact about Sima Qian

118
Ban Gu often appears on this list as well, which may come as a surprise, given that in most modern
contexts Sima Qian‟s literary reputation has overtaken his. Yet the Wenxuan 文選 contains multiple
selections of Ban Gu‟s prose, many more than of Sima Qian‟s.
119
XTS 201.5738.

36
is that he was the inventor of the historical genre known in modern Chinese as jizhuanti

紀傳體. This genre derives its name from two of the five sections of Shiji, namely the

“basic annals” (benji 本紀) and “arrayed traditions” (liezhuan 列傳). Ban Gu simplified

the jizhuanti form for historical writing as Sima Qian had conceived it and adopted it for

use in his own Hanshu. All later official dynastic histories would use variants of this

same basic form. Much ink has been spilled on the characteristics of jizhuanti writing,

the most imaginative treatment being Burton Watson‟s fully elaborated and oddly

impressive vision of a history of the United States of America as it would appear if

written in this genre.120

In this chapter, I examine readers‟ comments about the genre Sima Qian invented,

particularly in juxtaposition with the formal characteristics of various Classics,

particularly the Spring and Autumn Annals and the Shangshu. I argue that these

comparisons were made more for political reasons than from genuine structural insight.

There is evidence that Sima Qian might have intended for the structure of his work to

have a symbolic meaning, but commentators have been left to speculate on just what this

symbolic meaning might be. I review this debate as it appears in the interpretations of

Sima Zhen 司馬貞 (fl. 745), Zhang Shoujie 張守節 (8th century), and Liu Zhiji 劉知幾

(661-721). The final section of chapter 3 is devoted to individual sub-sections of the

Shiji, and to the debate over the extent to which Sima Qian inherited their genre

characteristics from pre-existing works, or whether he invented them himself.

In Part II, I move to an examination of two very different interpretive approaches

toward the Shiji. In chapters 4 and 5, I consider approaches that can be loosely described

120
Watson, Ssu-ma Ch’ien, 105-108.

37
as „autobiographical.‟ Generally the most prevalent approach to the Shiji in Western

scholarship today, autobiographical readings—in other words, taking Sima Qian‟s

statements about his own life as found in the “Self-Narration” and “Letter in Reply to

Ren An”, and using them as a key to the deeper meaning of the Shiji—actually began as a

form of devastating criticism against the Shiji. In chapter 4, I show that early readers,

from Ban Gu on, accused Sima Qian of polluting his history with subtle criticism

motivated by petty personal resentment. The fact that the “Letter in Reply to Ren An”

was included in the influential anthology, the Wenxuan 文選 (compiled ca. 530),

probably did much to improve Sima Qian‟s personal reputation. Still, even in the Tang,

autobiographical readings of the Shiji were either somewhat negative, or they avoided

focusing directly on Sima Qian‟s tragedy. It was only in the Song dynasty, as I discuss in

chapter 5, that autobiographical readings began to assume the positive aspect they have

today. Even then, whether readers approved or disapproved of autobiographical

transference from Sima Qian to the Shiji depended heavily on the context in which they

were writing.

Chapter 6 explores an alternate approach to reading the Shiji, one which argues

that Sima Qian‟s motivation in creating the Shiji was no more or less than the desire to

create a „true record.‟ I first review the development of the term „true record‟ as related

to the Shiji and other comparable works. I then consider Han and Six Dynasties uses of

the „true record‟ approach to defend the Shiji from accusations that it was a “defamatory

text” (bangshu 謗書). I then examine another aspect of the debate, characterized in the

Tang by the term „straight brush‟ (zhibi 直筆). This debate tended to focus on the tension

between the danger of writing the truth and at the same time the desirability of doing so.

38
Finally, I explore a Southern Song dynasty development in which certain thinkers

apparently attempted to use history as a reliable source of political judgement, but were

decisively refuted by Zhu Xi who vehemently insisted that the histories (Shiji among

them) could never replace the Classics as a source of truth. Because Zhu Xi was such an

influential figure, the result of his denunciation of this aspect of the Shiji was to swing the

pendulum back toward autobiographical readings in the private sphere. Such readings

would come to flourish in the Ming and Qing, but for now are beyond the scope of the

present study.

Part III explores the relationship between textual issues and Shiji authorship.

Chapter 7 addresses two issues that have been of interest to many scholars in the past

century: first, the idea of the Shiji as a multiply authored text rather than the monolithic

creation of a single author; and second, the idea that the transmitted Shiji is a badly

damaged text bearing less relation than generally supposed to the work that Sima Qian

actually authored. Though many of these issues are undecidable, I provide an analysis of

the rhetoric and motivations behind raising them. I show that, at least in part, the drive to

reassign portions of the Shiji to other authors arises from a desire to re-conceive or

displace the Sima Qian author-function.

Finally, chapter 8 explores the problems surrounding the dating of Sima Qian‟s

“Letter in Reply to Ren An.” The “Letter” is an indispensible but problematic source of

information about Sima Qian‟s life and motivations. I explore scholarly debates over the

authenticity, dating, and intention of the “Letter.” I conclude that while it will never be

dismissed as a forgery, we should also not confidently assume its total authenticity.

39
 In my conclusion, I bring together the various aspects and issues I have discussed,

sketching the history of interpreting the Shiji in periods I have not covered in detail.

Finally, I discuss the implications of this work for future studies of the Shiji.

40
 Part I

CONTEXTUALIZATION: SIMA QIAN’S PLACE IN THE TEXTUAL WORLD

“Liu Yin had seven sons. Five sons each received

instruction in one of the Classics. One son received
instruction in the [Writings of] the Honorable Senior
Archivist, and one son received instruction in the History of
the Han. Within one house, seven lines of study arose
together. As regards the learning of Beizhou, Yin‟s house
was its flourishing.” [劉殷有七子，五子各授一經，一子
授太史公，一子授漢書，一門之內，七業俱興，北州之
學，殷門為盛．]1

Liu Yin was an official for Liu Cong 劉聰 (Emperor Zhaowu 昭武, r. 310-318) of

the Chinese/Xiongnu state of Han Zhao 漢趙 (304-329). Neither his family nor his polity

could claim to be especially representative. I begin my discussion of contextualization

with this description, however, because as an idealized way of mapping out a text‟s

perceived significance, it is marvelously clear. How to educate one‟s sons, even when

one is blessed with seven, would have been a matter of crucial importance. The Five

Classics may have been an obvious choice, but the addition of the Shiji and Hanshu, at

the end as it were, shows that by the early fourth century these histories had begun to

form an alternative canon. The uneasy relationship between the Classics and the histories

forms a crucial background for our understanding of Sima Qian and the development of

his authorial reputation.

Naturally, the view reflected by the above quotation was not the only way the

Shiji was seen. In chapter 1, I begin by examining Sima Qian‟s own presentation of his

work and how he tried to make it fit into the textual world that existed in his time. Of

1
JS 88.2289.

41
course, an author does not always have the privilege of controlling his own immortality.2

Thus, I next examine early views of Sima Qian in the decades immediately following his

death. Third, I consider Ban Biao‟s and Ban Gu‟s influence on how Sima Qian was seen

and contextualized. After Ban Gu‟s compilation of the Hanshu, it almost immediately

became a foil for the Shiji, sparking a debate as to which text was superior. That debate

continues to the present day. Finally, I give an overview of the Shiji‟s context within the

historical tradition more generally.

Chapter 2 focuses largely on Tang and Song views of the Shiji from beyond the

confines of official history (though I do refer to the official history context for purposes

of comparison). I begin with the seemingly minor but ultimately influential role played

by the Shiji in the development of Tang dynasty Ancient-style prose, as exemplified by

Han Yu and Liu Zongyuan. I then show how the Ancient-style Prose movement of the

Northern Song further increased Sima Qian‟s importance to the project of composing

prose in the ancient style. In particular, influential Northern Song authors began to

construct literary genealogies, linking Sima Qian with Qu Yuan, Du Fu, or Han Yu. In

such genealogies, Ban Gu was often left behind.

In chapter 3, I turn to the issue of the Shiji‟s form or genre. For much of the

Shiji‟s history, the reputation of it as a text and of Sima Qian as its author was deeply tied

to issues of form: whether the Shiji‟s form was appropriate, and what meanings might be

encoded in it. In considering the relationship between the Shiji and other texts, readers

often paid a great deal of attention to the Shiji‟s formal characteristics as a basis for

2
I use the term in same sense as Milan Kundera did, meaning something like the almost irrevocable cast
which one‟s actions in life give to one‟s posthumous reputation. Kundera wrote, in describing Goethe, that
the great poet was “the administrator of his immortality… [a] responsibility [that] tied him down and
turned him stiff and prim” (Kundera, Immortality, 72).

42
comparison. In this chapter, I consider first how readers used the overall form of the Shiji

to contextualize it within the uneasy neighboring categories of „Classic‟ and „history.‟ I

then turn to the individual sub-sections of the Shiji and review the debate over their

possible precursors or analogues in the early textual tradition.

43
 Chapter 1

Original Conceptions and the New Historical Tradition

I begin my study of how Sima Qian has historically been read and interpreted by

giving an overview of the different contexts in which the Shiji was placed. These

contexts are not always explicit, but are often made apparent by the texts with which the

Shiji was compared and contrasted. I also consider the criteria that lay behind these

juxtapositions, the evaluations of the Shiji which resulted, and the influence of a given

juxtaposition (if any) on later readers.

First, it is worth considering the texts to which the Shiji compares itself, and by

extention the context in which it attempts to place itself. The Shiji qua textual

compilation displays a self-consciousness matched by few if any of its antecedents.

The second section of this chapter examines the Shiji through the eyes of its

earliest readers. These readers are not necessarily explicit about what kind of text they

thought the Shiji was, but the question was certainly being explored. The view of the

Shiji that eventually prevailed, at least during the Six Dynasties, was that its main

importance was as a predecessor to Ban Gu‟s (at that time) much more prestigious

Hanshu. Sima Qian and Ban Gu were seen as initiators of a new historical tradition

which eventually separated itself from the historical Classics (the Documents and the

Spring and Autumn Annals). Because of the textual overlap between the Shiji and the

Hanshu, and because they represented slightly different visions of history-writing, the

practice of contrasting the two texts became an extremely productive mode of

historiographical thought, one which began in the Six Dynasties and still continues today.

44
ORIGINAL CONCEPTIONS

It seems appropriate to begin with the context that Sima Qian himself seems to

have claimed for his work. The final chapter of the Shiji, the “Honorable Senior

Archivist‟s Self-Narration,” gives us a variety of choices, which may be separated into

four main approaches. The first approach, which arises from Sima Qian‟s remembered

conversations with his father, as well as his dialogue with Hu Sui, would be to place the

Shiji in the context of the Confucian Classics. More specifically, it would make the Sima

family enterprise a continuation of the work of Confucius—the creation of a latter-day

Classic. The second approach, based on the pedigree of the Sima family and Sima Qian‟s

discussion of sources texts, is to place the Shiji in the same tradition as scribal records

from the various states and earlier historical compilations. The third approach, suggested

by Sima Tan‟s “Essentials Points of the Six Schools” [六家之要指],1 and by the

intellectual antecedents Sima Qian claimed for both himself and his father, would situate

1
SJ 130.3288-3292. The translation of jia 家 as schools is extremely problematic. As Mark
Csikszentmihalyi and Michael Nylan have pointed out in “Constructing Lineages and Inventing Traditions
through Exemplary Figures in Early China,” the term often means individual experts. In the context of
Sima Tan‟s essay, Csikszentmihalyi and Nylan admit that the term refers to “six categories” of thinkers, but
resist the inference that “each of these six categories had a textual core” (67-68), let alone a “school” in the
sense institutionalized education. Instead they suggest that a jia as “a category was defined not by a
common founder, canon, or genealogy, but by particular governing „methods‟ or „techniques‟” (67).
Kidder Smith, in “Sima Tan and the Invention of Daoism, „Legalism,‟ et cetera,” argues that the jia in Sima
Tan‟s essay “indicated a conceptual area, a style of practice” and yet were also “ideal human types, not
simply concepts” (148). His article goes through each of the six jia, attempting to construct an argument
about what Sima Tan might have been referring to. Yet Smith‟s own evidence reveals that there is too
much missing to make definitive statements about what the jia did or did not have. Some like the Ru and
Mo did have a textual core. Others may have had a core of either textual or oral transmission that simply
did not survive. To say that the categories are defined only by methods or techniques is to ignore human
factors which must have existed (and about which we in fact do have at least some information for the
Warring States, Qin, and early Han). Techniques do not propagate in a vacuum—there is some medium
which by which their practitioners are related, and through which new practitioners are trained. One would
assume that the medium would be different for the different techniques or types of learning. In the use of
the word jia, the different media are brought together under the metaphor of “family” (though as the Sima
pedigree shows, the literal meaning may also play an important role). Direct translation would be
unnecessarily confusing, however. I consider the equivalent concept in idiomatic English to be “schools of
thought,” which need not refer to actual schools or even (necessarily) to a common textual core. However,
it does not succeed in capturing Smith‟s fairly compelling insight that an “ideal human type” is also part of
the semantic valence of the word.

45
the Shiji as the core text of a new intellectual school, with an implicit claim of its

superiority to all others.2 The final approach, which arises from Sima Qian‟s elliptical

narration of his disastrous involvement in the Li Ling affair and its consequences, would

place the Shiji in a more heterogeneous category: for lack of a better term I will call it the

“literature of suffering.”3

Of course none of these categories are mutually exclusive. In Sima Qian‟s view,

Confucius made one of the Classics, the Spring and Autumn Annals, by editing the scribal

records from the state of Lu, so that a “Classic” and a “history” were not necessarily

distinct entities. Similarly, figures associated with three of the categories (the Classics,

scribal records, and philosophical masters) appear in Sima Qian‟s list of “suffering

authors,” which I am considering to be a fourth „category.‟ In short, these four categories

of analysis, at least three of which correspond suspiciously well to three of the four

bibliographic divisions,4 were surely not four discrete functions that Sima Qian himself

intended his history to fulfill. Yet, insofar as he seems to have stated his various

ambitions for and conceptions of the text, these four categories encompass them

reasonably well.

2
In this regard, the Huainanzi can serve as a point of comparison. Mark Edward Lewis, for example,
discusses both works as examples of a universalizing tendency in “The Encyclopedic Epoch” (Writing and
Authority, 287-336).
3
See Qian Zhongshu‟s discussion of this trope in Guan Zhui Bian 管錐編 3.135-147, and Ronald Egan‟s
translation “Worldly Frustration and Literary Composition” on Limited Views, 35-40.
4
The four categories—Classics (jing 經), Histories (shi 史), Masters (zi 子), and Collected Works (ji 集)—
were introduced by the Suishu‟s 隋書 [History of the Sui] “Treatise on Classics and Records” [經籍志].
For a description how the divisions are categorized, see SuiS 32.906. Seeing the Shiji as in some sense
comparable to the Confucian Classics brings it close to the orbit of the jing category. It of course fits
naturally into the shi category, which where the Suishu places it. In the sense that it claims for itself the
distinction of being a new “school” of thought, it could also be seen as a zi text. Finally, although there is
less of a natural correspondence between ji and “the literature of suffering” perspective outlined above,
nonetheless, the ji category was the only place in the scheme where personal writings could be found. Thus,
in the sense that the Shiji was also a work of personal expression, it also could be seen as drawing near to
some of the works in the ji category.

46
The Shiji as a Classic on the Confucian Model

I will first consider Sima Qian‟s claim that the Shiji was intended to bear a special

relationship to the Six Arts [liu yi 六藝]5 that subsequently developed into the Confucian

Classics.6 I begin with a passage in the “Self-Narration” which raises the issue of the so-

called five hundred year sage cycle:

The Honorable Senior Archivist said, “The one who preceded me7 had a saying:
„Five hundred years after the Duke of Zhou died there was Confucius. It has now
been five hundred years since the death of Confucius. Is there no-one who can
bring back the enlightened age of the past, rectify the traditions of the Changes,
continue the Spring and Autumn Annals, and lay his foundations upon the
boundaries of the8 Odes and Documents, the rites and music?” Was this not his
ambition? Was this not his ambition? How can I, his son, dare to neglect his
will?‟” [太史公曰：「先人有言：『自周公卒五百歲而有孔子．孔子卒後至
於今五百歲，有能紹明世，正易傳，繼春秋，本詩書禮樂之際？』意在斯
乎！意在斯乎！小子何敢讓焉．」]9

The sage cycle idea is first seen in the Mencius, but even there it is somewhat

problematic. The matter is raised in two passages, Mencius 2B.13 and 7B.38. In the first

5
One presumes that to Confucius, the “Six Arts” referred to gentlemanly attainments of his time, listed in
the Zhouli 周禮 “Senior Minister of Works” [大司徒] as “rites, music, archery, charioteering, calligraphy,
and mathematics” (Zhouli 10.160). In the Shiji, this understanding exists side-by-side with the later
enumeration of the Six Arts, which took them to be “Rites, music, Documents, Odes, Changes, and Spring
and Autumn” (SJ 126.3197). There is also evidence that Dong Zhongshu and Jia Yi both understood the
term in this way. See Stephen Durrant‟s discussion of this issue, Cloudy Mirror, 47-69.
6
The convention of referring to the Classics as the “Six Arts” 六藝, still followed in the Hanshu‟s
“Treatise on the Arts and Writings” [藝文志], had already by the Eastern Han begun to be replaced by the
“Five Classics” [wu jing 五經]. The significance of this change went beyond the mere elimination of the
lost “Music” tradition. For a brief discussion, see Nylan, Five ‘Confucian’ Classics, 21.
7
Almost certainly referring to his father, Sima Tan.
8
Han Zhaoqi 韓兆琦, in his Shiji jianzheng 史記箋證, characterizes this phrase as being “somewhat
difficult” [略不順]. Its general sense is quite obvious, but scholars have perhaps paid too little attention to
the significance ji 際 in this context. Stephen Durrant‟s translation omits the last two characters: “take as
basis Poetry, Historical Documents, Ritual (Li) and Music (Yue)” (Durrant, Cloudy Mirror, 7). Burton
Watson‟s translation, which reads far more smoothly than mine, is “search into the world of the Odes and
Documents, the rites and music” (Watson, Ssu-ma Ch’ien, 50). I believe that ji 際 should be understood
more in the sense of boundary (as in the famous phrase from the “Letter in Reply to Ren An” 報任安書:
“investigate the boundary between heaven and man” [究天人之際], HS 62.2735). It could be argued that
the awkward phrase quoted above is in truth an important statement about the Sima family ambitions vis-à-
vis the Classics, though whether the boundaries in question are internal (among the four Classics mentioned)
or external (on the edges of those four) is in no way clear.
9
SJ 130.3296.

47
passage, the formulation is that “Every five hundred years a true King should arise, and

in the interval there should arise one from whom an age takes its name” [五百年必有王

者興，其間必有名世者]. Mencius then goes on to add, “From Zhou to the present, it is

over seven hundred years. The five hundred mark is passed; the time seems ripe” [由周

而來七百有餘歲矣。以其數則過矣，以其時考之則可矣].10 Confucius is not here

considered as part of the cycle of sovereigns (true kings), but is presumably “the one

from whom an age takes its name,” one who appeared in the interval. It is unclear what

Mencius, in saying that the five hundred year mark had passed, is impatient for—the

arrival of a true king? or the recognition that he himself bears some relation to this cycle

(whether as king or sage). He adds significantly, “As a matter of fact, heaven does not as

yet wish to bring peace to the Empire. If it did, who is there in the present time other

than myself?” [夫天未欲平治天下也，如欲平治天下，當今之世，舍我其誰也？]11

As this passage shows, reference to the five hundred year sage cycle can hardly avoid

being a thinly veiled claim for self-aggrandizement.

The other passage from the Mencius (7B.38) does clearly mention Confucius and

the figure “five hundred years,” but is less clear about the concept of a cycle. He simply

remarks that Yao and Shun were separated from Tang by that interval, as was Tang from

King Wen, and King Wen from Confucius. Mencius‟ main point there, however, is that

sageliness was transmitted down the line as a kind of cultural heritage. After Confucius,

he hints, there seems to be a danger of its dropping away: “In time we are so near to the

era of the sage while in place we are so close to his home, yet if there is no one who has

10
Mencius 2B.13 (Mengzi 4.85), trans. Lau, Mencius, 94.
11
Ibid., translation slightly altered.

48
anything of the sage, well then, there is no one who has anything of the sage” [去聖人之

世，若此其未遠也；近聖人之居，若此其甚也。然而無有乎爾，則亦無有乎爾].12

We must conclude that Sima Tan adapted rather freely from Mencius (and

perhaps other sources no longer extant). As Watson puts it:

Sima Tan…adapts the theory in a special way to apply not to rulers but to writers,
that is, from the Duke of Zhou, author of many of the Odes, to Confucius, author
of the Spring and Autumn Annals, to (he hopes) himself and his son. This passage
reveals the extent to which Sima Tan and his son regarded themselves as
peculiarly the heirs of Confucius and his model of historical writing.13

Sima Tan‟s comparison is both highly ambitious and rather free with the arithmetic, as

several scholars have pointed out.14

The five hundred year sage cycle is not Sima Qian‟s only claim that the Shiji

makes him an heir of Confucius. In the dialogue with Hu Sui, which also appears in the

“Self-Narration”, Sima Qian claims that he is “transmitting” ancient matters, not creating

a Classic as Confucius did with the Spring and Autumn.15 Yet many scholars who have

written about this passage consider this disavowal to be mere modesty.16 Confucius too

claimed to have been “transmitting, not creating” [述而不作].17 Earlier in his dialogue

with Hu Sui, the Sima Qian persona amply demonstrates his profound understanding of

the Classics, and expresses his own goals—his plan to fulfill his father‟s dying

command—in subtle, perhaps even ironic language.

12
Mencius 7B.38 (Mengzi 14B.264), trans. Lau, Mencius, 204.
13
Watson, Ssu-ma Ch’ien, 87. In quoting this passage, I do not mean to imply that I have any commitment
to the Duke of Zhou‟s authorship of the Odes or Kongzi‟s of the Chunqiu. Nor do I think that Watson held
any such beliefs, but merely meant that Sima Tan held them, a claim I find plausible.
14
E.g., Durrant, Cloudy Mirror, 7.
15
“What I call transmitting ancient affairs, and putting in order their generations and traditions, is not what
is called „creating‟, and it would be misguided for you to compare it to the Spring and Autumn” [余所謂述
故事，整齊其世傳，非所謂作也，而君比之於春秋，謬矣] (SJ 130.3299-3300).
16
See, for example, Durrant, Cloudy Mirror, 11; Puett, Ambivalence, 178.
17
Lunyu VII:1.

49
The Shiji as a Continuation of Ancient Scribal Tradition

The “Self-Narration” begins with a section detailing the Sima clan‟s pedigree.18

As Burton Watson notes, the earlier parts are an abbreviated paraphrase of a passage

from the Guoyu.19 In the time of the mythical emperor Chuanxu, the two ministers

Zhong and Li had charge of astronomical and terrestrial affairs respectively. There is a

vague statement of how their descendants continued to serve in the same capacity in the

reigns of Yao and Shun, and down through the Xia and Shang periods. Then the golden

mists of mythological time disperse slightly to reveal an actual name from the Zhou

period, a Lord Xiufu of Cheng. In the time of the Zhou King Xuan, the first Simas

appear: they are descendants of Lord Xiufu who had lost their holding and gained their

name. It is at this point that the text departs significantly from the Guoyu version, for it

claims that “the Sima clan had hereditary responsibility for the Zhou archival records”

[司馬氏世典周史].20

Sima Tan in his deathbed speech also emphasizes this point, underlining the

family connection with archival and astronomical responsibilities:

My ancestors were senior archivists for the house of Zhou. From high antiquity,
they were illustrious, meritorious, and renowned in that they were responsible for
astronomical affairs [in the courts of] Yu and Xia. Yet in later generations, there
was a decline. Will [the tradition] end with me? [余先周室之太史也．自上世嘗
顯功名於虞夏，典天官事．後世中衰，絕於予乎？]21
18
This section is generally referred to as a genealogy, but I use the term “pedigree” following a distinction
made by Raymond Geuss in the first chapter of Morality, Culture, and History. Geuss writes: “1) In the
interests of a positive valorization of some item 2) the pedigree, starting from a singular origin 3) which is
an actual source of that value 4) traces an unbroken line of succession from the origin to that item 5) by a
series of steps that preserve… or enhance… whatever value is in question” (1-5). It seems to me that, with
the exception of #4, this description fits the opening of the “Self-Narration” very well, and the clear effort
made in this section to gloss over the breaks in the line reveal a strong desire to establish #4 as well.
19
Watson, Ssu-ma Ch’ien, 202 n.3; see Guoyu 18.562.
20
SJ 130.3285.
21
SJ 130.3295.

50
Sima Qian is here placed in the context of a family tradition of historical records. Sima

Tan, at the end of his speech, exhorts Sima Qian to protect the preserved texts that are

apparently in the family‟s control: “I am deeply fearful that the archival writings of the

realm will be lost” [廢天下之史文，余甚懼焉].22

This is confirmed when Sima Qian inherits his father‟s office as Senior Archivist.

It is recorded that he “read the archival records [and?] the texts from the stone chambers

and the metal caskets” [紬史記石室金匱之書].23 The Shiji suoyin commentary explains

that the “stone chambers and metal caskets are both places where a state stores books”

[石室﹑金匱皆國家藏書之處]; contextually they are clearly linked to the archival

heritage of the Sima family. Sima Qian says nothing specific about the nature or

ownership of these sources, but implies much: that he and his family—and most of all,

his work on the Shiji—are linked to the ancient scribal tradition alluded to in texts like

the Zuozhuan,24 tracing its roots back to highest antiquity.

22
Ibid.
23
SJ 130.3296. Chou 紬 could also be understood as “to draw out” (usually silk), but Shiji Suoyin
commentator Sima Zhen argues that this is a borrowing for chou 抽, and quotes commentator Ru Chun‟s
如淳 (fl.230) glosses of the phrase as “to comprehend the historical events in old texts, arranging and
transmitting them”[抽，撤舊書故事而次述之]. Li Ciming 李慈銘 (1830-1894) argued, however, that the
word is a variant of zhou 籀, and means “to read texts”[讀書] (SKK 10.5199). There is more at stake in this
gloss than the contextual meaning of single character: at what point did Sima Qian begin the editorial
labors that would eventually result in the compilation of the Shiji? Ru Chun‟s gloss implies that he had
already begun editing, while Li Ciming‟s suggests he was only familiarizing himself with the texts that had
become available to him.
24
Examples of the heroism and/or erudition of the senior archivists (taishi 大史) in various states can be
found in the Zuozhuan (e.g., Xuan 2.5, CQZZ zhu 659-663). See also Burton Watson‟s informative
introduction regarding the figure of the pre-Han archivist, “Beginnings of Chinese Historiography”
(Watson, Ssu-ma Ch’ien, 70-100).

51
The Shiji as Philosophical Text

Sima Tan‟s essay, the “Essentials Points of the Six Schools” (六家要指) has often

been used as evidence that Sima Tan was a Daoist.25 The fact that Sima Qian describes

his father‟s education as including instruction not only in astronomy, but also in the Book

of Changes and “the theories of the Dao” (道論)26 does nothing to contradict the charge.

As many readers have pointed out, in Sima Tan‟s essay, the Daoist school (道家) is the

only one of the six to which no faults are ascribed. Certainly Sima Tan‟s essay reveals a

particular admiration for the Daoists, but Sima Tan‟s ambition seems to go beyond

merely praising one school over the others. By evaluating both the advantages and

shortcomings of all the schools, almost as a ruler might, Sima Tan in a sense placed

himself apart from, and above, all of them.27

To be sure, Sima Tan wanted his son to inherit the mantle of Confucius. But this

is not exactly the same as saying that he wanted Sima Qian to be a Ru.28 He complains of

the Ru that they “work very hard but with little result” [勞而少功], that the Six Classics

they revered had become so larded with commentaries that “one would not be able to

thoroughly comprehend their learning even in many generations [of study]; in a whole

lifetime one could not research [all the details] of their rites” [累世不能通其學，當年不

能究其禮]. 29 Surely this is not the kind of hopeless toil that Sima Tan would wish for

25
See for example, Li Changzhi, Sima Qian zhi ren’ge yu fengge, 28-31.
26
See SJ 130.3288.
27
For further discussion of Sima Tan‟s supposed Daoism as later related to Shiji authorship controversies,
see chapter 7.
28
The Lunyu suggests that the category of Ru predated Confucius (see Lunyu VI:13). Although the term
certainly did come to be associated with followers of Confucius, there is good evidence that Confucius was
available as a figure of wisdom to those outside the Ru tradition as well (see, for example, the story of Ai
Tai Tuo in Zhuangzi 156-157).
29
SJ 130.3290.

52
his son. Confucius, to the Simas, was the originator of those Six Classics: with his

editorial activities, he changed what it meant to be a Ru. But this is not to say that

Confucius was himself in exactly the same category as the Han Ru that his legend helped

create. Sima Tan, according to Sima Qian‟s retelling at least, seems to have hoped that

his son too could change the face of scholarship in his time. He wanted his son to be a

Confucius, not a Confucian.30

We know who Sima Tan‟s teachers were, but of Sima Qian‟s we know very little.

It has been suggested that he studied with Dong Zhongshu,31 since the “Traditions of the

Forest of Scholars” 儒林列傳 (Shiji ch.121) discusses Dong Zhongshu‟s pedagogical

peculiarities in detail,32 and furthermore Sima Qian seems to quote (or rather paraphrase)

Dong Zhongshu extensively in the dialogue with Hu Sui.33 Some have attempted, on this

account, to make Sima Qian a Gongyang scholar, and intellectually a member of Dong

Zhongshu‟s jia.34 Yet nowhere does Sima Qian openly acknowledge Dong Zhongshu as

his teacher. Though Sima Qian clearly respects Dong, I think it would be wrong to

overestimate the influence of Dong‟s thought upon Sima Qian, who clearly aimed to

make his own mark on the intellectual world.

30
See Li Changzhi, who argued that Sima Tan wanted Sima Qian to become “a second Confucius” (Sima
Qian zhi ren’ge, 63). Stephen Durrant, expanding on Li‟s point, discussed this issue in detail in “The
Frustration of the Second Confucius” (Cloudy Mirror, 1-27). Grant Hardy also emphasizes the connection,
writing that “the image of Confucius haunts the pages of the Shiji like the ghost of Hamlet‟s father…. One
can sense [Sima Qian‟s] desire to fulfill the role that his father had envisioned for him” (Bronze and
Bamboo, 116). See also Hardy, Bronze and Bamboo, 17.
31
Zhou Shouchang 周壽昌 (1814-1884) states outright in his commentary on the Hanshu version of this
passage that Sima Qian studied with Dong Zhongshu (Hanshu zhu bu zheng, qtd. SKK 130.21). See also Li
Changzhi‟s Sima Qian zhi renge yu fengge, 117 and Xiao Li‟s 肖黎 Sima Qian pingzhuan 司馬遷評傳, 36.
Durrant himself remains agnostic on the question (see Durrant, Cloudy Mirror, 58-59).
32
“He would lower the curtains [of his room] and lecture or recite [from behind them]. His disciples would
pass down his teachings to each other according to seniority, so that there were some who had never seen
his face” [下帷講誦，弟子傳以久次相受業，或莫見其面] (SJ 121.3127).
33
SJ 130.3297.
34
See, in particular, Jurij L. Kroll‟s “Ssu-ma Ch‟ien‟s Literary Theory and Literary Practice.”

53
 The Hanshu “Traditions of the Forest of Scholars” 儒林列傳 (Hanshu ch.88) also

mentions that Sima Qian “inquired about antiquity” from the Documents scholar Kong

Anguo, and that “Qian‟s writings record [material from] the Yaodian, Yugong,

Hongfan,Weizi, Jin Teng, and other chapters, with many Old Text explanations” [遷書載

堯典﹑禹貢﹑洪範﹑微子﹑金縢諸篇，多古文說].35 It is not clear, however, whether

Ban Gu is offering Sima Qian‟s having studied with Kong Anguo as an explanation for

his use of the Old Text documents, or whether on the contrary Ban Gu is inferring from

Sima Qian‟s use of those texts that he had studied with Kong Anguo. In either case,

nothing in the Shiji itself openly discusses Sima Qian‟s teachers.

Instead, Sima Qian gives himself an intellectual biography readily comparable to

that of Confucius.36 And in the concluding section of his “Self-Narration,” Sima Qian

makes the famous claim that the Shiji “completes the words of a single jia” [成一家之

言]37—where jia should be understood to mean both the Sima family and a whole new

school of thought, one which emphasized the lessons of history.

The Shiji and the Literature of Suffering

As mentioned above, Sima Qian suggested a last alternative context for his work,

one which in a sense includes at least some of the Classics but goes beyond them as well.

After the Li Ling affair, the “Self-Narration” implies, Sima Qian began to reconsider his

work on the Shiji. He now compared it to the works of writers in the past who had

suffered misfortune. The figures in the list are, as mentioned above, heterogeneous.

35
HS 88.3607.
36
See Durrant, Cloudy Mirror, 29-25.
37
SJ 130.3319.

54
Sima Qian reviews the apparent misfortunes of the writers of the Odes and Documents, of

King Wen who worked on the Changes, and of Confucius, who appears as compiler of

the Chunqiu. The list continues, however, with the poet Qu Yuan and his supposedly

autobiographical poem the Li Sao; Zuo Qiuming, the purported compiler of the Guoyu38;

Sunzi and his famous treatise on military strategy39; Lü Buwei with his encyclopedic

compilation the Lüshi chunqiu40; and Han Feizi, mentioned as author of two essays in

particular, the “The Difficulties of Persuasion” [說難] and “Solitary Resentment” [孤

憤].41 The list ends with a second mention of the authors of the Odes, even though they

were also included above, and a summary of what Sima Qian takes as the important

common thread binding all these diverse figures together: “These men all had pent up and

frustrated intentions, and were not able to carry out their Way. That is why they narrated

the affairs of the past, thinking of those who were to come” [此人皆意有所鬱結，不得

通其道也，故述往事，思來者].42

It may be that the confluence of misfortune and literary achievement was the only

purpose Sima Qian had in compiling this list.43 More could be read into his selection,

38
Zuo Qiuming is also the putative author the Zuozhuan, but only the Guoyu is mentioned here.
39
The Bingfa 兵法 [Military method], usually translated The Art of War.
40
In the list, this work is referred to merely as the Lü lan 呂覽, which forms the middle section (books 13-
20) of the text we know today as the Lüshi chunqiu 呂氏春秋. Knoblock and Riegel translate this section
as “The Examinations.”
41
These are chapters 11 and 12 in the Han Feizi text we have today.
42
SJ 130.3300.
43
Scholars who have understood the list to imply that misfortune preceded literary activity in each case
have raised numerous objections to the list on historical grounds. For example, Liang Yusheng 梁玉聲
takes Sima Qian to task for his mention of both Lü Buwei and Han Feizi (Shiji zhiyi 3.1470). I would agree
with Sun Deqian 孫徳謙 (1869-1935) that most commentators have misunderstood something about the
intended relationship between the misfortune and the textual production. Regarding the two above-
mentioned cases, for example, Sun writes, “They are saying that although [Lü] Buwei was exiled to Shu,
still the Lü Lan that he compiled is still transmitted by the world;, although Han Fei was imprisoned in Qin,
the world also transmits his “The Difficulties of Persuasion” and “The Sorrow of Standing Alone” [言不韋
雖遷蜀，而其所著之呂覽則世傳之。韓非雖囚於秦，而說難、孤憤世亦傳之] (Taishigong shu yifa,

55
however: did he mention these particular texts because he wanted them seen as potential

points of comparison for the Shiji?

Sima Qian also specifically mentioned the Li Sao in his chapter on Qu Yuan, and

his discussion there overflows with the empathy that the historian seems to have felt for

the poet: “To be trustworthy and find oneself doubted, to be loyal and find oneself

defamed—is it possible not to resent it? Qu Yuan‟s authorship of the Li Sao probably

arose from this resentment” [信而見疑，忠而被謗，能無怨乎？屈平之作離騷，蓋自

怨生也].44 It seems that in his misfortune, Sima Qian would have liked to have been

compared to Qu Yuan, though no one obliged him until much later.

The comparison with Zuozhuan and Guoyu is even clearer, insofar as those two

works were historical in nature, like the Shiji, and of course formed part of the Shiji‟s

source material. Readers familiar with the Shiji‟s multi-chapter depiction of the Chu-Han

war45 would also not be surprised to find Sima Qian pointing to Sunzi‟s Bingfa 兵法

[Military methods] as an inspiration and potential point of comparison: Gu Yanwu 顧炎

武 (1613-1682) once praised Sima Qian by saying that he “had the strategic disposition

of the realm inside his chest” [胸中固有一天下大勢], and that “probably since antiquity

there has never been an historical text which in describing military affairs gave as much

detail about the terrain as the [Shiji]” [蓋自 古史書兵事地形之詳，未有過此者].46

130). In Wai-yee Li‟s interpretation, Sima Qian “sometimes sacrifices factual accuracy to develop a new
conception of writing and to forge a special genealogy for his own enterprise” (“Authority,” 362-363).
44
SJ 84.2482.
45
For a detailed reading of the Shiji‟s overall presentation of this war, see Hardy, Bronze and Bamboo, 86-
113.
46
Ri zhi lu 日知錄 [Record of Knowledge Daily Gained], 27.737.

56
 The inclusion of Lü Buwei, however, reveals a different side of Sima Qian‟s

ambitions. The compilation that Lü sponsored, the Lüshi chunqiu 呂氏春秋 [Spring and

autumn of Master Lü] was noteworthy first and foremost for its completeness. If the Shiji

can be believed, Lü had the text “displayed at the market gate of Xianyang with a

thousand gold hanging above it, inviting any of the various lords, wandering scholars, or

visitors to add or subtract a single character from it; if they were able to do it they would

be given the thousand gold” [布咸陽[四]市門，懸千金其上，延諸侯游士賓客有能增

損一字者予千金].47 The encyclopedic aspect of the Shiji, too, is something that scholars

have tended to emphasize,48 and their arguments certainly raise the possibility that Sima

Qian‟s chapter on Lü Buwei may have been shaped partly by a desire to compare Lü‟s

project with encyclopedic or “macrocosmic” nature of Sima Qian‟s own work.49 Finally,

Han Feizi seems a puzzling case until we realize that later anti-Legalist sentiment has

obscured the extent to which Han may have been an acceptable model for sympathy and

admiration in Sima Qian‟s day. The Shiji quotes him approvingly more than once,50 and

places his story in the same chapter with Laozi‟s. Furthermore, Sima Qian expresses

personal regret at Han Feizi‟s unfortunate end, writing, “I only regret that Master Han

made the „Difficulties of Persuasion‟ but was simply unable to extricate himself [from his

47
SJ 85.2510.
48
As discussed below, Yang Xiong and Huan Tan compared the Shiji to the Huainanzi 淮南子 another
huge multi-authored compendium sponsored by Liu An 劉安, the King of Huainan. Scholars today have
also seen this aspect of the Shiji as being highly significant. See especially Mark Edward Lewis, who
writes that like the Huainanzi, “the Shiji also aimed to give a textual form to a world empire” (Writing and
Authority, 309); and Grant Hardy, who characterizes the Shiji as a “microcosmic model” of the world
(Bronze and Bamboo, 50-60).
49
See SJ 85.2511.
50
E.g., SJ 79.2425, 124.3181.

57
fate]” [余獨悲韓子為說難而不能自脫耳].51 Sima Qian, who despite his literary gifts

was also unable to extricate himself, may well have intended to draw the comparison.

In conclusion, the purpose of the suffering authors list is not merely that the

figures mentioned encountered tragedy. The fact that they produced works of greatness

is crucially important. The list is more than a canon of misfortune; it is also a canon of

achievements, achievements of a type which Sima Qian considered potentially

comparable to his own.

EARLY VIEWS OF THE SHIJI

Sima Qian‟s early readers seemed to view his work as being loosely associated

with the category known as the “various masters,” for the most part emphatically

different from the Classics. One of the earliest contexts in which the Shiji is mentioned

involves Liu Yu 劉宇 (d.19 BCE), the Prince of Dongping 東平:

[In 28 BCE, Liu Yu] came to court and sent a memorial to the throne requesting
the works of the various masters and the Writings of the Honorable Senior
Archivist. The emperor [Cheng, r.33-7 BCE] asked General Wang Feng 王鳳
[d.22 BCE] about it, and he replied, “I have heard that when the various lords
make official visits to the court, they investigate writings and rectify regulations.
They say nothing that is not according to propriety. Now the Prince of Dongping
has been fortunate enough to be allowed to come to court, but he does not think of
being orderly in conduct and respecting the rules, which are what prevents
dangerous slips. Instead he requests various writings, which is not the appropriate
to the duty of official court visits. Some writings of the various masters are
contrary to the methods of the Classics or go against the Sages. Others give honor
to ghosts and spirits and express belief in monsters and anomalies. As for the
Writings of the Honorable Senior Archivist, they contain the strategies of cunning
and leverage from the Warring States advocates of horizontal and vertical
alliances; [stories about] scheming ministers and bizarre policies from the very
beginning of the Han‟s rise; disasters and prodigies from the Office of Heaven;
information about terrain and strategic passes; and all of this would not be
appropriate in the hands of the various lords and princes. We cannot give them to

51
SJ 63.2155.

58
 him. The explanation for denying permission should say, [後年來朝，上疏求諸
子及太史公書，上以問大將軍王鳳，對曰：臣聞諸侯朝聘，考文章，正法
度，非禮不言．今東平王幸得來朝，不思制節謹度，以防危失，而求諸書，
非朝聘之義也．諸子書或反經術，非聖人，或明鬼神，信物怪；太史公書有
戰國從橫權譎之謀，漢興之初謀臣奇策，天官災異，地形阸塞：皆不宜在諸
侯王．不可予．不許之辭宜曰：]

The Five Classics were regulated by the Sage, and of the ten thousand
affairs there are none that are not exhaustively recorded there. If the
prince investigates and delights in the Way, and his advisors are all
classical scholars, and if day and night he diligently recites, that is enough
to rectify his person and give pleasure to his mind. Now petty debates are
ruinous to duty, and the lesser ways are not comprehensive. If one goes
far along this path, there is a fear that one will become bemired. None of
it is worth your attention. But as for the various things that will be of use
to your study of Classical methods, none will be kept from you. [五經聖
人所制，萬事靡不畢載．王審樂道，傅相皆儒者，旦夕講誦，足以正
身虞意．夫小辯破義，小道不通，致遠恐泥，皆不足以留意． 益於
經術者，不愛於王．]

When this reply was submitted, the Son of Heaven followed Feng‟s advice, and
did not grant [the Prince‟s request]. [對奏，天子如鳳言，遂不與．]52

This anecdote has often been understood by scholars to mean that access to the Shiji was

tightly controlled by the court, that in effect the Shiji was a banned book at this time.

When we consider more carefully the details of the case, however, we might be reluctant

to make such generalizations. There is actually no evidence that the Shiji had been

banned or controlled before this point. Chu Shaosun 禇少孫 (fl.1st c BCE), for example,

had read and enjoyed it (at least parts of it),53 and went on to supply continuations for

some of the chapters.54

To understand this incident, it is important to investigate its background

circumstances. Liu Yu was a son of Emperor Xuan (r.74-49 BCE) and received his

52
HS 80.3324-3325.
53
He commented, “I was fond of reading and surveying the Arrayed Traditions of the Honorable Senior
Archivist” [好覽觀太史公之列傳] (SJ 60.2114).
54
See discussion in chapter 7.

59
appointment to Dongping in 52 BCE. His biography in the Hanshu55 is entirely devoted

to his misbehavior, the reprimands he received from his brother Emperor Yuan (r.48-33

BCE), and his occasional faint-hearted attempts to reform. The prince was therefore the

young Emperor Cheng‟s uncle, and a known miscreant whose close kinship with the

previous emperor was the only thing that had heretofore saved him from serious

punishment.56 Probably any request he made would have been suspect.

Furthermore, had there been a precedent in place (if the Shiji was officially

banned in some way), Emperor Cheng would not have needed to rely on Wang Feng to

make the decision—and for that matter, Liu Yu might not have made such a bald request

in the first place. True, the answer Wang Feng supplied does seem to be a general

condemnation of the text, and is always taken as such. But we should remember that

Wang Feng too was a scheming general who could also be accused of having benefited

from the strategies and omens recorded in the Shiji.57 Certainly Wang Feng seems to

show a detailed knowledge of the Shiji‟s contents, suggesting that he could be counted

among the early readers of the text.

55
HS 80.3320-3326.
56
Indeed, the terms in which Wang Feng criticizes Liu Yu‟s behavior closely echo Emperor Yuan‟s earlier
policy toward the wayward prince, at one point repeating word for word a phrase that had been used by a
remonstrator on behalf of the former emperor—制節謹度—which I have tentatively translated as “being
orderly in conduct and respecting the rules.” On the earlier occasion Liu Yu had been enjoined to do this
“in order to assist the Son of Heaven” 制節謹度以翼天子 (HS 80.3321). Whether Wang Feng himself
chose to use the same words as the prior remonstration, or whether they were placed in his mouth by the
compiler of the Hanshu, either way it shows that Liu Yu‟s earlier transgressions formed part of the back-
story for his visit to court, and potentially relevant to the refusal to grant his request.
57
Wang Feng, like Liu Yu, was Emperor Cheng‟s uncle, but on the maternal side. In addition, he was
uncle to Wang Mang 王莽 (45 BCE-23 CE). Though he did not live to see his other nephew‟s usurpation
of the imperial throne—which brought an end to the Western Han—he was, at the time of the Liu Yu
incident, one of the most powerful officials in the land. In the words of Michael Loewe, Wang Feng “was
partly responsible for consolidating the powers that the Wang family were to hold for several decades,
culminating in Wang Mang‟s rule as emperor” (Biography, 520). Loewe also notes, “Wang Feng‟s rise to
power was taken as verification of a number of climatic or other phenomena which were seen as portents.
These included events such as the appearance of a comet or the outbreak of fires that were reported
between 43 and 27” (Biography, 521, cf. HS 97B.3982).

60
 In any case, what is important for the purpose of contextualizing the Shiji is how

it is here placed in sharp opposition to the Classics. The Classics, which are pronounced

to be the proper subject of Liu Yu‟s study, are portrayed as a tool for control and

attempted behavior modification. This is in contrast to other texts that were clearly more

entertaining and potentially more dangerous. That Liu Yu requests the Shiji together with

the various masters suggests that the Shiji was loosely associated with that category of

texts, but also shows that the Shiji was not considered to be one of these texts.

Wang Feng complained that the various master texts contradicted the sages and

the Classics, a criticism that would very soon be leveled against the Shiji as well. The

other problem with those texts, in Wang Feng‟s view, was their promotion of the

supernatural. The problems with the Shiji are slightly different. Clearly Wang Feng

feared that the Shiji would inspire the refractory Liu Yu to make trouble. The first two

aspects of the Shiji that he finds objectionable in this regard have to do with realpolitik of

the past—the Warring States and the post-Qin periods—both times when the central

government was weak. The other two aspects seem more current: disasters and prodigies

could be twisted for predictive or at least propagandistic advantage in the present,58 while

geographical information could become valuable military intelligence. We get the sense

that the Shiji might have been set apart from the various master texts because its potency

was so much greater.

58
Indeed, this was already happening with increasing frequency. See Bielenstein, “An Interpretation of the
Portents in the Ts‟ien-Han-Shu”; Kern, “Religious Anxiety and Political Interest in Western Han Omen
Interpretation”; Sukhu, “Yao, Shun, and Prefiguration.”

61
 A similar association of the Shiji with the various masters appears in Yang

Xiong‟s 揚雄 (53 BCE-18 CE) “Self-Narration,” as preserved in his Hanshu biography.59

When I look at the various masters, each uses his understanding to gallop in an
opposite direction from others. What it comes down to is slandering the Sage,
while their own works are bizarre and misguided, just crooked arguments and
lying words, which they use to bring chaos to current affairs. Though they are
only petty arguments, in the end they are capable of shattering the Great Way and
confusing the masses, causing people to have a weakness for hearsay, themselves
not being able to recognize its falseness. When it comes to the way the Honorable
Senior Archivist made records of the Six States and charted the Chu-Han period,
ending with the appearance of the unicorn, he was not in accord with the Sage,
and his judgments were rather different from those of the Classics. Therefore,
from time to time people have asked me questions, and I always use exemplary
sayings to respond to them. I have made a compilation of them here in thirteen
chapters, resembling the Lunyu, and I call it the Fayan. [雄見諸子各以其知舛
馳，大氐詆訾聖人，即為怪迂，析辯詭辭，以撓世事，雖小辯，終破 大道
而或衆，使溺於所聞而不自知其非也．及太史公記六國，歷楚漢，(記)[訖]
麟止，不與聖人同，是非頗謬於經．故人時有問雄者，常用法應之，譔以為
十三卷，象論語，號曰法 言．]60

As mentioned above, Wang Feng‟s complaint about the various masters had been their

failure to agree with the sage and Classics, and their predilection for the supernatural.

Yang Xiong, on the other hand, criticizes them most harshly for schismatic tendencies,

their failure to agree among themselves, and the potential damage this could do to the

intellectual unity of the empire. Yang‟s charge of “slandering the Sage” should be

understood in this context, for it is the Sage who, in Yang‟s view, represents the best

hope for unity.

Unlike Wang Feng, Yang Xiong does not criticize the Shiji for its content.

Instead he focuses on the problem of Sima Qian‟s judgments. Like the various masters,

59
Though Ban Gu does not actually say it is Yang Xiong‟s “Self-Narration,” that is the scholarly
consensus—see Loewe, Guide, 101. See also Che Xingjian 車行健, “Handai shuxu de tizhi” 漢代書序的
體制 [Structure of the Han Dynasty preface genre], which discusses this and other prefaces.
60
HS 87B.3580.

62
Sima Qian failed to agree with the Sage and the Classics,61 but there was otherwise

nothing wrong with his project. This is the judgment we would expect from Yang Xiong,

given that he himself also reportedly wrote a continuation of the Shiji (see below).

Furthermore, Yang Xiong offers his own Fayan 法言 [Exemplary sayings]62 as a

corrective to the Shiji‟s erroneous judgments: though the two works are very different in

form, it seems from Yang‟s “Self-Narration” that he envisioned some similarity of

function.

Yang Xiong‟s comments on the Shiji, all found in the Fayan, were extremely

influential for later readers of the Shiji. They will be discussed in more detail chapter 6.

Here, I will merely consider the other texts with which Yang compares the Shiji.

The first is the Huainanzi. The central government came into possession of the

text we now call the Huainanzi in 139 B.C.E.63 This means that Sima Qian could

potentially have seen it,64 though in writing of Liu An he made no mention of the

sponsoring or editing of such a work.65 More than hundred years later, however, the two

works were clearly seen as potentially comparable:

Someone asked, “Is it not so that Huainan and the Honorable Senior Archivist
possessed great knowledge? But how eclectic they were!” [I] said, “Eclectic,
how eclectic! When people go astray from having too much knowledge, then
their works are eclectic. Only the works of the Sage are not eclectic.”66 [或曰：

61
The statement that Shiji is potentially damaging because of its failure to agree with the Sage and the
Classics would later be paraphrased and popularized by both Ban Biao 班彪 (3-54) and his son Ban Gu, as
discussed below, and would eventually become one of the most common epithets associated with the Shiji.
62
The Fayan was completed around 9 CE (Loewe, Guide, 101).
63
Charles LeBlanc, “Huai nan tzu” in Loewe, Guide, 189.
64
Jin Dejian 金德建 argues that in fact he did (see Sima Qian suo jian shu kao, 349-361). I do not find his
arguments convincing, however. His main piece of evidence involves this quotation from Yang Xiong—he
argues that the juxtaposition of the two texts Yang, who lived very near in time to Sima Qian, implies that
Sima Qian saw Liu An‟s compilation.
65
See the “Huainan Hengshan liezhuan” 淮南衡山列傳 [Arrayed traditions of Huainan and Hengshan] (SJ
118.3075-3098).
66
I should note that Chen Zhi has a completely different interpretation of the last sentence, which he
glosses as, “Ordinary people consider these works to be eclectic, but a person of true attainment does not

63
 「淮南、太史公者，其多知與？曷其雜也！」曰：「雜乎雜！ 病以多知為
雜，惟聖人為不雜．」]67

The Huainanzi is seen today as belonging to the category of „various masters‟—the

closest early Chinese equivalent to „philosophy.‟68 The Shiji, on the other hand, is not

seen as philosophical in the slightest. The fact that Yang Xiong (or at least his

interlocutor) saw the two works as comparable points to the role of the Shiji in the above-

mentioned encyclopedic tradition—hinted at by Sima Qian himself when he placed Lü

Buwei on his list of suffering authors (see discussion above).69

In the Fayan, both Sima Qian and Liu An stand accused of eclecticism, in contrast

to the Sage. Yang Xiong‟s words have an underlying implication: that, in the words of

John Henderson “classics…contain nothing superfluous or insignificant.”70 Thus it is by

definition that the works of the Sage are not eclectic. Neither the Huainanzi nor the Shiji

could yet aspire to canonical status. Whereas the “great knowledge” of the Sage seemed

to Yang Xiong to be “comprehensive and all-encompassing,”71 the “great knowledge” of

Sima Qian and Liu An seemed merely eclectic because the latter two lack authoritative

status adequate to justify claims about the relevance of each and every word.72

consider them eclectic” (“Han Jin ren,” 224). Taken out of context, this is a fine reading of the passage.
But in context, especially in light of the introduction, it is clear that Yang Xiong (presumably a person of
true attainment in his own eyes) does consider the Shiji to be eclectic. Yang Haizheng‟s discussion of the
passage shows that she also understands the line in the way I have translated (Han Tang, 23-24). Chen‟s
reading shows the vulnerability of the laconic “exemplary saying” genre to slippage in interpretation.
67
FY 5.163.
68
In the Hanshu, the Huainanzi, together with the Lüshi Chunqiu, appear in the “miscellaneous” or
“eclectic” category (za 雜); see HS 30.1741. This is presumably due to the fact that their syncretic nature
makes them difficult to categorize under any one „school‟ of thought, which is how other „various masters‟
works are organized.
69
Note that Liu An does not appear on that list.
70
Scripture, Canon, Commentary, 121. This is presented as one of several commentarial assumptions
which appear in most if not all scriptural traditions.
71
Again, Henderson‟s phrase (Scripture, Canon, Commentary, 89). That a given canon is comprehensive
in this sense is, Henderson writes, “the most universal and widely expressed commentarial assumption.”
72
In the late Ming and Qing, when the Shiji had acquired a degree of canonical status, claims of total
relevance were sometimes made about it, at least with regard to certain core chapters. Mao Kun, for

64
 Having explored the common points of the Shiji and the Huainanzi, Yang Xiong

goes on in another passage to point out contrasts among a variety of works. Again, his

list includes both the Shiji and the Huainanzi:

The explanations in the Huainan are not as useful as the [Writings of the]
Honorable Senior Archivist. In the Honorable Senior Archivist, a Sage might find
something of value, but in the Huainan, there is very rarely anything worthy of
note. It is imperative to be a Ru! Sometimes getting it, sometimes not—that‟s the
Huainan. Elegant prose but scarce utility—that‟s Changqing [=Sima Xiangru].
Greatly fond but unrestrained—that‟s Zichang [=Sima Qian]. Zhongni
[=Confucius] was greatly fond—of moral duty. Zichang‟s great fondness is for
the unusual. [淮南說之用，不如太史公之用也．太史公，聖人將有取焉；淮
南，鮮取焉爾．必也，儒乎！乍出乍入，淮南也；文麗用寡，長卿也；多愛
不忍，子長也．仲尼多愛，愛義也；子長多愛，愛奇也．]73

Yang Xiong‟s reasons for choosing to compare these particular texts are a little obscure.

Perhaps he brought them together for purely chronological reasons: though the works of

Sima Xiangru predated the Shiji and Huainan by perhaps a generation, the three may

have been seen retrospectively as being products of the same era. Taking the two

passages together, we can conclude that Yang Xiong judged both the Shiji and the

Huainan to be overly eclectic, but considered them different in that the Shiji, at its best,

might appeal to a sage.

With the introduction of Sima Xiangru, another axis of comparison appears—that

of literary elegance. While the Huainanzi was inconsistent, the works of Sima Xiangru

were flashy but almost entirely lacking in usefulness. The Shiji, on the other hand, was

occasionally useful, but suffered a different problem, that Sima Qian‟s fondness for the

unusual exceeded his fondness for moral duty. If we put this together with Yang Xiong‟s

example, wrote, “Should you want to cut out or add a sentence or a character in any place [in the Shiji],
then it is like picking out one thread from a silk cloth—one would find it difficult indeed to do it” [於中欲
損益一句一字處，便如於匹練中抽一縷，自難下手] (SJYJJC 6.172).
73
FY 12.507.

65
“Self-Narration,” quoted above, we may infer that this is at least one of the ways in which

Yang Xiong sees Sima Qian as contradicting the Sage and the Classics.

The last comparison of Yang Xiong‟s that I will consider juxtaposes the Shiji with

texts that would eventually be included in the Classical canon:

Someone asked, “The Offices of Zhou?” [I] said, “They establish procedures.”
“Master Zuo?” [I] said, “It appraises and evaluates.” “Senior Archivist Qian?” [I]
said, “A veritable record.” [或問「周官」．曰：「立事．」「左氏」．曰：
「品藻．」「太史遷」．曰：「實錄．」].74

The first reference to the Offices of Zhou (otherwise known as the Zhouli 周禮75) is in the

Shiji “Treatise on the Feng and Shan” (ch.28). There it is one of the texts employed by

the various Ru in their attempts to design appropriate Feng and Shan rituals.76 By the

time of the Hanshu “Treatise on the Arts and Literature,” the Zhouli seemed to have

entered into the canon. Xun Yue ascribed this to Liu Xiang. As William Boltz writes in

Early Chinese Texts: a Bibliographic Guide:

Xun Yue states that Liu Xin also proposed calling the text Li jing, a title that is
occasionally used in reference to the Zhou li. Liu Xin sought to have the post of
an official scholar established for the Zhou li; since the name whereby the work is
entered in Hanshu ch.30 is Zhou guan jing, it may be said that from the time of
Liu Xin the Zhou li has been regarded as a classical text.

Though this outline of the Zhou guan‟s canonization does not specify exact dates, it does

seem that it took place during Yang Xiong‟s life-time. The Zuozhuan too was canonized

not long afterward as the third official commentarial tradition attached to the Chunqiu,

though this provoked a bitter debate.77 Of the three, only the Shiji failed to become

attached to the Classical canon. This raises some interesting questions about the

comparison. Did Yang Xiong and/or his interlocutor see all three of these works as
74
FY 10.413.
75
For evidence on this point, see Han Zhaoqi‟s discussion, Shiji jianzheng, 4.2025.
76
SJ 28.1397.
77
See, e.g., the Cloud Terrace debates of 28 BCE (HHS 36.1228-32).

66
potential additions to the Classical canon (two of which would ultimately succeed, while

the third, the Shiji, would not)? Or was the pronouncement of “true record” supposed to

convey to the reader that the Shiji is functionally different from the other two texts?78

The next generation of readers had a somewhat different perspective on the Shiji.

For them, Sima Qian was not such an original and dangerous thinker. In contrast with the

fad for politically-charged prophetic apocrypha, the Shiji must have seemed quite solid

by comparison.79 In the early Eastern Han restoration era, Sima Qian was presented

primarily as a compiler and arranger of impressively voluminous material.

Huan Tan 桓譚 (ca.43 BCE-28 CE), a younger contemporary of Yang Xiong‟s,

clearly portrayed him in this light. Huan Tan‟s Xin lun 新論 [New discourses] has been

lost, but fragments of it were preserved in a variety of places (including Shiji

commentaries). This selection is found in the Tang dynasty collection, the Yilin 意林

[Forest of intentions]:

Had Jia Yi not been degraded and disappointed, his literary elegance would not
have been produced. Had Liu An, the Prince of Huainan, not been noble,
successful, and wealthy, he could not have employed a host of eminent scholars to
compose a book. Had the Honorable Senior Archivist [Sima Qian] not been in
charge of texts and records, he would not have been able to put in order
everything from antiquity to the present. Had Yang Xiong not been poor, he
could never have written his Mysterious Words. [賈誼不左遷失志，則文彩不
發。淮南不貴盛富饒，則不能廣聘駿士，使著文作書。 太史公不典掌書
記，則不能條悉古今。揚雄不貧，則不能作玄言。]80

78
Further issues related to the profound and laconic pronouncement “true record,” what it might have
meant and what it would come to mean, will be explored in chapter 6.
79
See Itano Chōhachi, “The t’u-ch’en Prophetic Books and the Establishment of Confucianism” and Gopal
Sukhu op. cit.
80
Yilin 3.7; trans. adapted from Pokora, Hsin-lun, 18-19.

67
The comparison is restricted to Han figures, and here the ordering is chronological. I

would argue that to Huan Tan, this was the new canon, the great authors of recent times.

Each of the authors he mentioned had composed works appropriate to (and in fact

enabled by) their circumstances. The crucial aspect of Sima Qian‟s circumstances was

his access to texts and records—not that he was a tragic victim of the Li Ling affair. I

should also note that part of what motivates this list is Huan Tan‟s desire to elevate his (at

that time under-appreciated) hero, Yang Xiong, by bringing him into juxtaposition with

already famous authors.

Wang Chong 王充 (27-97?), another early reader of the Shiji, seemed also to have

seen Sima Qian as noteworthy primarily in his access to texts and in his ability as a

compiler. In “Surpassingly Rare” 超奇 (Lunheng 論衡 ch.39), Wang Chong explored the

value of textual production in explicit contrast to that of scholarship. He discussed each

type of textual production in what he considered ascending order of merit. He began by

mentioning that there are many people who were educated but failed to produce writings.

Barely higher than these were scholars who merely produced glosses or wrote memorials

making suggestions based on those glosses. Above them he placed Sima Qian and Liu

Xiang. He described the work of these two figures as “collecting and enumerating

historical facts of ancient and modern times, and narrating things that have happened” [抽

列古今，紀著行事]; he also praised the voluminousness of their writings. However, he

criticized them because “they relied on accomplished [facts] and merely record former

68
events, without producing anything from their own minds” [因成紀前，無胸中之造].81

By this standard, Wang Chong considered Lu Jia 陸賈 (240 BCE-170 BCE) and Dong

Zhongshu to rate more highly. Above them are Yangcheng Zichang 陽成子長 (for

writing the Yuejing 樂經 [Classic of music]) and Yang Xiong (for the Taixuanjing 太玄

經 [Classic of Supreme Mystery]): Wang Chong wrote that “no one but a man of almost

perfect talent could have produced [these works]” [非庶幾之才，不能成也],82 and he

compared them to the writings of Confucius. Highest of all is Wang Chong‟s evaluation

of Huan Tan, of whom he wrote that

The minds of lapidaries are surely more admirable than their precious stones, and
the wisdom of those who perforate tortoise-shells is closer to divine than that of
the tortoises. Similarly he who knows how to discriminate among the talents of
all scholars and assign a rank to each must be superior to those ranked. [采玉者
心羡於玉，鑽龜者知神於龜．能差眾儒之才，累其高下，賢於所累．]83

Has the profession of criticism ever received a more heartening endorsement? Alfred

Forke, an early translator of Wang Chong‟s work, found the assertion so strange that it

drew from him a highly indignant footnote: “This is evidently wrong. A critic must not

be superior to those he criticizes. They are in most cases much above him.”84 But Wang

Chong has no such assumption. He did not scorn criticism as a derivative product. If we

take seriously his metaphors of stones and tortoises, he meant that only through the
81
LH 39.607-608. Translation adapted from Forke II.297. Both Michael Nylan (“True Historian” 208) and
Stephen Durrant (“Agonistes”) have understood this, and a similarly worded statement in LH 82, to be
Wang Chong‟s complaint about Sima Qian‟s lack of emotion. The problem comes down to what Wang
Chong believed was “in the chest” (胸中). The question is difficult to decide without a thorough study of
early Chinese conceptions of the body, but in context it seems clear that Forke‟s rendering better reflects
Wang Chong‟s priorities. While Wang Chong may admire the pathos of Qu Yuan and Jia Yi, the nature of
his own Lunheng and many of his remarks suggest that what he values most highly is intellectual creativity.
Following this understanding, Wang Chong‟s criticism of Sima Qian, probably heavily tinged with a
deprived bibliophile‟s envy, is that the historian did little but copy the voluminous source materials he had
available to him, and came up with few ideas of his own.
82
LH 39.608, translation adapted from Forke II.297.
83
LH 39.608.
84
Forke II.298, nt. 1.

69
critic‟s labors can the full value of literary works can be revealed and appreciated. No

wonder, then, that the aspects of the Shiji that Wang Chong most appreciated (i.e., most

frequently cited), was the so-called “Honorable Senior Archivist” comments. These

often contain first person evaluations and critical judgments on the material.

Sima Qian—as archivist and compiler—again appears together with Liu Xiang in

“Defining Worthies” 定賢 (Lunheng ch.80). This chapter is a long discussion of how to

determine whether someone is a Worthy (xian 賢). One of the proposed criteria is: “May

those be called Worthies who possess a vast knowledge of things ancient and modern,

and who have memorized all the secret traditions and records?” [以通覽古今，祕隱傳記

無所不記為賢乎？].85 Wang Chong tentatively agrees that such people (whom today

we might call historians or archivists) might be considered worthies, but ranks them

below Ru (who themselves do not receive especially high praise). Why? Those with

historical erudition, Wang Chong considered, are

like heirs specially provided with everything. Being in possession of all the
writings left by generations of forefathers, they are able to complete their chapters
and works. They can peruse and recite [the rare texts they have access to] as
[easily as if they were] bureaucratic documents in their official charge [若專成之
苗裔，有世祖遺文，得成其篇業，觀覽諷誦，若典官文書].86

If the reference to Sima Qian‟s “Self-Narration” were not already clear, 87 Wang Chong

emphasized it by adding that these hereditarily privileged worthies “are like the

Honorable Senior Archivist and Liu Xiang who, being in charge of all the records, have

become famous for their great learning and vast erudition” [若太史公及劉子政之徒，

85
LH 80.143, trans. adapted from Forke II:143.
86
Ibid.
87
As mentioned above, Sima Qian begins his “Self-Narration” with a long exposition of how his ancestors
collected historical materials (SJ 130. 3285-3286), and also mentions looking at historical records after
inheriting his father‟s office (SJ 130.3296).

70
有主領書記之職，則有博覽通達之名矣].88 Wang Chong, whose eidetic memory was

apparently remarkable even in his day, was unimpressed by compilers‟ abilities to

manipulate the written material they already had on hand. As mentioned above, he

preferred the critical evaluative faculty of Huan Tan, whose Xinlun 新論 [New discourses]

he described as the model for his own Lunheng.89

In the “Responses of Objections” 對作 chapter (Lunheng ch.84), Wang Chong

expanded further on these issues. The entire chapter is composed of Wang Chong‟s

attempts to rebut criticisms of the Lunheng, so that it resembles a preface to a second

edition. The section I focus on here concerns the old problem of creation and

transmission:90 “Some say that the sages create, whereas the worthies transmit, and that,

if worthies create, it is wrong. The Lunheng and Zhengwu91 are creations, they think” [或

曰：聖人作，賢者述，以賢而作者，非也。論衡、政務，可謂作者] and thus they

object to them.92 Actually, Wang Chong argued that his own works were “neither

creations nor transmissions” [非作也，亦非述也]. The Five Classics are creations,

according to Wang Chong, while the Shiji, Liu Xiang‟s Xinxu 新序 [New narrations], and

Ban Biao‟s continuation of the Shiji were all merely transmitted. He then proposed a

third category, “Discussions” 論, into which he placed Huan Tan‟s Xinlun, among others.

88
LH 80.143, trans. adapted from Forke II.143.
89
Recall that Huan Tan too, in the passage quoted above, emphasized Sima Qian‟s access to the archives as
having been crucially important to the writing of the Shiji.
90
Kongzi had famously said, “I transmit but do not innovate” 述而不作 (Lunyu VII:1). When Confucius
was elevated to sagely status, the works attributed to him (i.e., the Classics, and especially the Spring and
Autumn Annals) came to be considered as creations. As mentioned above, we can see this in Sima Qian‟s
Shiji “Self-Narration,” where he denied that his book was a “creation.” Scholars today consider this mere
modesty, but Wang Chong took Sima Qian at his word and agreed with him, as we will see.
91
A work on government, now lost, which Wang Chong had written prior to composing the Lunheng.
92
LH 84.1180, translation adapted from Forke I.86.

71
He added that his own two works, the Lunheng and Zhengwu, were reflections of Huan

Tan‟s work—thus also “discussions” and not in fact “creations.”

We might conclude from the tenor of Huan Tan‟s and Wang Chong‟s comments

that the association between the Shiji and philosophical texts which had begun to develop

in Liu Yu‟s and Yang Xiong‟s generations was beginning to erode. The emphasis on

voluminous historical source materials and the description of Sima Qian‟s work as

transmission rather than creation or “discussion” (analysis) shows that the Shiji was

beginning to be read primarily as part of a new historical tradition.

THE NEW HISTORICAL TRADITION

People did not necessarily react to the Shiji as Sima Qian might have wished. Far

from being a unique, quincentenary classic, the Shiji almost immediately began to inspire

both continuations and imitations. This is to say that the Classics, already to some extent

a closed canon by Sima Qian‟s time, were not destined to expand in such a way that they

included the Shiji. Instead, the Shiji‟s most profound influence early on was as a

historical work. In the era directly after its compilation its effect seemed to have been to

convince people that recording history was an important and worthwhile enterprise. And

unlike the Classics, history had the advantage of being an ongoing project.

The Shiji had provided a model for a new way of writing of history, and almost

immediately a number of would-be historians seemed to have imitated it. Fan Ye 范曄

(398-445) wrote that because

after the Taichu period (104-101 BCE), [the Shiji] had lacunae or did not record
anything, later aficionados to some extent gathered up and patched together some
of the events of the time. These, however, were for the most part crude and

72
 mediocre, not worthy successors to [Sima Qian‟s] writings. [自太初以後，闕而
不錄，後好事者頗或綴集時事，然多鄙俗，不足以踵繼其書。]93

Liu Zhiji, writing on the same subject, mentioned by name fifteen such Han

“aficionados” (好事者) who wrote continuations of the Shiji. His list includes such

luminaries as Liu Xiang, Liu Xin, and Yang Xiong, as well as a dozen almost wholly

unfamiliar names.94 Not one of these continuations has survived in independent form.95

It was at this point that Ban Biao turned his hand to the task.

Ban Biao

Ban Biao was from an eminent family: His grandfather Ban Kuang had served as

Colonel of Picked Cavalry in the time of Emperor Cheng, while his father Ban Zhi had

been the Grand Administrator of Guangping during the time of Emperor Ai.96 During the

Xin dynasty he remained loyal to the Han, leaving the service of his employer Wei Ao 隗

囂 because the latter hinted at the possibility that the Han could be replaced.97 Later he

served Dou Rong 竇融, and through him came to the attention of the Guangwu Emperor.

The emperor admired Ban Biao‟s talent and summoned him to an audience, even offering

him a government position as Magistrate of Xu. Ban also served as clerk to the Minister

of Education. His real focus, however, lay in his single-minded concentration on

historical records.98 He was said to have compiled a continuation to the Shiji in “several

93
HHS 40A.1325.
94
STTS 12.338.
95
A trace of one of them remains in the Hanshu “Treatise on the Arts and Literature”: “Feng Shang‟s
continuation of the Honorable Senior Archivist, in seven chapters” [馮商所續太史公七篇] (HS 30.1714).
96
HHS 40A.1324.
97
HHS 40A.1323-24.
98
HHS 40A.1324.

73
dozen chapters” [數十篇];99 these are now presumed to have been absorbed into his son

Ban Gu‟s magnum opus, the Hanshu. Fan Ye‟s biography of Ban Biao has, however,

preserved another text, known as the “Brief Discussion of Previous Histories” [前史略

論], which is useful in understanding how the Shiji was contextualized in the early

Eastern Han. Ban Biao began his discussion with an overview of the entire historical

tradition from high antiquity to his own time:

The Odes and Documents reach back to the time of Tang and Yu and the Three
Dynasties. In that age, there were official archivists who had charge of written
records. When it came to the time of the feudal lords, each fief had its own
archivist. Thus Mengzi says, “The Taowu of Chu, the Sheng of Jin, the Chunqiu
of Lu—they all did the same kind of work.”100 In the time of [the Lu dukes] Ding
and Ai, Zuo Qiuming, a gentleman of Lu, discoursed upon and collected the
writings [of the time] and made the Zuoshi zhuan in 30 chapters. He also
compiled divergent and supplementary [material] and called it the Guoyu, in 21
chapters. From that time, the work of the Sheng and Taowu were obscured, and
the Zuoshi and Guoyu texts alone were in circulation. In addition, there are
records of the emperors and kings, dukes, nobles, ministers and officers from the
time of the Yellow Emperor down to the Spring and Autumn period. These were
called the Shiben, in 15 chapters. After the Spring and Autumn period, the seven
states all contended with one another. When Qin united the feudal lords, then
there was the Zhanguo ce in 33 chapters. When the Han arose and brought order
to all under Heaven, the Senior Palace Grandee Lu Jia made records of the
achievements of the time, and made the Chu Han Chunqiu in 9 chapters. [唐虞三
代，詩書所及，世有史官，以司典籍，暨於諸侯，國自有史，故孟子曰「楚
之檮杌，晉之乘，魯之春秋，其事一也」．定哀之閒，魯君子左丘明論集其
文，作左氏傳三十篇，又撰異同，號曰國語，二十一篇，由是乘、檮杌之事
遂闇，而左氏、國語獨章．又有記錄黃帝以來至春秋時帝王公侯卿大夫，號
曰世本，一十五篇．春秋之後，七國並爭，秦并諸侯，則有戰國策三十三
篇．漢興定天下，太中大夫陸賈記錄時功，作楚漢春秋九篇．]101

99
Ibid. Wang Chong describes the length of the work as “more than a hundred chapters” [班叔皮續太史
公書百篇以上] (LH 39.615), which is intriguing but difficult to judge as to accuracy.
100
Mengzi 8.146 (see Lau, Mencius 4B:21). There is a slight textual variant between this quotation and the
received Mengzi, which has: 晉之乘，楚之檮杌，魯之春秋一也. The two characters 其事 begin the next
sentence, setting up a parallel between “their events” 其事 and “their style” 其文.
101
HHS 40A.1325.

74
Ban Biao connects the work of court archivists to the earliest Classics, the Odes and

Documents. He envisions these Classics as owing their material to such official

„historians.‟ He also mentioned the Spring and Autumn, generally considered a unique

and singular work, in the context of other state annals. Thus Ban Biao presented at least

three of the Classics as arising from scribal/archival endeavors. He brought the record

down all the way to the Han, thus forming the background against which he wanted the

Shiji to be understood: the other works all are predecessors of the Shiji, on which the rest

of the essay is focused.

Rather than quote the entire text, I will give here only Ban Biao‟s list of Sima

Qian‟s sources—what he saw Sima Qian as having taken from the past—and his

comments about them:

In the era of Emperor Wu, the Honorable Senior Archivist Sima Qian selected
from the Zuoshi and the Guoyu, revised the Shiben and the Zhanguo ce, and relied
upon events from the various states of the Chu and Han eras. What Qian records
from the beginning of the Han and ending in the time of Emperor Wu is his
[highest] achievement. When it comes to selecting from the Classics and taking
excerpts from their commentaries, dividing up and promulgating the events
connected with the Hundred Schools, [Sima Qian‟s work] is really very sketchy
and not as good as the texts he bases it on. [孝武之世，太史令司馬遷採左氏、
國語，刪世本、戰國策，據楚、漢列國時事．遷之所記，從漢元至武以絕，
則其功也．至於採經摭傳，分散百家之事，甚多疎略，不如其本].102

As I will describe below, Ban Biao would not be the last to criticize Sima Qian for his

work on the ancient past and his use of the Classics. What is interesting is that Ban Biao

also criticized Sima Qian for his use of Warring States texts, and opined that Sima Qian‟s

work on the Han period is the most valuable part of the text. Implicitly, Ban Biao thus

emphasized the creation of a contemporary record, a focus on times closer to those in

which the historian himself lived. He presented history as a project whose greatest

102
Ibid.

75
significance lay not so much in being a record of the entire known past, but in achieving

ongoing accuracy.

Sima Qian and Ban Gu

Like Yang Xiong, Ban Gu criticized Sima Qian for failing to accord with the Sage

and with moral rightness as Ban Gu saw it. Nonetheless, Ban Gu‟s Hanshu owes much

to the Shiji. As many as 73 chapters overlap significantly between the two works, with

many copied almost word for word, significant alterations being made only in the

evaluations.103

The first thing to note is that the Hanshu “Treatise on the Arts and Literature”

categorized the Shiji under the Spring and Autumn school, a category which also included

the Gongyang, Guliang, and Zuo commentaries, the Zhanguoce, the Guoyu, and so on.104

Ban Gu‟s organizational scheme, probably inherited from the Han bibliographers Liu

Xiang and his son Liu Xin, offered various potential categories: the Classics,105 the major

or not so major philosophical schools,106 and a few miscellaneous genres at the end.107

For him, therefore, the Spring and Autumn category was the category for histories.

103
For discussions of this issue in English, see Lü Zongli “Problems,” 53 and Honey “Textual Criticism,”
69; in Chinese, see Xu Shuofang 徐朔方 Han lungao 史漢論稿 [Preliminary discussion of the Shiji and
Hanshu].
104
HS 30.1713-1714.
105
I.e., the Six Classics (Changes 易, Documents 書, Odes 詩, Rites 禮, Music 樂, and Spring and Autumn
春秋) as well as texts which would eventually become part of the Thirteen Classics canon: Lunyu 論語, the
Classic of Filial Piety 孝經, and “philology” 小學 (which would come to be represented in the canon by
the Erya 爾雅) (HS 30.1701-1724).
106
I.e., the Classicists 儒, Daoists 道, Yinyang 陰陽, Legalists 法, Logicians 名, Mohists 墨, Miscellaneous
thinkers 雜, Tillers 農, and Lesser Talk 小說 (HS 30.1724-1747).
107
I.e., Rhapsodies 賦, Songs and Poems 歌詩, Strategists 權謀, Military Strategists 兵形勢, another
section of Yinyang 陰陽, Military Technicians 兵技巧, Astronomers 天文, Calendrists 曆譜, Five Phases
五行, Diviners by Milfoil and Turtle 蓍龜, Miscellaneous Diviners 雜占, Geographers 形法, Numerists 數
術, Medical Classics 醫經, Classical Techniques 經方, Inner Room 房中, and Spirits and Immortals 神僊
(HS 30.1746-1780). Note that translations of all these terms are approximate and provided merely for the

76
 But what of Sima Qian himself? The Hanshu contains scattered mentions of him

which refer to his literary talent.108 In addition, there are two places in the Hanshu where

Ban specifically evaluates Sima Qian. One occurs in the “table of contents” section of

Ban Gu‟s “Self-Narration,” where each chapter receives a „small preface‟, as it were.

The preface for the Sima Qian chapter is as follows:

Alas Archivist Qian, punished because entangled in another‟s crime!109

Imprisoned he poured forth his resentment. With your intellect and your energy,
you wove together a flock of words. Antiquity and today here woven together,
completes a single school: in its essentials it is most illuminating. [烏呼史遷，薰
胥以刑！幽而發憤，乃思乃精，錯綜羣言，古今是經，勒成一家，大略孔
明．]110

The references to Sima Qian‟s misfortune will be discussed in more detail in Part III.

Here I merely point out that this Hanshu evaluation does implicitly place Sima Qian

among those who were inspired by misfortune to create great works—a category Sima

Qian himself proposed. Ban Gu‟s language also emphasized the monumental scope of

the Shiji (the “flock of words” and the stretch from “antiquity to today”), and recognizes

Sima Qian‟s claim to have established a new school of thought. He even calls Sima

Qian‟s work a classic (jing 經), though this term was certainly not restricted to the

Confucian canon. If we take this “small preface” at face value, we might conclude that

reader‟s convenience. For a detailed examination of this “Treatise” see Chen Guoqing‟s 陳國慶 Hanshu
Yiwen zhi zhu shi huibian.
108
See chapter 4 below. Note that the Hanshu “Yiwenzhi” also lists eight fu 賦 (rhapsodies) supposedly
from Sima Qian‟s hand (HS 30.1749).
109
This is a reference to a Shijing poem, “Yu wu zheng” 雨無正 (Mao #194), which laments the terrible
condition of government in the waning of the Zhou. The specific line alluded to is: “As for those who have
no crime, they too are entangled and ruined through injustice” [若此無罪，淪胥以鋪].
110
HS 100B.4257.

77
Ban Gu had quite a positive view of Sima Qian. However, as later discussion will show,

Ban Gu‟s approval was hardly unconditional.111

Sima Qian versus Ban Gu

After the appearance of the Hanshu, Ban Gu was Sima Qian‟s most frequent point

of comparison. Numerous scholars engaged in “Ban/Ma Contrast” (班馬異同),112 and it

is regarded as its own field of historiographical studies.113 A full treatment of this subject

is beyond the scope of my study, but given its importance to the history of the Shiji, it

seems essential to mention briefly some of the terms in which the two authors were

contrasted.

Early readers contrasted Ban Gu and Sima Qian according to their moral qualities.

Fu Xuan 傅玄 (217-278), for example, pointed out that Sima Qian‟s acknowledgement of

Sima Tan‟s contribution to the Shiji was quite unlike Ban Gu‟s failure to make any

similar acknowledgement:

Ban Gu depended on his father[‟s work] in order to be able to complete the

Hanshu, yet he suppressed this fact and did not speak of [Ban] Biao. [In this] he
was different indeed from [Si]ma Qian. [班固《漢書》，因父得成，遂沒不言
彪，殊異馬遷也。]114

111
The evaluation (zan 贊) included after the Sima Qian chapter of the Hanshu closely echoes Ban Biao‟s
“Brief Discussion of Previous Histories,” discussed above, though with minor changes.
112
Though the phrase 班馬異同 literally means “commonalities and contrasts between Ban Gu and Sima
Qian,” my translation emphasizes the element of contrast, because writers who address this topic rarely if
ever emphasize what the two authors have in common.
113
See “Ma Ban yitong cheng yi men xuewen” [馬班異同成一門學問] in SJYJJC 13.122-131 for an
excellent introduction to this subject. In the following discussion, I am indebted to the authors of that
article but differ significantly from them in my purpose and in many of my analyses as well.
114
This quotation, which is only preserved in fragmentary form in Ma Zong‟s Yilin, was pieced together by
Yan Kejun, Quan Jin wen, 526.

78
Fan Ye‟s Hou Hanshu 後漢書 evaluation of Ban Gu went farther, accusing Ban Gu of

hypocrisy: though Ban Gu criticized Sima Qian for foolishly falling into misfortune, he

himself eventually did the same—or worse in fact.115

Most of the field of “Ban/Ma Contrast,” however, is concerned with qualities of

literary style and editorial judgment. The comment that inaugurated the field is said to be

that of Wang Chong, who wrote of Ban Biao‟s Houzhuan that “its readers were of the

opinion that it was even superior to the Honorable Senior Archivist” [觀讀之者以為甲，

而太史公乙].116 Zhang Fu 張輔 (d. ca. 306) put forth a more substantive argument. He

wrote:

In what Qian wrote and transmitted, the words are brief and events
complete. He narrated three thousand years worth of events in only five hundred
thousand words. Ban Gu, on the other hand, narrated two hundred years worth of
events in eight hundred thousand words. The matter of prolixity and brevity is the
first point on which [Ban Gu] is not as good as [Sima] Qian. [遷之著述，辭約而
事舉，叙三千年事唯五十萬言；班固叙二百年事乃八十萬言，煩省不同，不
如遷一也。]
When a good archivist transmits events, what is good in them should be
enough to make one feel encouraged; what is evil should be enough to be make
one feel alert and forewarned; this is the constant Way of man. As for middling
and petty matters, he does not select them for inclusion. Yet Ban [Gu] wrote
everything down, and this is the second point on which he is not as good as [Sima
Qian]. [良史述事，善足以奬勸，惡足以監誡，人道之常。中流小事，亦無
取焉，而班皆書之，不如二也。]
The fact that [Ban Gu] defames and criticizes Chao Cuo, thereby doing
harm to the Way of the loyal minister, is the third point on which he is not as
good [as Sima Qian]. What Qian had already created, Gu merely followed along
in. The difficulty [of the one] and the ease [of the other] are very different. [毀貶
晁錯，傷忠臣之道，不如三也．遷既造創，固又因循，難易益不同矣．]117

115
HHS 40B.1386. For further discussion of this passage, see chapter 6 below.
116
LH 39.615, trans. adapted from Forke II.304.
117
JS 60.1640.

79
Though this comment appears to be a ringing endorsement of the Shiji, we cannot

therefore conclude that Sima Qian enjoyed ascendance over Ban Gu in Zhang Fu‟s time,

since the context of this remark is apparently an exposition of Zhang Fu‟s somewhat

eccentric opinions. The Jinshu makes a point of noting that Zhang Fu also thought that

“Guan Zhong was not as good as Bao Shu” [管仲不若鮑叔], that “Cao Cao was not as

good as Liu Bei” [魏武帝不及劉備], that “Yue Yi was inferior to Zhuge Liang” [樂毅減

於諸葛亮],118 and so on. Though most of these opinions might seem quite mainstream

today, this was not the case in Zhang Fu‟s time.

Some later scholars disagreed with Zhang Fu about using the criterion of brevity

to judge which text was superior. Liu Zhiji, for example, pointed out that for most of the

vast span of time covered by the Shiji, “the traces of events were brief and fragmentary”

[事跡殊略], so that “although Qian narrates three thousand years worth of events, it is

only the seventy or more years after the rise of the Han for which [the record] is detailed

and complete” [遷雖叙三千年事，其間詳備者，唯漢興七十餘載而已].119 As for

Zhang Fu‟s second point, Liu Zhiji commented that “Master Ban‟s Hanshu includes all of

the [Han portions of the] Shiji, yet he removed chapters like the „Diviners by Days‟120

and „Cang Gong,‟121 considering the events therein to be superfluous and verbose, not

118
Ibid.
119
STTS 16.473.
120
“Rizhe liezhuan” 日者列傳, Shiji ch.127. It is possible that this chapter may not have been extant in
Ban Gu‟s day. See discussion in Part III below.
121
“Bian Que Cang Gong liezhuan” 扁鵲倉公列傳, Shiji ch.105.

80
worthy of being part of the compilation” [班氏漢書全取史記，仍去其日者、倉公等

傳，以為其事煩蕪，不足編次故也].122

Liu Zhiji stated that in his opinion, though other writers tended to choose sides

between the Shiji and Hanshu, there was really no reason to do so. “These two works,

though each has its strong points and shortcomings, its achievements and failings in the

information it transmits, they are nonetheless in the same style, and could be considered

to be connected in type” [然此二書，雖互有修短，遞聞得失，而大抵同風，可為連

類].123

Sima Zhen, a contemporary of Liu Zhiji and major Shiji commentator, placed

more emphasis on the distinction between the two texts. He wrote:

Compared to Ban [Gu‟s Han]shu, [the Shiji] was subtle and had the ancient
[virtue of] substantiality.124 Thus the famous worthies of the Han and Jin did not
yet know to value it. But its goodness is like when Marquis Wen of Wei heard
old music and only feared he would doze off. [比於班書，微為古質，故漢晉名
賢未知見重，所以魏文侯聽古樂則唯恐臥，良有以也。]125

As regards the differences he saw between the Shiji and the Hanshu, Sima Zhen gave a

reasonably explicit analysis. He wrote:

Now, the Honorable Senior Archivist recorded affairs beginning from the
Yellow Emperor and coming down to the Tianhan period [100-97 BCE]. Though
he selected broadly from ancient writings, and transmitted and recorded the
various masters, within his text there is probably much that is defective and

122
Ibid. The Jin (金) dynasty critic Wang Ruoxu would later vehemently affirm Liu Zhiji‟s critique of
Zhang Fu, adding that
[Sima] Qian‟s record of events is sketchy and incomplete, yet there is a great deal of excess
verbiage. [Ban] Gu‟s recording of events is detailed and complete, and his excisions bring out
what is essential and appropriate. This is why Qian seems concise but is in fact excessive, while
Gu seems excessive but is in fact concise. [遷記事踈略而剰語甚多，固記事詳偹而刪削精當，
然則遷似簡而實繁，固似繁而實簡也，安得以是為優劣哉！] (Shiji bianhuo 15.98)
123
STTS 7.204.
124
A reference to the saying in Lunyu VI:18, which emphasizes the need for a balance between substance
(zhi 質) and refinement (wen 文).
125
“Preface to the Explication of the Hidden in the Shiji” 史記索隱序, 7.

81
 fragmentary. Sometimes he brings in strange reports to complete his accounts.
As a person, he is fond of the curious, but his phrases are concise; his accounts are
penetrating but his writing subtle. For this reason, there is much [in the Shiji] that
later scholars have not investigated completely. [夫太史公紀事，上始軒轅，下
訖天漢，雖博采古文及傳記諸子，其閒殘闕蓋多，或旁搜異聞以成其說，然
其人好奇而詞省，故事覈而文微，是以後之學者多所未究。]
As for Master Ban‟s text, it was completed in the latter Han. As [Ban]
Biao succeeded [Sima] Qian in his account, the style he employs is more lucid. In
this, he broadly selects a great number of worthies, and the various principles
completely fulfilled. Thus his meaning is abundant, his phrases patterned, and
this is why the various scholars of recent times have all come together in his
praise. [其班氏之書，成於後漢．彪既後遷而述，所以條流更明，是兼采眾
賢，群理畢備，故其旨富，其詞文，是以近代諸儒共行。]126

Sima Zhen‟s observation here is particularly interesting because of its focus on the

difficulty of even just reading the Shiji (as opposed to the relative ease of reading the

Hanshu). The reasons for this difficulty are three: first that Sima Qian‟s source materials

are themselves quite heterogeneous and antique in their language; second, the text of the

Shiji itself had suffered in transmission; and third, Sima Qian‟s predilection for subtle

writing meant that his intention was difficult to grasp fully. By contrast, the Hanshu‟s

subject matter was far more recent, its sources more homogeneous and reliable, and both

its style and message more coherent. If one chooses to see the field of Ban/Ma contrast

as a contest, most readers up to Sima Zhen‟s time had considered Ban Gu to be the

winner.

Of course to view it as a contest is less useful in the present context. A reader‟s

judgment between the two texts reveals as much about that reader as about the works in

question. To put it another way, in defining the differences between the two historians,

readers were actually displaying their own values. An interesting comment attributed to

Cheng Yi 程頤 (1033-1107) highlights this:

126
Ibid.

82
 The subtle nature and marvelous meaning of Zichang‟s [=Sima Qian] writings are
lodged outside the path of the written characters. In Mengjian‟s [=Ban Gu] prose,
the nature and meaning are entirely revealed along the path of the written
characters. When one reads the writing of Zichang, only those who insist on
going beyond what is being said can begin to get his meaning, and only those who
transcend the written characters can explain the fundamental principles. Master
Ban‟s literary composition can also be praised for its breadth and elegance, but if
one reads it more than once, its nature and phrasing are all used up. Zhang Fu
used the number of written characters to determine precedence, but how is this
sufficient for discussion Ban and Ma?! [子長著作，微情妙旨寄之文字蹊徑之
外。孟堅之文，情旨盡露于文字蹊徑之中。讀子長文必越浮言者始得其意，
超文字者乃解其宗。班氏文章亦稱博雅，但一覽之餘，情詞俱盡。張輔以文
字多寡為優劣。此何足以論班馬哉？！]127

Cheng Yi‟s judgment is a more elaborate version of Sima Zhen‟s, that the Shiji is

somehow more subtle and profound than the Hanshu, requiring more study but

potentially yielding greater rewards. As a philosopher of considerable sophistication,

Cheng Yi clearly appreciates the kind of text which rewards multiple readings and long

study. But Zhu Xi, a great admirer of the Cheng brothers, held quite a different opinion,

arguing that “the writings of the Honorable Senior Archivist are loose and

straightforward, while the writings of Ban Gu are tightly-constructed and recondite” [太

史公書疏爽，班固書密塞].128 The difference between Cheng‟s and Zhu‟s appraisals

suggests that what one considers praiseworthy depends on the kind of meaning one is

looking for. To put it in Cheng Yi‟s terms, Zhu Xi was not interested in “leaving the

path” of Sima Qian‟s prose to find the meaning allegedly lodged beyond it. He preferred

127
Cited in Jiao Hong‟s 焦竑 (1540-1620) Jiaoshi bi cheng 焦氏笔乘 2.37.
128
ZZYL 134.3202. As in many such concise and elliptical characterizations, the exact meaning and value
of the adjectival terms is open to question. This quotation seems to be the locus classicus for both
shushuang 疏爽 and misai 密塞. The fact that the former was later understood as positive (“energetic”)
and the latter potentially negative (“involved, abstruse”) has more to do with later relative valuations of the
Shiji and Hanshu that it does with the meaning of the words themselves. I have attempted to translate the
words in a way that reflects my overall understanding of Zhu Xi‟s opinion of the two texts, but readers
should certainly weigh the original characters for themselves.

83
the tightness of Ban Gu‟s prose to what he perceived as a careless looseness—potentially

a sense of underdetermination?—in Sima Qian‟s.

The Southern Song historian Zheng Qiao had different criteria. The “General

Preface” 總序 of his Tongzhi 通志 [Comprehensive treatise] juxtaposes praise for Sima

Qian‟s achievements with a bitter complaint against Ban Gu:

After the Chunqiu, only the Shiji displayed skill in the full scope of
literary creation. Unfortunately Ban Gu attacked [Sima Qian‟s] character, and
thus came to no thorough understanding of [Sima Qian‟s] meaning. After that,
the followers of Master Sima declined. [自春秋之後，惟史記擅制作之規模。
不幸班固非其人，遂失會通之旨，司馬氏之門户自此衰矣。]
Ban Gu was a puffed up and flowery sort of person, completely without
scholarly learning.129 His chief expertise was plagiarism. When Su Zong130
asked [Ban Gu] about the matters of creating ritual and composing music, Gu
responded that among the various classics masters in the capital there surely were
some who could understand these [matters]. Supposing that all ministers were of
his ilk, then how could one get advice from them? When it came about that the
various classics masters each set forth their [contributions], [Ban] Gu did no more
than steal Shusun Tong‟s “Ceremonies” in twelve pian131 in order to just knock
something together and get by. Supposing that all ministers were of his ilk, then
what could one gain from submitting memorials and disputations [to the throne]?
Su Zong knew how shallow and ignorant [Ban Gu] was, and therefore said to Dou
Xian (d.92): “You being fond of Ban Gu but disregarding Cui Yin132—it‟s just
like She Gong‟s fondness for dragons.”133 In his own time, Ban Gu‟s value had
already been determined, but what sort of compilation could a person of this sort
create? [班固者浮華之士也，全無學術，專事剽竊。肅宗問以制禮作樂之
事，固對以在京諸儒必能知之。儻臣鄰皆如此，則顧問何取焉。及諸儒各有

129
Those familiar with Ban Gu‟s work and reputation will surely take issue with this latter characterization.
However, in the discussion that follows I am attempting to give Zheng Qiao the most charitable possible
reading, that he is polemically proposing an alterate understanding of what it means to be learned.
130
I.e., Han Emperor Zhang (r.76-88). He is sometimes known as Han Su Zong Xiao Zhang Huangdi 漢肅
宗孝章皇帝, the Filial August Emperor Zhang, Solemn Ancestor of the Han.
131
Namely, the “Han Yi” 漢儀 [Ceremonies of the Han]
132
Cui Yin 崔駰 (d.92) was an Eastern Han figure famed for his literary skill. Later, Dou Xian did in fact
begin to associate with Cui Yin. See HHS 52.1718-1719.
133
The anecdote is found in HHS 52.1718-1719. The saying about She Gong‟s fondness for dragons can be
found in Liu Xiang‟s Xinxu 新序 5.190. Briefly, Duke Ai of Lu was famous for valuing talented retainers.
He had an audience with Zizhang, but did not treat him with proper courtesy. Zizhang complained of his
treatment by likening Duke Ai to She Gong, who was so fond of dragons that he decorated his entire hall
with them. But when a real dragon heard of his fondness and paid him a visit, She Gong was terrified. The
proverb is commonly used to describe superficial connoisseurship, a fondness for the semblance rather than
the reality.

84
 所陳，固惟竊叔孫通十二篇之儀以塞白而已。儻臣鄰皆如此，則奏議何取
焉。肅宗知其淺陋，故語竇憲曰：“公愛班固而忽崔駰，此葉公之好龍也。”
固於當時已有定價，如此人材將何著述？]134

Zheng Qiao‟s diatribe against Ban Gu makes it clear that he does not object to ad

hominem attacks in general, only to the terms of Ban Gu‟s attack against Sima Qian.

What is especially curious about Zheng Qiao‟s critique is his repeated emphasis

on Ban Gu‟s lack of originality, his “plagiarism” (as I have translated piaoqie 剽竊).

However, Zheng Qiao could not have meant plagiarism according to the way we

understand it, because earlier in the passage, he praised the Simas‟ ability to “unite the

words of the Odes, Documents, Zuozhuan, Guoyu, Shiben, Zhanguoce, and Chu Han

Chunqiu” [會詩、書、左傳、國語、世本、戰國策、楚漢春秋之言].135 Furthermore,

there were several points on which Zheng Qiao criticized Sima Qian (albeit gently), and

his critical discussion includes the observation that

[Sima] Qian‟s writings entirely employ writings of the past, but mix in common
expressions. His good point is that he could not set down his brush and scraper
while his gathering up [of materials] was not yet complete. Thus he says, “I do
not dare let the words of my ancestor(s) be lost, and thus I have transmitted events
of the past, and set in order their traditions. This is not what can be called
„creation.‟”136 [遷書全用舊文，閒以俚語，良由採摭未備，筆削不遑。故
曰：予不敢墮先人之言，乃述故事，整齊其傳，非所謂作也。]137

We might expect Zheng‟s conclusion to be that Sima Qian too should be more original,

instead of “entirely employing writings of the past.” Instead, it is the “mixing in of

134
Tongzhi 通志, “General Preface” 總序 2.
135
Admittedly, hui 會 could also be translated as “accord with” or even “have a thorough understanding
of.” Since the texts Zheng Qiao lists are traditionally understood as being Sima Qian‟s sources, though, it
makes sense to translate the character as something like, “bring together.”
136
Note that this is a paraphrase or approximate quotation from the “Self-Narration”, of which the current
standard version reads “there is no greater sin than to letting the words of my ancestor be lost. What I
mean by transmitting past affairs and setting in order their genealogies and traditions—this is not what can
be called creating. It is mistaken indeed for you to compare it to the Chunqiu” [墮先人所言，罪莫大焉．
余所謂述故事，整齊其世傳，非所謂作也，而君比之於春秋，謬矣] (SJ 130.3299-3300).
137
Tongzhi 通志, “General Preface” 總序, 2.

85
common expressions”—which presumably are Sima Qian‟s original contribution—which

become the target of Zheng Qiao‟s critique: “This is what might be called [another]

regrettable [characteristic of Sima] Qian, that he was not sufficiently orthodox” [所可為

遷恨者，雅不足也].138 It should also be noted that Zheng Qiao‟s own work, the

Tongzhi, consists of 200 chapters, the majority of which are completely derived from

earlier histories—only the twenty treatises (lüe 略) contain sufficiently original material

to be of interest to scholars today.

We must conclude from this either that Zheng Qiao is profoundly hypocritical or

that he is not accusing Ban Gu merely of lacking originality on the level of individual

passages or use of sources. If we give Zheng the benefit of the doubt and assume the

latter, it is still not easy to understand just what his accusation is. He hints at it in saying

that “All those who compose texts, though they select from the writings of their

predecessors, must themselves „establish the words of a single school‟” [凡著書者，雖

採前人之書，必自成一家言]. Still the crucial phrase, zicheng yijia 自成一家, is

notoriously ambiguous and has been since Sima Qian first coined it. The translation of

“school [of thought],” always a slightly uncomfortable one, is even less apt here than

elsewhere. Must everyone who writes a book found his own scholarly lineage?

The end of Zheng Qiao‟s diatribe against Ban Gu may give the best hope of

resolving these questions:

The greatest achievement of the Shiji lies in its ten tables, which are [to
the rest] as an official hat is to a set of clothes, as a root is to a tree, or as a
wellspring is to water. Ban Gu did not comprehend the “rows and columns,”139
138
Ya 雅 carries a sense of both elegance and orthodoxy, but this is difficult to convey in a single English
adjective.
139
This is to say, the principles behind the use of table form—what the rows and columns should contain
and why. The first use of this phrase, also known in the variant form 旁行斜上, is supposedly by Huan Tan

86
 but took people from antiquity to his own day and set up forced rankings and
classifications for them. Furthermore, he claimed that Han was a continuation of
Yao‟s line,140 and thus should follow Yao, criticizing the fact that when Qian
made the Shiji, [the Han] was placed next to the Qin and Xiang Yu. This is
groundless talk. Ever since [Ban Gu] limited his work to just the Han dynasty,
[the depiction of the Han] became unconnected from the Zhou and Qin,141
creating a gap between antiquity and the present. From Gaozu to Emperor Wu,
the first six generations in total, [Ban Gu] stole it all from [Sima] Qian‟s writings
without feeling any shame. From Emperor Zhao down to Emperor Ping, another
six generations in total, he got his material from Jia Kui and Liu Xin, and again
did not consider it a disgrace. Furthermore, there are also the finishing chapters
by Cao Dagu,142 so that what Ban Gu himself wrote was very little, and what in
the end came out of Ban Gu‟s own mind was no more than the “Table of People
in Antiquity and Today.”143 Other people have not [followed him in] that
absurdity. [史記一書，功在十表，猶衣裳之有冠冕，木水之有本原。班固不
通旁行邪上，以古今人物彊立差等，且謂漢紹堯運，自當繼堯，非遷作史記
厠於秦、項，此則無稽之談也。由其斷漢為書，是致周秦不相因，古今成間
隔。自髙祖至武帝，凡六世之前盡竊遷書，不以為慚。自昭帝至平帝凡六
世，資於賈逵劉歆，復不以為恥。况又有曹大家終篇，則固之自為書也幾
希，往往出固之胷中者，古今人表耳。他人無此謬也。]
In later generations many hands are involved in the writing of a work, so it
is like “depending on travelers for advice on building a house,”144 and they
plunder people‟s writings like one who “stops up his ears while stealing a bell.”145
It is all in imitation of what Gu started. If Gu‟s work was like this, and later
historians hurried [to imitate] Ban Gu‟s slapdash ways, is it even possible to

in the Xinlun, discussing the tables in the Shiji and the way they imitated the Zhou geneaological charts.
See Nanshi 南史 49.1223 and discussion in chapter 2 below.
140
For an excellent recent discussion of this intellectual phenomenon in the Xin and Eastern Han, see
Gopal Sukhu‟s “Yao, Shun, and Prefiguration,” Early China 30 (2005-2006): 91-153.
141
This phrase can and probably should be understood in two different senses. One is as a reference to
efforts to delegitimate the Qin and remove it from the cycle of dynasties, which Zheng Qiao also mentions
below. Another is a more general lack of historiographical continuity from one dynasty to the next. This
was a major complaint that Song dynasty writers of comprehensive histories (tongshi 通史) leveled against
single dynasty historiography. See de Crespigny, “Universal Histories” in Leslie et al. ed, Essays on the
Sources for Chinese History, 64-70.
142
Cao Dagu is Ban Gu‟s sister Ban Zhao. She married into the Cao family but was widowed early on.
Later she was taken into the imperial family as a tutor for the empress and other ladies. There she received
the sobriquet “Dagu.” She was said to have been involved in final stages of the Hanshu‟s compilation.
143
Hanshu ch.20. This peculiar table attempted a ranking of historical and contemporary figures according
to their moral qualities, and certainly represented a different interpretation from Sima Qian‟s of what a
table‟s function should be.
144
The origin of this expression is the Shijing poem, “Xiao Min” 小旻 (Mao #195): “They are like one
taking counsel with wayfarers about building a house/Which will consequently never come to completion”
[如彼築室于道謀，是用不潰于成] (trans. Legge, IV.332). Zheng Xuan explained the problem as being
the wayfarers‟ diversity of ideas (Mao Shi zhengyi 12.413).
145
An expression originating in the Lüshi chunqiu, “Zizhi”自知 chapter (LSCQ 24.1601). It is used to
describe a situation in which someone‟s attempt at deception is so obvious that he himself is the only one
who believes in it.

87
 evaluate their depth!? Qian is to Gu as a dragon to a pig. So why is it that the
various histories discard Qian and make use of Gu? [後世衆手修書，道傍築
室，掠人之文，竊鍾掩耳，皆固之作俑也。固之事業如此，後來史家奔走班
固之不暇，何能測其淺深？遷之於固如龍之於猪。奈何諸史棄遷而用
固？]146

To give Zheng Qiao the furthest possible benefit of the doubt, we must understand what

he means by “coming out of [one‟s] own mind.” As Zheng Qiao was surely aware, Wang

Chong had used this same phrase in complaining about Sima Qian’s lack of creativity.

But Zheng Qiao‟s standards and priorities differ from Wang Chong‟s.147

Zheng Qiao‟s concept of history placed a heavy emphasis on overall perspective,

the idea expressed by the word tong 通, a multivalent concept which includes

connotations of both “understanding” (thorough comprehension), and “connection” (of

the past with the present). The dragon-like quality he admired in Sima Qian was his

ability to bring the distant past and the very recent past all together in an unbroken net,

the ideal form of this endeavor being (from Zheng Qiao‟s point of view) the Shiji‟s ten

tables.

Ban Gu in every way failed to realize Zheng Qiao‟s ideal. His decision to limit

the chronological range of his history to the former Han was politically charged. The

atmosphere of Guangwu‟s court was one in which apocryphal texts of self-glorifying

“prophecy” were given the full weight of imperial patronage and interest. Even

Confucius was pressed into service as a sort of prophet, foretelling and preparing the way

for the Heaven-supported power of the Han.148 Part of this cosmological move involved

the disinheriting of the Qin as a legitimate unifying dynasty, hence the criticism of Sima

146
Tongzhi 通志, “General Preface” 總序 2-3.
147
Of course Wang Chong did not comment on the Hanshu, which was not completed until after Wang‟s
lifetime.
148
See Itano Chōhachi‟s “The t’u-ch’en Prophetic Books and the Establishment of Confucianism” (47-111).

88
Qian to which Zheng alluded. Part of the ideal of tongshi 通史 was to employ the

sobering perspective of all history as a cure for such intoxicating delusions as were

fostered by the „apocrypha‟ texts. To Sima Qian, and to Zheng Qiao, the lessons of the

Warring States and Qin periods were not to be forgotten.

Zheng Qiao‟s other complaint against Ban Gu related to the way he used the work

of his predecessors and contemporaries. Yet, as I emphasized above, Zheng Qiao was

not advocating against the use—even the extensive use—of sources. Concretely, Zheng

Qiao censured Ban Gu on two main points: the fact that he got the first half of his

material from Sima Qian, and the fact that he (purportedly) got the second half of it from

Liu Xin and Jia Kui. The two objections are perhaps different. In the first case, the

disgrace may have been “stealing” from Sima Qian for purposes of self-aggrandizement

vis-à-vis the Shiji. Ban Gu attacked his predecessor to discredit him on moral grounds,

and then superseded his work by incorporating the “best”149 part into his own book. In

doing so, of course, he was also replacing a comprehensive history (tongshi 通史) with a

single dynasty one (duandai shi 短代史), and so his readers suffered the concomitant

lack of perspective.

As for Ban Gu‟s getting the latter part of his work from Liu Xin 劉歆 (ca.53

BCE-23 CE) and Jia Kui 賈逵 (30-101), this charge is more difficult to substantiate.

Certainly Ban Gu‟s tremendously important “Treatise on the Arts and Literature” 藝文志

(Hanshu ch.30) was by his own admission developed from Liu Xin‟s “Seven Summaries”

七略, which in turn was based on Liu Xiang‟s 劉向 “Listings Arranged by Category” 別

錄. Of course no one insists that a bibliographic catalogue avoid building on the work of

149
At least according to Ban Biao—see foregoing discussion of HHS 40A.1325.

89
its predecessors, nor does Ban Gu fail to give credit to his sources in that case. As for

other borrowings from Liu Xin or Ban Gu‟s own contemporary Jia Kui, it is difficult to

identify them. Let us assume for a moment, however, that the charge is true. Perhaps the

most plausible analysis of the fault involved is that Ban Gu presumably had access to the

same materials as Liu Xin or Jia Kui. That he would use the work of his contemporaries

rather than doing his own work would then seem to confirm that “his chief expertise was

plagiarism.”

To conclude, what Zheng Qiao valued about Sima Qian was probably the idea of

one man being able to grasp the whole of history, and make use of the insight that such a

perspective would give him. Zheng Qiao does not see Ban Gu as bringing together

materials in the same way so as to gain perspective on the former Han (which is a

perfectly plausible alternative interpretation of Ban Gu‟s project). Perhaps he thought

Ban Gu failed to gain insight because he “stole” (qie 竊)—rather than “employed”—

sources as Sima Qian supposedly did. Though it is not easy to clarify the difference, we

can at least be sure it meant something important to Zheng Qiao in relation to the

historian‟s task.

“Ban-Ma Contrast” as an area of inquiry would continue to play an important role

in the historical criticism of the late imperial period.150 Indeed, some interesting

historiographical thinking was done in that context. At this point, however, I will step

back and consider how the Shiji fits into the larger view of the Chinese historical tradition

after Ban Gu.

150
For further discussion, see SJYJJC, vol.13, 122-131.

90
Continuing the Historical Tradition

In addition to the compilation of continuations, a major reaction to the Shiji was

an attempt to correct it or somehow reconcile it with the Classics. In particular, Qiao

Zhou 譙周 (ca. 200-270) is believed to have compiled his Gushi kao 古史考

[Investigations of ancient history] with such a purpose in mind. That work exists only in

fragments and the preface has not survived. Still, the Jinshu does record that

Qiao Zhou considered that when Sima Qian‟s Shiji wrote of the period of Zhou,
Qin, and before, he at some points adopted vulgar sayings and words of the
hundred schools, and did not exclusively rely on the correct Classics for his
evidence. This is why [Qiao] Zhou made the Gushi kao in twenty-five pian,
relying on the old Classics throughout, in order to rectify [Sima] Qian‟s errors.
[譙周以司馬遷史記書周秦以上，或採俗語百家之言，不專據正經，周於是
作古史考二十五篇，皆憑舊典，以糾遷之謬誤．]151

The passage occurs in the context of the “Traditions of Sima Biao,” who also “in turn

considered that [Qiao] Zhou‟s [work] was not quite perfect. He itemized 122 incorrect

points from the Gushi kao, drawing much of his evidence from the Jizhong chronicle”

[彪復以周為未盡善也，條古史考中凡百二十二事為不當，多據汲冢紀年].152

Clearly, “correcting” the Shiji was an ongoing endeavor, one in which the three major

Shiji commentators (Pei Yin, Sima Zhen, and Zhang Shoujie) also of course participated.

151
JS 82.2141. For an alternate translation of this passage, as well as a similar one by Liu Zhiji, see J.
Michael Farmer‟s The Talent of Shu, 100. Farmer‟s excellent and detailed discussion of the Qiao Zhou‟s
views on the Shiji in chapter 5 of that work, “Critical Approaches to Ancient History” (95-119) obviates the
need for extensive discussion here.
152
JS 82.2141. The Jizhong chronicle refers to the Zhushu jinian 竹書紀年 (see Farmer, Talent, 202 nt.40).

91
 Another interesting response to the Shiji was further compilations of “continuous

histories” 通史, which followed the Shiji model of extending beyond the boundaries of a

single dynasty. Liu Zhiji describes one such effort, though in quite critical terms:

Emperor Wu of the Liang also commissioned his various ministers to [make a

history beginning] at the Taichu [reign period of the Han] and ending with the Qi
dynasty. They completed a compilation called the Tongshi in 620 fascicles.153 In
their writings about the Qin and before, they in every case used the Shiji as a basis,
selecting from other teachings in order to broaden their stock of divergent
explanations. Coming to the Western and Eastern Han, they recorded in entirety
the annals and traditions of that time; yet when earlier and later [works] are [the
same and] both in circulation at the same time, it is like one stink added to
another…. On the whole, its [sub-sections] genres are all very much like those of
the Shiji; the only way in which it differs is just that it lacks tables, nothing more.
[至梁武帝，又敕其羣臣，上自太初，下終齊室，撰成通史六百二十卷。其
書自秦以上，皆以史記為本，而别採他説，以廣異聞；至兩漢已還，則全録
當時紀傳，而上下通逹，臭味相依。。。大抵其體皆如史記，其所為異者，
唯無表而已。]154

Continuous histories were unwieldy, but interest in them never wholly ceased, and

enjoyed a great revival of interest in the Song. Sima Guang‟s Zizhi tongjian 資治通鑒 of

course adopted the chronicle (編年) form, but Zheng Qiao‟s Tongzhi 通志, as discussed

above, was a close formal imitation of the Shiji.

The Shiji‟s primary place, however, was as one of the standard histories. Its

canonization as the first of the Three Histories 三史155 had already been accomplished by

153
As Shitong commentator Pu Qilong notes, Liu Zhiji has two sources for this description. The first is the
Liangshu “Traditions of Wu Jun” 吴均傳：“[Jun] was relieved of office. [Later], there was a edict
summoning to compile a Tongshi, beginning from the Three Thearchs and going down to the Qi dynasty.
Jun had already drafted the basic annals and hereditary households, [but] had not yet gotten to the arrayed
traditions when he died” [均免職。尋召撰《通史》，起三皇，迄齊代，均草本紀、世家功畢，列傳
未就，卒] (Slightly abbreviated version of Liangshu 49.699). The other source is the “Annals of Emperor
Wu” 武帝紀 in the same work, which reads: In the third year of Taiqing, the Tongshi was completed, and
he personally authored the evaluations and the preface. [The work] totaled six hundred chapters. [Emperor
Wu‟s] inborn nature was insightful and perceptive, and when[ever] he wielded his brush he wrote very
well” [太清三年，《通史》成，躬製賛序，凡六百卷。天情睿敏，下筆成章] (Liangshu 3.96).
154
STTS 1.18.
155
The third history being Fan Ye‟s Hou Hanshu.

92
Liu Zhiji‟s time. His appraisal of the place of the Three Histories in relation to the Five

Classics is interesting: the Shiji and Hanshu, he wrote, were “created as continuations of

the sages,156 but inferior to them. Thus the scholars of the world all speak first of the

Five Classics and then of the Three Histories” [繼聖而作，抑其次也。故世之學者皆先

曰五經，次云三史].157

Liu Zhiji was a man whose life was very much devoted to history. While faithful

to the Classics, he fully appreciated the need for records going beyond them. He had a

tremendous admiration for the sagely histories (the Documents and Spring and Autumn)

but recognized the need for a separate category for ongoing historical endeavors. As he

put it:

The Classics are like the sun, the histories like the stars. Now, when the bright
sun‟s light flows forth, then the arrayed stars rest their brilliance; but when the
sun sets between the mulberry and elm and evening comes, the constellations
appear clear and bright. Thus, if the writing in the Shiji is matched against the age
of the Book of Documents and the Spring and Autumn, its words would seem
shallow and vulgar, smacking of the back lanes and like drooping wings that
would not rise, or an out-of-tune fife unheard. Only when we come to the period
after the Warring States, distantly removed from the sages, can [the Shiji] reveal
its “spearpoints” at ease and unconstrained. [經猶日也，史猶星也。夫杲日流
景，則列星寢耀。桑榆既夕，而辰象粲然。故史記之文當乎尚書、春秋之世
也，則其言淺俗，涉乎委巷，垂翅不舉，懘籥無聞。逮於戰國已降，去聖彌
逺，然後能露其鋒穎，倜儻不羈。]158

To Liu Zhiji the Classics and the histories were separate and unequal. Nonetheless, the

histories (including the Shiji) had their place and would continue to do so. The standard

history canon would continue to grow, and by Liu Zhiji‟s time already included entries

156
I.e., the Documents and Spring and Autumn, which Liu Zhiji had just been discussing and to which he
gives pride of place as „sagely histories.‟
157
STTS 22.165. Trans. Stuart H. Sargent, “Understanding History,” 28.
158
STTS 22.165. Trans. Stuart H. Sargent, “Understanding History,” 28; I have altered his translation
somewhat however. He understands the comparison in the latter part of the passage to include the Hanshu,
which seems incorrect from the point of view both of the original language and the content of the works in
question.

93
for most of the Six Dynasties. Liu Zhiji had strenuous objections to the way that official

history writing was undertaken, but by his time it had become an institutional process

with its own momentum.159 For all its imperfections, it has continued even to the present

day.

159
For a description, see Denis Twitchett, Writing of Official History, esp. 3-30.

94
 Chapter 2

Sima Qian in the Realm of Literary Prose

I now turn to a very different context, one which ultimately had a profound effect

on Sima Qian‟s reputation and the fate of the Shiji: the evaluation of literary prose. From

the beginning, Sima Qian‟s prose style had been an admired aspect of his work. In

Hanshu references to Sima Qian, he was first and foremost a writer of great literary talent.

For example, the evaluation from “Traditions of King Yuan of Chu” 楚元王傳 (Hanshu

ch.36), in a discussion about talent, includes Sima Qian together five other figures:

Zhongni [=Confucius] pronounced: “How true it is that talent is difficult to

find!”1 After Confucius, there were many gentlemen who engaged in literary
composition, but only Meng Ke, Sun Qing [=Xunzi], Dong Zhongshu, Sima Qian,
Liu Xiang, and Yang Xiong [could be considered talented]. These several
gentlemen all have extensive knowledge and wide experience, a penetrating
understanding of matters both ancient and modern. Their words truly made a
contribution to their generation. [仲尼稱，材難，不其甚與！自孔子後，綴文
之士眾矣，唯孟軻﹑孫卿﹑董仲舒﹑司馬遷﹑劉向﹑揚雄．此數公者，皆博
物洽聞，通達古今，其言有補於世．]2

Another discussion of talent, this separating out various types of talents, placed both

Sima Qian and Sima Xiangru under the category of wenzhang 文章 (literary

composition).3 Admiration for Sima Qian‟s style languished somewhat during the Six

Dynasties and early Tang, when parallel prose was very much the admired form.

However, in the late Tang and, especially, the Northern Song, the importance placed

upon Sima Qian as a model of literary style by the successive Ancient-style Prose

movements (古文運動) would prove crucial to the later development of his later

reputation.

1
Lunyu VIII:20, trans. Lau, Analects, 95.
2
HS 36.1972.
3
HS 58.2634. Other available categories included: elegant classicism, conscientious conduct, simple
integrity, promotion of worthies, fixing statutes, jesters, and many more.

95
 Literary movements associated with the idea of Ancient-style Prose might be said

to have had high points in three different eras: the Tang (late eighth and early ninth

centuries), the Northern Song (mid and late eleventh century), and—to a lesser extent—

the Qing dynasty (eighteenth and nineteenth centuries) with the Tongcheng School 桐城

派. In the discussion which follows, I focus on the first two waves of the Ancient-style

Prose movements, in the Tang and Song.

ANCIENT-STYLE PROSE OF THE TANG

Yu-shih Chen gives a succinct characterization of the features of the Tang

Ancient-style Prose agenda: “a revolt against the prevalent parallel-prose tradition of the

Six Dynasties period and a revival of classical ideas in literature.”4 These projects

involved the creation of a new literary canon, and Sima Qian had a secure, if not

especially central, place in that canon.

Han Yu‟s most famous response to Sima Qian is the “Preface on Sending Off

Meng Dongye” [送孟東野序].5 Sima Qian‟s name is mentioned in it, but the more

important connection is that the theory of literary creation which Han Yu develops there

4
Images and Ideas in Chinese Classical Prose, 1. Chen‟s book is an excellent introduction to the Ancient
Prose Movements, organized as case studies of four major figures: Han Yu, Liu Zongyuan, Ouyang Xiu,
and Su Shi. For more detail on Han Yu and his self-conscious development of these ideas, see Hartman,
Han Yü and the T’ang Search for Unity. Ronald Egan‟s work on Ouyang Xiu (The Literary Works of Ou-
yang Hsiu) and Su Shi (Word, Image, and Deed in the Life of Su Shi) give valuable perspective on the Song
dynasty manifestation of the movement. For English translations of some of the exemplary prose pieces
from the Tang-Song Ancient Prose canon, see Shih-shun Liu‟s Chinese Classical Prose: The Eight Masters
of the Tang-Song Period.
5
Meng Dongye‟s given name was Jiao 郊 (751-814). He was a famous poet of the mid-Tang period, who
passed the jinshi exam at the age of 46. At age 50, he was assigned the office of District Defender (wei 尉)
in Liyang County 溧陽縣. He was talented, and had not met his time, thus was in a sorrowful mood.
When he was about to depart for office (i.e., 801), Han Yu wrote this to praise and comfort him, at the
same time revealing dissatisfaction with the way the court employed people. See Han Changli 4.349 n.1.

96
is, as Yang Haizheng and others have remarked,6 directly developed from Sima Qian‟s

theory of literature and suffering. The piece begins:

In general, when things do not obtain their rest, then they cry out. Grasses and
trees have no voice, but the wind stirs them up and they cry out. Water has no
voice, but the wind moves it to cry out. Their stirring is because something
incites them. Their rushing is because something obstructs them. Their seething
is because something scalds them. Metal and stone have no voice, yet when
someone strikes them they cry out. It is the same with people and their words:
only when we help it do we speak. Our singing is because of our longings. Our
wailing is because of what we have cherished. Indeed, everything that comes out
of our mouths and makes a sound is because something has not found its rest.
Music-making comes of having pent-up sorrows within that comes pouring out.
One chooses that which is best at crying out and borrows its cries. [大凡物不得其
平則鳴。草木之無聲，風撓之鳴；水之無聲，風蕩之鳴。其躍也或激之，其
趨也或梗之，其沸也或炙之。金石之無聲，或擊之鳴。人之為言也亦然；有
不得已者而後言，其歌也有思，有哭也有懷，凡出乎口而為聲者，其皆有弗
平者乎！樂也者，鬱於中而泄於外者也；擇其善鳴者而假之鳴。]7

The piece goes on to say that Heaven too “chooses what is best at crying out,” both in

nature (birds, wind, etc.) and among humans. Han Yu then gives a long list of famous

persons divided by period, including most (though not all) of Sima Qian‟s “suffering

authors,” and many others as well. For the Han dynasty, only Sima Qian, Sima Xiangru,

and Yang Xiong are mentioned as being “the best at crying out” [最其善鳴者].

At this point, Han Yu ceases to list any names, complaining that the writers of the

Six Dynasties “could not measure up to antiquity” [不及於古], and wondering, “Was it

that Heaven found their character so ugly that it did not look upon them? Is that why it

did not make those of them cry out who were best at crying out?” [將天醜其德莫之顧

6
Yang, Han Tang, 131-136.
7
Han Changli 4.348.

97
邪？何為乎不鳴其善鳴者也？]8 In the Tang, the list resumes, and includes Du Fu and

Li Bai, as well as Meng Jiao, the recipient of the piece.

Han Yu‟s famous expressive theory, of things “not obtaining rest and crying out”

[不平而鳴], clearly relates to Sima Qian‟s “outpouring of resentment and writing a text”

[發憤著書], though Han Yu‟s formulation is more general. More importantly, both

authors were constructing a kind of literary genealogy or lineage. As discussed above,

Sima Qian had created for himself a set of literary predecessors, a disparate set never

before joined together, and taken to be linked both by their misfortunes and their textual

creation.9 Han Yu was creating a target for his project of “returning to antiquity” 復古, a

model for the “new” Ancient style.10 Then too he was linking prestigious early

exemplars of the style to those in his own dynasty and among his contemporaries whom

he wanted to elevate.

Elsewhere, Han Yu gives the same list of Han dynasty prose masters, but adds Liu

Xiang as well: “Everyone in the Han could write fine prose, but only Sima Xiangru, the

Honorable Senior Archivist, Liu Xiang, and Yang Xiong did it superlatively” [漢朝人莫

不能為文，獨司馬相如、太史公、劉向、揚雄為之最].11 It is interesting to note how

this list differs from earlier ones we have considered. We do not find there Jia Yi or

8
Han Changli 4.350.
9
For comparison, see Jorge Luis Borges, “Kafka and his Precursors.” In that essay, Borges elucidates an
interesting truth about a writer‟s relationship to tradition: that part of a truly original writer‟s genius is to
unite through their similarity to himself a set of precursors we would not have otherwise considered
juxtaposing. As Borges writes: “If I am not mistaken, the heterogeneous pieces I have enumerated
[including works by Aristotle, Han Yu, Kierkegaard, Leon Bloy, and Robert Browning] resemble Kafka; if
I am not mistaken, not all of them resemble each other. This second fact is the more significant. In each of
these texts we find Kafka‟s idiosyncrasy to a greater or lesser degree, but if Kafka had never written a line,
we would not perceive this quality” (Labyrinths, 201). Though of course Borges‟ subject is Kafka‟s
idiosyncrasy, Sima Qian‟s tragedy has something of the same quality.
10
For more on Han Yu‟s return to antiquity, see Hartman, Han Yü and the T’ang Search for Unity, esp.
173-210.
11
“In Reply to a Letter from Liu Zhengfu”答劉正夫書, Han Changli 3.308.

98
Dong Zhongshu, and perhaps most strikingly, we do not find Ban Gu. Ban Gu‟s style,

which (especially in his writings other than the Hanshu) tended more toward ornateness

and parallelism than did Sima Qian‟s, was not as important an influence in Han Yu‟s

literary genealogy.

A somewhat more instructive list can be found in Han Yu‟s “Differentiation of

Advancement and Learning” [進學解]. Here Han Yu lists directly his literary influences:

Above, [I] take as a model Yao and Si,12 which are vast and boundless; the
“Announcements” of the Zhou and the “Pan” of the Shang,13 which are tortuous
and abstruse; the caution and rigor of the Spring and Autumn; the unrestrained
hyperbole of Master Zuo; the Changes, which are rarefied but exemplary; and the
Odes, which are correct but ornate. Below, I come down to Zhuangzi, the Li Sao,
and the records of the Senior Archivist. Ziyun [=Yang Xiong] and [Sima]
Xiangru are similar in their craftsmanship, though different in their details. [上規
姚姒，渾渾無涯，周誥殷盤，詰屈聱牙，春秋謹嚴，左氏浮誇，易奇而法，
詩正而葩，下逮莊、騷、太史所録。子雲、相如同工異曲。]14

The list is an interesting one. It includes four of the Five Classics, but juxtaposes these

with the works of Zhuangzi, Qu Yuan, and Sima Qian. Yang Xiong and Sima Xiangru

also appear, but almost as an afterthought. Again, Han Yu was creating the set of his

predecessors, the authors beside whom he wanted to stand. In doing so, however, he

elevated Sima Qian in an unprecedented way, giving him a place beside the Classics. Of

course, Han Yu probably did not mean to imply that the Shiji was on the same level as

the Classics. His influential juxtaposition, however, would in fact have the effect of

making that comparison more credible in later ages.

12
This refers to the Shangshu sections “Yu shu” 虞書 and “Xia shu” 夏書.
13
The Zhou “Announcements” refer to various Shangshu chapters in the “Zhou Documents” section, e.g.,
“Da Gao” 大誥, “Kang Gao” 康誥 and “Jiu Gao” 酒誥 etc. The Shang “Pan” refers to the “Pan Geng” 盤
庚 chapter of the same work.
14
Han Changli 1.67.

99
 Han Yu acknowledged Sima Qian as one among many influences, a minor one at

that, but others seemed to recognize a particular affinity between the two writers. The

fact that Bai Juyi 白居易 (772-846) described Han Yu as possessing “the air of Ban [Gu]

and [Si]ma [Qian]” [有班、馬之風] should probably be discounted as a conventional

expression, since Bai was recommending Han for a job in the Bureau of History at the

time.15 However, Liu Zongyuan‟s comment deserves somewhat closer examination.

Liu Zongyuan and Han Yu had a curious relationship with one another, one

characterized by a mixture of respect and hostility.16 Their exchange regarding the

writing of history will be examined in chapter 6 below. Here I will consider a letter that

Liu wrote to a third person, Wei Heng 韋珩. Wei Heng had received a letter from Han

Yu and sent it on to Liu Zongyuan. Therefore, Liu Zongyuan‟s letter is ponderously

titled “A Letter in Reply to Wei Heng‟s Having Shown me Han Yu‟s Letter, in which he

Defers to me in Literary Matters” [答韋珩示韓愈相推以文墨事書]. As the

commentator Yan Qi 閻琦 remarks, “Tuizhi‟s letter is not found in his collection, but one

can see from this [response] that it was written carelessly” [退之之書不見於集，而其略

粗見於此].17 Liu writes,

The letter of Tuizhi‟s [=Han Yu], which you enclosed to show me, says that he
wishes to defer to me in literary matters, and furthermore uses [me] to encourage
you. Talent like that of Tuizhi surpasses mine by several degrees, and it is by no
means appropriate for him to defer to me. It cannot be something he really
believes,18 and is certainly something he just used in order to make up his phrase.
The writers whom Tuizhi respects are Sima Qian and Yang Xiong. He and [Sima]
Qian are certainly about on the same level. As for works like [Yang] Xiong‟s

15
“Order appointing Han Yu gentleman of the Bureau of Review, compiler of the historiography institute”
韓愈比部郎中史館修撰制] (QTW 661.6723).
16
See Hartman, T’ang Search for Unity, esp. 52-57, 259-260.
17
LZYJ 34.881.
18
This construal is admittedly a stretch, but a more literal rendering does not fit the context.

100
 Great Mystery, Exemplary Sayings, or Four Sorrowful Rhapsodies,19 it is only
that Tuizhi has not yet written things like them. When he decides to write them,
they will be more magnificent, to the point where [Tuizhi‟s] writing will surpass
Yang Xiong‟s by a long way.... Suppose that [Yang] Xiong were to come [back],
it would still not be appropriate [for Tuizhi] to defer to him, let alone to me! [足下
所封示退之書，云欲推避僕以文墨事，且以勵足下。若退之之才，過僕數
等，尚不宜推避於僕，非其實可知，固相假借為之詞耳。退之所敬者，司馬
遷揚雄。遷於退之，固相上下。若雄者，如太玄、法言及四愁賦，退之獨未
作耳，决作之，加恢奇，至他文過雄逺甚。。。使雄來尚不宜推避，而况僕
耶。]20

A single line from this letter, specifying the writers Han Yu respects, is often quoted out

of context to discuss Han Yu‟s literary influences. However, Liu Zongyuan‟s startling

claim that Han Yu and Sima Qian were “about on the same level” is generally not quoted

or taken seriously. I mention it here, however, because I want to take into account the

full context of Liu Zongyuan‟s remark, the tone in which he makes it.

To be sure, Liu Zongyuan admired Han Yu very much or he would not compare

him favorably with the two great Han dynasty writers, even in jest. At the same time, he

was clearly irritated with the breezy way in which Han Yu has “deferred” to him—

presumably because Han Yu himself could not be bothered to write a real reply.

Therefore, Liu‟s praise of Han is quite deliberately exaggerated. Furthermore, in

comparing Han Yu to Yang Xiong, Liu writes that the latter‟s works “are rather limited

and stagnant, nothing like the wild and unbridled, reckless way in which Tuizhi

composes” [短局滯澁，不若退之猖狂恣睢，肆意有所作]. This compliment has, so to

speak, a sting in its tail. Turned upside down, it suggests that Yang Xiong had mastered a

sort of discipline and restraint which Han Yu could not yet command. Is the reference to

Sima Qian a mere throwaway then? I am inclined to say it is not. Sima Qian too was

19
These are also known simply as the “Four Rhapsodies,” which is how they are referred to in the
Hanshu‟s chapter on Yang Xiong (HS 87B.3583).
20
LZYJ 34.882.

101
sometimes said to have an “unrestrained” quality in his prose,21 and great number of later

scholars would later draw comparisons between the Sima Qian and Han Yu (see below).

I would argue then that Liu Zongyuan was suggesting an analogy between Han Yu and

Sima Qian on the one hand, and between himself and Yang Xiong on the other.22

For his own part, Liu Zongyuan also discussed Sima Qian in relation to other

writers and to Liu‟s own work. His “Preface to Liu Zongzhi‟s Writings in the Manner of

the Western Han” [栁宗直西漢文類序]23 exalts the writings of Western Han literary

figures in contrast to those who came after or went before:

In the early times of the Shang and the Zhou, writings were brief and rustic. From
the Wei and Jin on down, [writings] have been clashing and full of ornamentation.
Only under the house of Han was a balance attained, and in the Eastern [reign] of
the house of Han, it had already declined. In the time of Emperor Wen, Master
Jia first illuminated the Classicist learning techniques, and Emperor Wu was
especially fond of this. People like Gongsun Hong, Dong Zhongshu, Sima Qian,
and Sima Xiangru gave rise to a [literary] style and elegance that increased and
flourished until it spread to all the realm. From the Son of Heaven down to the
high lords, ministers, officers, gentlemen, and common people, all were able to
comprehend it. [殷、周之前，其文簡而野，魏、晉以降，則盪而靡，得其中
者漢氏。漢氏之東，則既衰矣。當文帝時，始得賈生明儒術，武帝尤好焉。
而公孫弘、董仲舒、司馬遷相如之徒作，風雅益盛，敷施天下，自天子至公
卿大夫士庻人咸通焉。]24

Liu Zongyuan is here proposing the Western Han as a pinnacle of literary style,

comparable to the classical stage which lies between the archaic and the baroque in

Heinrich Wöfflin‟s theory of artistic cycles.25 As usual, Sima Qian finds his place among

Western Han paragons, but the company he keeps there is rather interesting. The list

begins by singling out Jia Yi as a successful promoter of the (Confucian) Classicist

21
E.g., Ban Gu wrote that Sima Qian “galloped across past and present” [ 馳騁古今] (HS 62.2737); Qin
Guan‟s “Sima Qian” poem describes him as buji 不覊 [unbridled] (Quan Song Shi 全宋詩 1053.12065).
22
It should be noted that Liu Zongyuan was a great admirer of Yang Xiong, and in fact wrote a
commentary on the Fayan [Exemplary Sayings].
23
LZYJ 21.575-577.
24
LZYJ 21.577.
25
See his highly influential if now much problematized Principles of Art History (1915).

102
learning. It then continues with two other eminent Classicists: Gongsun Hong, who is

credited with developing an early version of the government service examination based

on knowledge of the Classics,26 and Dong Zhongshu, who famously proposed a ban on

non-Ru learning in the imperial academy.27 In short, this would seem to be a Confucian

genealogy as much as a literary one. Yet Sima Qian and Sima Xiangru are not usually

seen as paragons of Confucian ideology. Perhaps to Liu Zongyuan the most important

facet of these writers as a group is that they wrote in a style that was easy to understand.

Yang Haizheng argues that the main emphasis of the last sentence is the writers‟ ability

to influence all strata of society;28 given the context, however, I am more inclined to

believe that Liu Zongyuan was first and foremost making a point about classical (gu 古)

prose style. The fact that “everyone” was able to comprehend the style of these Han

figures29 is what led him to choose them as exemplars, in contrast to the ornamented

prose of the intervening centuries, which was presumably far less comprehensible to all

readers.

Liu Zongyuan‟s best-known opinion of Sima Qian‟s writing is that it was a

paragon of “purity” [潔]. This comment is found in two related letters. The more

detailed of these is the “Reply to a Letter from Wei Zhongli,30 Discussing the Way of the

Teacher” [答韋中立論師道書],31 wherein Liu embarks on a fairly long discussion of his

own literary influences:

26
See his famous memorial on the subject of education and government service (SJ 121.3118-3120).
27
HS 56.2496-2524.
28
Han Tang, 133.
29
No doubt Liu‟s view is somewhat exaggerated.
30
Wei Zhongli appears to have no biography in the Tang histories, but must have been an interesting
person since Liu Zongyuan answered his query in such length and detail. He passed the examinations in
Yuanhe 14 (820).
31
LZYJ 34.871-874.

103
 I base my [literary practice] on the Documents, seeking their substantial simplicity.
I base it on the Odes, seeking their constancy. I base it on the Rites, seeking their
appropriateness. I base it on the Spring and Autumn [Annals], seeking their
judgement. I base it on the Changes, seeking their movement. These are [the
works] from which I take in the fundamentals of the Way. [In my writing], I [also]
consult the Master Guliang in order to sharpen its spirit. I consult Mengzi and
Xunzi in order to make its branches flourish. I consult the Laozi and Zhuangzi to
relax its straightness. I consult the Guoyu to broaden its interest. I consult the Li
Sao to achieve its depth. I consult the Honorable Senior Archivist to bring out its
purity. These are what I keep nearby in order to improve and harmonize myself
with them, and use them in my writing.” [本之書以求其質，本之詩以求其
恒，本之禮以求其宜，本之春秋以求其斷，本之易以求其動，此吾所以取道
之原也。參之穀梁氏以厲其氣，參之孟荀以暢其支，參之莊老以肆其端，參
之國語以博其趣，參之離騷以致其幽，參之太史公以著其潔，此吾所以旁推
交通而以為之文也。]32

This discussion has distinct similarities to Han Yu‟s “Differentiation of Advancement

and Learning” (discussed above), which also assigned specific literary values to each of

the Classics. Liu Zongyuan, however, gave more details regarding the values of the non-

Classical works as well. An interesting feature of this list, in contrast to Han Yu‟s, is the

inclusion of the Guliang zhuan and Guoyu, as well as the non-inclusion of the Zuo zhuan.

The Guliang zhuan is further highlighted in another letter of this sort, which Liu

clearly wrote shortly after the one discussed above, the “Reply to a Letter from Scholar

Chen, Lord Yuan, Declining to be Addressed as his Teacher” [報袁君陳秀才避師名

書].33 Here Liu Zongyuan gives a shortened version of the “great books” list he gave to

Wei Zhongli, adding that “the larger theory is all in [my] reply to Wei Zhongli‟s letter, so

now if you go [see him] you can look at it” [其大説具荅韋中立書，今以往可觀之].

This second, shortened version contains a few interesting differences however:

The most fundamental thing is conduct, that is, in first making one‟s rectitude
sincere. Beyond that, one should first read the Six Classics, and next the Lunyu
32
LZYJ 34.873
33
LZYJ 34.880-881. He mentions that he has also written to Wei Lizhong, and Scholar Chen can go to him
and read about it if he wants.

104
 and Meng Ke‟s writings, which are also both canonical sayings. One should also
to some extent assimilate the phrases of Zuoshi, Guoyu, Zhuang Zhou, and Qu
Yuan. Master Guliang and the Honorable Senior Archivist are most lofty and
pure; one can use them both freely. As for other texts, wait until your writing is
successful and study them at some later date. [以行為本，在先誠其中。其外者
當先讀六經，次論語、孟軻書，皆經言；左氏、國語、荘周、屈原之辭，稍
采取之；穀梁子、太史公甚峻潔，可以出入；餘書俟文成異日討也。]34

Here Liu Zongyuan‟s list is prescriptive rather than descriptive. The Classics are heavily

emphasized, while the Zuozhuan, Guoyu, Zhuangzi, and works of Qu Yuan are slightly

de-emphasized. Interestingly, the Guliang and Shiji seem to occupy a middle ground:

they do not have the profound moral importance of the Classics, but they seem to also

lack whatever flaws Liu Zongyuan perceived in the other four works which led him to

qualify his recommendation.35 Finally, it is worth noting that, in this letter, the canon Liu

Zongyuan defined was a fairly closed one: the student is discouraged from reading farther

afield until his style has developed more fully.

As mentioned above, Liu Zongyuan once compared Han Yu to Sima Qian. It

seems that Han Yu returned the favor, for Liu Yuxi 劉禹錫 (772-842), who edited Liu

Zongyuan‟s works, mentions this in his “Preface”:

Han Tuizhi [=Han Yu] of Chang Li wrote an epitaph for him, and also sent a
letter of condolence which said, “Alas, that a person like this should be so
unfortunate! I once described his writing as being vital and profound, elegant and
invigorating, quite like that of Sima Zichang [=Sima Qian]. Cui [Yin] and Cai
[Yong]36 fall very far short of him.” Huangfu Shi of Anding, who himself has
nothing to be modest about as far as literary writing is concerned, also considered
Tuizhi‟s words to be correct. [昌黎韓退之誌其墓，且以書來弔曰：哀哉，若

34
LZYJ 34.880.
35
The exact meaning of churu 出入 in this context is not entirely clear. I have assumed that Liu intends a
spatial metaphor, something to the effect that one can “come and go in these texts” freely and/or frequently,
the way one can have no reservations about visiting a worthy acquaintance, for example. I admit, however,
that this is not the only possible interpretation.
36
Cui Yin was compared favorably to Ban Gu by the emperor the anecdote alluded to above. Cai Yong 蔡
邕 (133-192) was also a tremendously skillful writer, who lived at the very end of the Eastern Han. For
more regarding his relationship to Sima Qian, see chapter 4 below.

105
 人之不淑。吾嘗評其文，雄深雅健似司馬子長。崔蔡不足多也。安定皇甫湜
於文章少所推讓，亦以退之言為然。]37

Clearly there is a merely conventional element to such comparisons: to say someone

writes like Sima Qian was just to say that they wrote extremely well. But it is also the

case that Han Yu‟s comparison was also made deliberately for the purpose of

highlighting certain literary traits. Both Liu Zongyuan and Han Yu would later be

compared to Sima Qian very frequently, and the criteria for such comparisons were

shaped by the Ancient-style Prose agenda that the two writers would come to be

associated with. As modern scholar Yu Zhanghua has pointed out, three main terms

eventually became associated with Sima Qian‟s prose style: “vital and invigorating” (雄

健), “lofty and pure” (峻潔), and “tactful and indirect” (婉曲).38 Of these, the first two

clearly derived from Han Yu and Liu Zongyuan respectively.

Though I have examined Han Yu‟s and Liu Zongyuan‟s comments on Sima Qian

in considerable detail, it should be noted that objectively speaking neither had especially

much to say about him. A perhaps more significant aspect of the two influential Tang

Ancient-style Prose writers‟ relationship to Sima Qian is the clear evidence of stylistic

borrowing uncovered by later scholars.39 A full treatment of that issue is beyond the

scope of this study, however. Through whatever means, the importance of Han Yu and

Liu Zongyuan for Sima Qian was that they brought him into the Ancient-style Prose

canon. For the Ancient-style Prose movements of later times, which looked up to the two

37
QTW 605.6111. Han Yu‟s original letter has not been preserved in his complete works, but given that
Liu Yuxi was his contemporary, I see no reason to doubt the authenticity of the quotation. The piece‟s
absence is noted and discussed in Zhang Hao‟s 張淏 (fl.1216) Yungu zaji 雲谷雜紀 2.23.
38
Yu Zhanghua et al., Shiji jiaocheng, 304 ff.
39
See, for example, Yu Zhanghua, “Tang Song badajia yu Shiji.”

106
Tang prose masters as founders and teachers, this inclusion was of great significance in

their relationship with the Shiji.

ANCIENT-STYLE PROSE OF THE NORTHERN SONG

According to Su Shi, Ouyang Xiu 歐陽修 (1007-1072) was “the Han Yu of his

day.”40 Certainly, Ouyang Xiu‟s foundational role in the Song Ancient-style Prose

movement was solidified by the examination he oversaw in 1057, in which he famously

failed everyone who did not write their essays in the ancient style.41 Given his

tremendous influence on the Northern Song literary scene, it is no surprise that Ouyang‟s

attitude toward Sima Qian is quite representative both of the admiration and the

hesitation that Song dynasty literati tended to feel toward Sima.

In the realm of prose style, Ouyang Xiu not only expressed his admiration for

Sima Qian but explicitly aimed to imitate him. He ends his “Traditions of Sang Yi” 桑懌

傳, for example, with the following meditation on the Shiji‟s style of narration:

I certainly enjoy passing down people‟s stories, and especially love Sima Qian‟s
fine “Traditions.” The things he wrote are all extraordinary, exemplary, rare, and
restrained, so that scholars delight in reading them. I wanted to study his [method
of] composition, but found it strange how few contemporary people are like those
that [Sima] Qian wrote about. I suspected that [Sima] Qian had a particular
virility in his prose, a fine strength in his narration, but that the ancients were not
necessarily as [he portrayed them]. When I found out the story of Sang Yi, then I
realized that some among the ancients were [as their “Traditions”say], and that
[Sima] Qian‟s writing contained no lies. I [also] realized that there are certainly
[such people] alive today [too], but we simply do not know all about them. What
[Sang] Yi did was strong indeed, though I do not know if my prose can, like
[Sima] Qian‟s writing, cause people to delight in reading [it]. Thus I have
arranged and set it down. [余固喜傳人事，尤愛司馬遷善傳，而其所書皆偉烈

40
Su wrote in his preface to Ouyang‟s works, “Master Ouyang is the Han Yu of today” [歐陽子今之韓愈
也] (Su Shi wenji 10.316), describing this as a generally-held opinion.
41
For a good description of this event, see Ronald Egan, The Literary Works of Ou-yang Hsiu, 27 ff.

107
 奇節，士喜讀之。欲學其作，而怪今人如遷所書者何少也，乃疑遷特雄文，
善壯其說，而古人未必然也。及得桑譯事，乃知古之人有然焉，遷書不誣
也，知今人固有而但不盡知也。譯所為壯矣，而不知予文能如遷書使人讀而
喜否？姑次第之。]42

Ouyang Xiu‟s small afterword actually makes a profound statement about the relation

between style and content. He first suggests the possibility that the prose can exaggerate

the content—Sima Qian was such a great writer that his history just made people seem

more wonderful than they really were. Ouyang Xiu then solves the question by

introducing Sang Yi, the subject of his piece: because of Sang Yi, Ouyang now knows

that Sima Qian‟s writing was true to its subjects (or rather, the subjects were true to the

writings). This is to say that the existence of Sang Yi proves a point both about people‟s

heroic qualities and about the truthfulness of Sima Qian‟s prose; because Sang Yi exists

in the present day, we can know that such people also existed in Sima Qian‟s time, and

thus that his record was true.

It is not my intention to over-read what is mostly just a clever compliment to Sang

Yi. Still, I think that what Ouyang Xiu “learns” from Sima Qian goes beneath the surface

here. In saying that he “wanted to study [Sima Qian‟s method of] composition” [欲學其

作], he is saying that he learned more than technique. He also implicitly claims to have

learned a lesson about choosing one‟s subject—about the qualities that are worth writing

about. Though rare, they could be found in antiquity and can also be found today. One

can conclude that with the proper choice of subject, there is no need for stylistic tricks or

ornamented phrases, no need to exaggerate.

42
Ouyang Xiu Quanji 66.971-972.

108
 In stating his admiration for Sima Qian, Ouyang Xiu used emotional terms. He

was “fond of” (愛) Sima Qian‟s “Traditions,” which scholars “delight in” (喜) reading.

Intellectually speaking, however, Ouyang Xiu had reservations about Sima Qian‟s

historical methods and principles. For example, in his “Preface to the Genealogical Chart

of the Thearchs and Kings” [帝王世次圖序], he complained that people like Sima Qian

“devoted themselves to [collecting] as much information [as possible], considering this

the highest goal” [務多聞以為勝者], and then “were not the slightest bit selective,

fearing only that something would be left out” [無所擇而惟恐遺之]. In his Shi ben yi 詩

本義 [Fundamental Principles of the Odes], Ouyang Xiu also wrote, regarding King Wen:

“Sima Qian‟s Shiji and the various apocrypha and mandate prophecies have a tremendous

number of strange and erroneous sayings. Fundamentally, they are intended to praise

King Wen and honor him, but in fact they are words that, put together, are very much to

his detriment” [司馬遷史記及諸讖緯符命，怪妄之說不勝其多。本欲譽文王而尊

之，其實積毁之言也].43 Ouyang Xiu‟s doubts about the Shiji‟s reliability are clearly

reflected here by the fact that he places the Shiji together with works like the apocrypha

and mandate prophecies.44 He did not necessarily mean that the Shiji had that kind of

political agenda—more likely he meant that Sima Qian‟s inclusion (and thus tacit

legitimation) of such strange sayings could lead to phenomena like the apocrypha or

worse.

The problem of reliability will be treated in detail in chapter 6 below. I introduce

it here only to point out the importance of context: Ouyang Xiu greatly admired Sima

43
Shi ben yi 10.1B.
44
In contrast, Ouyang himself proposed that scholars not go beyond what he understood to be Confucius‟
own portrayal of the early Zhou.

109
Qian as a prose stylist, but had serious doubts about him as an interpreter or transmitter of

antiquity. The reasons for this were related to the complaint of Yang Xiong, Ban Biao,

and Ban Gu—that the Shiji did not accord with the Classics. Where the Shiji and the

Classics disagreed about ancient history, Ouyang Xiu had no choice but to affirm the

version found in the Classics.

Other writers also make it clear that for the Song Ancient-style Prose movement,

as with its Tang precursor, the Shiji formed an integral part of a would-be Ancient-style

Prose writer‟s curriculum. Wang Zhengde‟s 王正徳 (fl.12th c.) Yushi lu 餘師錄

[Records from an abundance of teachers] gives an anecdote showing the importance that

Zeng Gong 曾鞏 (1019-1083), a student of Ouyang Xiu‟s and a major Song Ancient-style

Prose figure, placed upon the Shiji:

When Chen Houshan45 first took his writings and showed them to Master
Nanfeng [=Zeng Gong], the Master read them and asked, “Have you ever read the
Shiji or not?” Houshan replied, “Indeed, I have been reading it since I was very
young.” Nanfeng said, “Not like that. To do it properly you must also put away
other books, and read nothing but the Shiji for two or three years until you are
very familiar with it.” Houshan did as Nanfeng told him, read it, and afterward
again brought his writings to show Nanfeng. Nanfeng said, “Now this is [good]
enough.” [陳后山初携文巻見南豐先生，先生覽之，問曰：曾讀《史記》
否？后山對曰：自幼年即讀之矣。南豐曰：不然，要當且置它書，熟讀《史
記》三兩年爾。后山如南豐之言，讀之，後再以文巻見南豐，南豐曰：如是
足也。]46

Whether this anecdote is true literally or just in spirit, it suggests that Zeng Gong thought

highly of the Shiji as an instructional text for younger Ancient-style Prose writers like

45
I.e., Chen Shidao 陳師道 (1053-1101). He eventually became a member of Su Shi‟s inner circle and was
a well-known Northern Song writer and poet.
46
Yushi lu 1.10b. It is perhaps worth noting a certain similarity between this anecdote and Lunyu XVI:13.

110
Chen Shidao. Furthermore, simply reading the text was not enough: a concentrated

reading and thorough familiarity was required.47

Like Ouyang Xiu, however, Zeng Gong showed a considerable ambivalence

toward the Shiji in the context of history-writing. In his “Preface to the Table of Contents

of the Southern Qi History” [南齊書目錄序], he wrote:

Now, among those who acted as historians after the Three Dynasties, writing like
Qian‟s cannot but be called exquisite and magnificent, [the work of] an
exceptional gentleman. And yet on the other hand, you can also say of it that
[Qian‟s] insight was insufficient to encompass the Principles of ten thousand
matters, that his Way was insufficient to be suitable for use in the realm, that his
wisdom was insufficient to penetrate the meaning of what is difficult to
understand, and that his writing was insufficient to express circumstances that are
difficult to show clearly. How could that be!? Probably as regards the lofty
perfection of the sages and worthies, there are aspects to which Qian was certainly
unable to attain in full purity, or manifest for later generations. Thus he did not
succeed in conforming with them [i.e., the sages]. If Qian‟s attainments and
failures were like this, how much moreso those of others?! [夫自三代以後為史
者，如遷之文亦不可不謂雋偉拔出之材，非常之士也。然顧以謂眀不足以周
萬事之理，道不足以適天下之用，智不足以通難知之意，文不足以發難顯之
情者。何哉？葢聖賢之髙致，遷固有不能純達其情而見之於後者矣，故不得
而與之也。遷之得失如此，况其他邪？]48

Zeng Gong‟s writing is turgid but his view is clear: the Shiji was important enough to

warrant exclusive and prolonged study, but it did not rise to the same level as the Classics.

Sima Qian wrote well but he did not reach the lofty perfection of the sages.

47
A conversation from the Zhuzi yulei makes reference to Chen Shidao‟s familiarity with the Shiji. In that
case, Zhu Xi differs from Zeng Gong in his opinion about how studying the Shiji affects one‟s prose style:
Someone asked, “How about the Shiji?” [Zhu Xi] said, “The Shiji is not something one should
study. One does not succeed in studying it [properly], and instead just gets worse. It is still not as
good as understanding a standard model of writing.” They asked about Houshan (=Chen Shidao)
learning from the Shiji. He said, “Houshan‟s writing is extremely standard, almost too standard.
However, he made a great many fragmented sentences, and this is [because] he learned from the
Shiji. [問：「史記如何？」曰：「史記不可學，學不成，劔顛了，不如且理會法度文
字．」問後山學史記．曰：「後山文字極法度，幾於太法度了．然做許多碎句子，是學史
記．」] (ZZYL 139.3320-3321)
48
Zeng Gong ji 11.188.

111
 The so-called “Three Su,” i.e., Su Xun 蘇洵 (1009-1066) and his two sons, Su Shi

蘇軾 (1037-1101) and Su Zhe 蘇轍 (1039-1112), were of key importance to the Song

Ancient-style Prose movement. They admired the Shiji in varying degrees, but each also

displayed the same sort of ambivalence toward it as did Ouyang Xiu and Zeng Gong.

One of Su Xun‟s most well-known and influential essays, the “Shilun” 史論

[Discussion of History], attempts to reconcile the Classics and the histories by comparing

and contrasting their differing functions. Without directly addressing potential points of

conflict between the Shiji and the Classics, Su points out that “a history is not a constant

rule for ten thousand generations” [史非萬世之常法],49 in a sense implying that the Shiji

need not be held up to the standard of the Classics. On the other hand, writing about

what he considers the finest of the histories up to his time—the Shiji and the Hanshu—he

does venture to suggest that “at times they measure up to Zhongni‟s inherited intentions”

[時得仲尼遺意],50 and by implication can function as Classics.

Against this generally positive picture, the third section of the “Shilun” details a

great number of faults committed by both Sima Qian and Ban Gu. It begins:

Someone asked, “When you discuss history, it is fine for you to pick out the
models of concealment and hidden meaning [employed by] Zhongni
[=Confucius], [Sima] Qian, and [Ban] Gu. Zhongni is not someone we can
criticize. But I only mean to say that [Sima] Qian and [Ban] Gu are not sages.
Can they, like Zhongni, be without a single fault one can point to?” [I] would say:
[Sima] Qian delighted in miscellaneous theories, and did not attend to what the
Way permitted or forbade. [Ban] Gu prized flattery and hypocrisy, and gave too
little importance to a principle held to the death. By and large, this is how one
lays out the critique [of their writings]. If one also wants to take into account
trifling points in picking out their faults, there are so many it is impossible to
mention them all. [或問：子之論史，鈎抉仲尼、遷、固濳法隱義，善矣。仲
尼則非吾所可評，吾意遷、固非聖人，其能如仲尼無一可指之失乎？曰：遷

49
Jiayou ji 9.229.
50
Jiayou ji 9.232.

112
 喜雜説，不顧道所可否；固貴諛偽，賤死義，大者此既陳議矣。又欲寸量銖
稱以摘其失，則煩不可舉。]51

The Tang Ancient-style Prose movement had, for stylistic reasons, often placed the Shiji

in juxtaposition with the Classics. In emphasizing that Sima Qian and Ban Gu were not

sages, Su Xun carefully draws a line between taking the great histories as a model for

stylistic imitation and revering them as flawless exemplars. The phrasing of his

hypothetical interlocutor skirts the edge of an interesting issue: Zhongni—and the

Classics generally—are above criticism, so one can only “select their good points and

follow them.” As for their bad points, it would be impossible even to point them out (let

alone “correct” them), unless to argue that they were the result of some interpolation. As

authorial figures, then, Sima Qian and Ban Gu are real human beings with both good and

bad points, closer to readers and imitators than the distant and perfect sages.

Su Xun, like Ouyang Xiu, had certain misgivings about the reliability of Sima

Qian‟s historical work. There is no doubt that in the realm of prose style, though, Su Xun

too took Sima Qian as a model. In a letter to Tian Kuang 田況 (1003-1061), he described

what he had been reading and why:

For the past several years, I have retired to live in the mountain wilderness. When
one separates himself and forever renounces [the world], every day becoming
further separated from ordinary customs of one‟s day, one thereby succeeds in
putting a great deal of force into literary composition. The gentleness of the
Shijing poets‟ sorrow, the profundity of the Chuci poets‟ spirit, the warmth and
simplicity of Mengzi and Han Feizi, the virility and firmness of [Sima] Qian and
[Ban] Gu, the conciseness of Sun Wu—I throw my writing in the same direction
as these, and there is nothing that fails to satisfy. [數年來，退居山野，自分永
棄，與世俗日疎闊，得以大肆其力於文章。詩人之優柔，騷人之精深，孟、
韓之温淳，遷、固之雄剛，孫呉之簡切，投之所嚮，無不如意。]52

51
Jiayou ji 9.236.
52
“Letter respectfully presented to Tian of the Bureau of Military Affairs” [上田樞密書], Jiayou ji 11.318.

113
Certainly Su Xun‟s is another list of great books that aspiring stylists should study and

know well. But in addition there is a strong sense of breaking with the present in order to

return to the past. Sima Qian forms an important, though not necessarily dominant, part

of that past.

A curious letter from Su Xun to Ouyang Xiu reveals something more about how

Sima Qian‟s reputation functioned during that time. The letter, which in Su Xun‟s Jiayou

ji 嘉祐集 is entitled “The second letter respectfully presented to Ouyang of the inner

Hanlin” [上歐陽内翰第二書], begins with a variation on the lament placed in Confucius‟

mouth by Sima Qian, namely that “A gentleman dreads that he will die without his name

being known by people” [君子病沒世而名不稱焉].53 Su Xun wrote, “Now we are born

in droves and go quickly and unremarkably to our deaths; and not one in ten million has

become known or left a mark on history” [今夫羣羣焉而生，逐逐焉而死者，更千萬

人不稱不書也].54

Su then follows with what I might call a Ru scholarly genealogy, tracing a

succession from Confucius to Mengzi and Xunzi, thence to Yang Xiong, and finally to

Han Yu. Sima Qian is notably absent from this genealogy because, despite the greatness

of his achievement, his moral character was somewhat in question—or as Su Xun himself

put it in his “Shilun” (quoted above), he “did not attend to what the Way permitted or

forbade.” Su Xun‟s reason for making this list, as becomes clear later in the letter, is that

Ouyang Xiu had previously commented to him: “Your „Discussion of the Six Classics‟

53
SJ 47.1943.
54
Jiayou ji 12.334.

114
might as well have been written by Xun Qingzi” [子之《六經論》，荀卿子之文也].55

Before reminding Ouyang Xiu of this remark, however, Su Xun adds another anecdote:

I, Xun, am a single poor commoner, and up until now have been the most useless
of my generation. I have longed to become famous for a single talent or have one
fine deed of mine recorded, yet have been unable to attain it. Truly how much
less, as regards the talent that the [aforementioned] four masters displayed in their
writings, would I dare hope to be a ten thousandth as good.56 A short time ago,
Zhang of Yizhou57 looked at my writing, and pronounced that it resembled that of
Sima Zichang. I was not pleased, and demurred. A commoner, having his
writing compared to Sima Qian by a noble lord, being displeased and refusing the
compliment—is this not considerable departure from natural human feelings? In
truth, I was afraid that the people of the realm would not believe [Lord Zhang],
and furthermore was anxious that he would be unable to stand by his words, and I
would therefore just be laughed at again by the vulgar people of this age. [洵一窮
布衣，於今世最為無用，思以一能稱，以一善書而不可得者也。况夫四子者
之文章，誠不敢冀其萬一。頃者張益州見其文，以為似司馬子長。洵不悦，
辭焉。夫以布衣，而王公大人稱其文似司馬遷，不悦而辭，無乃為不近人
情？誠恐天下之人不信，且懼張公之不能副其言，重為世俗笑耳。]58

This anecdote shows (and there is a great deal of evidence to confirm it) that saying

someone “wrote like Sima Qian” was a fairly common compliment. It was high praise,

to be sure, but Su Xun questioned the sincerity of such casual comparisons.

The continuation of the letter, though it has more to do with Ouyang Xiu‟s

comparing Su Xun to Xunzi, makes the same point with a very human poignancy of

disappointment:

All my life I have made patterned words, seeking that, among the thousands and
tens of thousands of people, my name should at least be somewhat known to later
generations—and yet, I could not attain it. Now in a single day I am suddenly
catapulted into the ranks of those four masters.59 When has the world ever seen

55
Jiayou ji 12.334.
56
These “four masters,” namely, Mengzi, Xunzi, Yang Xiong, and Han Yu, were discussed in the previous
section of the letter.
57
Zhang Fangping 張方平 (1007-1091). He referred to here as Zhang Yizhou because he was posted to
Yizhou, the then-capital of Sichuan province, and succeeded in calming serious unrest that had been
brewing there.
58
Jiayou ji 12.334.
59
See note 56, above.

115
 such a thing? I suppose that these words60 were spoken offhand and carelessly.
Now you, Director, praise the prose style of Shilu,61 and for poetry favor Zimei62
or Shengyu,63 but I have never heard you make such comparisons between [any of
them and one of the four great masters]. I suppose that you were joking [when
you said it about me]. Still, I was foolish and did not think carefully about it, and
spent days writing out a fair copy of my essays, awaiting only your request so that
I could send them straightaway. When I had inquired several times, and been put
off with a demur each time, you finally said, “I have no time to read it.” So I
withdrew and went back to my isolation, not daring to present myself to you again.
With great shame I said to a friend, “It was true then, he was simply joking!” [平
生為文，求於千萬人中使其姓名髣髴於後世而不可得，今也一旦而得齒於四
人者之中，天下烏有是哉？意者其失於斯言也。執事於文稱師魯，於詩稱子
美、聖俞，未聞其有此言也。意者其戲也。惟其愚而不顧，日書其所為文，
惟執事之求而致之。既而屢請而屢辭焉，曰：“吾未暇讀也。”退而處，不敢
復見，甚慙於朋友，曰：“信矣，其戲也！”]64

Su Xun, at this time still an unknown, finds his writing compared to that of Sima Qian

and Xunzi yet declares himself not pleased. The praise seems empty to him because

frequent use has, so to speak, devalued the currency.

In this study I too will henceforth pay little attention to passing comparisons (of

which there are a great many) between a given writer‟s style and that of Sima Qian.65

Instead I focus on the revealing aspects of writers‟ attitudes toward Sima Qian. Of Su

Xun‟s sons, the more famous, Su Shi, had no special fondness for the Shiji, though

scholars have shown his thorough familiarity and frequent use of the text.66 Su Zhe,

however, wrote a fine letter romanticizing the connection between Sima Qian‟s youthful

travels and his prose style:

60
Namely, the comparison between Su Xun and Xunzi, quoted above.
61
Yin Zhu 尹洙 (1001-1047).
62
Du Fu 杜甫 (712-770).
63
Mei Yaochen 梅堯臣 (1002-1060).
64
Jiayou ji 12.334-5.
65
Perhaps the most interesting of these are two versions of a story about Wang Anshi comparing Su Shi‟s
“Biaozhong guan bei” 表忠觀碑 [Stelae inscription for Expressing Loyalty Prospect] to Sima Qian in its
form (體). The two versions differ as to just which specific chapter Wang Anshi is likening it to however.
See Yu Zhanghua, “Tang-Song ba da jia,” 136.
66
Su Shi‟s attitude toward the Shiji is explored more thoroughly in chapter X. For evidence of the Shiji‟s
stylistic influence on Su Shi, see Yu Zhanghu‟s “Tang Song ba da jia yu Shiji,”136-137.

116
 By nature I am fond of writing, and have pondered on it deeply. In my view,
writing is formed by vital force. While the ability to write is not acquired through
[mere] study, vital force can be cultivated. Mencius said, “I am good at
cultivating my overflowing vital force.”67 Today we see that his writing was
broad, substantial, grand and profound, pervading heaven and earth, in direct
proportion to his vital force. The Senior Archivist [Sima Qian] toured the world,
saw all the famous mountains and great rivers within the four seas and associated
with the heroic men of Yan and Zhao. His writing was therefore unconventional,
and its vital force rather extraordinary. Did Mencius and the Senior Archivist
ever, holding a writing-brush, learn to write as they did? Their vital force filled
them up within, and overflowed into their countenance, moving in the words they
spoke, and revealing itself in their writings, all without their being conscious of it.
[轍生好為文，思之至深，以為文者，氣之所形，然文不可以學而能，氣可
以養而致。孟子曰：我善養吾浩然之氣。今觀其文章，寛厚宏博，充乎天地
之間，稱其氣之小大。68太史公行天下，周覽四海名山大川，與燕、趙間豪
俊交游，㳺故其文疎蕩，頗有竒氣。此二子者，豈嘗執筆學為如此之文哉？
其氣充乎其中而溢乎其貌，動乎其言而見乎其文，69而不自知也。]70

Here Sima Qian is paired with Mengzi. Though there is no claim that Sima Qian is

Mengzi‟s equal, neither is there any explicit contrast drawn between them. According to

Su Zhe, they had in common an inner forcefulness and an ability to express it in words,

almost without conscious effort.

The letter quoted above was probably written when Su Zhe was a young man.

Thus part of his purpose in using Sima Qian as an example was to link himself with that

67
See Mengzi zhushu 2B.
68
There seems to be some resonance, in the wording of this letter, with Han Yu‟s “Letter in Reply to Li Yi”
答李翊書,which also discusses the relationship between vital force and one‟s writing and the importance
of nourishing one‟s talent (see Han Changli 3.256).
69
This last is a clear reference to the Shijing “Great Preface”:
Feelings moving inwardly and are embodied in words. When words are insufficient for them,
recourse is had to sighs and exclamations. When sighs and exclamations are insufficient for them,
recourse is had to prolonged utterances of song. When those prolonged utterances of song are
insufficient for them, unconsciously the hands begin to move and the feet to dance. [情動於中，
而形於言，言之不足，故嗟歎之，嗟歎之不足，故永歌之，永歌之不足，不知手之舞之足
之蹈之也。] (Mao Shi Zhengyi 1.13, trans. Legge, She-king, 34)
70
“Letter respectfully presented to Defender-in-Chief Han of the Bureau of Military Affairs”[上樞密韓太
尉書], Su Zhe ji 22.381. My translation is based on that of Shih-shun Liu, Chinese Classical Prose, 295-
297. I have made a number of modifications, particularly in the discussion of qi 氣, which Liu translates as
“spirit”. I do not feel that “spirit” really captures the tangibility that the word and have tentatively altered it
to “vital force,” but that also does not truly satisfy.

117
great literary talent, through the convenient means of comparing Sima Qian‟s “youthful

travels” (壯遊) to his own journey to the capital, which is described in the latter part of

Su Zhe‟s letter.71

Later in his life, Su Zhe compiled a work entitled Gushi 古史 [History of

Antiquity]. It contains 60 chapters, all but seven of which overlap with existing Shiji

chapters, and is very substantially indebted to the Shiji for its material. In short, it is a

reorganized and edited retelling. Each chapter in the Gushi ends with a long “Master Su

said” [蘇子曰] section in clear imitation of the Shiji‟s “The Lord Grand Scribe said” [太

史公曰], though the judgments are often quite different. Clearly, Su Zhe‟s purpose in

rewriting the pre-Han portions of the Shiji was to „repair‟ their moral deficiencies and

perceived lack of accord with the Classics.72 Thus in his “Preface” to the Gushi, Su Zhe

attacks Sima Qian‟s historical writing:

The Honorable Senior Archivist was the first to alter the chronological method [of
history-writing] and made [instead] the annals, the hereditary households, and the
traditions. He recorded everything since the Five Emperors and Three Kings, and
no one in later generations was able to alter it. But his character was shallow and
vulgar, and he was not really learned. He was careless, and gullible. In the period
of the Han emperors Jing and Wu, the old text of the Documents, the Mao
Commentary on the Odes, and Master Zuo‟s Spring and Autumn were all not
included in the [curriculum] of official learning. There were few people in that
generation who were able to read them. Thus, when [the Shiji] records the affairs
of Yao, Shun, and the Three Dynasties, it in no case attains the Sage‟s intentions.
As for the Warring States period, the various masters and rhetoricians each made
their own texts, and many of them added or subtracted ancient events in order to
make credible the theories of their own particular time. Yet [Sima] Qian believed
them all, and what is worse, selected from colloquial sayings and legends in order
to change the old explanations given in the writings of antiquity. [太史公始易編
年之法為本紀、丗家、列傳，記五帝三王以来，後世莫能易之。然其為人淺
近而不學，疎略而輕信。漢景武之間，《尚書》古文、《詩》毛氏、《春

71
Su Zhe was also making use of a growing Song dynasty trope of romanticizing travel, which would
eventually exercise a certain influence on readers‟ reactions to the Shiji.
72
Compare to the efforts of Qiao Zhou and Sima Biao.

118
 秋》左氏皆不列於學官，世能讀之者少，故其記堯舜三代之事，皆不得聖人
之意。戰國之際，諸子辯士各自著書，或增損古事以自信一時之說，遷一切
信之，甚者或采丗俗相傳之語以易古文舊說。]73

To summarize, Su Zhe criticized the Shiji on three different points. First, he attacked

Sima Qian‟s personal failings, his character (其為人). Second, he pointed out that the

orthodox canon as he knows it was not available (or perhaps not comprehensible) to Sima

Qian, and so the Shiji fails to match the Sage‟s intentions simply for lack of information.

Third, and perhaps most damning, he claimed that Sima Qian was swayed by the

differing and self-serving accounts of antiquity as promulgated by various (non-

Confucian) masters.

Su Zhe‟s critical tone here contrasts strikingly with his earlier enthusiasm. Of

course, the letter was written when Su Zhe was a young man, whereas the Gushi was a

work of his maturity, written nearly thirty years later. Furthermore, Su Zhe needed to

justify his rewriting of Sima Qian‟s work: if there was nothing wrong with Sima Qian‟s

portrayal of ancient history, then what need for Su Zhe‟s project?74 These factors aside,

it still seems reasonable to conclude that, as with Ouyang Xiu and Zeng Fan, Su Zhe‟s

comments reveal a contextual difference in the evaluation of Sima Qian‟s work. As a

stylist, he was wonderful. As a historian, he left much to be desired.75

73
QSW 2076.261.
74
The recognition of this circumstance does not spare Su Zhe the ire of modern Chinese critics, however.
Yu Zhanghua remarks indignantly, “This critique of Su Zhe‟s has many problems, and furthermore his
words do not fit with reality. When he says that Sima Qian is „shallow and vulgar, not really learned,
careless, and gullible,‟ his criticism does not avoid excessive harshness, and one finds it quite unbearable”
(Shiji xintan, 190).
75
Zhang Lei‟s “Letter Respectfully Submitted to Academician Zeng Zigu” 上曽子固龍圖書 (Zhang Lei ji,
56.844-845) and “Discussion of Sima Qian” 司馬遷論 (Zhang Lei ji 41.664-665) together provide striking
and very similar example of the same phenomenon. Since I discuss these two works in more detail in
chapter 5, however, I will not cite them here.

119
 Su Shi was the third and most famous of the “Three Su.” His fame in many ways

eclipsed that of his illustrious teacher Ouyang Xiu, and indeed even that of the entire

Song Ancient-style Prose movement, of which he is considered the foremost exemplar.

Though he had little to say about the Shiji, an anecdote from his Dongpo zhilin highlights

an interesting trend that seems to have begun in the Song, namely the juxtaposition of

Sima Qian with Du Fu:

I once asked, “What do lychees resemble?” Someone said, “They resemble

longans.” The guests sitting there all smiled at his shallowness. Actually lychees
do not resemble anything. I said, “Lychees are like scallops.” Hearing this they
were all rather confused, and I did not explain further. Yesterday I ran into Bi
Zhongyou.76 I asked him, “Who does Du Fu resemble?” Zhongyou said, “He
resembles Sima Qian.” I was delighted and did not respond. Really it was a
match for what I had said previously. [僕嘗問：「茘枝何所似？」或曰：「似
龍眼。」坐客皆笑其陋。茘枝實無所似也。僕曰：「茘枝似江瑤柱。」應者
皆憮然。僕亦不辨。昨日見畢仲游。僕問：「杜甫似何人？」仲游云：「似
司馬遷。」僕喜而不答，葢與曩言會也。]77

The comparison between lychees and longans was presumably rejected because of its

obviousness. Yet the comparison Su Shi proposed, between lychees and scallops, was

also not wholly successful—at least immediately—because the two things seem too

dissimilar. It is the same with Du Fu and Sima Qian: on the surface the poet would seem

to have very little in common with the historian.

Yet Su Shi had a reason for recording this anecdote, and it deserves further

consideration. He stated that “lychees do not resemble anything,” and yet he asked his

guests what they resembled. The point he was making is therefore about the nature of

comparison. Anyone who has eaten longans can attest that lychees in fact do resemble

76
Bi Zhongyou 畢仲游 (1047-1121). Su Shi had a great appreciation for his writing. Indeed, when Bi
took the examination to enter the Institute of Academicians (xueshi yuan 學士院), his fellow examinees
included a number of Su Shi‟s protégés (Huang Tingjian, Zhang Lei, Chao Buzhi), but Su singled out Bi‟s
writings as first in rank.
77
Su Shi wenji 蘇軾文集 6.2363.

120
longans, very closely. The comparison between them is quite useful if someone had

eaten longans but not lychees. In a context where everyone is familiar with both,

however, it is completely uninteresting. Thus the guest is mocked for his “shallowness”

(陋)—he spoke as if to an audience of ignorant people.

The comparison with scallops is of a different order: it is the connoisseur‟s

comparison, an analogy designed exclusively for an audience of a select few. At the time

Su Shi made his remark about scallops, it was not successful, but Su Shi did not explain

or try to persuade his listeners. The reason is not that he is skeptical about the value of all

comparison (though that is what one might be tempted to conclude from the remark that

“lychees do not resemble anything”). The reason is that a comparison meant for insiders

is similar to a joke: if you have to explain it, it is no longer interesting. The comparison

of unlike things brings out unexpected aspects of both. Either the comparison has the

effect of delighting and stimulating the mind, or it fails because its hearer is not

sufficiently sophisticated to appreciate it. Whichever the case, there is nothing more to

say. This is why Su Shi‟s response is to remain silent.

The juxtaposition between Du Fu and Sima Qian is also a comparison for

connoisseurs. To compare similar things is to point out their differences: the entire field

of Ban/Ma Contrast (discussed in chapter 1 above) gives an ample demonstration of this.

But to compare unlike things is emphasize their similarities. Both lychees and scallops

are translucent white, soft and yet firm, moist and toothsome, perhaps also exotic and

expensive. What, then, do Du Fu and Sima Qian have in common?

121
 The question is like a riddle that has intrigued scholars ever since. Su Shi‟s

version, being so clever, is perhaps the best known comparison. But other Song figures

explored the juxtaposition as well. Tang Geng 唐庚 (1071-1121) wrote:

After the Six Classics, there was Sima Qian. After the three hundred and five
pieces, there was Du Zimei [=Du Fu]. One cannot study [writing] from the Six
Classics, and there is no need study it there. Thus, composing prose one should
study Sima Qian, and in composing verse one should study Du Zimei. One must
be constantly reading these two texts [i.e., the Shiji and the works of Du Fu]; of
them one might say, “How could a I go a single day without these gentlemen?”78
[《六經》已後，便有司馬遷，《三百五篇》之後，便有杜子美。《六經》
不可學，亦不須學，故作文當學司馬遷，作詩當學杜子美，二書亦須常讀，
所謂「何可一日無此君」也。]79

In analyzing Su Xun‟s “Discussion of History” above, I mentioned the implicit claim that

“Zhongni is not someone we can criticize” and that therefore Sima Qian and Ban Gu are

more appropriate models for aspiring writers. Here that claim is made explicit. What

makes Tang‟s statement particularly interesting is that he singled out Sima Qian‟s prose

as the best model for imitation. For him Sima Qian was not (as he was for Liu Zongyuan

and Han Yu) just one among many writers of prose. Tang Geng instead gave Sima Qian

first place, as it were—always excepting the Classics. Thus one possible way of

understanding the similarity between Sima Qian and Du Fu is that each is the best model

for his own genre.

With the compilation of the highly influential Wenxuan at the court of Liang, Ban

Gu had in some sense parted company from Sima Qian. His place in the Wenxuan vastly

78
This is a saying ascribed to Wang Huizhi, the son of Wang Xizhi. Wang Huizhi had such a deep
fondness for bamboo that he once pointed to the bamboo and asked, with rhetorical flourish, “How could a
I go a single day without these gentlemen?” (JS 80.2103). The allusion is apt in its surface meaning, and
may also suggest a comparison between the two authors and the bamboo, a plant much invested with
symbolic significance.
79
SJYJJC 6.241.

122
overshadowed Sima Qian‟s,80 and there is no avoiding the conclusion that at that time he

was the more important writer, even outside the realm of historical compilation. Owing

to the influence of the Ancient-style Prose movement, however, the situation in the Tang

and after slowly reversed: the Ancient-style Prose movement embraced Sima Qian and

was far more reluctant to accept Ban Gu. Though the disagreements between “Ban”

proponents and “Ma” proponents was far from settled, Sima Qian gained a great deal of

ground in the Northern Song. He was adopted as an ancestor of the Ancient-style Prose

lineage in a way that Ban Gu was not. Thus we find Su Shi‟s good friend and influential

student Huang Tingjian giving advice on writing which includes the suggestion that the

writer should “be better versed in the writings of Sima Zichang and Han Tuizhi” [更熟讀

司馬子長、韓退之文章].81

Of course there is space for many models of prose style from antiquity, but Sima

Qian had in some sense made his way to the top. If Ouyang Xiu was the Han Yu of the

Song, Sima Qian became, in retrospect, the Han Yu of the Han.

80
See discussion in chapter 4 below.
81
“Da Hong Ju Fu shu”答洪駒父書, cited in SJYJJC 6.168.

123
 Chapter 3

The Shiji’s Form in Context

It is almost universally acknowledged that one of Sima Qian‟s major authorial

achievements was the invention of an entirely new form of historical writing, the

pentapartite form known today as jizhuanti 紀傳體 [annals-traditions form]. Most

traditional readers readily ascribed the creation of this form to Sima Qian, sometimes

accompanied by the caveat that it was only with Ban Gu that the form was perfected. In

this chapter, I consider how the “Shiji genre” was seen as being related to other texts, first

the Classics, and specifically the Shangshu and Chunqiu. There were other extra-textual

relationships to be considered as well: commentators attempted to establish symbolic,

numerical correspondences between sections of the Shiji and astronomical or calendrical

concepts.

In examining the Shiji‟s place within the overall development of historical genres

up to the end of the Song dynasty, two further issues emerge. First, there was the rivalry

between the Shiji‟s genre, known today as jizhuanti 紀傳體 [annals and traditions form],

and the chronicle form, biannianti 編年體. A second debate, which cross-cut the first,

was over whether a history should treat a single dynasty (known in modern Chinese as

duandaishi 斷代史) or whether it should attempt, as Sima Qian did, to encompass

many—or even all—different dynasties (“comprehensive history” tongshi 通史).

Next, I review each of the Shiji‟s five sections (wu ti 五體): the Basic Annals 本

紀, Hereditary Households 世家, Tables 表, Treatises 書, and Arrayed Traditions 列傳.1

1
I translate each of the five terms tentatively, in part based on current conventions. Given that much of my
discussion will involve nuanced—and occasionally contradictory—explication of the terms in question, no

124
Late imperial and modern readers would become more concerned with the problem of

possible precursors for individual Shiji sections. Many of these are lost, making the

debate hypothetical and conclusions difficult to come by. Still, the evidence that does

remain sheds some light on the models Sima Qian might have had available to him and

even to some extent what was behind the specific choices he made in developing his form.

THE OVERALL FORM OF THE SHIJI

Writing in the Tang dynasty, Huangfu Shi 皇甫湜 (jinshi 806) expressed the

consensus view that Sima Qian invented the annals-traditions genre of historical writing.

“The ancient histories were arranged chronologically,” he wrote. “It was the Han

archivist Sima Qian who first altered their system and made annals and traditions, which

have been passed down even until today, and there has been no basis for altering them”

[古史編年，至漢史司馬遷始更其制，而為紀傳，相承至今，無以移之].2 In contrast,

Wang Tong 王通 (584-618) was said to have pronounced, “The failure of history3 began

with [Sima] Qian and [Ban] Gu” [史之失，自遷固始].4 Wang Tong‟s remark seems

strange and jarring in light of the generally-accepted view today: that Sima Qian and Ban

Gu were the first true Chinese historians, and that far from destroying the Chinese

historical tradition, they played a major role in creating it.

The original context of Wang Tong‟s remark is cryptic. He went on to say about

the two early historians that “their recording of events is voluminous but the essence

one translation can adequately render the full array of meanings these terms take on. Where I believe that a
given writer‟s conception of a term is too different to be encompassed by the above English translation, I
do not hesitate to supply a different term, but will also provide Chinese characters to avoid confusion.
2
“Biannian jizhuan lun,” in Chizheng wenji 2.129. Many other authors made similar statements.
3
That is, of historical writing, and in particular the change in the genre thereof. See discussion below.
4
Zhongshuo 2.40.

125
therein is paltry” [記繁而誌寡].5 Genre is the most marked way in which Sima Qian‟s

history differed from its predecessors. Furthermore, it is a crucial feature that Sima Qian

and Ban Gu had in common—being the authors of the first and second major annals-

traditions style histories. Lü Zuqian 呂祖謙 (1137-1181) certainly read Wang Tong‟s

remark as relating to genre, commenting, “If [Wang Tong] is criticizing the loss of the

ancient form of history-writing [i.e., chronological organization], then [his judgment] is

fitting” [譏其失古史之體，則當矣].6

Both Huangfu Shi and Wang Tong were reacting to an aspect of Sima Qian‟s

authorship that is usually seen in a positive light today: his invention of a new genre for

historical writing, his claim to have established a new “school” of historical study. The

new form was immediately successful, and perhaps that success in itself was the cause of

anxiety among some early readers (Wang Tong among them, perhaps)—especially those

who were committed to the primacy of the Classics.

The Shiji and Canonical Forms of History

The earliest surviving comments on Sima Qian‟s new genre (other than Sima

Qian‟s own) were made by Ban Gu‟s father Ban Biao, who described the genre as

follows:

When Sima Qian arranged the emperors and kings, then he called it “basic
annals.” [If it was] lords and marquises who passed down their lands, then he
called it “hereditary households.” For officers and gentlemen who were
5
I am by no means certain of this translation. Ji 記 and zhi 誌 are so closely related in meaning that it is
difficult to understand how they can be used contrastively. However, some contrast is clearly intended.
My translation benefits from the modern Chinese gloss of Zheng Chunying (Wen Zhongzi zhong shuo yi
zhu, 40), although it seems to me that Zheng goes to far in translating zhi 誌 as “the essence of classical
learning” [經學的精髓]. However difficult to interpret, the remark clearly relates in some sense to the
presentation of the material.
6
Dashi ji jieti 12.134b. Lü then goes on to defend the Shiji (while criticizing the Hanshu) on other grounds.

126
 exceptionally outstanding, he called it “arrayed traditions.” [司馬遷序帝王則曰
本紀，公侯傳國則曰世家，卿士特起則曰列傳．]7

As Fan Wenlan has noted,8 Ban Biao‟s description implied (without explicitly stating)

that the form of the Shiji was Sima Qian‟s own invention.

Nonethless, Ban Biao had much to say about the relationship between the Shiji

and the Classics. He wrote, “When it comes to selecting from the Classics and taking

excerpts from their commentaries, dividing up and scattering in matters from the

Hundred Schools, [Sima Qian‟s work] is really very sketchy and not as good as the texts

he bases it on” [至於採經摭傳，分散百家之事，甚多疏略，不如其本]. Ban praised

the Han portions of the text, however, writing that “in what [Sima] Qian recorded, his

real achievement lay in [the period] from the beginning of the Han to [where the text]

ends in the time of Emperor Wu” [遷之所記，從漢元至武以絕，則其功也].

Ban Biao‟s underlying argument was in favor of “single dynasty history” 斷代史

over “comprehensive history” 通史. He did not criticize Sima Qian‟s formal innovation

as such, but neither did he see the new historical form as having the need or ability to

replace the version of the past found in the Classics (for the Spring and Autumn period

and before) or even the Hundred Schools texts (for the Warring States era). As I will

discuss further in Part II below, Ban Biao did not emphasize Sima Qian‟s role in the

creation of a new school of historiographers. Rather he saw archival writings as a long-

standing tradition stretching back to the official archivists from the Three Dynasties and

7
HHS 40A.1327. Curiously, Ban Biao did not describe in detail the other two sections of the pentapartite
form, the “tables” 表 and the “treatises” 書, though he had mentioned them in passing earlier in the essay.
Did he consider them peripheral? Or was it merely that he did not intend his continuation to extend to
these sections?
8
Zhengshi kaolue, 16.

127
before, and connected to the state archives of the Warring States described by the Mengzi,

the Zuozhuan, and the Guoyu.

Ban Biao also had clear personal reasons for criticizing the grandiosity of Sima

Qian‟s overall project while praising the Han portions of it: he was engaged in writing a

continuation to the Shiji, which he called “these latter chapters” (此後篇).9 If Ban Biao

had identified Sima Qian‟s comprehensive gathering and ordering of all known history as

his “great achievement,” Ban‟s own labors would have seemed a mere afterthought. By

judging that the pre-Han portions of the Shiji were “not as good as the texts they were

based on,” Ban Biao was claiming that a historian‟s proper task lay more in creation of a

new text than in the recompilation of pre-existing ones; that a historian‟s proper subject

was the recent past. His view of a historian‟s work had more in common with the court

archivists (shi 史) described in the Zuozhuan, who wrote down significant events as they

happened (and sometimes paid with their lives for their candor).10 On the other hand,

Ban Biao did not condemn the new form Sima Qian had invented, and indeed he himself

adopted a modified version of it.11

If Ban Biao saw the Shiji as belonging to the court scribe tradition, it was

therefore necessarily related to that tradition‟s most exalted member, the Chunqiu. This

understanding of the Shiji was also reflected in the Hanshu “Yiwenzhi”, and hence shared

9
Note that Wang Chong also mentions Ban Biao‟s continuation of the Shiji in several different chapters of
the Lunheng (LH 39.613, 61.869, 83.1171). Ban Biao‟s work is not extant today, but is presumed to have
formed the core of the Hanshu, now attributed to his son Ban Gu.
10
For a discussion of the dangers a historian faces, see Han Yu‟s “Letter to Liu Ke” (Han Changli 5.473-
474), discussed in chapter 6 below. The curious closing lines of Ban Biao‟s essay also seem related to the
dangers of the historian‟s métier: “A tradition says, „He who murders an archivist suffers extreme
[misfortune]. The principles underlying the Chunqiu are that fair and comprehensible, straight and
direct‟”[ 傳曰：「殺史見極，平易正直，春秋之義也」] (HHS 40A.1327). Given the instability of the
times in which Ban Biao wrote—the last days of the Wang Mang era and the early restoration—these
words must have been included as a kind of defensive gesture.
11
For Ban Biao‟s perspective on the Shiji, see HHS 40A.1325-1327.

128
by Liu Xiang and Liu Xin as well: at least in the Hanshu catalogue, the Shiji is classified

under the Chunqiu section of the “Six Arts” 六藝 (i.e., the Classics). Those concerned

with compiling histories in the Han and beyond took note of Sima Qian‟s formal

innovation, and annals-traditions style was frequently (though not invariably) the

preferred option. There is little evidence that, early on, writing an annals-traditions

history was seen as somehow disrespectful toward the Chunqiu. Nonetheless, historians

continued to compile chronologically organized works as well as annals-traditions

histories; examples of the former include Xun Yue‟s 荀悅 (148-209) Hanji 漢紀 [Annals

of the former Han] and Yuan Hong‟s 袁宏 (328-376) Hou Hanji 後漢紀 [Annals of the

latter Han].

Fan Ye 范曄 (398-445), author of the now-official account of the Eastern Han,

complained about such works when he wrote his comment on the genre issue:

In the Chunqiu, the writing is both general and brief. It is easy to lose the shape
of events. This is the shortcoming of those [works] now written in imitation of it.
The annals and traditions represent a major innovation by the Archivist [i.e., Sima
Qian] and the Bans, [enabling them] to weave together a single age, making the
deeper significance of events both comprehensive and complete, well-suited to
study by later [readers]. These are what make [the annals-traditions form]
superior, and thus I have continued [to employ it] in my account. [春秋者，文既
總略，好失事形，今之擬作，所以為短．紀傳者，史、班之所變也，網羅一
代，事義周悉，適之後學，此焉為優，故繼而述之．]12

It is intriguing that Fan Ye here criticized the Sagely classic, at least from a formal point

of view, as an inappropriate model for later histories. The original context of Fan‟s

comment has unfortunately been lost, and it survives only as part of a disquisition by the

later historian Wei Dan 魏澹 (fl.580). Perhaps in quoting Fan, Wei Dan ungenerously

12
Quoted by Wei Dan in his own discussion of the same issue; see Suishu 58.1419.

129
omitted some reassuring affirmation that the Chunqiu‟s deeper meaning, at least, was

without flaw.13

What was it about the Chunqiu that made it, in Fan Ye‟s view, “lose the shape of

events”? Perhaps he was commenting on the same frustration Hayden White has

described with regard to medieval Western annals: “We are likely to be put off by the

annalist‟s apparent failure to see that historical events dispose themselves to the

percipient eye as „stories‟ waiting to be told, waiting to be narrated.”14 A chronicle does

not so readily take the form of a story,15 while Sima Qian‟s annals-traditions form

actively lend themselves to various types of narratively and thematically oriented

organization.

A chronicle is linear, arranging events along a single axis of time. In comparing

annals-traditions histories to the weaving of a net, Fan Ye was alluding to Sima Qian‟s

own description of his work as “[gathering] in a net the neglected and lost old knowledge

of the realm” [罔羅天下放失舊聞].16 But while Sima Qian probably meant that he was

gathering various traditions together in a single work (like fish in a net), in order to

13
Intent on advancing his own insight into the issue (discussed below), Wei Dan might well have found it
convenient to exaggerate Fan Ye‟s disrespect toward the Sage. Still, it is unlikely that he invented or
knowingly falsified the argument he attributed to Fan Ye.
14
“The Value of Narrativity”, 6. Concerned with a self-conscious investigation of the entire phenomenon
of narrativity in historical writing, White continues, “Surely a genuinely historical interest would require
that we ask not how or why the annalist failed to write a „narrative‟ but rather what kind of notion of reality
led him to represent in the annals form what, after all, he took to be real events.” In examining the possible
advantages of the annalist‟s methodology, he suggests that “it seems eminently rational and, on the face of
it, rather prudent in both its manifest desire to record only those events about which there could be little
doubt as to their occurrence and its resolve not to interpellate facts on speculative grounds or to advance
arguments about how the events are really connected to one another” (9). Again, it seems to be just these
elements which led Fan Ye to complain that the annalistic Chunqiu-style histories failed to capture “the
shape” of events.
15
It made be argued that the chronologically-organized Zuo zhuan is full of wonderful stories. Leaving
aside the question of whether the Zuo zhuan was originally conceived as a chronological record, any reader
will admit that the Zuo‟s tendency to present whole narratives is more at odds with its chronological
structure than in accord with it.
16
SJ 130.3319.

130
preserve them from the ravages of time, Fan Ye makes the metaphor more serious and

abstract. He seems to envision an entire era (not just its texts) being captured by a

transformed historical genre. In that genre, multiple time-lines could be read in parallel,

giving an added dimension to the portrayal of events, a more complete sense of their

“shape.” If Sima Qian was implicitly likening himself to a fisherman scooping up fish,

Fan Ye portrayed him as being the first designer of the net, as well as a fisherman who

used that net to great effect.

In the early Tang, Emperor Gaozu commissioned Wei Dan to improve upon Wei

Shou‟s 魏收 (506-572) Weishu 魏書. Wei Dan produced the now-lost Weishi 魏史 in

ninety-two juan. His biography in the Suishu 隨書 explains that “there were many

differences between Wei Dan‟s structural principles and Wei Shou‟s” [澹之義例與魏收

多所不同].17 Several are then laid out, in what seems to be an excerpt from the

“ „Discussion of History‟ or „Organizational Principles of History‟ in one juan” [史論及

例一卷], which Wei Dan was said to have produced as part of the Weishi.18

Of these differences, the fifth is most relevant to the current discussion. There

Wei Dan claims to have made a discovery about the Shiji‟s relationship to the Classics:

Where Hu Sui put forth his questions and Sima Qian answered them, the
principles [of the Shiji] were already entirely [laid forth therein]. Those who later
transmitted it still did not grasp this. The idea of Dong Zhongshu and Sima Qian
is fundamentally this, that the Shangshu was a register of a peaceful and
prosperous age, while the Chunqiu was a model for a tumultuous and disordered
time. Since the underlying natures of rise and decline are not the same, so should
the textual creations also differ. [If the realm] is well-governed and stable, then
the accounts are direct and respectful of the enlightened [sovereign]. If the era is
one of disorder, then the phrases at once both reveal and conceal.19 Different
17
Suishu 58.1417.
18
Ibid.
19
This may be compared to the Great Preface of the Shijing, where the qualities of musical pieces are also
categorized according to the qualities of the government under which they are produced: “The sounds of a

131
 roads mandate [one‟s choice of] category,20 and neither [category] relies upon or
imitates the other. This is what [the Shiji] means where it states: “When the Way
of the Zhou was cast away, the Chunqiu arose there.”21 And: “When Yao and
Shun flourished, the Shangshu recorded it.” [The Shiji continues]: “Since the Han
arose, the calendar has been corrected and the color of the court clothing has
changed. Though I sing the praises of this sagely virtue with all my humble
strength, I still cannot express it fully. When I speak of „transmitting past events‟
and you sir compare it to the Chunqiu, you are terribly mistaken.” 22 This being
so, it is clear, then, that the annals-traditions form comes from the Shangshu,
rather than being modeled after the Chunqiu. [壺遂發問，馬遷答之，義已盡
矣．後之述者，仍未領悟．董仲舒、司馬遷之意，本云尚書者，隆平之典，
春秋者，撥亂之法，興衰理異，制作亦殊．治定則直敘欽明，世亂則辭兼顯
晦，分路命家，不相依放．故云「周道廢，春秋作焉」、「堯、舜盛，尚書
載之」，是也．「漢興以來，改正朔，易服色，臣力誦聖德，仍不能盡，余
所謂述故事，而君比之春秋，謬哉」．然則紀傳之體出自尚書，不學春秋，
明矣．]23

In the Hanshu “Yiwenzhi”, as mentioned above, the Shiji was classified as being

part of the Chunqiu line. In that same treatise, Ban Gu reminded readers that in ancient

times, there had been two archival officials, the Archivist of the Left and the Archivist of

the Right: “The Archivist of the Left recorded words, and the Archivist of the Right

recorded events. Events make up the Chunqiu[-type texts]; the words make up the

Shangshu” [左史記言，右史記事，事為春秋，言為尚書].24 It seems, then, that Ban

Gu considered Sima Qian to be, at least symbolically, an Archivist of the Right, a

recorder of events in the Chunqiu tradition.

well-ordered age are peaceful and then joyful, for its governance is harmonious. The sounds of a chaotic
age are resentful and then furious, for its governance is discordant. The sounds of a ruined state are
mournful and then longing, for its people are in difficulty” [治世之音安以樂，其政和，亂世之音怨，以
怒其政乖，亡國之音哀以思，其民困] (MSZY “Great Preface,” 14).
20
The character translated as “category” is jia 家, more usually understood as “family” or “school of
thought.” I translate it “category” here because Wei Dan is clearly referring to the Shangshu and Chunqiu
divisions in the Hanshu “Yiwenzhi”.
21
Not in today‟s Shiji; see discussion below.
22
SJ 130.3299.
23
Suishu 58.1419.
24
HS 30.1715. Note that the Archivists of the Right and Left are also mentioned in the Liji, “Jade
Pendants” chapter (Liji 29.545) as performing the functions Ban Gu imputes to them, though the
association with the Chunqiu and Shangshu does not seem to be present in that context.

132
 Wei Dan, however, tried to make the case that the Shiji belonged in the “left side”

Shangshu line. His argument relied on a different categorization of the two Classics.

While Ban Gu focused on their different content (narratives versus speeches), Wei Dan

chose to emphasize the different conditions in the time periods when the two Classics

originated. He used quotations from the Shiji to support his contention that Sima Qian, at

least, saw the two Classics the same way he (Wei Dan) did. The first phrase about the

origin of the Chunqiu, however, is not found in today‟s Shiji. Possibly it is a paraphrase

from the small preface to the “Hereditary Household of Chen She” (Shiji ch.48)25;

possibly the phrase appeared in the version of the Shiji that Wei Dan knew, but later

disappeared. In either case, Wei Dan makes “Shangshu-style” works (including the Shiji)

into appropriate models for a flourishing age, while seeing “Chunqiu-style” histories

(perhaps intending those that were organized chronologically) as better-suited to an age

of decline.

This is a superficially plausible defense of Wei Dan‟s own choice to employ the

annals-traditions form. It was also potentially an attack on those who preferred

chronicles. Yet it is difficult to see that the Shiji really has much in common with the

Shangshu, either formally or as regards the characteristics of the time period(s) each work

covers. Wei Dan accepted Sima Qian‟s dialogue with Hu Sui at face-value, but it is also

possible to read the same text as having ironic or ulterior meaning. In short, Wei Dan

attempted to reinvent the Shiji‟s genealogy vis-à-vis the Classics, but he did so on fairly

shaky grounds.

25
“Jie and Zhou lost the Way, and Tang and Wu arose. Zhou lost its way and the Chunqiu arose” [ 桀﹑紂
失其道而湯﹑武作，周失其道而春秋作] (SJ 130.3310). Alternatively, the last three characters could be
translated “the Spring and Autumn period arose,” though in the context of bibliographic discussion,
chunqiu is more likely to be understood as the text rather than the period. Furthermore, the period takes its
name from the text, and thus any reference to the period is also, underlyingly, a reference to the text.

133
 Liu Zhiji 劉知幾 (661-721) may have been the most important scholar of

traditional Chinese historiography, and much of his magnum opus, the Shitong 史通

[Comprehensive discourse on history], is deeply concerned with the issue of form. The

first two chapters offer slightly contrasting insights into how Liu Zhiji saw the Shiji as a

model of historical writing, and its relationship to other such models.

The Shitong‟s chapter 1, “The Six Schools”, certainly bears some relation to Sima

Tan‟s essay “Essential Points of the Six Schools” (anthologized in the last chapter of the

Shiji). Sima Tan‟s subject was schools of thought, while Liu Zhiji wrote about schools of

historical writing, but Liu evaluated the strengths and weaknesses of his subjects in much

the same way as Sima Tan did. Liu Zhiji‟s six schools are:

1. Shangshu 尚書
2. Chunqiu 春秋
3. Zuozhuan 左傳
4. Guoyu 國語
5. Shiji 史記
6. Hanshu 漢書

Liu Zhiji‟s classification criteria deserve a discussion in their own right. However, for

the present I discuss only how the relationship between the Chunqiu and Shiji is

portrayed in this chapter.

Writing about the six different histories as separate jia 家, Liu Zhiji was still

careful to point out the relationships between them. So, in writing about the Chunqiu

school, Liu already raised the issue of the Shiji‟s relationship with it: “We come to the

Honorable Senior Archivist making the Shiji: he began by making basic annals for the

Sons of Heaven. If we investigate his ancestral precepts, he would seem to be taking the

134
Chunqiu as a model” [至太史公著史記，始以天子為本紀，考其宗旨，如法春秋].26

Liu Zhiji thus to some extent supported the Hanshu “Yiwenzhi” classification of the Shiji

(and especially the basic annals) as being closely related to the Chunqiu.

Yet Liu Zhiji gave the Shiji its own separate category (“school”). This was in part

because he was affirming Sima Qian‟s claim to have “completed the words of a single

school,” but also in part because he was concerned as much with the followers of a

school as with its founders. So in discussing the development of the tradition that began

with the Shiji, Liu Zhiji explained how it diverged from the Chunqiu:

Thenceforth, those who served as official historians all used [Sima Qian‟s] model.
However, with the passage of time and arrival of different generations, the form
[in the sense of genre] was also altered. In writing about events, there is very
little hidden criticism in their words, and neither do they make evaluative
judgments upon events. Thus it is what [Si]ma Qian [would have?] called setting
in order ancient events and nothing more.27 How could one then compare it to the
Chunqiu? [自是為國史者，皆用斯法。然時移世異，體式不同。其所書之事
也，皆言罕褒諱，事無黜陟，故馬遷所謂整齊故事耳，安得比於春秋哉！]28

In short, the Shiji has its own school because the dynastic histories which came

afterwards did not live up to model Sima Qian himself had set.

The interesting implication of Liu Zhiji‟s discussion, however, is that Sima Qian’s

work may in fact have been recognizably related to the Chunqiu, at least in its “ancestral

precepts” (宗旨), if not in its form. What disqualifies the Shiji‟s successor works from

the Chunqiu category is that their words rarely “contained hidden criticism”, i.e., baohui

褒諱 (understood to mean something like “appearing to praise while actually criticizing”),

26
STTS 1.8.
27
A paraphrase of a statement from Sima Qian‟s dialogue with Hu Sui: “What I mean by transmitting
ancient events is putting in order the genealogies and traditions; it is not what is called „creating‟” [余所謂
述故事，整齊其世傳，非所謂作也] (SJ 130.3299). Sima Qian‟s speech ends, of course, by disclaiming
comparisons between his work and the Chunqiu.
28
STTS 1.8.

135
and they failed to indicate whether or not their subjects‟ actions were meritorious.

Though absent from the Shiji‟s successors, these are qualities that the tradition does tend

to ascribe to the Shiji itself. Thus, although Liu paraphrased Sima Qian‟s modest self-

description—that he was “setting in order” events and nothing more—in fact, Liu did not

take Sima Qian to be accurately describing the Shiji, only its successors.

In addition to the above divisions, which are more concerned with tracing a kind

of textual lineage, Liu Zhiji also, in “Two Genres” 二體 (the second chapter of the

Shitong), presented a different type of division. There he gave his version of the history

of history, but with particular emphasis on tracing historical genres current in his day:

Since the era of the Three [Thearchs] and Five [Emperors], as regards writings
there were Registers and Barrows: long ago, remote indeed. One cannot discuss
them in detail. [Historical records] from [the time of] Tang and Yu down to the
Zhou: these are the Old Text Revered Documents. However, that age being one
of purity and substantial simplicity, the writing is accordingly concise and
abbreviated. If one seeks for a complete genre therein, it is certainly deficient.
Once [Zuo] Qiuming had made his commentary on the Spring and Autumn, and
Zichang [=Sima Qian] had composed the Shiji, the genres of official history were
therein completed. Those who came after continued to produce [similar work],
following their predecessors mechanically, pretending to change or enlarge
[something], altering the names or titles, but their scope had its limitations, and
who has ever been able to transcend [Zuo Qiuming and Sima Qian]?! Probably
Xun Yue29 and Zhang Fan30 were in the party of [Zuo] Qiuming, while Ban Gu
and Hua Jiao31 were of Zichang‟s branch. There are only these two schools, and
each has something to boast about. If [I] had to distinguish their strengths and
weaknesses, it is in fact possible to discuss them. [三、五之代，書有典、墳，
悠哉邈矣，不可得而詳。自唐、虞已下迄于周，是為古文尚書。然世猶淳
質，文從簡略，求諸備體，固已闕如。旣而邱明傳春秋，子長著史記，載筆
之體，於斯備矣。後來繼作，相與因循，假有改張，變其名目，區域有限，
孰能踰此！蓋荀悅、張璠，邱明之黨也；班固、華嶠，子長之流也。唯二
家，各相矜尚。必辨其利害，可得而言之。]32

29
Compiler of the Hanji 漢紀.
30
Compiler of a Hou Hanji 後漢紀, now lost.
31
Compiler of a Hou Hanshu 後漢書, now lost.
32
STTS 2.27.

136
The first thing to note is that Liu Zhiji rejected Wei Dan‟s dichotomy, which would have

the so-called Shangshu form be appropriate for times of peace, while the Chunqiu form

was reserved for times of unrest. Instead, he considered both the Shangshu, and (by

implication) the Chunqiu, to fall short of being complete forms of historical writing.

In Liu Zhiji‟s conceptual categories, we can see the reflection of the new Suishu

“Jingji zhi” bibliographical categorization system: a consciousness that history was

coming into its own as a division, separate from if not equal to that of the Classics. Thus

for Liu Zhiji, it is not Kongzi‟s Chunqiu but Zuo Qiuming‟s that stands opposite Shiji as

the major representative of chronological form. Though the Zuozhuan to this day never

fully separated from the category of “Classic”, at least between Zuo Qiuming and Sima

Qian there was something of a fair contest. It had become increasingly impossible to find

fault with the Supreme Sage,33 but Zuo Qiuming was not immune to criticism.

Furthermore, the Zuozhuan has a level of detail lacking in the Chunqiu but comparable to

that in the Shiji.

In “Two Genres”, Liu Zhiji also suggested that the liezhuan chapters of the Shiji

were like a commentary to the benji section:

The annals as a genre are like the Classic of the Spring and Autumn, binding
together the days and months to complete the years and seasons, making a record
of the ruler above in order to reveal the unity of the state…. When there was
some great event worthy of recording, it appears under the year and month [in
which it happened]. When [Sima Qian] wrote about a matter in full detail
[though], he put that in the associated traditions. This is his principle. [蓋紀之為
體，猶春秋之經，繫日月以成歲時，書君上以顯國統．．．有大事可書者，
則見之於年月；其書事委曲，付之列傳．此其義也．]34

33
Su Xun expressed a similar sentiment in his “Discussion of History” 史論: “Zhongni is not someone we
can criticize. [But]…[Sima] Qian and [Ban] Gu are not sages” [仲尼則非吾所可評…遷固非聖人]
(Jiayou ji 9.237-238).
34
STTS 4.37-38.

137
Admittedly, Liu Zhiji‟s explication was not very explicit. Based on his discussion,

however, it was not difficult for the reader to take the next step: namely, making an

analogy between the Chunqiu sanzhuan [Three commentarial traditions of the Spring and

Autumn] and the liezhuan section of the Shiji.

Various scholars accepted and elaborated on Liu Zhiji‟s understanding. For

example, Zhang Xuecheng wrote:

As for the original principle behind the use of ben [in the benji section], Sima
Qian‟s intention was to take the Spring and Autumn Annals as a model. However,
the Zuoshi, Gong[yang] and Gu[liang] were traditions each made by a separate
specialist, while [Sima] Qian‟s [text] was just the work of one person. In addition
[to the basic annals], he wrote treatises, tables, and arrayed traditions, in order to
make the weft. Thus, in adding to the ji the word ben, he was just making clear
that the ji were the warp. [原其稱本之義，司馬遷意在紹法《春秋》．顧《左
世》、《公》、《榖》，專家各為之傳，而遷則一人之書，更著書、表、列
傳以為之緯，故加紀以本，而明其紀之為經耳．]35

The appealing image that Zhang Xuecheng calls up is that of the Shiji as a whole cloth,

with classic (jing 經) and commentary (zhuan 傳) both created by a single writer. It

made the Shiji—as an authorial creation—seem somehow more complete than the

Chunqiu. Again, Liu Zhiji never said this explicitly, but as a possible development it was

inherent in his suggestion.36

35
WSTY 7.703. Others who accept Liu Zhiji‟s theory about the Shiji‟s genre include Chen Shih-hsiang,
who wrote: “In naming his „biographies‟ zhuan, Sima Qian was holding fast to the earlier sense of the word,
that the individual lives he depicted were mere illustrations of the greater events and ideals of the times;
and his liezhuan therefore stands in a subservient position to his „imperial annals‟ (pen-chi) in a sense not
too different from that of the Gongyang zhuan to the Chunqiu” (“An Innovation”, 50).
Denis Twitchett‟s “Chinese Biographical Writing” also contains an explicit statement of the idea,
which he understood to be conventionally accepted: “On the basis of materials in the Shiji a reasonably
good case may be made out for a parallelism between the Spring and Autumn Annals and the benji sections
on the one hand, and the three „traditions‟ and the liezhuan on the other. The assessment of the relative
roles of the two categories, with its implicit evaluation of their reliability, became an article of faith with
the official historians of later centuries” (97).
Durrant, Cloudy Mirror, 155, n.28 also reviews the debate. Mark Edward Lewis makes the same
point in Writing and Authority (308-313) without citing previous scholarship.
36
Zhang Xuecheng and other later scholars make it sound as if the annals-traditions form were the clear
winner, but in earlier times (especially the Song through Ming), this was still very much an open question.

138
 Huangfu Shi 皇甫湜 (jinshi 806), a student of Han Yu, wrote a long essay

specifically addressing the “two genres” delineated by Liu Zhiji. In it, he gives a general

sense of the verdict on Sima Qian and genre issue as it stood in his time:

Those who through generations have discussed [the genre issue] have considered
Qian as being lacking in restraint and exercising arbitrary private intent, throwing
away the ancient models. [Furthermore, they say that] the annals-traditions style
is redundant and verbose, not as good as the chronological style. [歴代論者，以
遷為率私意，蕩古法，紀傳煩漫，不如編年。]37

Having set up his opponent‟s position, Huangfu Shi then proposed to provide a more

nuanced discussion:

I, Shi, believe that when judging whether something is in harmony with the
Sagely Classics, one should consider the heart of the matter and not just surface
traces. Arriving at the right genre for a good history lies in what suits [the content,
or the times], and not in exact imitation. [Whether to use] chronological or
annals-traditions style should depend only the needs of the time, and where the
talents [of the historian] are strongest. What need is there for a constant rule? [湜
以為合聖人之經者以心不以迹，得良史之體者在適不在同。編年紀傳繫于時
之所宜，才之所長者耳，何常之有？]38

This is Huangfu Shi‟s overall thesis, that the genre of history need not be formally

identical to those of the Classics (and in particular the Chunqiu). In short, the issue of

genre should not, by itself, be used to condemn a work as being contrary to the Classics.

Liu Zhiji must also have believed something close to this, though he did not state it so

clearly or explicitly.

Though Huangfu Shi began from the position that either of the two forms is

acceptable, the underlying argument he developed is similar to Fan Ye‟s, that in practice

Sima Qian‟s form is superior. He argued this point in relation to both the pre-Han and

post-Han traditions:

37
Chizheng wenji 2.129.
38
Ibid.

139
 When a chronology records events, it is bound by its ordering scheme,
constrained to juxtapose [events] that are quite separate from one another. It
forces [the historian] to focus on the larger outline, and be very brief in his
narration of events. This actually causes quite a lot of gaps in the record, and a lot
of writings are left out. Thus it is necessary to compile a separate record in order
to supplement the work with speeches and exhaust all the roots and branches of an
affair. This is why, in making a Chunqiu there must also be a Shangshu; outside a
Zuozhuan, also one must make a Guoyu. [編年記事，束于次第，牽于混并，
必舉其大綱，而簡于序事，是以多闕載，多逸文，乃别為著録，以備書之語
言而盡事之本末。故《春秋》之作，則有《尚書》，《左傳》之外，又為
《國語》。]39

.…Chronologically organized histories have fallen into disuse, and there is

probably a reason for it. Only Master Xun‟s Hanji and Master Pei‟s Song Lue
insist on forcing a return to antiquity, both being made chronologically. However,
there is much that is left out as regards fine words and eloquent speech, detailed
events and thorough explanations. If one reads over the standard histories, only
then is one able to be completely enlightened—for in every chapter of them [one
finds] clear revelations about the secrets of success and failure. [編年之史遂
廢，葢有以也。唯荀氏為《漢紀》，裴氏為《宋略》，強欲復古，皆為編
年，然其善語嘉言，細事詳說，所遺多矣。如覽正史，方能備明，則其密漏
得失章章于是矣。]40

These two passages show that, to Huangfu Shi, the major failing of chronological form is

its tendency to lack detail, particularly in the realm of speeches and of in-depth

explanations. A thoughtful reader might complain about the way he used the Chunqiu

and Shangshu in his argument, for in terms of content there is little or no overlap between

them. Clearly it is the form that concerns him, however: in this case a pure narration of

events (Chunqiu form) suffers from the lack of speeches and detailed accounts of

episodes (Shangshu form).

The Zuozhuan/Guoyu contrast is, if anything, more forced. The Zuozhuan

certainly does not lack for long speeches; a reader unsympathetic to Huangfu Shi‟s

overall argument could easily point out that it already represents an acceptable hybrid

39
Chizheng wenji 2.130.
40
Ibid.

140
form. The Guoyu does, however, contain speeches that the Zuozhuan lacks, as well as

longer versions of speech that also appear in the Zuozhuan. Furthermore the Guoyu,

being organized by state, appears more sensitive to distinctions among regional histories

that the Zuozhuan tends to obscure.

In addition to criticizing the weaknesses of chronological histories, Huangfu Shi

gives a highly favorable account of Sima Qian‟s project in creating the new annals-

traditions form:

It should be possible to reunite the Archivist of the Left with the one on the Right,
and to fit the outer traditions to the inner ones! [Yet] if they are put together than
it was too complicated; if they are separated then differences arise; if one is left
out then there is a deficiency; Zichang [=Sima Qian] regretted that it had to be so.
Thereupon, he reformed the old classics and inaugurated a new technique, making
annals, traditions, tables, and treatises. He narrated a given matter from start to
finish. Both the inside and the outside were expressed. He more or less reached
an appropriate middle ground, and so his work has been passed down and
immortalized. [可復省左史于右，合外傳于内哉！故合之則繁，離之則異，
削之則闕，子長病其然也，于是革舊典，開新程，為紀為傳為表為志，首尾
具叙述，表裏相發明，庶為得中，將以垂不朽。]41

Sima Qian‟s great insight, as portrayed by Huangfu Shi, was to recognize the desirability

of putting previously disparate elements of history together in one work. The Archivists

of the Left and Right, as discussed above, represent one type of dichotomy: the division

between speeches and events. “Outer” and “inner” traditions are less familiar in this

context, but seem to refer to the division between a bare Chunqiu-style narration (inner

tradition) and a fuller narrative that clarifies causes, provides details or explanations, and

makes explicit judgements (outer tradition).

A chronology overloaded with the full weight of speeches, details, explanations

and judgments would be complex in the extreme. (Again, the massive, multi-layered

41
Ibid.

141
Zuozhuan would seem a good example of the problem.) Instead of creating separate

works or leaving things out, Sima Qian brought into being a new form, one which

purported to solve the problems encountered by previous histories. He made a form

which was conducive to recording all types of material.

Huangfu Shi considered it important to point out that the Shiji still retains the

advantages of chronological histories, while managing to avoid their defects:

Is the making of chronologies not fitting events to days, fitting days within
months, fitting months into the seasons, and fitting seasons into years? When
Master Sima made his annals, he took Xiang Yu as the inheritor of Qin, and he
took Empress Lü as the continuer of [the Han dynasty]. He indeed considered
that, in the succession of years, one could not discard any—that the timeline could
not suffer a gap—and that is why he wrote [as he did]. [編年之作，豈非以事繫
日，以日繫月，以月繫時，以時繫年者哉！司馬氏作紀，以項羽承秦，以吕
后接之，亦以歴年不可中廢，年不可闕，故書也。]42

The Xiang Yu and Empress Lü annals were amongst the most troubling to readers, and

Huangfu Shi‟s explanation of them will be discussed in more detail below. What is

important here is his perception that the basic annals section is a complete chronological

history. Liu Zhiji too was partly expressing this intuition when he said that the Chunqiu

was Sima Qian‟s model in making the basic annals.

Huangfu Shi had no idea that the chronological form of history, which had “fallen

into disuse” in his time, was about to make a tremendous comeback—most notably with

Sima Guang‟s Zizhi tongjian 資治通鑑 [Comprehensive mirror for aid in government].

In the Southern Song, there was also Lü Zuqian‟s unfinished Da shi ji 大事紀 [Record of

major events].43 Both works make extensive use of Shiji material, as well as frequent

42
Chizheng wenji 2.129.
43
The Zizhi tongjian begins in the last years of the Eastern Zhou King Weilie 威烈 (r.425-402 BCE), and
ends in 959 CE, just before the founding of the Song dynasty. The Da shi ji begins where the Chunqiu
leaves off, in 479 BCE, and ends during the reign of Emperor Wu, in 90 BCE.

142
reference to the Honorable Senior Archivist‟s perspective. (I will not address here the

issue of how (or whether) such works overcame the problems that Huangfu Shi raised.)

By the early Ming, Zhu Xi‟s Zizhi tongjian gangmu 資治通鑑綱目 was officially

considered the epitome of historical studies.44 This was a drastically abbreviated and

more specifically moralizing version of Sima Guang‟s work. Its official elevation would

seem to represent the triumph of chronological history over annals-traditions form, as

well as a triumph of continuous history over dynastic history. Indeed, Li Tao‟s 李燾

(1115-1184) Xu zizhi tongjian 續資治通鑑長編 continued Sima Guang‟s work through

the Northern Song, and other continuations would follow.

On the other hand, dynastic histories in annals-traditions style continued to be

compiled under official auspices for each subsequent dynasty. This even includes a Qing

shi 清史 [History of the Qing dynasty], whose compilation is actively continuing in the

present day. Quite simply, the annals-traditions form was too useful to discard, however

problematic its canonical pedigree. And as long as that tradition continued, the Shiji

would stand at its beginning and Sima Qian would stand as its creator.

Structure as Symbol: A Commentarial Strategy

In the previous section, I considered readers‟ views on the Shiji‟s formal structure

specifically as related to the Classics of the so-called Confucian canon. Comparing the

Shiji to the Chunqiu or to the Shangshu (in form but then also, by implication, in function)

was one way of justifying the quasi-canonical status that the Shiji was gradually

44
See discussion in Elman, “History in Policy Questions”, esp. 190-201.

143
acquiring. A few readers, however, pursued a different strategy, claiming implicitly that

the genre Sima Qian created made the Shiji into its own, new kind of Classic.

The foregoing discussions tended to understand the genre of the Shiji to be

annals-traditions form, thus primarily emphasizing only two of the five sections: the

basic annals (benji 本紀) and the arrayed traditions (liezhuan 列傳). Since these were the

only sections of the Shiji also common to every later dynastic history, they naturally

received the most attention. However, for those who wanted to elevate the status of the

Shiji specifically, as opposed to that of the annals-traditions form and the other histories

that employed it, the whole of the Shiji‟s five-part structure would have to be significant.

In particular, the number of chapters in each sub-section, together with the general nature

of their contents, was given symbolic meaning.45

There was no explicit justification for this interpretive move. It is perhaps

connected to a commentarial strategy John Henderson has identified in commentators on

the Confucian Classics: attempting to establish “systems of correspondence between the

books of the Confucian canon on the one hand and a universal moral, cosmological,

historical or mental order on the other.” This strategy enabled commentators to maintain

that “the canon was based in the nature of things… [and] was not, in other words, just a

congeries of ill-sorted fragmented ancient writings.”46 Of course the canonical status of

the Shiji was not as well-established as that of the Classics. Yet in the work of those who

would seemingly like to see it elevated at least to the status of a secondary canon, we can

in fact identify what Henderson called the two most common commentarial assumptions

45
Mark Edward Lewis also discusses this issue in Writing and Authority, 308-313, but to a very different
purpose. Thus I consider it worthwhile to revisit the evidence as part of my larger argument.
46
Scripture, Canon, and Commentary, 48.

144
about the canonical texts, namely that they are “comprehensive and all-encompassing”47

and that they are “well ordered and coherent, arranged according to some logical,

cosmological, or pedagogical principles.”48 Henderson describes how some Chinese

commentators believed that the “fiveness” of the Five Confucian classics “was not

arbitrary, but a matter of metaphysical (or perhaps numerological) necessity” while others

prefer to dwell on their “sixity.”49 The more complex numerical possibilities for the Shiji

become the subject of interpretive interest perhaps because they represented an

opportunity to show that the Shiji too possessed a profound order and completeness.

Potential symbolic meanings underlying the Shiji‟s structure are first hinted at by

Sima Qian himself. His “Self-Narration” chapter contains brief descriptions of each of

the Shiji‟s five sections, but only the one which describes the “Hereditary Households”

specifically connects the number of chapters in the section with an implied symbolic

meaning:

The twenty-eight constellations revolve around the North Star. Thirty spokes
share a single hub. They circle endlessly. Top-level ministers, the arms and legs
of the ruler, are the same as these. They loyally and sincerely carry out the true
doctrine and thereby serve the ruler. [Therefore], I make thirty “Hereditary
Households.” [二十八宿環北辰，三十輻共一轂，運行無窮，輔拂股肱之臣配
焉，忠信行道，以奉主上，作三十世家．]50

Stephen Durrant points out in his discussion of this passage that it alludes to both the

Lunyu (II:1)51 and the Laozi (ch.11).52 Perhaps it was the appealing syncretic possibilities

47
Ibid., 89.
48
Ibid., 106.
49
Ibid., 49.
50
SJ 130.3319. Trans. Durrant, Cloudy Mirror, 30.
51
“Rule by virtue can be compared to the Pole Star, that commands the homage of the multitude of stars
without leaving its place” [為政以德，譬如北辰，居其所而眾星共之] (trans. Lau, 63, slightly modified).
52
“Thirty spokes share a single hub” [三十輻共一轂].

145
afforded by such a parallel which prompted Shiji‟s emphasis on the number of chapters

here.

The Tang commentator Sima Zhen, however, took the numerological idea much

further. His version of the Shiji‟s symbolic structural meaning comprehends all five

sections:

The twelve “basic annals” represent the complete twelve-year circle of the Year-
star (i.e., Jupiter).53 The eight “treatises” having eight chapters are modeled after
the eight segments of heaven‟s seasons. The ten “tables” imitate the hard and soft
characteristics of the ten days. The thirty “hereditary households” can be
compared to a month having three [ten day] weeks. The seventy “arrayed
traditions” are taken from the retirement age for the elderly [usually seventy years
old]. The hundred and thirty chapters represent [the twelve months plus] the
intercalary period that make up a year. [觀其本紀十二，象歲星之一周，八書有
八篇，法天時之八節，十表放剛柔十日，三十世家比月有三旬，七十列傳取
懸車之暮 齒，百三十篇象閏餘而成歲。]54

In building his theoretical construct, which likens the Shiji‟s structure to the various

divisions of time, Sima Zhen does not adopt the symbolism suggested by Sima Qian

himself regarding the “Hereditary Households.” True to the goal suggested by the title of

his commentary, “Seeking the Hidden”, Sima Zhen seems to be suggesting a different

unstated purpose in Sima Qian‟s numerical scheme. There are two types of questions we

might ask about this scheme. First, was Sima Zhen (and later Zhang Shoujie as well)

merely over-reading the numbers, searching at random for some calendar-related

correspondence—or did he perceive a meaningful set of correspondences which fit

53
Sima Zhen‟s analysis here may be partially inspired by the awareness that ji 紀, which means “annals” in
a historiographic context, is also used to refer to astronomical periodicity of various types, as in the
Shangshu “Hongfan” 洪範 [Great Plan] chapter: “There are five ji: the first is the year, the second the
month, the third the day, the fourth the constellations, the fifth the calendrical calculations” [五紀：一曰
歲，二曰月，三曰日，四曰星辰，五曰厤數] (Shangshu 12.171). It is also worth noting that the Shiji
gives jixing 紀星 (Annals Star, or Mark-Star) as an alternative name for Jupiter (SJ 27.1317). I am
indebted to Daniel Morgan for these observations (personal communication, January 3, 2010). For details
on the identification of the suixing 歲星 with Jupiter, see Needham, Science and Civilisation, 3.402-3.
54
“Bu Shiji xu” 補史記序 [Supplementary preface to the Shiji], qtd. SKK 10.5323.

146
together into a larger symbolic structure? Second, did Sima Qian himself, in writing the

Shiji, intend a meaningful set of correspondences? And if so, were they the same as, or

different from, the ones his Tang commentators proposed?

It is difficult to make any overall sense of the correspondences as Sima Zhen

explained them. Perhaps this is what prompted Zhang Shoujie, whose commentary

frequently seems to be responding to Sima Zhen‟s,55 to modify and refine the symbolic

scheme. Both commentators agree that the basic annals correspond to the months of the

year,56 and that the total number of chapters, one hundred and thirty, correspond to the

months of the year with an intercalary remainder. However, Zhang Shoujie elaborated on

the next three sections, while keeping Sima Zhen‟s basic idea:

[Sima Qian] made the ten tables to represent the ten firm and yielding days of
heaven, in order to record the cycles of the successive feudal generations. He
made the eight treatises as a figure of the eight seasonal partitions of a year, in
order to record heaven and earth, the sun and the moon, the mountains and rivers,
and ritual and music. He made the thirty hereditary households as a figure of the
thirty days in a month, or as the thirty spokes joined at a single hub, in order to
record the loyalty, filiality, gains, and losses of households with hereditary
emolument, and of prime ministers and crucially important servants of the throne.
[作表十，象天之剛柔十日，以記封建世代終始也．作書八，象一歲八節，
以記天地日月山川禮樂也．作世家三十，象一月三十日，三十輻共一轂，以
記世祿之家輔弼股肱之臣忠孝得失也．]57

It is interesting to note how Zhang Shoujie attempted to reconcile Sima Zhen‟s version of

the “Hereditary Household” description with Sima Qian‟s original. In the process, a

single symbolic correspondence becomes a double one: the thirty hereditary households

are both the spokes sharing a hub and the thirty days of a month.

55
See Cheng Jinzao, “Shiji zhengyi yu Suoyin guanxi kao” [史記正義與索隱關係考] (Investigation of the
connection between the Shiji zhengyi and suoyin).
56
Though Sima Zhen had made reference to the Year-star, the implication is still connected with a twelve-
part division of the year. Zhang Shoujie instead makes specific reference to the twelve months of the year.
Perhaps Sima Zhen was reluctant to associate months (“moons” 月 in Chinese, often bearing feminine
associations) with rulers whose proper symbolic correlate—in the Tang at least—would have been the sun.
57
“Lun shi li” 論史例 [Discussion of historical form], Shiji back-matter,13.

147
 In describing the treatises, too, Zhang Shoujie joins the language of Sima Qian‟s

own description with Sima Zhen‟s symbolic concept. Sima Qian had written:

Ritual and music decrease and increase, pitches and calendrical systems are
altered and changed. Military might, mountains and rivers, ghosts and spirits, the
very boundary between heaven and man—[in all these] there is decline, continuity,
and change. [Thus] I made the eight treatises. [禮樂損益，律曆改易，兵權山
川鬼神，天人之際，承敝通變，作八書。]58

Although Zhang Shoujie‟s summary mentions things that Sima Qian‟s does not (the

heaven-earth pairing, the sun and moon) and lacks some things that Sima Qian does

mention (pitches and calendrical systems, military might, the heaven-man pairing), there

is enough in common between the two lists that we might conclude they are related.

Describing the last section, the liezhuan, Zhang Shoujie departed completely from

Sima Zhen‟s interpretation:

[Sima Qian] made the seventy arrayed traditions as a figure of the seventy-two
days in a phase [a fifth of a year]: he was expressing the fact that seventy would
be the complete number, with the remaining two days being the figure of the
intercalary remainder. [The traditions were made] in order to record kings and
lords, generals and ministers, heroes and worthies, all of whom established
accomplishments and made a name for themselves in the sub-celestial realm, and
could thus be ordered and arrayed. [作列傳七十，象一行七十二日，言七十者
舉全數也．餘二日象閏餘也，以記王侯將相英賢略立功名於天下，可序列
也．]59

Sima Zhen‟s proposal that the seventy traditions should correspond to a civil servant‟s

retirement age was intriguing because it fit the common readerly intuition that the

liezhuan were intended to be human-centered, narratives of individual lives. Zhang

Shoujie‟s alternate proposal, that (with a little room for error) the number seventy

corresponded to a fifth of a year, fit better with the calendrical model but is somewhat

more difficult to understand as a meaningful division.

58
SJ 130.3319.
59
Ibid.

148
 The table below summarizes Sima Zhen‟s scheme and Zhang Shoujie‟s alterations:

Table 1: Proposed Symbolic Structure for Shiji

Shiji Number of Sima Zhen’s Zhang Shoujie’s “Days” per

Section chapters Scheme Scheme chapter60
Basic
12 One circuit of Jupiter Months in the year 30
Annals
Tables 10 Hard and soft days Hard and soft days 36
Treatises 8 Seasons in the year Seasons in the year 45
Hereditary Days in a month OR
30 Days in a month 12
Households Spokes in the wheel
Arrayed Age of retirement (in Phase (one-fifth of a
70 5
Traditions years) year)
Twelve “decades” Twelve “decades” plus
Entire Text 130 -
plus intercalary period intercalary period

An explication of Zhang Shoujie‟s proposed symbolic correspondences is

possible, though the results prove painfully flat-footed. The whole social world of the

Shiji would be represented by a one year time span, divided up in five different ways.

(That is to say, each section represents a different way of dividing up a year.) Within

each section‟s year, the number of days allotted to each chapter represents, not a literal

measure of time, but rather the status of the chapter‟s subject within the social hierarchy.

The Sons of Heaven, as the most important people in the hierarchy, each receive thirty

days. The hereditary aristocrats, as the next most important, each receive twelve days.

Those represented by the Traditions each receive five days.61 The treatises, which deal

with social institutions, are each apportioned forty-five days. This shows that the

importance of a given institution transcends that of individual reigns or dynasties.

Similarly, the tables each have thirty-six days—exceeding the allotment for Sons of

60
Calculations based on a 360 day year with intercalary approximations where convenient.
61
Of course some chapters are concerned with individuals, others with groups or family lines. The
proposed symbolic structure is an idealization, however, and necessarily glosses over these differences.

149
Heaven—because they represent trans-historical patterns, giving the reader perspective

on how dynasties (families, individuals) fall and rise.

Some such overall scheme must have been at the back of Zhang Shoujie‟s mind

when he departed from Sima Zhen‟s interpretation of the “Traditions.” At the same time,

he was clearly uneasy about whether the scheme could actually be projected back onto

Sima Qian‟s actual intentions: this is probably why he attempted to graft paraphrases and

quotations from Sima Qian‟s own structural summary onto his explication. Sima Zhen,

on the other hand, probably did not conceive of such an overall symbolic scheme: his

proposed correspondence for the “Traditions” simply cannot be made to fit. Yet the “age

of retirement” idea is a more sensitive interpretation than most of the others: seventy

years old being the maximum age a person would typically remain in public life, it is also

the amount of time during which one has a chance to earn special merit—to deserve the

type of immortality that a Shiji biography could confer.62

It is Zhang Shoujie‟s summary that reveals what was probably real purpose

behind this symbolic theorizing—to grant the Shiji one important characteristic of any

canonical text, that of being a complete figure for the world, with nothing lacking and

nothing superfluous:

The Honorable Senior Archivist made these five grades—and not a single one can
be discarded—to [trace] the unifying principles of heaven and earth, to exhort and
awaken and admonish and warn, and as a model for those who would come after.
[太史公作此五品，廢一不可，以統理天地，勸嚔箴誡，為後之楷模也．]63

62
The so-called ethnographic chapters—the “Xiongnu liezhuan” 匈奴列傳 (SJ ch.110), “Nanyue liezhuan”
南越列傳 (SJ ch.113), “Dongyue liezhuan” 東越列傳 (SJ ch.114), “Chaoxian liezhuan” 朝鮮列傳 (SJ
ch.115), “Xinan yi liezhuan” 西南夷列傳 (SJ ch.116), “Dayuan liezhuan” 大宛列傳 (SJ ch.123)—are as
usual an awkward exception, but they have long held a peripheral position in the minds of the Shiji‟s
traditional readers, at least as regards considerations of form.
63
“Lun shi li” 論史例 [Discussion of historical form], Shiji back-matter, 13.

150
It is worth emphasizing Zhang‟s claim that each of the five sections had a vital part to

play. Of the histories that succeeded the Shiji and took its basic form as a model, not one

retained all five parts (the most frequent omission being the problematic “Hereditary

Household” section). To make Sima Qian the true founder of the annals-traditions

form—and not just an imperfect precursor to Ban Gu—it was necessary to show that it

was Sima Qian’s conception that was complete.

In actually reading the Shiji text, one quickly discovers that the kind of numerical

correspondences Sima Zhen and Zhang Shoujie wanted to find are mostly absent. The

idea that some chapters appear in one section rather than another simply to make up the

correct number is a possible explanation for some curious anomalies. On the whole,

though, the symbolic scheme is far less satisfying to contemplate than the section by

section interpretations discussed below. It is not impossible that Sima Qian (or even

perhaps some later editor) chose the number of chapters in each section with care, to at

least create the appearance of a significant scheme.64 But it is probably safe to say that

Sima Zhen‟s and Zhang Shoujie‟s efforts represent a commentarial over-reading of the

structure.

Neither did this circumstance go unnoticed. The cantankerous Jin scholar Wang

Ruoxu 王若虛 (1174-1243) reacted against both the sense that the symbolic readings

were forced and the grandiosity of the Shiji commentators‟ implied claims for the text:

The texts of the [Lun]yu and Meng[zi] originally had no ordering of their
chapters.65 Yet some simple-minded people make up forced arguments about it.

64
A similar issue arises with the numbers which appear in the titles of several of the Tables, and the
headings of the Tables generally. See below.
65
In modern Chinese pianci means “table of contents”, and although this is an anachronistic translation of
Wang Ruoxu‟s language, he probably did intend something like “a deliberate overall scheme of ordering
chapters,” both in terms of the chapters‟ ordering and relation to one another, and in terms of larger, multi-
chapter sub-structures.

151
 This is already not worth taking into account. Now Sima Zhen, in transmitting
the Shiji, believed that the twelve “basic annals” were the figure of one revolution
of the Year-star, [etc.]… —an absurd idea, a very strained interpretation. To dare
to suggest this sort of thing, is it not extreme indeed!? [《語》、《孟》之書，
本無篇次，而陋者或強論之，已不足取。司馬貞述《史記》，以為十二本紀
象歲星之一周，…妄意穿鑿，乃敢如此，不已甚乎？]66

Wang Ruoxu was an exceedingly harsh critic of the Shiji. When he criticized Sima Zhen

for daring to come up with a symbolic structural scheme, his primary concern is not that

it is a betrayal of Sima Qian‟s intent. Instead, he implies that the symbolic claims were

made in imitation of similar claims on behalf of the Lunyu and Mengzi (it is difficult to

ascertain whether this is accurate or not). Bad enough, Wang implies, for such

presumptuous ideas to attach to the texts of revered Confucian thinkers, but to ascribe

such things to a mere historian like Sima Qian—who was also a morally suspect servant

to his dynasty—well, that was simply going too far.

As I suggested above, Wang Ruoxu‟s basic understanding was probably correct.

From a textual historian‟s point of view, however, his argument is weak. Both the Lunyu

and Mengzi show signs of being selected texts, collections of sayings that originated in

various chronological periods, almost certainly not put together by the putative speaker

himself. The Shiji, on the other hand, was self-consciously produced as a text, complete

with “table of contents” and word count. It would be more likely than the Lunyu or

Mengzi to have some kind of overall intended structure—not less.

INTENTION AND INVENTION IN THE SHIJI‟S FIVE SECTIONS

Up to this point, I have discussed the form of the Shiji as a whole, and how

readers tried to relate it to other texts or structures. There is also room for a finer-grained

66
Hunan Yilao ji 31.194.

152
discussion, however, of how each of the five separate sections of the Shiji were seen as

individually relating to one another and to other texts. I should begin, though, by

considering the question, what do I mean by form or genre in the context of the

individual sections of the Shiji? The issue is non-trivial, especially when considering the

Chinese discursive realm (both traditional and modern) whose categories do not line up

exactly with these words in English. The concept I am referring to when I use the words

“form” or “genre” is most often referred to in modern Chinese as tili 體例 (stylistic rules,

layout), or sometimes also tizhi 體制 (system of organization).67 In literary Chinese it is

usually just ti 體, but there seem to be some exceptions (such as Liu Xie‟s use of shi 式,

discussed below). As much as possible, I attempt to distinguish stylistic influence from

the borrowing of material (qucai 取材). It is a truism that “form follows content”;

content will inevitably have some part to play in my discussion. However, where the

Shiji‟s material came from, and why it was selected, are separate issues which I will not

attempt to discuss thoroughly here.

A less obvious distinction should also be made: between title (mu 目 or mingmu

名目) and genre. It is important to understand the influence of a long-standing Chinese

tradition with regard to text titles, namely, that the token (i.e., title of an individual work)

very frequently contains an indicator of the type as well. For example, the first of the

Shiji‟s standard commentaries is Pei Yin‟s Shiji jijie 史記集解 [Collected explanations of

the Shiji]. This type also includes He Yan‟s 何晏 (d.249) Lunyu jijie 論語集解

[Collected explanations of the Lunyu], Du Yu‟s 杜預 (222-284) Chunqiu jingzhuan jijie

67
The semantic overlap between the two terms is considerable. Cheng Jinzao 程金造 (e.g., Shiji guankui,
1 ff.) used the term tili 體例, while Zhang Dake prefers tizhi 體制 (e.g., Shiji yanjiu, 191 ff.; Shiji wenxian
yanjiu, 130 ff.), but the topic of discussion in both cases is the same.

153
春秋經傳集解 [Collected explanations of the Chunqiu classic and its commentaries], and

many others. These commentaries have in common the genre (or sub-genre)

characteristic of being anthologistic, and this is reflected by the shared element in their

titles, “collected explanations”. (By contrast, the third of the standard Shiji commentaries,

Zhang Shoujie‟s Shiji zhengyi 史記正義 [Correct explanation of the Shiji], belongs to a

different sub-genre of commentary, that of “correct meanings”.) Different readers might

debate what is intended or expressed by the typological element of a title, but none would

deny that to use such an element constitutes a claim that the work belongs in a certain

formal category.68

Because of this tradition, it is tempting to see a one-to-one relationship between

title-types and genres (or sub-genres): the invention of a new title constitutes the

invention of a new genre (or sub-genre), and invention of a new genre would be marked

by the use of a new title-type. This over-simplified picture may capture some aspect of

the later Chinese literary tradition, but remains problematic for early periods (i.e., the

Han and before). The Shiji pre-dated explicit, self-conscious genre theory. The text itself

contains evidence that Sima Qian had begun thinking about the problem of historical

genres, and this in itself was a type of innovation. But the Shiji‟s own discussion is

neither systematic nor detailed. Available evidence suggests that both title-types and the

genres to which they referred were in a state of flux. Thus, while it would be over-

rigorous to ignore completely the evidence given to us by title-types, it would also be

68
For a contemporary example, one might consider the differing claims made by scholarly studies whose
titles end in lun’gao 論稿 [Preliminary argument regarding…], tanyuan 探源 [Origins of…], yanjiu 研究
[Research on…], and xinzheng 新證 [New evidence about…], etc.

154
rash to assume that because two works share a title-type it implies that the two works in

question have the same form.

With this caveat in the background, I will turn to the difficult, even impossible,

question: to what extent should the various parts of the Shiji be seen as Sima Qian‟s

innovations? A number of scholars have certainly been concerned with the problem,

which takes on different aspects for each of the five sections. As is often the case when

available evidence leaves the answer under-determined, various theories clearly reflect

the biases of their proponents but should not on that account be ignored.

Benji 本紀

There are twelve basic annals. The first (SJ 1) deals with five mythical emperors

of highest antiquity. The next four (SJ 2-5) each record the initially approximate but

increasingly precise chronologies of a separate dynastic ruling house: i.e., the Xia, Shang,

Zhou, and pre-unification Qin. The sixth chapter also concerns the Qin, but is

specifically devoted to the First Qin Emperor, his successor, and the collapse of the Qin

imperium. The seventh and eighth chapters both narrate the period of the Chu-Han

conflict, but from the different perspectives of the two major contenders for power, Xiang

Yu and Liu Bang (respectively). The ninth chapter narrates the rise and fall of Liu

Bang‟s wife Empress Lü. Finally, the last three chapters (SJ 10-12) are devoted to the

post-Lü restoration Emperors Wen, Jing, and Wu.

As discussed in the last chapter, the basic annals section can be interpreted as

providing a complete chronicle of the entire period covered by the Shiji.

155
 It has often been pointed out that the Shiji “Arrayed Traditions of Dayuan” 大宛

列傳 (SJ 124) quotes from a text called Basic Annals of Yu 禹本紀.69 On this evidence,

Zhao Yi 趙翼 (1727-1814) argued that “prior to the Han, there was a separate text called

the Basic Annals of Yu, and it is exactly this upon which [Sima] Qian based [his own

basic annals]” [漢以前別有禹本紀一書，正遷所本耳].70 Opinions have been divided

regarding the correctness of Zhao Yi‟s conclusion. Some scholars, like Takigawa

Kametarō, have concluded that “the title-type „basic annals‟ had been in existence since

antiquity” [本紀之目，自古有之].71

Zhang Dake discounts any relationship between the Basic Annals of Yu and the

Shiji “Basic Annals,” arguing that

In fact the Basic Annals of Yu were mentioned by Sima Qian together with the
Classic of Mountains and Seas, and [Sima Qian] points out that these are texts
that discuss records of strange matters. It is only that it has the name “basic
annals”; actually it has nothing in common with the “basic annals” that transmit a
record of the affairs of emperors and kings. [其實司馬遷所言《禹本紀》與
《山海經》相提並論，指出是言志怪之書，僅有《本紀》之名而已，與載述
帝王事跡的《本紀》風馬牛不相及。]72

69
The Shiji passage is as follows:
“The Honorable Senior Archivist said: In the words of the Basic Annals of Yu, „The [Yellow] River comes
out of the Kunlun [mountains]. The height of the Kunlun [mountains] exceeds 2500 li, and this is where
the sun and the moon hide [in turn while the other] is shining its light. Upon these mountains are the
Sweetwine Spring and the Jasper Pool. Now Zhang Qian has been sent as an envoy to the Daxia and
explored the source of the [Yellow] River. Where did he see any trace of what the Basic Annals call the
Kunlun?! Thus in speaking of the mountains and streams in the nine prefectures, the Shangshu is closest to
[the truth]. As for the strange things found in the Basic Annals of Yu and the Classic of Mountains and
Seas, I do not dare speak of them.” [太史公曰：禹本紀言「河出崑崙．崑崙其高二千五百餘里，日月
所相避隱為光明也．其上有醴泉﹑瑤池」．今自張騫使大夏之後也，窮河源，惡睹本紀所謂崑崙者
乎？故言九州山川，尚書近之矣．至禹本紀﹑山海經所有怪物，余不敢言之也．] (SJ 123.3179)
70
Gaiyu congkao 陔餘叢考 5.86.
71
Shiki kaichū kōshō (SKK) 10.5323. Fang Bao, Fan Wenlan, and Cheng Jinzao also use this reference to
argue against the notion that Sima Qian created the “basic annals” form. See “Zhengshi kaolue”, Fan
Wenlan quanji 2.16.
72
Shiji yanjiu, 191-192.

156
In short, according to Zhang Dake, the similarity between the Basic Annals of Yu and

Sima Qian‟s “Basic Annals” would have been in title-type only.

Jin Dejian‟s 金德建 search for other citations of the Basic Annals of Yu shows

that, while various writers do mention a text that might be the same or related, it is

referred to as Yu shou di ji 禹受地記 [Record of Yu‟s receipt of land], Yu da zhuan 禹大

傳 [The great tradition of Yu], or (in the Hanshu “Yiwenzhi”) as Da Yu 大 [Great

Yu]—but nowhere else as Yu benji 禹本紀.73 This suggests that the title-type benji was

not an integral aspect of the “story of Yu” text-family (if a single family it was), and that

the content was, as Sima Qian‟s citation suggests, more geographical than annalistic in

nature. To Cheng Jinzao, the difference in title-type proves that “I am afraid that these

[others] are not the text which the Honorable Senior Archivist referred to as the Basic

Annals of Yu” [此恐非太史公所謂《禹本紀》之書].74 Yet as both Jin Dejian and

Zhang Dake convincingly argue, the quotation from the Basic Annals of Yu found in the

Shiji text sounds more like the type of content found in the Classic of Mountains and

Seas—both in terms of geographical information and strange or exotic things found in

such locales—than it resembles the Shiji‟s own basic annals.

One early suggestion is that the Shiji‟s twelve “basic annals” 本紀 were inspired

by the Lüshi chunqiu‟s 呂氏春秋 [Spring and autumn of Master Lu] twelve “almanacs”

紀.75 This suggestion seems to have originated with Liu Xie 劉勰 (ca. 465-520), who

wrote “the Senior Archivist Tan, taking the Lü lan as a model, linked them with the

73
Sima Qian suo jian shu kao, 408-409. Yan Shigu glosses as an archaic form of 禹. See also HS
30.1742.
74
Shiji guankui, 12.
75
Also referred to in Chinese as yueji 月紀. In translating the Lüshi chunqiu‟s ji 紀 as “almanacs” here, I
follow the convention of Knoblock and Riegel‟s complete translation (2000).

157
appellation ji” [太史談取式《呂覽》，通號曰紀].76 Zhao Yi, however, pointed out the

obvious fact that “the twelve monthly almanacs of the Lü lan are not specifically accounts

of the affairs of emperors and kings”—that being the primary attribute of the Shiji basic

annal chapters.77 It is true that the similarity would seem limited to name alone, that the

title-type ji 紀 was perhaps still flexible in Sima Qian‟s time, probably even more so in

Lü Buwei‟s: as Zhang Dake argued, expanding on Zhao Yi‟s point, “the individual

sections of the Lüshi chunqiu78 are all short essays; it is just that they are called by

different names, not at all that they differ from one another in their basic substance” [呂

氏春秋的各體都是短篇的論文，只是名稱不同，并無本質的區別].79

Liu Xie cannot have been unaware of these facts. One wonders why, then, he

even made the connection? If we put his remark back in context, we can see that he

viewed the issue of how the Honorable Senior Archivists80 should title the section as

being a problematic and difficult one:

When Han annihilated the Ying and Xiang,81 it was a martial accomplishment of
several years. Lu Jia looked to antiquity and made the Chu-Han Chunqiu.
Thereafter, Senior Archivist Tan grasped the bamboo slips as his hereditary [duty].
Zichang [=Sima Qian] carried on [Sima Tan‟s] intention, examining and putting
in sequence the accomplishments of emperors. If he had called these dian in
comparison with Yao, then the ranks of rulers would be intermixed with mere

76
WXDL 16.573. Cf. a similar but more extensive comparison by Zhang Xuecheng 章學誠 (WSTY 702-
705).
77
Jin Dejian was more outspoken, complaining that Liu‟s theory is “just a fantasy, completely without any
evidence to back it up” [毫無根據的想像罷了] (Sima Qian suo jian shu kao, 410). Zhang Dake also
dismisses the comparison (Shiji yanjiu, 193).
78
I.e., the Ji 紀 [Almanacs], Books 1-12; the Lan 覽 [Examinations], Books 13-20; and the Lun 論
[Discourses], Books 21-26.
79
Shiji yanjiu, 193.
80
It is clear from Liu Xie‟s discussion that he considered Sima Tan to be the inventor of the genre. This
opinion would be echoed by Fang Bao 方苞 (1668-1749) (see discussion in Zhang Dake, Shiji yanjiu, 54-
59).
81
I.e., the Qin dynasty (Ying was the hereditary name of the Qin ruling house; Qin Shihuang was also
known as Ying Zheng 嬴政) and Xiang Yu 項羽 (possibly also including the uncle who raised him, Xiang
Liang 項梁).

158
 worthies. If he had titled them jing in imitation of Kongzi, then [the problem is]
that his writings were not those of the Mysterious Sage. [漢滅嬴項，武功積年、
陸賈稽古，作《楚漢春秋》；爰及太史談，世惟執簡；子長繼志，甄序帝
勣．比堯稱典，則位雜中賢；法孔題經，則文非玄聖．]82

Lu Jia had called his text a chunqiu, but Sima Tan could not—the reason being, in Liu

Xie‟s view, that it would have been presumptuous to imply a comparison with the Sage.83

Nor could he call his work dian in imitation of the Shangshu: many of the rulers he wrote

about were in no way worthy of a title-type that was firmly associated with the sage kings

of old. It was for this reason, according to Liu Xie, that he used the Lüshi chunqiu‟s

pattern and made ji.

In connecting the Shiji and the Lüshi chunqiu, the word Liu Xie used was not ti 體,

but shi 式. It is hard to know whether or not Liu Xie understood the two words as

roughly synonymous. It is possible that he meant shi as a more superficial aspect of form,

perhaps even referring merely to title or title-type. If that was the case, we could

understand his analysis to be: as far as form is concerned, the Shiji‟s benji closely

resemble Lu Jia‟s Chu-Han Chunqiu. However, because of intertextual complications to

which Sima Tan and Qian were especially sensitive, they borrowed the title-type from the

Lüshi chunqiu.

Another potential influence on the form of the benji chapters was probably the

now-lost text called Shiben 世本 [Geneaological origins]. Though readers from Ban

Biao on have agreed that the Shiben was a source for the Shiji, Qin Jiamo 秦嘉謨 has

82
WXDL yizheng 16.573.
83
Although by Han times “chunqiu” was already an established title-type used to refer to historical texts, to
use this title for a work that also duplicated the time period chronicled by the sagely Classic would perhaps
have been seen as inappropriate.

159
made the farthest-reaching claims regarding the relationship, explicitly suggesting that

the ben 本 in both Shiben and benji should be taken as related, and that the title-types

from other Shiji sections were also derived from the Shiben.84 It does seem possible,

given what we know of the Shiben‟s contents.85 The Shiben appears in the Hanshu

“Yiwenzhi” as having fifteen chapters; the Hanshu treatise also notes: “The official

archivists of antiquity recorded the various feudal lords and great officers from the

Yellow Emperor down to the Spring and Autumn” [古史官記黃帝以來訖春秋時諸侯大

夫].86 Sima Zhen, citing Liu Xiang, gives a slightly variant description, adding that “it

records the ancestry, posthumous names, personal names, and sobriquets for emperors

and kings, feudal lords, down to ministers and officers, ever since the time of the Yellow

Emperor” [錄黃帝已來帝王諸侯及卿大夫系謚名號].87 The chronological range seems

significant, since it matches at least one of the Shiji‟s explicit beginning points.88

Furthermore, the “Basic Annals” of the Shiji seem to contain numerous quotations from

the Shiben. While the Shiben was probably a major source—and possibly stylistic

influence—for the benji section of the Shiji, it clearly did not approach the level of

narrative detail for which the Shiji‟s first twelve chapters are justifiably appreciated.

Finally, there being twelve “basic annals” naturally suggests a connection with the

twelve dukes of Lu, who provide the chronological framework for the Chunqiu. Fan

Wenlan states this outright, writing, “the “basic annals” are twelve [in number] because
84
See Shiben jibu “Zhu shu lun shu” 諸書論述 [Various texts discuss and transmit (the Shiben)] 1b: “Note
that the Writings of the Honorable Senior Archivist selects from the Shiben, and in his creating and
establishing title-types for his chapters—such as „basic annals‟, „hereditary households‟, and „arrayed
traditions‟—for all of them he relied on the Shiben” [按太史公書采世本其創立篇目如本紀如世家如列
傳皆因世本].
85
For further arguments and a reconstruction of the text, see Qin Jiamo‟s Shiben jibu.
86
HS 30.1714.
87
Shiji jijie “Preface”, Sima Zhen‟s note (5).
88
See SJ 130.3300.

160
in fact they are modeled after the twelve dukes of the Spring and Autumn” [本紀十二，

實效法《春秋》十二公而作].89 Liu Zhiji, as mentioned above, had already implied a

connection with the Chunqiu in suggesting that “the „annals‟ as a genre are like the

„canon‟ of the Spring and Autumn” [紀之為體，猶春秋之經].90 The connection is not

based merely on the coincidence of the numbers, but also on the form of a continuous

year-by-year record of events in a reign or dynasty, common to the Spring and Autumn

and at least the latter eight “basic annals” of the Shiji.

Almost as soon as readers began to discuss their understanding of the five

sections, there also appeared complaints and arguments about what chapters did and did

not fit the form. Sima Qian was perceived to have broken his own rules (poli 破例)—at

least as these rules were conceived by later readers. Some readers dismissed the

irregularities in the Shiji as evidence of carelessness, or, more generously, lack of a fully

thought out formal system. Other readers saw these genre aberrations as having profound

and subtle meaning, a fully intended message from an author who bore creative

responsibility for the form as well as the content.

Of the “basic annals”, the most problematic is clearly the “Basic Annals of Xiang

Yu” (SJ ch.7). Xiang Yu, who contended with Liu Bang for the empire after the collapse

of the Qin, did not hold power long enough or thoroughly enough to lay claim to the kind

of legitimacy possessed by other rulers chronicled in the “basic annals.” Why then is he

found in the “basic annals” section? Ge Hong suggested that “Xiang Yu is included in

the basic annals because [Sima Qian] considered that being placed in a high position did

89
“Zhengshi kaolue”, 16.
90
Shitong 2.37. The fruitful comparison Liu drew in this passage will be discussed in more detail below.

161
not have any connection with one‟s virtue” [項羽列於本紀，以為居髙位者非關有徳

也].91 This may be, but it does not explain why Sima Qian considered the position Xiang

Yu did obtain to merit an annal.

Liu Zhiji was particularly adamant in his objection to the Xiang Yu chapter, not

merely because Xiang Yu did not seem to merit an annals, but because in his view the

chapter itself was formally unsuitable for the section in which it had been placed.

Sima Qian wrote a “Traditions” chapter for King Xiang, but used for it the name
of “Basic Annals.” It is not only that Xiang Yu was a usurpatious thief, who
ought not be [considered] in the same [category] as the Sons of Heaven. If you
further extend [your consideration] to its narration of events, [the Shiji “Xiang
Yu” chapter] is all the phrased as a “Traditions” chapter. If you seek to consider
it an annal, you simply cannot do it. Someone said, When [Sima] Qian made the
annals of the Five Emperors, the Xia, and the Yin, these are also an arrangement
of events and nothing more. You never thought there was anything strange about
that. How is it that you only find fault with the Xiang annals? I responded, It is
not so. Now, the Xia and Yin in relation to the Five Emperors, are calendrically
diachronic, inheriting from one another, [similar] passing it down from son to
grandson. Though [full information] about the chronology is not always available,
what harm is there in making an annals?! But as for someone like Xiang Yu, his
affairs arise from what remained of the Qin, and he himself died at the beginning
of the Han…. Now the Xia and Yin annals do not draw in other matters. When
[Bo] Yi and [Shu] Qi remonstrated with the Zhou, though it was in truth taking
place in the days of [the Yin ruler] Zhou, it is separated off into an arrayed
tradition, and not put in the chapter on the Yin. In the case of the Xiang annals,
superiors and inferiors are recorded side by side; ruler and minister are
interspersed and muddled. It has the name “annals” but the form of a “traditions,”
and thereby becomes downright laughable. [如項王立傳，而以本紀為名，非唯
羽之僣盜，不可同於天子；且推其序事，皆作傳言，求謂之紀，不可得也。
或曰：遷紀五帝、夏、殷，亦皆列事而已。子曽不之怪，何獨尤於項紀哉？
對曰：不然。夫五帝之與夏、殷也，正朔相承，子孫遞及，雖無年可著，紀
亦何傷！如項羽者，事起秦餘，身終漢始。。。 且夏、殷之紀，不引他
事。夷、齊諫周，實當紂日，而析為列傳，不入殷篇。項紀則上下同載，君
臣交雜，紀名傳體，所以成嗤。]92

91
Xijing zaji 4.125.
92
STTS 2.46-47.

162
Liu Zhiji placed more emphasis on the stylistic issue, and tended to minimize the

importance of the gap between the Qin and Han.

This gap was of primary concern in Huangfu Shi‟s explanation of both the Xiang

Yu annals and another problematic chapter in the annals section, the “Basic Annals of

Empress Lü” (SJ ch.9). If having an annal is a mark of dynastic legitimacy, how would it

be possible to justify Empress Lü—a woman, and not a Liu family heir—being given one?

Huangfu Shi proposed an explanation:

When Master Sima made the annals, he took Xiang Yu as the inheritor of Qin,
and he took Empress Lü as the continuer of [the Han dynasty]. He also
considered that in the succession of years one could not discard any of them, that
the timeline could not suffer a gap. That is why he wrote [as he did]. [司馬氏作
93
紀，以項羽承秦，以吕后接之，亦以歴年不可中廢，年不可闕，故書也。]

Wang Yun 王惲 (1227-1304) expanded on Huangfu Shi‟s idea, writing:

When the Qin had already fallen but the Han had not yet been established, in all
the realm no one was able to unite and order it. Yet [the realm] cannot be without
a ruler for a single day. Moreover, in the matter of enfeoffing and establishing
kings and lords, the governance proceeded from [Xiang] Yu. If one sets aside
Xiang Yu, who would be the ruler? [方秦已亡，漢未立，天下莫有收屬，不可
一日無君，況封建王侯，政由羽出，舍羽孰主哉？]94

Wang Yun mentioned as an auxiliary point how “governance proceeded from Xiang Yu”,

perhaps referring to a theory by Lin Jiong 林駉 (Southern Song, dates unknown).95 This

explanation, that in the Shiji, annals are assigned through consideration of de facto rather

than de jure power, would later become the favorite one. Liu Xianxin 劉咸炘 (1896-

1932), for example, wrote,

93
“Biannian jizhuan lun,” in Chizheng wenji 2.129.
94
“Qian Gu ji zhuan bu tong shuo” 遷固紀傳不同說 [Arguments regarding the differences between Qian‟s
and Gu‟s annals and traditions], in Qiujian xiansheng daquan wenji 秋澗先生大全文集 [The great
complete writings of Master Qiujian], 45.462.
95
His theory states essentially the same point, that because Xiang Yu had the power to parcel out land, he
deserved a annal. See SJYJJC 6.109.

163
 The basic annals are the outline for the entire work, and are concerned with where
the strategic power of the age was collected. There are no distinctions, therein,
among king, lord, emperor, or empress. Thus when the Senior Archivist created
the genre, both Xiang Yu and Empress Lü are given annals. [本紀者一書之綱，
惟一時勢之所集，無擇於王、伯、帝、后。故太史創例，項羽、呂后皆作
紀。]96

It is difficult to say whether considerations of actual power were the original motivating

factor for Sima Qian‟s assignment of the annals. A reader less favorably inclined might

suspect that he chose the subjects as he did merely to make up the symbolic number

twelve. But while Sima Qian‟s idiosyncratic choices were not emulated by later histories,

they were observed with a certain interest and admiration, and read as potentially bearing

a deeper message.

Biao 表

Among the Shiji‟s five sections, the one whose predecessors are most clearly

acknowledged is the “Tables” 表 section. In the small preface to the first of the Tables

(Shiji ch.13), Sima Qian wrote,

Being that the Three Dynasties are of high [antiquity] indeed, the chronicle of
their years cannot be investigated. I have for the most part selected from
genealogies and old stories. Based on these, I have sketched out and extrapolated
from them, and thus made the First Table of Generations of the Three Dynasties.
[維三代尚矣，年紀不可考，蓋取之譜牒舊聞，本于茲，於是略推，作三代
世表第一。]97

It is clear that these “genealogies” (譜牒) are an independent source used in the

compilation of the Shiji, and not (as is the case with the mention of shijia 世家, discussed

below) a reference to the Shiji chapter itself. The pre-existing genealogies seem to have

96
SJYJJC 6.111.
97
SJ 130.3303.

164
provided both an inspiration for the form and the information that made up the content.

As Huan Tan 桓譚 (ca.43 BCE-28 CE) wrote, noting the similarity between the first table

and genealogical sources that presumably still existed in his time: “The Honorable Senior

Archivist‟s „Table by Generations of the Three Dynasties‟ is written crosswise and up

and down, parallel to and imitating the genealogies of the Zhou” [太史公三代世表旁行

邪上，並效周譜].98

The small preface to the “Third Table” (Shiji 15) also mentions these genealogical

documents: describing the political complexity of Zhou‟s declining years, the preface

complains that “the various lords seized control of governance, and there are things

which the Spring and Autumn does not chronicle” [諸侯專政，春秋有所不紀]. But, it

adds, “the genealogies give a brief outline” [譜牒經略].99

In addition to the pudie 譜牒 texts, which I have translated above as

“genealogies,” Shiji chapter 13 mentions dieji 諜記, which I have translated as

“genealogies and records.”100 These probably form a slightly larger category, one which

included, but also extended beyond, the pudie 譜牒. Other sources Sima Qian mentions

consulting include “the calendrical charts and the succession101 of recurrent cycles of the

five Virtues” [曆譜諜、終始五德之傳] (which he says were all dated but not mutually

98
Cited in Liangshu 50.716.
99
SJ 130.3303. The word jinglue 經略 is potentially ambiguous. In my translation I have followed Han
Zhaoqi‟s gloss that jing 經 in this context means “system” 統緒 or “outline” 綱領. He explains the phrase
as a whole by saying that “Texts of the genealogical type have only an outline, so that recording and
narration becomes very simple” [譜牒之類的書光有一个綱領，記述更為簡單] (Shiji jianzheng 9.6386).
100
In this I have followed Sima Zhen‟s understanding. He seems to understand die 諜 as close in meaning,
if not identical, to die 牒, which he defines as “a document which chronicles the posthumous names of
ancestors” [紀系謚之書也] (SJ 13.488).
101
Following Sima Zhen‟s comment (SJ 13.488, nt.2) I read 傳 as chuán, “to pass down or transmit”, rather
than as zhuàn, “commentary” or “tradition”.

165
consistent in their dating); the “Ancestral Connections of the Five Emperors” 五帝繫

諜;102 and the Shangshu.103 Sima Zhen‟s note on this passage explains:

The Da Dai Li[ji] has chapters on “The Virtuous Power of the Five Emperors”
and “Linear Sequence of the Emperors.”104 Probably the Honorable Senior
Archivist took the charts from these two chapters and the Shangshu, assembled
their information and chronicled in a systematic table the reigns from the Yellow
Emperor onward. [大戴禮有五帝德及帝繫篇，蓋太史公取此二篇之諜及尚
書，集而紀黃帝以來為系表也．]105

The “Table by Years of the Twelve Feudal Lords” draws heavily on the canon of

the Spring and Autumn, both in scope and content. As Mao Kun wrote, “Kongzi made

the Spring and Autumn, and the Honorable Senior Archivist was able to rely on it to

compile into a table the traces of the twelve feudal lords‟ roots and branches, their

flourishing and decline” [孔子作春秋，而《太史公》得因之以表十二諸侯本末盛衰

之跡也].106 However, the preface to this table suggests that there were sources even

more closely related to what would become the third Shiji table: “When the Honorable

Senior Archivist read the Spring and Autumn [period] calendrical charts and documents

and came down to King Li of Zhou, he never failed to cast away the text with a sigh” [太

史公讀春秋曆譜諜，至周厲王，未嘗不廢書而歎也].107

What is particularly interesting about this table, however, is that its preface goes

on to criticize its various precursors, and to explain how this table differs from them:

The Honorable Senior Archivist said: Those who teach the Classics [are too
concerned with] moral judgments on [historical events], while galloping
persuaders get carried away with their rhetoric; neither strive to bring together
recurrent cycles. Calendrists select their years and months, while numerologists
102
My translation follows that of Mémoires historiques I.cxlii, n.224.
103
SJ 13.488.
104
Chapter 62 and 63 in today‟s recension of the Da Dai Liji.
105
SJ 13.488.
106
SJYJJC 6.314.
107
SJ 14.509.

166
 glorify the divine progressions. However, a genealogical chart only records
generations and posthumous names. Its words are brief, for the hope is to see in
one glance the various essentials and difficulties. Therefore did I chart the twelve
feudal lords, from the Gonghe regency (841-828 BCE)108 down to Kongzi. The
table shows clearly that which is criticized by the learned of the Spring and
Autumn and Guoyu. [Also] the great principles behind flourishing and decline are
broadly laid out in this chapter, created as a distillation for the sake of those who
would study and master the ancient writings. [太史公曰：儒者斷其義，馳說者
騁其辭，不務綜其終始；曆人取其年月，數家隆於神運，譜諜獨記世謚，其
辭略，欲一觀諸要難．於是譜十二諸侯，自共和訖孔子，表見春秋、國語學
者所譏，盛衰大指著于篇，為成學治古文者要刪焉．]109

Essentially, Sima Qian110 hoped in this table to unite the brevity and clarity of the

genealogies he had access to with the profundity of the Spring and Autumn, at the same

time omitting the heavy-handed emphasis on moralizing that he seems to be criticizing in

the Classics teachers. The mention of Guoyu is clearly meant to correspond with the

“galloping persuaders” (as the Spring and Autumn corresponded with the Classics

teachers). I would argue that it is brought in here to emphasize the value of geographical

organization, which the Guoyu employs, though as a text the Guoyu was considered to be

marred by its excessive emphasis on rhetoric. In short, Sima Qian claimed that the value

of his table went beyond that of the Spring and Autumn Classic, since the explicit

geographical focus of that work was the state of Lu with other states in some sense taking

a subsidiary role.

Something can be said about the sources for the first three tables of the Shiji.

There is much less comment on precursors for the other tables. Fan Wenlan noted that

The Hanshu “Bibliographical Treatise”, in the School of the Calendrists [section],

has “Calendrical Charts of Han Reign Periods [back through] the Yin and

108
The Gonghe regency marks the first solid absolute date in Chinese history, namely, 841 BCE when the
Zhou King Li (r. 857 BCE-841 BCE, d. 828 BCE) was driven out of the capital and into exile. Lord He 和
of the state of Gong 共 served as regent during this period. See Cambridge History of Ancient China, 344-
345.
109
SJ 14.511.
110
Or perhaps Sima Tan?

167
 Zhou”111 (17 chapters), “Genealogies of Emperors, Kings, and Various Lords”
(20 chapters), “Annual Charts of the Emperors and Kings Since Antiquity” (5
chapters). [漢書藝文志曆譜家有漢元殷周諜曆十七卷帝王諸侯世譜二十卷古
來帝王年譜五卷]112

Fan uses this as part of his evidence that the biao 表 genre as seen in the Shiji had come

down from antiquity. It is not clear, however, whether the above-mentioned works pre-

date the Shiji. The first, in particular, seems closely related to the politicized issue of the

Han genealogy, which did not reach its height until decades after Sima Qian‟s death. 113 It

is probable that the “Month Table of Chu-Han” (SJ 16) owes something to Lu Jia‟s

work—perhaps envisioned as standing in similar relation to it as the “Twelve Feudal

Lords” (SJ 14) table does to the Chunqiu. For the other tables, commentarial interest in

their formal precursors has been less pronounced.114

Shu 書

As regards the treatises (shu 書), the only predecessors commentators assign to

them are the Classics: Liu Zhiji suggested that both Sima Qian‟s and Ban Gu‟s treatises

“for the most part imitated the Ritual Classics” [多效禮經],115 while Sima Zhen

comments that the word shu is “a general name for the Five Classics and Six Records”

111
Chen Guoqing 陳國慶 notes that yuan 元 here should be understood as jiyuan 紀元 (reign period).
Shen Qinhan 沈欽韓 (1775-1832) speculated about this work, “This takes the Han reign periods and
extrapolates back through the Yin and Zhou dynasties. It is like the „quarterly calendar‟ [devised during
the reign of Emperor Zhang] starting with the third year of Filial Emperor Wen‟s Houyuan reign, at which
point the year was gengchen; forty five years before, when the year was yiwei, it was the first year of the
Han dynasty” [此以漢元上推殷、周，猶四分曆起於孝文皇帝後元三年歲在庚辰。上四十五歲，歲在
乙未，則漢興元年也] (Hanshu Yiwenzhi zhushi huibian, 205).
112
Zhengshi kaolue, 17.
113
See Gopal Sukhu “Yao, Shun, and Prefiguration,” 91-153.
114
For an intriguing discussion of the Shiji tables as related to excavated texts with similar properties, see
Griet Vankeerberghen, “The Tables (biao) in Sima Qian‟s Shiji”, esp. 295-301.
115
STTS 3.56-57.

168
[五經六籍總名也].116 If one considers Liu Zhiji‟s chapter on this form, as well as the

later progress of historical writing in China, both show that the genre of the treatise

experienced its most significant developments after the Shiji, and that Sima Qian was

always considered the originator (if sometimes a highly imperfect originator) of the form.

One problem with the treatise section of the Shiji is that it is the section most

badly damaged by the ravages of time. According to Zhang Yan‟s 張晏 (3rd c. CE) list,

three of its eight chapters were among the ten missing. Though they have since been

found or replaced, one of the chapters mentioned by Zhang Yan, a “Treatise on the

Military” 兵書, is not even found in today‟s Shiji.117 The other two show signs of

extensive or even wholesale copying from other sources.118

Concerning this problematic section of the Shiji, the most interesting theory about

its origins comes from Zhang Xuecheng. He suggested that the treatise section was

formally and conceptually related to the “Master” (zi 子) texts of the Warring States and

early Han, especially the Guanzi, Lülan [Lüshi chunqiu], and Honglie [=Huainanzi]. He

adds, however, that in Sima Qian‟s case, “the accounts in the eight treatises were

organized by category” [八書之敘述為類列].119 Zhang suggested the following specific

correspondences:

116
SJ 23.1157.
117
SJ 130.3321 and HS 62.2724-25.
118
For a discussion of these issues and the ten missing chapters generally, see Part IV.
119
WSTY 7.811-812.

169
 Table 2: Zhang Xuecheng’s Proposed Correspondences
(* One of the 10 “missing” chapters)

“Master” Chapter (#) Shiji Treatise (#)

Text
Guanzi “Wu Xing” 五行 (41) “Feng Shan shu” 封禪書 (28)
管子 “Feng Shan” 封禪 (50) “Pingzhun shu” 平準書 (30)
“Di Shu” 地數 (77)
Lüshi chunqiu “Chi Yue” 侈樂 (5.3) “Lü shu” 律書 (25)*
呂氏春秋 “Shi Yin” 適音 (5.4) “Li shu” 禮書 (23)*
“Gu Yue” 古樂 (5.5) “Yue shu” 樂書 (24)*
Huainanzi “Tian Wen” 天文 (3) “Tian Guan shu” 天官書 (27)
淮南子 “Di Xing” 地形 (4) “Lü shu” 律書 (25)*
“Bing Lue” 兵略 (15)

It is not clear whether Sima Qian had access to the Huainanzi, which he never mentions

in the Shiji. Another murky issue is the extent to which the “missing chapters” were a

later production. But side-stepping for the moment these bibliographic issues,120 Zhang

Xuecheng‟s idea has much to recommend it. Sima Qian explicitly wanted to encompass

not just the Classics but the large compendia texts as well.121 He may well have

envisioned this section as an opportunity to preserve and display technical knowledge

found in texts like the Guanzi, Lüshi chunqiu, and Huainanzi, knowledge that would not

ordinarily be seen as fitting into the province of historical texts.

Shijia 世家

A more controversial case is that of the shijia 世家 [hereditary households]. The

word shijia is first found in the Mengzi, but there it refers not to a text but to the

120
These will be raised again in Part III, below.
121
I.e., “Putting in order the miscellaneous discourses of the hundred schools” [整齊百家雜語] (SJ
130.3319-3320).

170
household itself, translated by James Legge as “an ancient and noble family.”122 As Zhao

Yi noted, the Shiji does clearly use the word to refer to a text—namely in the evaluation

at the end of the “Hereditary Household of Wei” 衛世家 (Shiji ch.37): “The Honorable

Senior Archivist said: When I read the words of the hereditary household…” [太史公

曰：余讀世家言…].123 Zhao Yi concluded that “this is a genre, „hereditary household‟;

it originally existed ever since antiquity” [是古來本有世家一體].124 Takigawa

Kametarō pointed out, however, that “the three characters shijia yan [世家言] also appear

both in the Hereditary Household of Guan and Cai, and in the Hereditary Household of

Chen and Qi. It is the Honorable Archivist referring to his own text” [世家言三字，又

見管蔡、陳杞各世家。史公自稱其書也].125

Yet while Takigawa‟s observation regarding the other uses of the phrase shijia

yan seems reasonable, the one cited above from Shiji ch.37 still strikes readers as

problematic. Though Zhang Dake agrees with Takigawa‟s judgement,126 William

Nienhauser et al. note at this point in their translation: “It seems odd that Ssu-ma Ch‟ien

would refer to his own writing in this way.”127 They then cite Liang Yusheng‟s theory

that the chapter was Sima Tan‟s work (see discussion below). In any case, Zhao Yi was

probably incorrect in his suggestion that the “hereditary household” Sima Qian read was

122
Mengzi 3B.10, Legge, Works of Mencius, 286. Takigawa also cites the Mengzi and glosses the term as
“families who held hereditary offices” [世祿之家] (SKK 5.2062). The Shiji also uses the term in this sense,
for example, “Suo Zhong said, „Among the younger sons of the hereditary families and people of means,
there are those who [arrange] cockfights, or the racing of dogs and horses, who [spend their days] hunting
and playing games of chance, and cause turmoil among the common people” [所忠言：世家子弟富人或
鬥雞走狗馬，弋獵博戲，亂齊民] (SJ 30.1437).
123
SJ 37.1605.
124
Nian er shi zha ji 廿二史劄記 1.3.
125
SKK 5.2062. The passages to which Takigawa was referring are found on SJ 35.1570 and SJ 36.1585.
They are discussed in detail below.
126
Shiji yanjiu, 192.
127
Grand Scribe’s Records V.1.261 n.115.

171
some ancient source—also called a “hereditary household”—whence Sima Qian got the

title-type and the form. The material Sima Qian claims to have been reading all appears

in the Shiji‟s own chapter 37, and the particular passage so closely parallels Zuozhuan

Huan 16128 that it is difficult to believe Sima Qian did not use the Zuo as his source for it.

Would Sima Qian describe reading his own writing? If so, it would be a

potentially interesting insight into the composition process behind the “Honorable Senior

Archivist says” comments. Alternatively, was Sima Qian reading a chapter that had

previously been composed by his father, Sima Tan? In either case, it would make good

sense for Sima Qian to read (or re-read) the completed chapter before delivering his

verdict. In that respect, it does not seem at all “odd.”

Liezhuan 列傳

The most influential of Sima Qian‟s five forms would prove to be the liezhuan 列

傳 (or simply zhuan 傳), variously translated into English as “traditions,” “memoirs,” or

“biographies.” The difficulty with translation is profound rather than incidental. As far

as anyone knows, the Shiji is the locus classicus of the term liezhuan 列傳. Prior to Sima

Qian‟s time, the word zhuan 傳 referred to a commentary, or exposition, as the

counterpart of a classic (jing 經). This led thinkers like Liu Zhiji to explain the liezhuan

form in terms of classic and commentary, as discussed above.

Liu Zhiji‟s insight does not seem to exhaust the particular characteristics of Sima

Qian‟s invention however. The Shiji‟s liezhuan differ from other known zhuan in that

(for the most part) they take as their subject the lives and/or deeds of particular

128
Yang, CQZZ zhu, 145-147.

172
individuals. This use of the title-type was generally thought to have originated with Sima

Qian. In Nianer shi zhaji, for example, Zhao Yi wrote:

In ancient books, zhuan referred to records of events, discourses set forth [in
writing], and explications of the Classics. It was not a [title-type] devoted
exclusively to the recording of an individual‟s life-events. Its being used
specifically to record one person per zhuan started with [Sima] Qian. [古書凡記
事立論及解經者，皆謂之傳。非專一人事跡也。其專記一人為一傳者，則自
遷始。]129

Some modern scholars, though, found reason to question this traditional attribution.

Cheng Jinzao mentions two main pieces of evidence suggesting that a genre similar to the

Shiji‟s liezhuan existed prior to Sima Qian‟s time.130

Cheng Jinzao begins by differentiating three different possible meanings of the

term zhuan: records in general (書傳), classical commentaries (經傳), and historical

accounts (史傳). He claims that all three uses preceded Sima Qian. His first piece of

evidence is the Mu tianzi zhuan 穆天子傳, a text discovered in a Warring States tomb

during the Jin dynasty (approximately 281 CE). Most scholars accept that although “no

other versions… than that of the Chin scholars [is] available in traditions sources of

literature [nor] in recent archaeological excavation,” the greater part of the surviving text

is nonetheless very likely datable to around 350 BCE.131 Since the Mu tianzi zhuan is

focused on the life and activities of a single individual, King Mu (r.956-918 BCE),

Cheng‟s argument runs, it belonged to an ancient biographical tradition predating Sima

Qian. This argument could hold true even if Sima Qian had no knowledge of the Mu

129
Nianer shi zhaji 1.4.
130
Note that Fan Wenlan also cites the same two pieces of evidence, but without specific discussion
(Zhengshi kaolue, 20).
131
Rémi Mathieu, “Mu t‟ien tzu chuan” in Loewe, ed., Guide, 342-346.

173
tianzi zhuan itself, since there could be other texts in that same biographical tradition

which were extant in the Han but have vanished since.

There is one problem with this line of reasoning: while the Mu tianzi zhuan itself

might well have been an authentic Warring States text, there is no guarantee that its title

was. Some Warring States excavated texts do bear titles, but many others do not. The

title “Mu tianzi zhuan” could easily have been added by Guo Pu, who wrote a

commentary to the text, or by any other editor. Jin dynasty scholars were certainly

already familiar with the Shiji, and in general with the convention which attached the

title-type zhuan to the record of an individual. It would not be at all strange, then, for

them to retroactively apply the title-type to the newly discovered text.

Cheng‟s second piece of evidence is more interesting. It comes from the first of

the liezhuan, the “Bo Yi liezhuan” (Shiji ch.61). That chapter begins with a long and

important discursive section, which concludes, “I grieve for Bo Yi‟s aspiration; I see the

unanthologized ode and it seems different [from what Kongzi purportedly said about Bo

Yi].132 The zhuan says…” [余悲伯夷之意，睹軼詩可異焉。其傳曰…].133 What

follows is a brief biography of Bo Yi and his brother Shu Qi, from their origins in Guzhu

to their death on Shouyang. But how should we understand the word yue 曰? Is Sima

Qian introducing this biography as a quotation, material borrowed from some pre-

existing Bo Yi zhuan?

Sima Zhen certainly seems to have understood the phrase in this way, for he

commented: “ „His tradition‟ probably refers to the Han Shi waizhuan or the Lüshi
132
It is possible that this might be an over-translation of yi 異, which Burton Watson renders merely as
“very strange” (Chapters, 12). To read it as referring back to Kongzi‟s pronouncement comes from Sima
Zhen‟s gloss, SJ 61.2123, which states as much. Donald Holoch also understands the line as I do
(“Melancholy Phoenix,” 175).
133
SJ 61.2122.

174
chunqiu” [「其傳」蓋韓詩外傳及呂氏春秋也].134 The explanation is problematic,

however, for as Cheng Jinzao points out, “neither the Lüshi chunqiu nor the Han Shi

waizhuan have this passage” [呂氏春秋及韓詩外傳皆無此文].135 Cheng subscribes to

his teacher Gao Buying‟s explanation, that “I am afraid that the [above] two texts are not

what the Honorable Archivist was citing. Probably, there was some separate biography

that recorded their affairs, and that is why it says, „His tradition says,‟” [二書恐非太史公

所據。蓋別有傳記載其事，故曰其傳曰].136

Both Sima Zhen‟s and Gao Buying‟s theories seem problematic. As Wang Ruoxu

complained:

I simply do not understand the two characters „the zhuan says‟. According to the
Suoyin, it refers to the Lüshi chunqiu and the Han Shi waizhuan. But when [Sima]
Qian recorded the affairs of the ancients, when is it not taken from various pre-
existing works? If it were really as [Sima Zhen‟s] explanation [says], then [why
does] this one alone mention a zhuan?! [傳曰二字，吾所不曉。索隐云，謂吕
氏春秋韓詩外傳也。信如是說，則遷記古人事，孰非摭諸前書者，而此獨稱
傳乎。]137

Though Wang might be accused of exaggerating Sima Qian‟s use of pre-existing material

(or perhaps not), his argument does make some sense. The passage in the “Traditions of

Bo Yi” is set off rhetorically in a way that resembles a quotation, and Sima Zhen‟s

comment led later readers to understand it as such. Cheng Jinzao, whose underlying

motive was to show that each of the five Shiji sub-genres was based on a pre-existing

134
SJ 61.2123.
135
Shiji guankui, 27.
136
Shiji bielu 史記別錄, cited in Shiji guankui, 29. A.C. Graham promotes a more specific theory, arguing
that there was a Nongjia 農家 [“Tiller”] version of the story that preceded both the Lüshi chunqiu and Shiji
versions (“The Nung-chia „School of the Tillers‟”, 80-84). A passage in chapter 28 of the Zhuangzi gives a
brief biography of Bo Yi and Shu Qi, which contains much of the same information as the Shiji zhuan,
although it differs in wording and in emphasis (Zhuangzi , 771-772).
137
Hunan Yilao ji 11.78. Also cited Shiji guankui, 30.

175
form,138 would also, of course, have a vested interest in reading it as a reference to some

lost biographical text. But as Wang Ruoxu remarked, Sima Qian did not see the need to

mark off any of his other quotations in this way.

On the other hand, the “Bo Yi liezhuan” is clearly different from any of the other

chapters in the section. Pu Qilong argued that the Bo Yi chapter “should be seen as a

general preface and survey of the seventy arrayed traditions” [當作七十列傳總序觀].139

He offers little justification, but the long preface which precedes the biography proper,

and the long comment which follows it, are so much more extensive and weighty than the

short narrative that they enclose. It seems to me that those portions of the chapter are, as

Pu Qilong says, functioning as “a general preface and survey,” while the narrative marker

qi zhuan yue 其傳曰 serves to mark off the part of the chapter which is actually the zhuan

proper. In fact, it does not much matter how closely the biographical material which

follows resembles pre-existing sources. The important point lies in the way Sima Qian is

using this remarkable chapter to develop a new meaning for the word zhuan. In short,

like the reference to the text of the shijia 世家 in the “Hereditary Household of Wei”

chapter, it may be that this use of zhuan is actually self-referential and self-consciously

different from uses of the same word in other early texts.

The above discussion shows that as far as form is concerned, the Shiji had

numerous inter-textual relationships and influences. How readers perceived these

relationships had a non-negligible effect on how they contextualized and categorized the

Shiji. The extent to which individual sections had genre-precursors is controversial.

138
His conclusion admits as much: see Shiji guankui, 30-31.
139
STTS 27.213.

176
However, Takigawa expressed the scholarly consensus when he wrote that “it was an

innovation of the Honorable Archivist to put [basic annals] together with treatises, tables,

hereditary households, and arrayed traditions” [但與書、表、世家、列傳竝稱，自史

公創也].140

140
SKK 10.5323.

177
 Part II

AUTOBIOGRAPHY OR TRUE RECORD?

In Part I, I contextualized Sima Qian‟s authorship of the Shiji „from without‟ by

looking at how the Shiji was compared to other texts, and „from within‟ by talking about

how the Shiji‟s structure was read. I turn now to the two most prominent strategies of

approaching the Shiji, each proceeding from a different set of assumptions about the

nature of Sima Qian‟s authorship.

In her controversial article, “Sima Qian, A True Historian?” Michael Nylan

divided the field of Shiji studies into two factions, which she called the “lyrical-romantic”

and the “social-scientific.” Nylan‟s characterization captures something profoundly true

about Shiji studies (and perhaps the Shiji itself), not just in the past century but—suitably

qualified—throughout the text‟s history. It is an interpretive dilemma that may at times

be glossed over but has never yet been solved. Nylan purports to have arrived at a

solution, but I think I am not alone in finding it unsatisfying. Her so-called religious

reading, though exceedingly close at times to Willard Peterson‟s “cultural history”

hypothesis, attempts to resolve the dilemma by adopting and reconciling both sets of

assumptions. She arrives at a reading that is plausible but in no sense radical, and at the

end of the day no more compelling than any other reading.

I pursue a different approach in the chapters which follow. I do not produce a

“reading” of the Shiji, but instead trace the history and development of each of the two

major interpretive approaches. I begin in chapters 4 and 5 with what I call

“autobiographical readings.” These correspond roughly to Nylan‟s “lyrical-romantic”

approach, but in fact are far more general. The basic characteristic of this approach is to

178
treat the Shiji as a sort of roman à clef, using aspects of Sima Qian‟s life (real or

imagined—and where the line should be drawn is not always clear) to interpret the Shiji.

I argue that although the popularity of these readings experienced a great flourishing in

the Ming dynasty, they had already assumed their defining and enduring characteristics

by the Song.

In chapter 6, I turn to the second approach, which for want of a better term I will

call “true record” interpretations. Again, Nylan‟s corresponding term, “social-scientific

readings,” is in some ways an apt characterization for modern instances of the

phenomenon, but is again too narrow and anachronistic to apply to the entire history of

the idea‟s development. The term “true record” is a problematic translation for an

equally problematic Chinese term, shilu 實錄. There are two advantages in retaining this

term, however: first, its locus classicus is a very early description of the Shiji, and second,

its range of meanings in both traditional and modern Chinese is broad enough and

flexible enough to capture the whole range of readings which can be defined in contrast

(usually) to the autobiographical ones.

Fundamental to my methodology is the belief that the next step toward a better

understanding of the Shiji is a deconstruction and emphatically not a synthesis. Anyone

can produce a reading of the Shiji, or some part thereof, by arbitrarily adopting a subset

of conventional assumptions about Sima Qian and what his project might have been.

These readings are sometimes very interesting, but in the end they are all just as the

arbitrary as the assumptions that generate them. In showing that the conventional

assumptions themselves—which seem to most readers like self-evident truths about the

Shiji—are themselves arbitrary and historically contingent, I go beyond issues specific to

179
the Shiji or even the pre-modern Chinese context, and lay the groundwork for a new

consideration of authorship in general and its profound dependence on the agency of

readers.

180
 Chapter 4

Autobiographical Readings and the Early History of the Sima Qian Romance

His passionate sorrow and agitated energy

had no outlet, and therefore he lodged it
all in his writing. [其憤懣不平之氣無所
發泄，乃一切寓之於書].
--Qin Guan 秦觀 (1049-1100)

In 110 BCE, Emperor Wu of the Han set out to perform the feng 封 sacrifice, the

most solemn imperial sacrifice to the cosmic spirits of Heaven. For reasons unknown, he

left behind a certain minor official whose duties had led him to expect that he would take

part in the ritual. That official, Sima Tan, “was filled with resentment and lay on the

point of death.” From his deathbed Sima Tan charged his son Sima Qian with the

fulfillment of a great project, editing generations‟ worth of historical materials collected

by the Sima family and transforming them into a Classic worthy of comparison to the

work of Confucius. And then Sima Tan succumbed to his frustration and breathed his

last.

This was the first tragedy of Sima Qian‟s life, at least as he himself tells the story.

The episode is often overshadowed by the even more terrible drama of the Li Ling affair.

We have it, again according to Sima Qian‟s account, that the general Li Ling fought a

valiant battle against the Xiongnu (an ethnic group on the northwestern borders of the

empire then engaged in hostilities with the Han). In the end, Li Ling was defeated and

captured alive, which was considered a great betrayal and humiliation. Sima Qian spoke

up to defend Li Ling, but the emperor was infuriated and had Sima Qian thrown in prison,

where he was eventually punished with castration.

181
 These are the two great tragedies of Sima Qian‟s life. It remains an open but

much-debated question how these tragedies affected the great historical work which

immortalized Sima Qian in Chinese tradition—namely, the Shiji 史記 [Archivist‟s

Records], also known as the Taishigong shu 太史公書 [Writings of the Honorable Senior

Archivist].

In this section, I offer a history of how readers have seen this relationship. I

include mainstream interpretations, but also much that has since been rejected as dubious.

These latter were often influential in their own time, and their effects on today‟s view of

the Shiji may still in some sense be felt, which seems justification enough for including

them. Though there is surely much I have missed (and much that has been lost), I have

aimed for the most complete possible picture. I pay particular attention to the readers‟

views of what actually happened to Sima Qian (they are not always identical), what they

thought he was trying to do, and—insofar as I can reconstruct—why they might have

believed and/or professed these interpretations.

Among the Shiji‟s readers, there has always been a range of responses to Sima

Qian‟s personal tragedy. First, readers vary as to how much sympathy and approval they

evince for Sima Qian: some condemn him while others seem very much to identify with

and admire him. Second, they disagree as to how critical the Shiji is toward the Han

dynasty, and as to what the nature of that criticism is (is it true or false? is it quite obvious

or subtle and indirect?). Third, they disagree as to what Sima Qian‟s underlying purpose

was and whether or how it was affected by the tragic events: did he write from a desire

for revenge, to relieve his feelings and justify himself, or simply to attain the immortality

with which he was so preoccupied? Where a reader might fall along these three scales

182
depends upon his era, upon his personal circumstances and character, and not least upon

the context in which he is expressing himself. As much as possible, I will take all of

these into consideration.

SOURCES FOR SIMA QIAN‟S BIOGRAPHY

The first question to ask is, what are the sources for Sima Qian‟s biography? As

Burton Watson noted, “practically all we know of Sima Qian is what he chose to tell us….

later readers have been able to do little more than mull over, and occasionally confuse,

the evidence.”1 There are two texts generally accepted to be from Sima Qian‟s hand that

have served as the main sources of information about his life: the “Taishigong zixu” 太史

公自序 [Honorable Senior Archivist‟s Self-Narration] (SJ ch.130/HS ch.62), and the

“Bao Ren An shu” 報任安書 [Letter in Reply to Ren An] (HS ch.62/WX ch.41). Certain

passages from other chapters of the Shiji and, less well known, a poem entitled “Bei shi

bu yu fu” 悲士不遇賦 [Rhapsody lamenting unappreciated gentlemen] supplement these

to some extent, but with greater uncertainty.2 Finally, there are quasi-fictional anecdotes

which were rarely or never accorded the honor of being called “reliable,” and whose

1
Watson Ssu-ma Ch‟ien, 40-41.
2
The so-called taishigong yue 太史公曰 [Honorable Senior Archivist says] passages are discursive pieces
which appear at the end (and occasionally the beginning or middle) or most Shiji chapters. There is some
uncertainty as to whether the taishigong of these passages refers to Sima Tan or to Sima Qian. The “Bei
shi bu yu fu” is another difficult problem as regards authenticity. The earliest text of it extant today is quite
late (appearing in the Yiwen leiju 藝文類聚 [Categorized compendium of arts and literature]), though there
are references to it by Tao Qian 陶潛 (365-427) [“Gan shi bu yu xu” 感士不遇賦序, Tao Yuanming jijian
zhu 5.431], Liu Xie 劉勰 (ca. 465-520) [Wenxin diaolong zhuding 47.446], and Huangfu Shi 皇甫湜 (jinshi
811) [QTW 685.7024]. A translation of the extant portion of Sima Qian‟s fu can be found in Hightower
1954, “The Fu of Tao Ch‟ien.” In any case the “Bei shi bu yu fu” has never figured large as an interpretive
key to the Shiji. Perhaps, since it is always paired with Dong Zhongshu‟s quite similar piece on the same
theme, it struck readers as too conventional to be an authentically personal document.

183
original sources are largely unknown, but seem unlikely to have actually occurred as

written. In the discussion below, I will review the major sources for the Sima Qian story.

The Honorable Senior Archivist‟s Self-Narration

The final chapter of the Shiji has been considered by some to be “the classical

example” or model of Chinese autobiography.3 Indeed, in this chapter Sima Qian does

write of much that might find a place in a Western-style autobiography—his childhood

and place of origin, his own youthful travels, details about his life‟s work and his tragic

misfortune. Yet the careful reader will note that the overall intention of the “Self-

Narration” is not strictly autobiographical. 4 Stephen Durrant, while placing it under the

heading of “autobiographical writings” nonetheless claimed that Sima Qian “never wrote

an autobiography per se” and considers that the real function of the “Self-Narration” is

“explaining the genesis and function of [Sima Qian‟s] comprehensive history.”5 In short,

the story of the Shiji and that of its creator were intertwined from the beginning, a fact

which surely influenced the development of Shiji autobiographical readings.

Burton Watson usefully divides the “Self-Narration” into nine sections, but I

consider that there should actually be ten: 1) Genealogy; 2) The Biography of Sima Tan;

3
See, for example, Wolfgang Bauer‟s “Time and Timelessness,” in which he discusses autobiographical
forms in general, and cites Sima Qian‟s “Self-Narration” and “Letter” as classical examples or models. He
adds that this text of Sima Qian‟s “became so famous that the term „[author‟s] self-narration‟ zixu 自序
took also the meaning „autobiography‟ even if it was not added to a book” (Bauer, “Time and
Timelessness,” 23). For more detailed discussion of autobiography in China, see Bauer, Das Antlitz Chinas.
4
Indeed Pei-yi Wu, in his work on Chinese autobiography, excludes Sima Qian‟s “Self-Narration” entirely
from his definition of autobiography, arguing that Sima Qian‟s and other authorial Self-Narrations to
historical works “lend weight to the supposition that Chinese historiography posed probably the most
formidable obstacle to what is nowadays valued in an autobiography—a personal voice, a private point of
view, or any self-revelation” (Confucian‟s Progress, 43). But Wu is perhaps holding Han dynasty
conventions to a standard that would seem nonsensical to Han people themselves. Sima Qian‟s “Self-
Narration” was an unusually personal and expressive document for its time, and is by no means as devoid
of self-revelation as Wu claims.
5
Durrant, “Self as the Intersection,” 34.

184
3) The Discussion of the Essentials of the Six Schools; 4) The Early Years of Sima Qian;

5) The Death of Sima Tan; 6) Sima Qian Becomes Senior Archivist; the Revision of the

Calendar; 7) A Discussion of the Spring and Autumn Annals and the Shiji; 8) The

Misfortune of Sima Qian; 9) Table of Contents of the Shiji; 10) Concluding Remarks of

Sima Qian.6 Here, I will focus on sections 5-8, which give a very specific “creation

story” for the Shiji.

The narrative of Sima Qian‟s youthful travels (in section 4) transitions abruptly

into the story of Sima Tan‟s death:

It was in that year [110 BCE] that the Son of Heaven first established the Feng
sacrifice of the house of Han. But The Honorable Senior Archivist [Sima Tan]
was forced to stay behind at Zhounan and could not take part in the ceremony.
He was filled with resentment and lay on the point of death. [是歲天子始建漢家
之封，而太史公留滯周南，不得與從事，故發憤且卒。] 7

The narrative stops short of blaming the emperor for Sima Tan‟s decline. And yet, as

Fang Bao 方苞 (1668-1749) would later point out, there seems to be some ulterior motive

in the construction of these few short phrases.8 Also worth noting is the use of fafen 發憤,

a phrase which would later become associated with Sima Qian‟s creation of the Shiji. As

the rest of the “Self-Narration” and the “Letter” would reveal, fafen (literally “pouring

forth one‟s resentment”) was, for Sima Qian, closely associated with the act of creation.

And it is with Sima Tan‟s resentment-unto-death that the creation of the Shiji is first

openly mentioned in the “Self-Narration.” “When you become Senior Archivist,” said

6
Watson makes these divisions as part of his annotated translation of the entire chapter (minus the table of
contents). See Watson, Ssu-ma Ch‟ien, 42-57. I have divided his part 4 into two parts to better suit the
demands of my discussion.
7
SJ 130.3295.
8
See Fang Bao 方苞, “Shu taishigong zixu hou” 書太史公子序後 [After copying out the Honorable
Senior Archivist‟s Self-Narration], in Fang Bao ji 2.59-60.

185
Sima Tan (at least in his son‟s recollection)9, “you must not forget what I have desired to

expound and write.” Tan went on to place his project in the context of the Classics and

historical necessities (as discussed in Part I) and concluded:

Enlightened rulers and worthy lords, faithful ministers and gentlemen resolved to
die for their principles—I have been Senior Archivist, and yet I have failed to set
forth a record of them. Regarding the neglect and loss of the archival writings in
the world, I am deeply fearful about it. You must remember and think of this!
[明主賢君忠臣死義之士，余為太史而弗論載，廢天下之史文，余甚懼焉，
汝其念哉！]10

Qian, in tears, “requested completely to discourse upon the ancient knowledge which has

been collected and passed down by our ancestors” [請悉論先人所次舊聞].11

After Sima Tan died and the mourning period was complete, Sima Qian inherited

his father‟s office. He began to familiarize himself with the source materials available to

him, but it seems his work on the Shiji did not properly begin until after the new calendar

was announced (in 104 BCE).12 The “Self-Narration” then presents a long dialogue with

one of Sima Qian‟s colleagues on the calendar project, an official named Hu Sui 壺遂.13

Hu Sui asks first why Confucius made the Spring and Autumn Annals. Borrowing

and paraphrasing from the theories of Dong Zhongshu 董仲舒 (179 BCE-104 BCE), the

„Sima Qian‟ of the dialogue delivers a learned discourse on the motivations behind and

9
Stephen Durrant has argued convincingly for the strong psychological likelihood that Sima Qian‟s
account of his father‟s death would have been altered by the passage of time and the trauma of subsequent
events. See Durrant, Cloudy Mirror, 8-10.
10
SJ 130.3295.
11
SJ 130.3295. An alternative—perhaps equally likely—understanding of xianren 先人 is not “our
ancestors” but just Sima Tan himself.
12
Sima Qian‟s mention of his work on the calendar may well be another argument for his authority to
undertake the creation of the Shiji. From a certain perspective, the entire “Self-Narration” can be read as
being composed of such arguments, claims that he Sima Qian a suitable and authoritative person to
undertake the authorial task.
13
Hu Sui also appears—together with Sima Qian—in the Hanshu description of the calendar reform of 104
BCE. See HS 21.974-975. According to another Hanshu account, he was a gentleman of such ability and
character that the Emperor wanted to make him a prime minister (xiang 相), but Hu Sui passed away before
this could happen (see HS 52.2406).

186
potential effects of the Spring and Autumn Annals, both by itself and in comparison with

the other Classics. I will mention here only one small but provocative phrase: among

other things Confucius is said to have undertaken in the Spring and Autumn Annals, Sima

Qian claimed that he “criticized the Son of Heaven” [貶天子]. This bold characterization

of the Spring and Autumn Annals was perhaps not acceptable in the Eastern Han—in the

Hanshu version of Sima Qian‟s “Self-Narration,” included as part of the chapter on Sima

Qian, these three characters are not present.14 Yet it is important to remember that Sima

Qian used them, and perhaps conceived of both himself and Confucius as having the right

(even the duty) to criticize their rulers.15

In the next part of the dialogue, Hu Sui interrupts Sima Qian‟s panegyric on the

Spring and Autumn Annals in order to draw an explicit comparison between it and Sima

Qian‟s own writings:

In Confucius‟ time, there was no enlightened ruler above, and below worthy men
were not employed. Therefore he made the Spring and Autumn Annals…. But
you, sir, live in a time when there is an enlightened emperor above, while all
below are men fit for their positions. The myriad affairs are all accomplished, and
everything is arranged appropriately. Now in your writings, what is it you are
trying to reveal? [孔子之時，上無明君，下不得任用，故作春秋。。。 今夫
子上遇明天子，下得守職，萬事既具，咸各序其宜，夫子所論，欲以何
明？]16

From a rhetorical point of view, this is a clever trick on Sima Qian‟s part. He puts the

comparison between his writing and the Spring and Autumn Annals into the mouth of his

14
This, at least, is Li Wai-yee‟s interpretation, which strikes me as reasonable. See Li, “Authority,” 360
n.34.
15
That Sima Qian‟s characterization of the Spring and Autumn Annals was problematic can be shown by
the fact that even in the twentieth century, Li Li 李笠 (1894-1962) would insist, “The three characters [貶
天子] are a corruption. When Confucius made the Spring and Autumn Annals, it was to support the ruler
and subordinate ministers, clarifying hierarchical divisions. That is why [Sima Qian] says [here], „advance
the affairs of kings.‟ Criticizing emperors was not his intent” [三字衍。孔子作春秋，所以扶君抑臣，明
上下之分，故曰達王事也。貶天子，非其義矣] (SKK 10.5201).
16
SJ 130.3299, trans. Watson, Ssu-ma Ch‟ien, 53.

187
respected, unimpeachable, and probably already deceased colleague. Then his persona in

the dialogue proceeds to both affirm and deny it with the words, Wei wei, fou fou (唯唯否

否，不然), which must be something like “well yes but then again no, it is not so.”17

With a fervor that almost seems to approach irony, „Sima Qian‟ gushes praise for

the glories of high antiquity as celebrated in the Classics—the Spring and Autumn Annals,

he adds, “does not confine itself solely to remonstration and ridicule” [非獨刺譏而已

也]—and seamlessly transitions into a description of the triumphs achieved in his own

age. The problem, Sima Qian‟s persona in the dialogue concludes with a sycophantic

flourish, is not any fault in the ruler or lack of opportunity for talented would-be ministers.

The problem now is merely the failure of officials (himself included) to adequately

express and describe for posterity the splendor of the age. It is for this purpose, Sima

Qian writes, that he undertakes the monumental task with which his father charged him.

Then, famously, he both claims and disclaims the relationship between his work and that

of Confucius, concluding, “What I call transmitting ancient affairs…is not what is called

„creating‟, and it would be misguided for you to compare it to the Spring and Autumn”

[余所謂述故事。。。非所謂作也，而君比之於春秋，謬矣].18

17
The Shiji jijie quotes Jin Zhuo‟s 晉灼 gloss, “wei wei is modest agreement; foufou is that it is not a
thorough understanding” [唯唯，謙應也．否否，不通者也] (SJ 130.3300), which understanding is more
or less captured by Watson‟s translation: “Yes, yes. What you say is quite right, but you misunderstand my
purpose” (Watson, Ssu-ma Ch‟ien, 54). The compilers of the Hanshu cidian 漢書辭典 comment on this
phrase, “wei wei: a sound of response, showing that one agrees…. Weiwei foufou, buran: here, after the
weiwei he adds on the two negating terms foufou and buran in order to emphasize his connotation of
contradiction. This is the Honorable Senior Archivist contradicting Hu Sui‟s point of view” (Hanshu
cidian, 637). It seems possible, given the multiple levels of meaning and potential irony in this dialogue,
that considerable ambiguity might have been intended here. Which parts of Hu Sui‟s argument Sima Qian
was really affirming, and which parts denying, is a judgment left to the reader.
18
SJ 130.3299-3300. As Stephen Durrant points out, “By stating that he only transmits and does not create,
Sima Qian appears humbly to reject the comparison with Confucius, but he is in reality only affirming it”
(Durrant, Cloudy Mirror, 11), c.f. Lunyu VI:1, “I transmit and do not create” [述而不作].

188
 As in the great battle narratives of the Zuozhuan, where the preliminaries are all

and the actual clash occurs backstage (as it were),19 Sima Qian‟s actual work on the Shiji

is described in a single sentence, a mere six characters: “Then the Honorable Senior

Archivist discoursed upon and arranged his text” [於是論次其文].20 It is as if (but only

as if) the die had been cast, the authorial intention fixed and clearly elucidated. Seven

years pass in the blink of an eye, and then, in the same breath, “he met with the

catastrophe of the Li Ling affair and was hidden in darkness, bound in black ropes” [遭李

陵之禍，幽於縲紲].21 Like the association of the emperor‟s Feng sacrifice and Sima

Tan‟s death, this textual juxtaposition must surely have influenced readers to form a link

between Sima Qian‟s writing and his misfortune (whether or not Sima Qian actually fully

intended this association).

The “Self-Narration” contains no details about the Li Ling affair. It merely

presents Sima Qian, imprisoned and despairing. “This is my fault, this is my fault,” he

laments, “my body is mutilated and I will be unable to serve” [是余之罪也夫！是余之

罪也夫！身毀不用矣].22 This is the only direct reference he gives to his castration in

the “Self-Narration.” He then turns yet again to the authors of the Classics, this time

emphasizing their misfortunes in association with their creations (discussed in Part I

above).

One point worth revisiting in this context is the first example Sima Qian raised:

“Those in the Odes and Documents who were troubled and in distress [yinyue] desired to

19
See Egan, “Tso Chuan,” 334-335.
20
SJ 130.3300.
21
Ibid.
22
Ibid.

189
fulfill the longings of their aspirations” [夫詩書隱約者，欲遂其志之思也].23 The two

characters yinyue 隱約 have been a source of some debate among the commentators. I

have followed Yan Shigu‟s reading (from the Hanshu version): “隱 is to be troubled. 約

is to feel wronged” [隱，憂也。約，屈也],24 as did Burton Watson. Watson defends his

choice of Yan Shigu‟s reading with reference to the context established by the author list:

“The more common interpretation is „subtle and terse,‟ …this seems out of place when

Qian‟s entire theory is that literature is motivated by suffering and dissatisfaction.”25

The “more common interpretation”—“subtle and terse”—was established by the

Tang commentators on the Shiji. Sima Zhen glossed the line as saying, “their meaning is

hidden and subtle, and their words are terse” [意隱微而言約也].26 The problem is not

merely a philological one. The choice of gloss has deep implications for how one

chooses to interpret the Shiji. Does one want to emphasize, as Watson does, the

emotional aspect of Sima Qian‟s theory of literature? Or does one want to emphasize the

possibility that subtle meaning (often criticism) is embedded in great works of literature,

both the Classics and (by implication) the Shiji as well? Readers of different eras would

tend to make different choices, and this is one of the ambiguities which enable them to do

so.

The passage ends, however, with an unambiguous statement of an expressive

theory that would come to be associated with Sima Qian.27 Referring to the writers of the

23
Ibid.
24
HS 62.2722.
25
Watson, Ssu-ma Ch‟ien, 212 n.77.
26
SJ 130.3300.
27
For more on expressive theories in traditional Chinese literature, see James J.Y. Liu‟s Chinese Theories
of Literature, 67-87. Though Liu does not mention Sima Qian directly in this context, his discussion is

190
past, Sima Qian says: “These men all had pent up and frustrated intentions, and were not

able to carry out their Way. That is why they narrated the affairs of the past, thinking of

those who were to come” [此人皆意有所鬱結，不得通其道也，故述往事，思來者].28

It was in the same spirit, the reader is to infer, that Sima Qian wrote his own work. With

the beginning of the “Table of Contents” (section 9), we have a sense that the Shiji

genesis story is complete. It has had six major stages: the death of Sima Tan (impetus);

Sima Qian‟s reading and work on the calendar (preparation); discussion with Hu Sui of

the Shiji‟s aim and relation to the Classics (justification); first writing period; Li Ling

affair (misfortune/second impetus); and—it may be inferred—a second writing period,

this one undertaken under the stimulus the Sima Qian‟s own misfortune.

Let us consider what the “Self-Narration” does not give us. We know that Sima

Tan died of resentment at not being allowed to assist at Emperor Wu‟s Feng sacrifice, but

we do not know why he was not invited. We know that Sima Qian did seven years worth

of writing prior to the Li Ling affair, with which he somehow became disastrously

entangled, but we are told nothing about the manner of this entanglement, either in this

chapter or anywhere else in the Shiji.29 We know that Sima Qian was mutilated but not

how or why, and we know that after his misfortune he “transmitted a record of the past,”

presumably undertaking further writing and/or revision of the Shiji. Finally, the “Self-

Narration” tells us nothing of how or when Sima Qian died.

most informative about the background conditions which may have informed readers‟ evaluations of Sima
Qian.
28
SJ 130.3300.
29
A brief note about Li Ling is attached to the end of the “Traditions of General Li” 李將軍列傳, Shiji
ch.109. It mentions his capture by the Xiongnu and describes the execution of his family, but is silent on
Sima Qian‟s involvement.

191
 As any reader familiar with Chinese literary history will know, many of these

questions (excepting the last) are answered in the other autobiographical document that

Sima Qian ostensibly produced, the “Letter in Reply to Ren An” [報任安書]. There is no

evidence, however, that the “Letter” circulated prior to the compilation of the Hanshu 漢

書 (i.e., until during or after the lifetime of Ban Gu, in the first century CE).30 For more

than a hundred years, therefore, the “Self-Narration” version of the Shiji‟s background

was perhaps the only one that readers had access to.

The Letter in Reply to Ren An

The “Letter in Reply to Ren An” first appears anthologized in the Hanshu “Sima

Qian liezhuan” 司馬遷列傳 [Arrayed traditions of Sima Qian, HS ch.62], with a slightly

different version appearing in the Wenxuan some five hundred years later. It has been

called “the most influential letter of the Han dynasty,”31 a text that “could be said to stand

at the head of epistolary prose.”32 Yet its textual history prior to its appearance in the

Hanshu is completely unknown. As Bernard Fuehrer pointed out, “we do not know

whether Sima Qian actually sent the letter to Ren An and whether the latter received it in

prison or not.”33 Even its date of composition is subject to serious dispute.34 Yet no

30
How it came to survive at all until it came into Ban Gu‟s hands is in fact a mystery to which the
surviving historical record provides no solution. See below.
31
Min 2001, 81.
32
Zhang 2004, 8.
33
Fuehrer, “Court Scribe,” 175-176.
34
Burton Watson provides a helpful discussion of this controversy (Watson, Ssu-ma Ch‟ien, 94-198). As
he explains, one theory, originating from the Tang dynasty Wenxuan commentator of Lü Xiang 呂向, was
that the letter was written in 91 BCE, during the Wugu 巫蠱 (witchcraft) Affair (Liu chen WX 41.765; for a
description of the Wugu Affair, see Loewe, Crisis and Conflict, 37-90). Another theory, advanced by
Wang Guowei 王國維, was that the letter was written in 93 BCE, as this date more closely matches Sima
Qian‟s description of his recent activities. Burton Watson, who reviews extensive arguments for both sides,
ends by endorsing the one advanced by Wang Guowei. Many Chinese scholars, however, still follow Lü

192
scholar has succeeded in mounting an effective or convincing challenge against this text‟s

authenticity.

Although the “Letter in Reply to Ren An” was certainly available to readers even

before it was re-anthologized in the Wenxuan,35 it did not, in early times, play the same

kind of role that it often does today, as major interpretive key to the Shiji. Still, it seems

worthwhile to briefly review the information that the “Letter” gives about Sima Qian‟s

life. Since the “Letter” has a smoother narrative flow than the “Self-Narration,” it is

much more difficult to divide in discrete parts. I would consider, however, that there are

roughly seven different sections:

1) Reply to Ren An‟s original letter and explanation of refusal to help (HS
62.2725-2727);
2) Outline of Sima Qian‟s life (HS 62.2727-2729);
3) Li Ling affair (HS 62.2729-2730);
4) Consequences of the Li Ling affair (HS 62.2730);
5) Decision not to commit suicide, with historical examples (HS 62.2732-2733);
6) Desire to write the Shiji, with historical examples (HS 62.2735);
7) Summary (HS 62.2736).

Because the separate sections of the “Letter” do not lend themselves well to individual

analysis, I will focus instead on three major themes which would later become important

in readers‟ use of the “Letter” as a key to interpreting the Shiji.

The first of these is the importance of (and disappointment in) friendship. This

theme first arises in section 1, where Sima Qian is explaining his reasons for refusing to

help Ren An. In introducing the “Letter,” Ban Gu had referred to Ren An as Sima Qian‟s

“old friend” (guren 故人). Conventional wisdom has it that a man‟s duty toward a true

Xiang in believing that the letter was written in 91 BCE. For a more detailed discussion of the “Letter‟s”
dating and authenticity, see chapter 8 below.
35
Yan Shigu‟s Hanshu commentary preserves pre-Wenxuan comments on the “Letter,” some of which are
specific enough to make it clear that these commentators were commenting on the phrase of the “Letter”
and not just on individual words in other contexts. See, for example, comments by Fu Qian 服虔 (fl.2nd c.
CE) and Ru Chun (fl.221-65) on HS 62.2726-2727.

193
friend is so great that it should extend even to self-sacrifice. Sima Qian acknowledged

this duty when he wrote:

A proverb says, “For whom will you do it? and whom will you get to listen?” So,
after Zhong Ziqi died, Bo Ya never again played the zither. Why? “A gentleman
will sacrifice himself for one who knows him, as a woman will make herself up
for one who delights in her.” [諺曰：誰為為之？孰令聽之？蓋鍾子期死，伯
牙終身不復鼓琴。何則？士為知己用，女為說己容。]36

Yet both before and after this acknowledgement, Sima Qian disqualifies himself from

such a duty. “Were I to act,” he writes, “I would only incur blame; wanting to help, I

would do only harm” [動而見尤，欲益反損].37 And again, “Someone like me, whose

very body is marred and mutilated… in the end could achieve no glory, and is suitable

only to for being laughed at and bringing shame upon himself” [若僕大質已虧缺。。。

終不可以為榮，適足以發笑而自點耳].38

Sima Qian did once engage in an action of this nature however—he spoke before

the emperor on behalf of Li Ling. In the beginning of section 3, Sima Qian describes his

relationship with Li Ling as quite different from true or affectionate friendship: “Li Ling

and I... were never on friendly terms. Our tastes were different, and we never drank wine

together or encounted with one another the joys of deep affection. However, I observed

that he had the character of an outstanding officer” [夫僕與李陵。。。素非相善也，趣

舍異路，未嘗銜盃酒接殷勤之歡。然僕觀其為人自奇士].39 Sima Qian probably took

such pains to distance himself from Li Ling because part of his crime was “acting as a

roving persuader” (游說) for Li Ling, that is, pleading his case based for private

36
HS 62.2725.
37
Ibid.
38
Ibid.
39
HS 62.2729.

194
reasons.40 Thus, Sima Qian emphasizes Li Ling‟s good qualities and courageous acts, not

any personal fondness that might have existed between the two men. It was for the sake

of these objective merits, Sima Qian claims, that he was willing to risk his life.

The theme of friendship arises a third time in section 5, where Sima Qian

describes the consequences of the Li Ling affair. “My family was poor. Neither our

money nor our goods were sufficient funds to ransom me.41 Of my friends, not one

would save me; nor did anyone in the Emperor‟s retinue or among his intimates say so

much as a single word on my behalf” [家貧，財賂不足以自贖，交遊莫救，左右親近

不為壹言].42 Having risked all for Li Ling, Sima Qian found that there was no one

willing or able to risk everything in order to save him, and so he was subjected to the

shame of mutilation.

Indeed, for all that the first passage about friendship quoted above is a famous

statement of its importance, the rest of the “Letter” seems seriously to question the role of

friendship in the political arena. Though Ren An is (or at least seems to be) Sima Qian‟s

friend, Sima Qian refused to follow his advice.43 On the other hand, Sima Qian did act

on behalf of Li Ling, who was not his friend. And when Sima Qian needed help, none of

his so-called friends (presumably including Ren An) were willing to step forward.

40
HS 62.2730.
41
I.e., to buy commutation of the sentence, as was legally possible at that time.
42
HS 62.2730.
43
It is often believed that Ren An‟s request is actually a plea for help. See, for example, Bao Shichen‟s 包
世臣 (1775-1855) interpretation:
„Promoting worthies and recommending scholars‟ was not a phrase from Shaoqing‟s original letter.
The Honorable Archivist is not mentioning the fact that Shaoqing was pleading for rescue, and
thus uses these four characters to allude to the intention of the original letter…. [Sima Qian]
decided not to die because the Shiji was not completed. The body of the Honorable Archivist is
none other than the body of the Shiji…. The Honorable Archivist could die for Shaoqing, but the
Shiji must not be thrown away for Shaoqing‟s sake. [推賢薦士,非少卿來書中本語，史公諱少卿
求援，故以四字約來書之意。。。 不死者，以史記未成之故，是史公之身，乃史記之
身。。。史公可為少卿死，而史記必不能為少卿廢也。] (Qtd. SKK 10.5260)
Regarding the chronological difficulties of this theory, see chapter 8 below.

195
Readers would later link Sima Qian‟s statements on friendship in the “Letter” with

several chapters in the Shiji, most notably the “Yanzi liezhuan” 晏子列傳 [Arrayed

traditions of Yanzi, Shiji ch.62], the “Youxia liezhuan” 遊俠列傳 [Arrayed traditions of

the roving warriors, Shiji ch.124], and the “Huozhi liezhuan” 貨殖列傳 [Arrayed

traditions of the merchants, Shiji ch.129].44

Another major theme in the “Letter” is that of suicide. This is prefigured first in

Sima Qian‟s description of Li Ling (section 3) as a subject who would “go forth to face

ten thousand deaths, without the least thought for his own life, hurrying [to solve] the

difficulties of his lord” [出萬死不顧一生之計，赴公家之難].45 Sima Qian argued that

Li Ling‟s bravery in fearlessly facing death in battle should outweigh the fact that he

allowed himself to fall into the hands of the enemy: his earlier courage proved that he had

been captured, not because he was too cowardly to kill himself, but because “he hoped to

have the opportunity to requite his debt to the Han” [欲得其當而報漢].46

After his disastrous attempt to intervene on Li Ling‟s behalf, Sima Qian too was

put in a position where suicide would have been the expected course. In section 5, he

justified his decision not to kill himself, first because both his father and himself were too

insignificant:

It would be like one hair off nine oxen, no different from [the death of] an ant or
mole cricket. No one would compare me with those men who were able to die for
their cause, but would merely consider that my wisdom was exhausted and my

44
For a discussion of these interpretations, see below.
45
HS 62.2729.
46
HS 62.2730.

196
 crime extreme.47 [若九牛亡一毛，與螻螘何異？而世又不與能死節者比，特
以為智窮罪極].48

Famously, Sima Qian added, “Each man has only one death. Of deaths, there are some as

weighty as Mouth Tai, and others that are light as a goose feather. The difference is in

how one uses it” [人固有一死，死有重於泰山，或輕於鴻毛，用之所趨異也].49

Yet that is not Sima Qian‟s only argument against suicide. He emphasizes the

depth of his shame at having been castrated, points out the inevitably debilitating effects

of imprisonment, and raises many historical examples. He argues, therefore, that

“bravery and cowardice are a matter of circumstance” [勇怯，勢也] and that once one

has sunk low enough to consider suicide, it is already too late to save one‟s honor.

Though admitting his own cowardice, Sima Qian insists that it was not what stayed his

hand. “If even a slave or lowly maidservant is able to „open a channel,‟50 how much

more someone like me, who has no other choice!?” [且夫臧獲婢妾猶能引決，況若僕

之不得已乎].51 Sima Qian then goes on to offer a further justification for his decision—

the desire to finish the Shiji.

In both his own case and Li Ling‟s, Sima Qian described a situation in which

suicide was the expected course, but the person in question eschewed it in pursuit of a

higher cause. There are a number of similar examples in the Shiji, some of which Sima

Qian himself mentions in the “Letter”:

The Earl of the West was an earl, yet he was imprisoned at Youli. Li Si was chief
councilor, yet he suffered all the five punishments [SJ ch.87]. Huaiyin was a king,
but he was put into fetters at Chen [SJ ch.92]. Peng Yue [SJ ch.90] and Chang Ao

47
This was indeed the judgement Ban Gu seems to have arrived at in any case; see below.
48
HS 62.2732.
49
Ibid.
50
For a discussion of this expression, see Durrant, Cloudy Mirror, 18-19.
51
HS 62.2733.

197
 [SJ ch.89] faced south and called themselves „the lonely one,‟52 but they were
both arrested and put into prison, convicted of crimes. The Marquis of Jiang
executed the Lü family; he was powerful enough that he could have overthrown
the Five Hegemons, yet he was imprisoned in the Interrogation Room [SJ
ch.57].53 [The Lord of] Weiqi was a great general, yet he wore the red clothing
and was bound with three fetters [SJ ch.107]. Ji Bu was chained up as a slave for
Zhu Jia [SJ ch.100], and Guan Fu suffered shame at Jushi [SJ ch.107]. All these
men achieved the positions of kings, lords, generals, or councilors, and their fame
reached to neighboring states. But when they committed crimes and sentence was
passed upon them, not one was able to „open a channel‟ and end his own life. [且
西伯，伯也，拘牖里；李斯，相也，具五刑；淮陰，王也，受械於陳；彭
越﹑張敖南鄉稱孤，繫獄具罪；絳侯誅諸呂，權傾五伯，囚於請室；魏其，
大將也，衣赭關三木；季布為朱家鉗奴；灌夫受辱居室．此人皆身至王侯將
相，聲聞鄰國，及罪至罔加，不能引決自財．]54

This list, whose members nearly all figure in the Shiji, seems like a direct invitation to

engage in autobiographical readings, linking Sima Qian‟s own situation to his portrayal

of characters in the Shiji. The curious thing is really that no such readings are recorded

until at least the Song dynasty (see chapter 5 below).55

The third and final theme I will discuss, which runs throughout the “Letter,” is

that of the link between misfortune and literary creation. This theme had already been

raised in the “Self-Narration” above, and several of the same passages relating to it occur

in both texts. However, the “Letter” is a much more personal document and differs from

the “Self-Narration” in emphasis.

52
The self-designation of rulers.
53
According to Hanshu commentator Ying Shao 應劭, the qingshi was “a place where one confessed one‟s
crimes” [請罪之室] (HS 48.2259).
54
HS 62.2733.
55
As Stephen Durrant also points out, Sima Qian‟s decision not to die may also be related to a fascination
with those who did choose commit suicide, as shown in the dramatic Shiji narratives of Wu Zi Xu 伍子胥,
Qu Yuan 屈原, Xiang Yu 項羽, Li Guang 李廣, and others (Cloudy Mirror, 18-19).

198
 The first hint of this theme is in section 1 of the “Letter,” an allusion to the poem

“Yuan you” 遠遊 [Wandering afar], now part of the Chuci collection.56 Here Sima Qian

wrote, “I am oppressed and miserable, and have no one to speak to” [抑鬱而無誰語].57

Sima Qian further emphasized that, because of his mutilation, he had no more hopes for

friendship or achievement in the present.

After describing the punishment that resulted from the Li Ling affair (in section 4),

Sima Qian added that he and Li Ling were a laughingstock and sighed, “Alas, alas! This

matter is not easy to explain in detail to ordinary people” [悲夫！悲夫！事未易一二為

俗人言也].58 This statement is ambiguous. It could be a knowing nod to his

understanding friend Ren An, or on the other hand it could be Sima Qian breaking off his

narrative by jeering that Ren An is also merely an “ordinary person.” In either case, the

implication is that the “ordinary person” would disapprove of Sima Qian‟s failure to

commit suicide and be unable to understand the redemptive possibility offered by literary

creation.

More explicit is section 6, in which Sima Qian defends his choice to live on and

write his book: “I regret that I have things in my heart which have not been expressed

fully, ashamed to think that after I depart the world my writings will not be known to

[people who] come after” [恨私心有所不盡，鄙沒世而文采不表於後也].59 He went

on to write that only extraordinary men were able to produce works that immortalized
56
David Hawkes, in the preface to his translation of “Yuan you,” analyzes the authorship controversy
surrounding the work, and concludes that “the combination of Daoist mysticism with an enthusiasm for
Chu poetry is a hallmark of the little group of poets and philosophers…under the patronage of Liu An,
Prince of Huai-nan” and that “everything about Yuan you points to authorship by a member of this group.”
The line that Sima Qian seems to be alluding to is “Fallen on a time of foulness and impurity,/Alone with
my misery I had no one to confide in” [遭沈濁而汙穢兮，獨鬱結其誰語] (CCBZ 5.163, Hawkes, 193).
57
HS 62.2725.
58
HS 62.2730-32.
59
HS 62.2733.

199
them. He listed them (as in the “Self-Narration”) and concluded, “These men all had

pent up and frustrated intentions, and were not able to carry out their Way. That is why

they narrated the affairs of the past, thinking of those who were to come” [此人皆意有所

鬱結，不得通其道，故述往事，思來者].60

As his “Letter” has already made clear, Sima Qian too has been unable to

accomplish what he wished. In section 2 he listed four ways in which he had been unable

to serve his ruler.61 And in section 3 we see him also unable to help the general he

admires, Li Ling. Wracked with shame, he was also discouraged about any future service,

and the “pent up frustratration” [所鬱結] is entrusted to his “inefficacious writings” [無

能之辭].62 These, he hopes, “passed down among people and will penetrate the cities

and capital” [傳之其人，通邑大都].63 Again, he added that the reasons behind his

decision could not be understood by just anyone: “These kinds of things can be said to a

truly wise man, but are difficult to explain to people of the common run” [此可為智者

道，難為俗人言也].64 Sima Qian placed his hope in the enlightened readers of the

future, stating firmly that “it is necessarily only after the day of one‟s death that right and

wrong can at last be determined” [要之死日，然後是非乃定].65

As discussed in Part I, Sima Qian‟s bid to have the Shiji included among the

master-works of the past would often fail as far as later readers were concerned. His

60
HS 62.2735.
61
To summarize: he has not supplied the ruler with clever strategies, nor recommended talented men, nor
won military victories, nor attained wealth and glory (see HS 62.2727).
62
HS 62.2735.
63
Ibid.
64
Ibid.
65
HS 62.2735.

200
effort to connect the Shiji‟s composition with his own personal tragedy and frustration,

however, was destined to succeed—perhaps too well.

EARLY AUTOBIOGRAPHICAL READINGS

The earliest references to the Shiji show no tendency toward autobiographical

readings. Sima Qian‟s life, and in particular his misfortune, are either not well enough

known or simply not deemed relevant to the interpretation of the Shiji. The first example

of something that even remotely resembles an autobiographical reading merely serves to

demonstrate the extent to which Sima Qian was not yet seen as a tragic author-figure.

This selection (also quoted in chapter 1 above) is from Huan Tan‟s 桓譚 (ca.43 BCE-28

CE) lost work, the Xin lun 新論 [New discourses]:

Had Jia Yi not been degraded and disappointed, his literary elegance would not
have been produced. Had Liu An, the Prince of Huainan, not been noble,
successful, and wealthy, he could not have employed a host of eminent scholars to
compose a book. Had the Honorable Senior Archivist [Sima Qian] not been
responsible for keeping records, he would not have been able to put in order
everything from antiquity to the present. Had Yang Xiong not been poor, he
could never have written his Mysterious Words. [賈誼不左遷失志，則文彩不
發。淮南不貴盛富饒，則不能廣聘駿士，使著文作書。 太史公不典掌書
記，則不能條悉古今。揚雄不貧，則不能作玄言。]66

The overall thrust of this passage is that the creation of a literary work depends not just

on the qualities or talents of its author, but also—and perhaps more importantly—on

external circumstance. The second interesting thing about this passage is the fact that in

Huan Tan‟s perspective, Sima Qian‟s misfortune (the notorious “Li Ling disaster”) did

not play a significant role in Sima Qian‟s textual production. Instead, it is Jia Yi 賈誼

(200 BCE-168 BCE) who was “degraded and disappointed,” and so wrote as he did. It
66
Yilin 3.7; trans. Pokora 1975, 18-19. I have altered Pokora‟s translation of “to know” for 悉. My own
understanding of the word in this context as “all of.”

201
was Yang Xiong 揚雄 (53 BCE-18 BCE) who was poor, and therefore produced a work

of genius. Sima Qian, on the other hand, more closely resembles Liu An 劉安 (179

BCE-122 BCE), in that their positions afforded them certain advantages.67 In short, it

seems that Sima Qian‟s claims about his family and official position made a much deeper

impression on Huan Tan than did his misfortune.

Ban Biao

Ban Biao 班彪 (3-54) was an approximate contemporary of Huan Tan: though he

was born a generation later, they died around the same time. More importantly, they both

produced writings during the reign of Emperor Guangwu. According to his biography in

the Hou Hanshu 後漢書 [History of the latter Han], Biao lived an unsettled existence in

the aftermath of the short-lived Xin dynasty, throwing his lot in first with Wei Ao 隗囂

and then with Dou Rong 竇融 (both leaders struggling for control of the chaotic situation).

Eventually, the new emperor heard of Ban Biao‟s talents and summoned him to an

audience, even offering him a government position as Magistrate of Xu. It is uncertain

for how long Biao served in this position, if he served at all (the wording of his biography

might be interpreted to mean that he declined it outright). Afterwards, he seems to have

concentrated single-mindedly on historical records, including a continuation of Sima

Qian‟s historical work, that was probably a major source for his son Ban Gu‟s 班固 (32-

92) great work, the Hanshu.68

67
It is also worth noting that Liu An was essentially forced by Emperor Wu to commit suicide. See SJ
118.3093-3094.
68
For Ban Biao‟s historical work, see HHS 40A.1324 and discussion in Part I above. According to Fan Ye,
Ban Gu “considered that Biao‟s continuation of the earlier history [i.e., of Shiji] was not yet very detailed,

202
 Fan Ye‟s Hou Hanshu gives a chapter to Ban Biao and Ban Gu, and includes Ban

Biao‟s evaluation of the Shiji. While on the whole it is fairly balanced, Ban Biao has

some extremely critical things to say:

When [Sima Qian] discourses upon techniques and learning, he reveres Huang
Lao and slights the Five Classics. In his narration on the merchants, he makes
light of benevolence and duty, and considers poverty shameful. In speaking of
the roving warriors, he denigrates modest restraint and values vulgar
achievements. These are places where he greatly obscures and harms the Way,
and thus came the guilt that incurred such extreme punishment. [其論術學，則
崇黃老而薄五經；序貨殖，則輕仁義而羞貧窮；道游俠，則賤守節而貴俗
功：此其大敝傷道，所以遇極刑之咎也。]69

Ban Biao‟s critique would be borrowed by Ban Gu and adapted into the most oft-cited

condemnation of the Shiji in the long history of its reception. However, what is striking

about his remark is the last phrase. It is not clear what exactly he thought the connection

was between the moral flaws he observes in the Shiji and the punishment Sima Qian

endured. But he seems to allude, and indeed to subscribe, to a view that the Shiji was

somehow the cause of Sima Qian‟s punishment rather than a response to it.

Wei Hong

A more detailed version of this view has been preserved by another contemporary

of Huan Tan and Ban Biao, Wei Hong 衛宏 (fl. 1st c. CE). Wei Hong was a talented and

well-educated scholar. According to the Hou Hanshu “Rulin liezhuan”, he arrived in the

capital during a sort of scholarly renaissance brought about by the Guangwu Emperor

following the turmoil of the restoration.70 He studied the Mao Odes [毛詩] with the

and therefore investigated and contemplated its hidden essence, out of the desire to accomplish his father‟s
task” [以彪所續前史未詳，乃潛精研思，欲就其業] (HHS 40A.1333).
69
HHS 40.1325-1327.
70
HHS 79A.2545.

203
scholar Xie Manqing 謝曼卿 and even wrote his own preface for them, which, according

to the Hou Hanshu was “especially good at reaching the aims of the Airs and the

Elegantiae, and is still extant today” [善得風雅之旨，于今傳於世].71 Later he also

studied the Old Text Documents [古文尚書] with the Senior Minister of Works72 Du Lin

杜林, making a set of glosses and explications of it. According to the Hou Hanshu, he

had students of his own and “from then on ancient learning flourished greatly” [由是古

學大興]; “Emperor Guangwu considered him a righteous gentleman” [光武以為議郎].73

The Hou Hanshu also mentions that Wei Hong “made the Han Jiu Yi [Han paragons of

former days] in four chapters, in order to record matters of the Western Capital” [作漢舊

儀四篇，以載西京雜事]. The Han Jiu Yi, in addition to some of Wei Hong‟s

rhapsodies, hymns, dirges, etc., was still extant in Fan Ye‟s time.74

It is the Han Jiu Yi that Pei Yin 裴駰 (fl.438) quotes near the end of his

commentary on the “Self-Narration.” He notes:

Wei Hong‟s Hanshu Jiu Yi [sic] comments, “When Sima Qian made the „Basic
Annals of Emperor Jing,‟ he expended great passion in discussing [Emperor
Jing‟s] shortcomings and Emperor Wu‟s excesses. Emperor Wu was infuriated.
He destroyed and cast away [those annals]. Later, [Sima Qian] was sitting in

71
HHS 79B.2575. Many are of the opinion that the preface referred to here is actually the well-known
Great Preface. See Nylan, Confucian Classics, 83.
72
See Bielenstein, Bureaucracy, 15-17.
73
HHS 79B.2576.
74
The Han Jiu Yi appears as having 4 chapters in the Suishu 隨書 “Jing ji zhi” 經籍志 [History of the Sui,
Treatise on Classics and Records] (Suishu 33.969) and also in both Tang histories (JTS 46.2006, XTS
58.1487). It is quoted not just in Pei Yin‟s Shiji jijie commentary but also by both Tang Shiji “Three
Scholars” commentators, Sima Zhen and Zhang Shoujie. In the Songshi 宋史 “Yi wen zhi” 藝文志
[History of the Song, Treatise on Literature and the Arts], it appears as having 3 juan (Songshi 25.5131). In
the Yuanshi 元史 [History of the Yuan] and Mingshi 明史 [History of the Ming] it is mentioned in passing
but need not have been extant for the uses to which it was put (YS 72.1787, Mingshi 307.7902). The Siku
quanshu zongmu 四庫全書總目 [General catalogue of the Four Treasuries Complete Books] mentions only
a few fragments, one of which, called the Hanguan jiuyi 漢官舊儀 in only one juan, may well have been
some part of the same book (SKQS “Zongmu” 82.1088). That text was not included in the Siku quanshu
and has now apparently been lost.

204
 attendance and put in a good word for Li Ling. Ling surrendered to the Xiongnu.
Because of this Qian was sent to the Silkworm Chamber. Resentful words were
spoken, and [Qian] was sent to jail where he died.” [宏漢書舊儀注曰：司馬遷
作景帝本紀，極言其短及武帝過，武帝怒而削去之。後坐舉李陵，陵降匈
奴，故下遷蠶室。有怨言，下獄死。] 75

The modern scholar Yu Jiaxi 余嘉錫 has pointed out, in no uncertain terms, that this

story contradicts the account given in Sima Qian‟s own “Letter in Reply to Ren An”:

Wei Hong was a person of the early Eastern Han…. In his time, the Ban father
and son had not yet completed their text [the Hanshu], and Yang Xiong‟s and
others‟ continuations of the Writings of the Honorable Senior Historian were also
not transmitted broadly. Hong had no supporting evidence and nothing to rely on.
Thus, the text he made more or less just records words passed about in lanes and
alleys [gossip]…. Thus the saying that Emperor Wu was angry and destroyed his
[own] Basic Annals belongs to a dubious tradition. We cannot, just because [Wei
Hong] is a person of the Han, consider [his account] trustworthy. [衛宏東漢初
人。。。 其時班氏父子書未成，楊雄等續太史公書蓋亦傳播未廣，宏無所
據依，故其所著書，頗載里巷傳聞之辭。。。則其言武帝怒削本紀，自屬訛
傳，不可以其漢人而信之也。]76

In addition noting a further example of what he considers to be Wei Hong‟s

unreliability,77 Yu also quotes a number of other scholars‟ objections to the story as

well.78 But before discarding this tradition completely, we should perhaps consider what

it can tell us.

First, it has some resonance with Ban Biao‟s remark about the Shiji being the

cause of Sima Qian‟s punishment. Second, even after the Hanshu (“Letter”) account of

75
SJ 130.3321.
76
“Wangpian,” 17-18.
77
This example concerns the Documents scholar Fu Sheng, who was said to have preserved the tradition
through the Qin bibliocaust but was too old to go to court and expound his teachings. Instead of
summoning him, therefore, the court sent Chao Cuo to learn from him. This incident is narrated twice in
the Shiji (SJ 101.2745-2746, SJ 121.3125-3126). Commenting on the first occurrence, in the chapter on
Chao Cuo, Zhang Shoujie includes a further detail from Wei Hong‟s “Preface to the Imperially Sanctioned
Old Text Revered Documents” 詔定古文尚書序, namely, “[Fu Sheng] was more than ninety years old and
he could no longer speak clearly. He asked his daughter to transmit his words and teach Cuo” [年九十
餘，不能正言，言不可曉，使其女傳言教錯] (SJ 101. 2746). While this added layer of indirection is
hardly heartening for scholars of the Documents, it is difficult to see it as especially unreliable or unlikely
(indeed, it seems more plausible than the Sima Qian story).
78
See below.

205
Sima Qian‟s misfortune was widely known, variants of Wei Hong‟s story continued to

circulate and be transmitted, especially within the official histories and their

commentaries (as will be discussed in the following chapter). Third, the fact that the “Jin

shang benji” 今上本紀 (Basic Annals of the Current Emperor, Shiji ch.12) really has

been missing since the Han dynasty (and remains so, unlike the other “missing” chapters)

no doubt gave the story a further raison d‟être and at least superficial believability.79

Yu Jiaxi takes pains to point out that though Wei Hong lived much closer to Sima

Qian‟s time than we do, we should not automatically consider his account reliable. That

old texts are not necessarily historically accurate is an important consideration; the people

of the Han were as susceptible as anyone to the power of an exciting story. At the same

time, one wonders how Yu Jiaxi can conclude with such certainty that Wei Hong had no

supporting evidence and nothing to rely on, that he was recording mere gossip. One

problem is that Wei Hong‟s Han Jiu Yi is no longer extant, as mentioned above. Another

is that, for such fragments as do survive, it is difficult to judge just what Wei Hong‟s

sources may have been. And even supposing that his sources were merely “words passed

about in lanes and alleys,” such words may also be of interest, at least to the twenty-first

century historian.

This anecdote tells us that a person of the early Eastern Han considered it

plausible that a) Emperor Wu read some portion of the Shiji; b) the portion that he read

was openly critical of himself and his father; c) he was enraged enough to destroy what

he had read but not to have Sima Qian immediately put to death; d) he used the Li Ling

affair as an excuse to have Sima Qian punished. Most intriguing though are the final two

79
Yu Jiaxi also notes this in discussing a later version of the same story (see “Wangpian,” 20).

206
parts of the story: e) resentful words were spoken; and f) Sima Qian died in jail. As the

Qing scholar Gui Fu 桂馥 (1736-1805) pointed out, one would expect that if Sima Qian

had died in jail, the Hanshu would have mentioned it.80 Liang Yusheng 梁玉繩 adds that,

according to the “Letter,” Sima Qian held the office of Zhongshuling 中書令 [Prefect of

Palace Writers]81 after his punishment, and it is difficult to imagine Emperor Wu giving

him such a position if Wei Hong‟s story were true as written.82 Indeed, it is difficult to

imagine Emperor Wu, if he had indeed become enraged, hesitating to punish Sima Qian

and needing to wait for an excuse to do so. Scholars are probably correct in doubting

Wei Hong‟s account—but it is interesting to note that they do so mainly on the basis of

the “Letter” as an alternate source of information. The Shiji itself, including the “Self-

Narration,” does not entirely rule out Wei Hong‟s version of events.83

Wang Chong

The final pre-Hanshu reader of the Shiji I will consider is the first century thinker

Wang Chong 王充 (27-ca.97). He refers to Shiji frequently and sometimes extensively in

his Lunheng 論衡 [Discourse weighed in the balance]. Like Huan Tan, he seemed to

80
See his Wanxueji 晚學集 [Collection of late studies] (alternate title: Wanxue wenji 晚學文集) ch.4, qtd.
“Wangpian,” 18.
81
According to Bielenstein, this must be an abbreviation for the full title, zhongshu yezhe ling 中書謁者令
[Prefect of Palace Writers and Internuncios], an office involved with clerical assistance provided to
emperors who preferred “to conduct some of their work in the more relaxed atmosphere of the private
living quarters” but required eunuch-secretaries in order to do so with proprietary (Bielenstein, Bureacracy,
49).
82
From Shiji zhiyi 史記志疑 [A record of doubts about the Shiji], qtd. “Wangpian,” 18.
83
Inspired by Guo Moruo‟s 郭沫若 1956 article “Guanyu Sima Qian zhi si” 關于司馬遷之死 [Regarding
Sima Qian‟s death] some scholars in the PRC have toyed with the idea that Wei Hong was more reliable
than Yu Jiaxi gave him credit for. For a review of some of the arguments in the case, see Yuan
Chuanzhang 袁傳璋, “Wei Wei Hong zhi Sima Qian „xia yu si shuo‟ bian wu bu zheng” 為衛宏之司馬遷
下獄死說辨誣補證 [Disputing slander and supplementing evidence regarding Wei Hong‟s “theory that
Sima Qian died in jail”].

207
think of the Shiji more as a triumph of archival compilation than as a text brimming with

emotional power. In “Chaoqi” 超奇 [The Surpassingly Rare, Lunheng 39], Wang Chong

explores the value of literary production in explicit contrast to scholarship. He discusses

each type of literary production in what he considers ascending order of merit. First,

there are the many people who are educated but fail to produce any writings. Barely

higher than these are scholars who merely produce glosses and write memorials that

make suggestions based on those glosses. Above them he places Sima Qian and Liu

Xiang. He describes the work of these two figures as “collecting and enumerating

historical facts of ancient and modern times and narrating things that have happened” [抽

列古今，紀著行事]. He praises the voluminousness of their writings, but criticizes

them because “they rely on accomplished facts and merely record former events, without

producing anything from their own minds” [因成紀前，無胸中之造].84

Wang Chong‟s most extensive engagement with the Shiji occurs in the “Huo xu”

禍虛 [Wrong notions of misfortune] chapter of the Lunheng, and has considerable

relevance as a kind of autobiographical reading of the Shiji. The general argument of the

“Huo xu” chapter is quite a radical one, that misfortune is not visited upon people as a

result of immoral actions but is merely a result of chance or fate. Among the numerous

84
LH 39.607-608; Forke II, 297. Michael Nylan seems to understand this (and a similarly worded
statement in Lunheng 82) as Wang Chong‟s complaint about Sima Qian‟s lack of emotion (Nylan 1998,
208). The problem comes down to what Wang Chong believed was “in the breast” (胸中). The question is
difficult to decide without a thorough study of early Chinese conceptions of the body, but in context it
seems that Forke‟s rendering better reflects Wang Chong‟s priorities. While Wang Chong may have
admired the pathos of Qu Yuan and Jia Yi, the nature of his own Lunheng and many of his remarks suggest
that what he valued most highly was intellectual creativity. Following this understanding, Wang Chong‟s
criticism of Sima Qian, probably heavily tinged with a deprived bibliophile‟s envy, would be that the
historian did little but copy the voluminous source materials he had available to him, coming up with few
or no original ideas.

208
examples Wang Chong used to discuss this problem is an extended series drawn from the

Shiji.

Wang Chong begins with the “Meng Tian liezhuan” 蒙恬列傳 [Arrayed traditions

of Meng Tian, Shiji 88]. In that chapter, the Second Qin Emperor sends an envoy to

Meng Tian, ordering him to commit suicide and thus preempting his potential objection

to the Second Emperor‟s succession and the execution of Tian‟s brother Meng Yi 蒙毅.

Meng Tian gives a long speech in protest, which the envoy refuses to relay to the

Emperor. Had Meng Tian been a man of a different character, it might have been a

moment (like many others in the Shiji) when the course of history hung in the balance.

But we can see that Tian is not a man of destiny, for he sighs as he says, “How have I

offended against heaven? Why should I die having done nothing wrong?” [我何罪於

天，無過而死乎？].85 It is the sigh that tells us he is a loyal servant, not a defiant rebel.

After deliberating for some time, Meng Tian then concocts an indictment against

himself, for which both Sima Qian and Wang Chong criticize him: “In fact, I Tian do

deserve to die for my crime. Starting at Lintao and joining up to Liaodong, I built a wall

more than 10,000 li long. How could I not have cut through the veins of the earth in the

process? That must be my crime” [恬罪固當死矣。起臨洮屬之遼東，城壍萬餘里，

此其中不能無絕地脈哉？此乃恬之罪也].86 The Honorable Senior Archivist takes

issue with Meng Tian‟s self-accusation on two different points: first he praises the results

of Meng Tian‟s construction work, giving no credence to the idea that cutting the veins of

85
SJ 88.2570. Wang Chong‟s version has a slight variant, 我何過於天，無罪而死乎 (LH 21.275).
86
SJ 88.2570. Again, Wang Chong‟s versions of this and the subsequent passage contain minor variants.

209
the earth was a crime. However, he adds harshly, Meng Tian and his brother did deserve

to die because:

They did not take the opportunity to remonstrate forcefully, nor did they relieve
the distress of the common people, nor did they care for the old and rescue the
orphaned, nor did they labor to bring harmony to the masses. Instead they
flattered [the emperor‟s] desires and undertook this construction. It was for this
that the brothers suffered execution, and is it not fitting!?” [不以此時彊諫，振百
姓之急，養老存孤，務修 庶之和，而阿意興功，此其兄弟遇誅，不亦宜
乎].87

Wang Chong criticized both Meng Tian‟s and Sima Qian‟s interpretation of the situation.

His argument against Meng Tian‟s self-accusation is tortuous and muddled by textual

corruption, but it is possible to paraphrase it as follows: if Meng Tian met with his

misfortune because he had committed a crime against the earth, then what crime had the

earth committed against heaven,88 such that it deserved to the misfortune of having its

veins cut by Meng Tian? Since it is impossible to find consistency in Heaven‟s supposed

responsiveness to human action, Wang Chong argues that Heaven does not actually

respond at all—making Meng Tian‟s self-accusation meaningless.

As for Wang Chong‟s argument against Sima Qian‟s assessment, it comes near to

being an autobiographical reading of the Shiji. Wang Chong writes that if we follow

Sima Qian‟s judgment on Meng Tian, then “those who ought to remonstrate and do not,

will be punished with death” [夫當諫不諫，故致受死亡之戮],89 or, in more general

terms, “the misfortune suffered tells against a person” [所任非其人].90 In other words, if

87
SJ 88.2570.
88
Following Huang Hui‟s 黃暉 proposed emendation of 天 for 人.
89
LH 21.276.
90
LH 21.276; trans. adapted from Forke I.168. It is difficult to understand the literal sense of this phrase. I
have tentatively followed Forke‟s translation of 任 as suffer, both here and in the next passage, but this
would have to be seen as an extensional meaning of the original “to carry or bear a burden.” Another
possibility would be to understand it as “to take responsibility for,” but, while that would be more faithful
to the meaning of the character, it is harder to make sense of in context.

210
someone suffers a misfortune, it follows that they have done something wrong.

Attacking this position, Wang Chong brings in the circumstances of Sima Qian‟s own life:

“Sima Qian himself had to suffer for Li Ling in the Silkworm Chamber” [身任李陵，坐

下蠶室] and hence,

If Sima Qian censures Meng Tian for not having strongly remonstrated with his
sovereign, wherefore he incurred his disaster, then there must have been
something wrong about himself likewise, since he was put into the Silkworm
Chamber. If [on the other hand, Sima Qian] himself was not wrong, then his
criticisms of Meng Tian are wrong. [非蒙恬以不彊諫，故致此禍，則己下蠶
室，有非者矣。己無非，則其非蒙恬，非也。]91

Wang Chong seems to consider this argument a reductio ad absurdum for Sima

Qian‟s position—he expects Sima Qian‟s life and his philosophical views to be consistent

with one another. Furthermore, Wang Chong‟s reasoning implies an assumption that the

Shiji was composed or put in final form after the Li Ling incident, not allowing for the

possibility that Sima Qian may have formed his opinion as stated in the Meng Tian

chapter before his own disaster befell him. (Or that it might have been Sima Tan who

formulated this judgment.)

Like the other figures discussed above, Wang Chong seems to depend only on the

Shiji “Self-Narration,” and does not betray any knowledge of the “Letter.” For in the

“Letter” we read that Sima Qian, however debatable his sincerity, gives an interpretation

of what his misstep had been in the Li Ling incident. It was not that Sima Qian believed

he was wrong about Li Ling himself—his long narrative in defense of the disgraced

general precludes that possibility—but he explained the unfortunate consequences of his

91
LH 21.276; trans. adapted from Forke I.168. I have largely followed Forke‟s translation here but have
changed “warm room” to Silkworm Chamber for the sake of a more literal translation. In either case, the
reference is to the place where the punishment of castration was carried out.

211
own act by saying, “I was not able to make myself fully understood” [未能盡明].92 Even

the way in which Wang Chong referred to Sima Qian‟s tragedy in the passage quoted

above (merely saying that he was sent to the Silkroom Chamber, without further details)

does not necessitate a sure knowledge of the “Letter,” though it does suggest some

knowledge outside the “Self-Narration.”93

Wang Chong‟s next move is to bring in another text, “Bo Yi Liezhuan” 伯夷列傳

[Arrayed traditions of Bo Yi, Shiji 61], quoting the passage about Yan Yuan dying young

and Robber Zhi living to a ripe old age. Wang Chong‟s use of this passage is difficult to

understand unless we realize that he was reading it differently from the way we read it

today. Yan Yuan died young, Sima Qian says, and adds, “Is this how heaven requites a

good person?” [天之報施善人，其何如哉].94 And in speaking of Robber Zhi‟s living a

long life despite being the epitome of wickedness, Sima Qian writes, “What virtue was

being revered?!” [是遵何德哉].95 (I.e., how did he deserve the long life granted to him?)

Readers have interpreted this to be an example of Sima Qian‟s tortured doubts about the

rightness of heaven‟s judgements, a highly emotional protest against the injustice of

earthly existence.96

92
HS 62.2730.
93
The text of the “Self-Narration” does contain a reference to Sima Qian‟s mutilation, when he says of
himself, “my body is mutilated and I will be unable to serve” [身毀不用矣] (SJ 130.3300), but it does not
specify that the mutilation was castration. The case must have been generally known as part of the lore of
the Western capital, however: see discussion of the Wei Hong anecdote (which includes the “Silkworm
Chamber” detail), above.
94
SJ 61.2124-2125.
95
SJ 61.2125.
96
See for example Durrant, Cloudy Mirror, 23-25. C.f. Holoch, “Melancholy Phoenix,” which de-
emphasizes the emotional content of the chapter and instead calls it “a carefully structured radical argument
whose rhetoric reproduces the process of thought rather than crystallizing its result” (Holoch, “Melancholy
Phoenix,” 172). Nonetheless, Holoch arrives at essentially the same interpretation of Sima Qian‟s point as
Durrant does.

212
 Yet Wang Chong took Sima Qian‟s words here as a fine proof and argument for

the very position Chong was espousing in his own chapter, namely that misfortune and

conduct are unrelated. “As it says here, Yan Hui should not have died so prematurely”

[若此言之，顏回不當早夭], Wang Chong says, meaning, if misfortune were

exclusively a punishment for wrong-doing, Yan Yuan would not have died prematurely—

for Yan Yuan (in all the known stories about him) never did anything wrong. Wang

Chong adds, “Not to wonder at Yan Yuan‟s premature death, but to say that Meng Tian

deserved to die, is inconsistent” [不怪顏淵不當夭，而獨謂蒙恬當死，過矣].97 The

peculiarity of this statement draws a protest from the Lunheng commentator Huang Hui,

who suggests that the “not” should be stricken from the text. For, Huang writes,

“Wondering at the early death of Yan Yuan is just exactly what the Honorable Archivist

is doing!” [史公正怪顏淵早夭也].98 But Huang Hui has not thought carefully about the

logic of this emendation.

According to Wang Chong, Sima Qian‟s belief—as expressed in the “Meng Tian

liezhuan”—is that a person‟s misfortune arises as a result of his wrong-doing. Now,

there is a consensus belief that Yan Yuan is by definition guiltless, a person who had

done no wrong. Therefore, this ought to put Sima Qian into consternation—he ought to

wonder at it, as in fact we and Huang Hui think he does. But if Sima Qian does wonder

at it, then there is none of the inconsistency that Wang Chong is pointing out. Therefore,

whatever the facts about Sima Qian‟s work itself, we have to conclude that, in Wang

Chong‟s reading, Sima Qian is not wondering at Yan Yuan‟s early death. Instead, for

Wang Chong, Sima Qian is providing a persuasive argument that there cannot be an

97
LH 21.277; Forke I, 168.
98
LH 21.277.

213
exclusive relationship between misfortune and misconduct. Conveniently, that puts the

Sima Qian of the “Bo Yi liezhuan” chapter neatly into Wang Chong‟s own camp—but

also makes Sima Qian‟s vacillation in the “Meng Tian” chapter all the more frustrating

for Wang Chong.

It is easy to criticize Wang Chong‟s reading of the Shiji here as excessively

unsympathetic and forced. But in doing so, we risk overlooking the valuable information

he can give us about how the Shiji may have been read in Han times. Wang Chong‟s

critique is the beginning of a demand for consistency in the Shiji as a whole, a technique

of reading one chapter against another.99 Furthermore, Wang Chong‟s biographical

reading, however unsympathetic, assumed that the author‟s life should accord with the

views expressed in his text.100 Earlier writers, including Sima Qian himself and also

Huan Tan, had shown themselves aware of a relationship between the author‟s life and

his works. But Wang Chong made a stronger case, not merely recognizing the

consistency between the two, but actually demanding it.

Ban Gu Reads Sima Qian

In the Hanshu, the “Letter in Reply to Ren An” is followed by Ban Gu‟s

evaluation of Sima Qian‟s work and his life as a whole. In this evaluation, Ban Gu writes:

99
Grant Hardy, who considers Sima Qian‟s inconsistencies to be a pedagogical tool (e.g., Hardy, Bronze
and Bamboo, 206) would probably consider Wang Chong a bad reader of the text, unable to grasp Sima
Qian‟s “invitation to think through history with him” (ibid). Yet if it was impossible for a Han dynasty
reader as astute and intelligent as Wang Chong to grasp Sima Qian‟s lesson, could the lesson really have
been as deliberate as Grant Hardy argues it is? Or are inconsistencies not so much intentionally instructive
as just a natural side-effect of the generic structure?
100
A similar impulse may have been behind Sima Qian‟s construction of the much criticized “Kongzi
shijia” 孔子世家 [Hereditary household of Confucius]. His occasionally portraying the Sage in a negative
light may have been an effort to find consistency among sources supposedly from Confucius‟ own hand but
in reality reflecting considerable admixture of various types of philosophical views. For critiques of the
Confucius chapter and possible explanations for its characteristics, see the discussions in Durrant, Cloudy
Mirror, 29-46 and Hardy, Bronze and Bamboo, 154-168.

214
 Alas! One with Qian‟s breadth of knowledge and experience still failed to
understand how to keep himself from harm. Having suffered the extreme penalty,
he was sorrowful and poured forth his resentment, and his writing is also most
believable. The original traces of this self-inflicted suffered and shame are found
in the “Chief Eunuch” of the Lesser Odes.101 But what the Great Ode says,
“Enlightened and wise, he keeps himself from harm” 102—this indeed is difficult.
烏呼！ 以遷之博物洽聞，而不能以知自全，既陷極刑，幽而發憤，書亦信
矣．跡其所以自傷悼，小雅巷伯之倫．夫唯大雅既明且哲，能保其身，難
矣 ！]103

Though Ban Gu here expresses admiration for the “Letter,” his remarks still suggest a

tendency to criticize rather than pity Sima Qian for his misfortune. He seems to take

Sima Qian‟s humility104 at face value, blaming Sima Qian for bringing disaster upon

himself. He implies that if one is wise and perspicacious, one should be able to avoid

such difficulties.105

Ban Gu‟s most well-known criticism of Sima Qian, often referred to as “the three

faults of the Honorable Senior Archivist” (太史公三失) actually occurs earlier in the

same evaluation. It is a very close paraphrase of Yang Xiong‟s 楊雄 (53 BCE-18 CE)

101
The poem referred to is “Xiang Bo” 巷伯 (Mao #200). The poem, written by a eunuch (surely an
intended point of similarity), contains eloquent denunciations of slanderers.
102
The poem referred to is “Zheng Min” 烝民 (Mao #260). It is said to be a poem in praise of the virtuous
Zhou minister Zhong Shanfu. The relevant lines are (in Legge‟s translation), “Intelligent is he and
wise,/Protecting his own person;/Never idle, day or night,/In the service of the One man” (既明且哲，以
保其身。夙夜匪解，以事一人). It is interesting, though perhaps coincidental, that both the Odes Ban Gu
referred to are among the small set of Shijing poems whose authors name themselves.
103
HS 62.2738.
104
Near the end of the “Letter,” Sima Qian writes, “I met this misfortune because of the words I spoke. I
have brought upon myself the scorn and mockery even of my native village and I have soiled and shamed
my father‟s name. With what face can I ascend and stand before the grave mound of my father and
mother?” [僕以口語遇遭此禍，重為鄉黨戮笑，汙辱先人，亦何面目復上父母之丘墓乎] (HS 62.2736;
Watson, Ssu-ma Ch‟ien, 66).
105
Despite Ban Gu‟s critical evaluation of Sima Qian in Hanshu 62, the latter cuts a brave enough figure in
the same work‟s chapter on Li Ling (Hanshu 54). It is worth noting that there Ban Gu makes Sima Qian
speak to his sovereign in much the same language as the “Letter” and (as mentioned above) almost
certainly got his material from there. For a more detailed discussion of the parallels involved, see chapter 8
below.

215
criticism106 and of Ban Biao‟s107. Nonetheless, it was in Ban Gu‟s formulation—and as

originating from him—that this famous critique stood in later history:

His judgments stray rather often from those of the Sage. In discussing the Great
Way, he puts Huang-Lao first and slights the Six Classics. In his introduction to
the “Roving Warriors” [SJ ch.124], he disparages lofty gentlemen in retirement
and promotes heroic scoundrels. In narrating the “Merchandise and Prices” [SJ
ch.129], he honors those who were skilled at making a profit and heaps shame on
poverty and low station. It is these points that mar his work. [其是非頗繆於聖
人，論大道則先黃老而後六經，序遊俠則退處士而進姦雄，述貨殖則崇勢利
而羞賤貧，此其所蔽也。 ]108

As we will see in Song dynasty essays, Ban Gu‟s judgment had so much prestige that

almost any defender of Sima Qian would feel the need to refute it.

Still, Ban Gu‟s evaluation of Sima Qian is even harsher—and more interesting for

present purposes—in another text, this one written for presentation directly to the

Emperor. The “Dian Yin” 典引 [Extension of Constant Models] is a fuming 符命, an

extremely formal and lofty genre written on imperial command and generally concerning

the affairs of sages and worthies.109 This particular example appears in the Hou

Hanshu110 and in the Wenxuan.111 Only the latter, however, includes Ban Gu‟s own

preface, which is what is interesting in this context.112 Below, I give a translation of the

first two parts of preface:

106
Yang Xiong‟s criticism, found in the preface to the Fa Yan 法言 [Model sayings], also complains that
Sima Qian “was not in accord with the Sage, and his judgments were rather different from those of the
Classics” [不與聖人同，是非頗謬於經] (HS 87B.3580).
107
HHS 40A.1325, quoted above.
108
HS 62.2737-2738.
109
David Knechtges translates this term as “Mandate through Prophetic Signs” (Wenxuan I.22), but does
not discuss the genre in detail.
110
HHS 40B.1375-1385.
111
WX 48.2158-2166.
112
The Hou Hanshu excerpts only a part of this preface, which reads, “When Xiangru wrote about the feng
and shan, it was beautiful but not dignified. When Yang Xiong wrote „In Praise of the Xin,‟ it was
dignified but not true” [相如封禪，靡而不典，楊雄美新，典而不實] (HHS 40B.1375). Fan Ye adds,
with perhaps a hint of sarcasm in light of Ban Gu‟s exaggerated (false) modesty, “Probably [Ban Gu]
himself would say that he had attained the utmost” [蓋自謂得其致焉] (ibid).

216
 Minister Gu said, “In the seventeenth year of Yongping [74 CE], I,
together with Jia Kui, Fu Yi, Du Ju, Zhan Long, Chi Ming, and others,113 was
summoned by the emperor and awaited further orders at Yunlong Gate114. There
the junior palace eunuch Zhao Xuan, holding the text of the [Shiji] “Basic Annals
of Qin Shihuang,” asked me and the others: “In words of the evaluation passed
down by Senior Archivist Qian, is there anything that is incorrect?” I replied, “In
this evaluation, Jia Yi‟s „Essay on the Faults of Qin‟ says, „Suppose Shiying had
had the talent of even a mediocre ruler, and had only obtained average assistants,
then it would not have been right for the sacrifices of Qin to have been cut off.‟
These words are not correct.” At this, I was summoned to enter, and [the emperor]
conveyed the following request: “I have heard that you think this discussion [of
Jia Yi‟s] is wrong. Shall I therefore summon you to an audience and ask about
your meaning so that I may be enlightened [about this]?” I provided answers
[based on] what I had previously heard and my understanding of the situation. [臣
固言：永平十七年，臣與賈逵、傅毅、杜矩、展隆、郗萌等，召詣雲龍門，
小黃門趙宣持《秦始皇帝本紀》問臣等曰：太史遷下贊語中，寧有非耶？臣
對：此贊賈誼過秦篇云，向使子嬰有庸主之才，僅得中佐，秦之社稷未宜絕
也．此言非是．即召臣入，問：本聞此論非耶？將見問意開寤耶？臣具對素
聞知狀．]
The edict therefore read: Sima Qian compiled a text that completed the
words of a single family and glorified his name in later generations. Later,
because he himself suffered punishment, he turned to subtle writing and
piercing satire, denigrating and detracting from his own generation. He was
not a gentleman with a sense of duty. Sima Xiangru‟s conduct was corrupted and
unchaste, and his phrases were all florid and fluttering, unsuitable for any
practical function. But after he had fallen critically ill, he still remained loyal.
For when the sovereign sought to collect up his writings, in the end he obtained a
work that praised and passed down meritorious virtue, speaking of the feng and
shan sacrifices, and this proves that he was a loyal servant. From this sort of
worthiness, [Sima] Qian is far removed indeed. [詔因曰：司馬遷著書成一家之
言，揚名後世，至以身陷刑之故，反微文刺譏，貶損當世，非誼士也．司
馬相如洿行無節，但有浮華之辭，不周於用，至於疾病而遺忠，主上求取其
書，竟得頌述功德，言封禪事，忠臣效也．至是賢遷遠矣．]115

There are a few things worth noting about the context of this piece. First, it is official in

every sense of the word. It gives us a rare and fascinating window into how an emperor

113
On these supporting actors in Ban Gu‟s little drama, Li Shan wrote: “The Hou Hanshu says, „Jia Kui,
whose zi Jingbo, served as Palace Attendant.‟ The Seven Abstracts [by Liu Xin] says, „Palace Gentleman
of the Masters of Writing Zhan Long of Beihai.‟ Although the Seven Abstracts was compiled in the era of
[Emperors] Ai and Ping, it is possible that Zhan Long lived into the Yongping period” [後漢書曰：賈逵，
字景伯，為侍中．七略曰：尚書郎中北海展隆．然七略之作，雖在哀、平之際，展隆壽或至永平之
中] (WX 48.2158).
114
In the south part of the imperial capital of Luoyang.
115
WX 48.2158, emphasis added.

217
might initiate discussion of texts during the Han. The text in question (Shiji 6, “Basic

Annals of Qin Shihuang”) was physically present. Ban Gu (and presumably the others as

well) were asked to give their opinion on one specific part of it. Finally, Sima Qian—

who had endorsed the opinions Jia Yi had expressed in the anthologized “Essay on the

Faults of Qin”—is apparently held responsible for them. Jia Yi‟s essay, as far as moral

responsibility was concerned, might as well have been written by Sima Qian.116

It is easy to see why Ban Gu‟s answers pleased the Emperor. Though Sima Qian

had much to say in criticism of the Qin, his narratives of the short-lived unification does

not wholly condemn that dynasty which preceded and was replaced by the Han.117

Furthermore, there are certain textual hints that people of Sima Qian‟s era may have had

an even more positive view, at least of central Qin figures like Li Si.118 What the Eastern

116
This tendency to hold Sima Qian responsible for every source he quotes is also shown in an entry from
the Fayan, where someone says, “Sima Zichang has a saying, that the Five Classics cannot match the Laozi
for conciseness, for „generations of scholars could not master their study, nor could a man in his whole
lifetime thoroughly comprehend all their rules‟” [司馬子長有言，曰五經不如老子之約也，當年不能極
其變，終身不能究其業] (FY 7.222). The last part of the saying is actually from Sima Tan‟s essay in the
“Self-Narration” (I have used the translation in Watson, Ssu-ma Ch‟ien, 45), and not by Sima Qian at all.
Still, is attributed to him, and Yang Xiong responds quite as if he had said it himself.
117
In the preface to the “Liu guo nianbiao” 六國年表 (Chronological Table of the Six States, Shiji 15),
Sima Qian wrote:
Scholars, influenced by what they have heard, see that the Qin occupied the position of emperor
for only a short period, and they fail to examine the beginning and end of the matter. Hence they
refer to the Qin only as an object of ridicule and decline to say anything more about it. This is as
ridiculous as trying to eat with one‟s ear, and lamentable indeed. 學者牽於所聞，見秦在帝位日
淺，不察其終始，因舉而笑之，不敢道，此與以耳食無異．悲夫！(SJ 15.686; trans. Watson,
Qin, 87).
Sima Qian‟s own view of the Qin defies easy categorization, but he seems to have at least considered it to
have been an important object of study.
118
Sima Qian‟s evaluation of Li Si begins with the interesting complaint that:
People all claim that Li Si exerted the utmost loyalty and in spite of that suffered death by the five
penalties. But if we examine the root of the affair, we find that it is very different from this
popular appraisal. If he had truly been such a man, then his accomplishments would have been
worthy to rank with those of the Duke of Zhou and the Duke of Shao. [人皆以斯極忠而被五刑
死，察其本，乃與俗議之異．不然，斯之功且與周﹑召列矣] (SJ 87.2563; trans. Watson, Qin,
206).
This evaluation is more complex than it might at first seem. Though Sima Qian questioned the positive
view of Li Si that “people all” held, he clearly had a deep admiration for Li Si‟s accomplishments. The

218
Han emperor wanted to hear, however, was a wholesale condemnation of the short-lived

Qin (which would also be an affirmation of the dynastic legitimacy of the Han). One

might even suggest that in the Qin, the emperor could have seen an analogy with the

short-lived Xin dynasty of Wang Mang. Is it true that if Wang Mang had “had the talent

of even a mediocre ruler, and had only obtained average assistants,” the Xin dynasty

would never have fallen? History records no such speculation, but the emperor‟s choice

of this particular line from the Shiji certainly opens the possibility that he was sensitive to

such a comparison—and that he craved the sort of ringing endorsement of Han

legitimacy that the highly sycophantic “Dian yin” eagerly provides. 119

But first, Ban Gu does not miss his chance to denigrate his predecessor and rival,

Sima Qian. The next part is framed as “an edict” (a pronouncement from the emperor)

but surely either came from Ban Gu‟s brush or was at least heavily influenced by his

views. In particular, if we compare the emphasized passage above with the quotation

from the Song dynasty figure Qin Guan with which I began this section, we see that the

idea is much the same: Sima Qian suffered punishment and therefore the Shiji contains an

indirect meaning. Ban Gu, however, betrays nothing of the sympathy for Sima Qian that

later writers (such as Qin Guan) would evince. He presents the post-Li Ling affair

products of Sima Qian‟s brush as almost poisonous, subtly damaging to the interests of

implicit comparison between the Qin and the sage rulers of the Zhou, even if it is the dissimilarity that is
being emphasized, is nonetheless quite striking.
119
Another document, said by commentators to have also come from Ban Gu‟s hand, has become attached
to the end of the “Qin Shihuang benji” 秦始皇本紀 [Basic Annals of the First Qin Emperor, Shiji ch.6]. It
is a thorough attack on the Qin, which begins,
The Zhou era was at an end. Benevolence did not follow in proper succession [a reference to a
five phase theory which delegitimized the Qin]. Qin took up its position, but Lü Buwei‟s bastard,
Zheng [i.e., the First Qin Emperor, whose parentage was sometimes said to be uncertain], was
cruel and tyrannical. However, in the thirteenth year, he used the feudal lords and united the realm.
His nature was extremely dissolute… [周曆已移，仁不代母．秦直其位，呂政殘虐．然以諸侯
十三，并兼天下，極情縱欲…] (SJ 6.291)

219
the Han. By contrast Sima Xiangru, however disorderly his personal life, is portrayed a

loyal servant to his dynasty due to the essay he wrote on the Feng and Shan sacrifices.120

Ban Gu was hardly an objective bystander. As regards Han history his work was

in some ways a competitor to Sima Qian‟s. It must have been he who decided to turn his

father‟s project—a Shiji continuation—into an independent work which incorporated

much of the Shiji‟s Han material. The process of deciding what to change and what to

leave the same must also have tested his judgment about what constituted being “a

gentleman with a sense of duty” to the dynasty. As Zhang Taiyan 章太炎 (1869-1936)

wrote of the “Dianyin”: “Ban Gu probably had no choice but to write this. He knew that

[subtle criticism] was something the ruler hated, and he wrote this in order to avoid

disaster [himself]” [班固蓋不得已而作，知為人主所忌，要以避禍].121 Ban Gu‟s ploy

seems to have worked. The Hanshu is by no means free of this sort of subtle criticism,

but no one called it a bangshu 謗書 (that is, a defamatory text injurious to the ruler‟s

reputation). The Shiji, on the other hand, was dogged by that epithet, and cursed—later,

blessed—with the reputation of being a hidden critique of its times.

The main thing to note here is that Ban Gu provides a full statement of the

premise with which I began this section on autobiographical readings, the very

groundwork for such readings: “Because [Sima Qian] himself suffered punishment, he

turned to subtle writing and piercing satire, denigrating and detracting from his own

generation.” Ban Gu, very much aware of the “Letter”, follows Sima Qian‟s own version

of the story, reversing the causality suggested by Ban Biao and Wei Hong: for Ban Gu,

the punishment came first, the satire afterwards. Unfortunately, Ban Gu gave no specific

120
See SJ 117.3063 ff.
121
Qtd. in Shao Yiping 卲毅平, “Han Mingdi zhaoshu,” 65.

220
information about what he considered “subtle writing and piercing satire” to be. Surely it

was not limited to the single remark in an essay not even from Sima Qian‟s hand?

Given the context of the “Dian yin,” it would be almost impossible for Ban Gu to

produce a sympathetic reading of Sima Qian‟s story there. Nor, officially speaking, was

it easy for anyone after Ban Gu to be sympathetic to an historian who was condemned in

such terms—Sima Qian was guilty, on the one hand, of disloyalty to his ruler, and, on the

other, of causing moral incoherence because he disagreed with the sage. As quoted

above, already by Ban Gu‟s time the premise of the autobiographical reading of Shiji was

fully formed. Yet it looks very different from the autobiographical (“lyrical-romantic,” in

Michael Nylan‟s terminology122) readings that we are accustomed to today. For in order

to ensure that the reader‟s sympathies did not stray to the wrong places, Ban Gu made

sure to add a particularly damning judgement upon Sima Qian, that “he was not a

gentleman with a sense of duty.”

Historical texts up to this point—the Zuozhuan, and even moreso the Guoyu and

the Spring and Autumn Annals of Master Yan—had emphasized that it was the duty of a

subject to criticize his ruler. But Sima Qian, as Ban Gu presented him, was doing

something other than dutiful remonstrance. The “subtle writing” that Ban Gu accused

him of was construed as an attempt to escape the ruler‟s notice. It was seen as a text

directed not at the ruler but, in a self-serving way, at posterity.

SIX DYNASTIES DEVELOPMENTS

Wei Hong‟s version of Sima Qian‟s tragedy and death flourished during the Six

Dynasties, despite being more or less contradicted by the “Letter in Reply to Ren An.”
122
Nylan, “True Historian,” 205.

221
One resultant development was the idea that the Shiji was a “defamatory text” (bangshu

謗書), which is closely associated with the story of Cai Yong‟s 蔡邕 (133-192) death. I

introduce the “defamatory text” notion below, but discuss it in more detail in chapter 6.

The Death of Cai Yong

The story of Cai Yong‟s death changed had an important influence on the way the

Shiji was interpreted. The story is of course set in the late Eastern Han, but we know it

only from Six Dynasty texts. It concerns the brilliant literary artist Cai Yong. Anecdotes

about him in the Shishuo xinyu 世說新語 [New account of tales of the world]123 marvel

at his talent but do not mention his death in 192 CE. According to his chapter in the Hou

Hanshu, however, the circumstances of his death were as follows:

When [Dong] Zhuo was executed, Yong was at a meeting [called by] the
Minister of Works Wang Yun. In a completely involuntary manner, [Yong]
sighed as he spoke of [Zhuo‟s death], and there was a change in his
countenance.124 Yun suddenly roared at him, saying, “Dong Zhuo was a great
villain in this land, and nearly toppled the house of Han. You were a servant of
the [Han] ruler, and the appropriate behavior would have been to share our anger
123
E.g., SSXY 9:1 (499), 26:20 (840).
124
Cai Yong served under Dong Zhuo‟s administration as a General of the Household Gentlemen 中郎將
(trans. Bielenstein, Bureaucracy, 24-27). According to Cai Yong‟s Hou Hanshu chapter, “[Dong] Zhuo
valued Yong‟s talent and learning, and treated him very generously. Every time [Zhuo] assembled [guests]
for a feast, he would as a usual procedure order Yong to play the zither and also participate in the
discussion. Yong, for his part, always had something to correct or add” [卓重邕才學，厚相遇待，每集
讌，輒令邕鼓琴贊事，邕亦每存匡益] (HHS 60B.2006). A further anecdote, perhaps posthumously
concocted by defenders of Cai Yong‟s reputation who did not like to see him so contented with a notorious
usurper, puts a less harmonious cast on the relationship:
Zhuo most often did exactly as he pleased [without listening to others]. Yong resented the fact
that his suggestions were often not followed. He said to his cousin Gu, “Lord Dong‟s character is
rigid and the [way] he goes about things is incorrect. In the end, it will be difficult for him to
succeed. I would like to flee east to Yanzhou, but the road is long and it will be hard to get
through. How would it be if I hid away east of the mountain 124 and awaited [the outcome of
things]?” Gu said, “Your appearance is different from that of an ordinary person. Whenever you
go out people gather together and fill [the streets], in order to look at you. Because of this, would
it not indeed be difficult to conceal yourself?” Yong therefore desisted. [卓多自佷用，邕恨其言
少從，謂從弟谷曰：「董公性剛而遂非，終難濟也．吾欲東奔兗州，若道遠難達，且遯逃
山東以待之，何如？」谷曰：「君狀異恆人，每行觀者盈集．以此自匿，不亦難乎？」邕
乃止．] (ibid.)

222
 [at Zhuo]! But instead, you treasured the way he treated you personally, and
forgot your higher obligations! Now when heaven has executed the villain, you
perversely grieve for him. How is this not rebelling along with him, then?!”
Thus [Yun] had [Yong] arrested and given to the Commandant of Justice125 for
the punishment of his crime. Yong pleaded guilty to everything but begged to
have his face tattooed and his feet chopped off [rather than being executed], that
he might continue his work on the history of the Han. The great officers all pitied
him and sought to rescue him, but they were unable to do it. [及卓被誅，邕在司
徒王允坐，殊不意言之而歎，有動於色．允勃然叱之曰：「董卓國之大賊，
幾傾漢室．君為王臣，所宜同忿，而懷其私遇，以忘大節！今天誅有罪，而
反相傷痛，豈不共為逆哉？」即收付廷尉治罪．邕陳辭謝，乞黥首刖足，繼
成漢史．士大夫多矜救之，不能得．]
The Grand Commandant126 Ma Midi took horse and hastened to where
Yun was, saying “Bojie (=Cai Yong) is one of the rarest talents in this wide world,
and knows much about the affairs of the Han. Should he be allowed to complete
his continuation of the Han history, it would become a great classic of the age.
Furthermore, he is loyal and filial and pure, and you have arrested him without
him having committed any particular crime. If you execute him, will you not lose
the respect of the people?” Yun said, “Formerly, Emperor Wu failed to kill Sima
Qian, allowing him to make his defamatory text, and pass it down to later
generations. Now when the throne of our land is in decline, and its ritual vessels
are unsteady, we cannot let this toadying servant take up his brush and attend
upon our young ruler. There would be no gain in [the emperor‟s] sagely virtue,
and [the reputation] of our own party would once again be endangered by his
satires.” Midi retired and told someone, “Surely Lord Wang is not long for this
world! Good men are a guiding line for the state, and regulations are its model.
If he annihilates the guiding line and discards the model, how can he continue for
very long!?” So it was that Yong died in prison. [太尉馬日磾馳往謂允曰：「伯
喈曠世逸才，多識漢事，當續成後史，為一代大典．且忠孝素著，而所坐無
名，誅之無乃失人望乎？」允曰：「昔武帝不殺司馬遷，使作謗書，流於後
世．方今國祚中衰，神器不固，不可令佞臣執筆在幼主左右．既無益聖德，
復使吾黨蒙其訕議．」日磾退而告人曰：「王公其不長世乎？善人，國之紀
也；制作，國之典也．滅紀廢典，其能久乎！」邕 遂 死 獄 中 ．]127

Reading this anecdote with Sima Qian‟s story in mind, the first thing we notice is that Cai

Yong‟s disaster is a kind of figural echo of Sima Qian‟s. Like Sima Qian, Cai Yong was

not openly disloyal to the person in power; it was merely that he expressed too much

sympathy for someone perceived as a dangerous traitor. Like Sima Qian, Cai Yong was

125
Translation of the title, tingwei 廷尉 from Bielenstein, Bureaucracy, 38-39.
126
Taiwei 太尉, trans. ibid., 12-14.
127
HHS 60B.2006.

223
willing to accept shameful mutilating punishments rather than death—in hopes of being

able to write a historical text.

But here the resemblance ends. Compared to Sima Qian, Cai Yong was happier

in his friends and unhappier in his fate. Sima Qian complained, “Of my friends, not one

would save me” [交遊莫救].128 Cai Yong, on the other hand, had many would-be

rescuers, including Ma Midi. Yet Sima Qian lived on, wrote his history, and died

apparently unmourned (certainly unremarked by history). Cai Yong, on the other hand,

died without ever producing his history, but “of the nobles and classicist scholars, there

was not one who did not weep for him” [搢紳諸儒莫不流涕].129

Let us consider, for a moment, the case of Cai Yong by itself. Wang Yun‟s

misgivings about Cai Yong‟s potential historiographical efforts seem to be twofold: first,

that whatever he would write could have a potentially destabilizing effect on the already-

deteriorating political situation, and second, that Cai Yong might write something that

would damage Wang Yun and his group‟s current and future reputation. In short, having

made an issue of Cai Yong‟s involuntary regret for Dong Zhuo, Wang Yun cannot let

him go unpunished. However, any punishment short of death would leave Wang Yun

and his group—even the young emperor himself—exposed to the power of a historian‟s

revenge. That Wang Yun was willing to disregard the pleas of many officials of the time

and execute “one of the rarest talents in this wide world” clearly shows how much he

feared this revenge—and by extension, how successful Sima Qian‟s presumed efforts in

that regard had come to be considered.

128
HS 62.2730.
129
HHS 60B.2006.

224
 We need not believe in the literal historicity of this anecdote to learn something

from it about the changing view toward the Sima Qian story and the idea of a subtly

satirical Shiji. Placing Wang Yun (not a Heaven-appointed emperor) in the position of

authority, and portraying the sympathy of all the nobles as being on the side of Cai Yong,

the story then proposes the comparison with Sima Qian. Ban Gu had contextualized

Sima Qian as a subject who had clumsily failed in his duty to his ruler and then, being

punished, had vented his misplaced resentment upon the dynasty under which he lived.

Fan Ye (and others, see below) redefined the terms. Where Ban Gu had seen failed

remonstrance and disloyal resentment, the Cai Yong anecdote emphasizes spontaneous

expression of emotion and the opposition between fallible authority130 and valuable, not-

to-be-squandered talent.

Sima Qian as the Recluse‟s Other

Debates on Sima Qian that took place within the realm of Six Dynasties

historiography questioned the appropriateness of his critiques, but not the appropriateness

of his involvement in the official realm. Yet Sima Qian and his story seem to have

played a minor role in another type of debate as well, where he served as a foil or contrast

to recluses, figures on the verge of disengagement.

That this came to be Sima Qian‟s reputation is not surprising. The “Youxia

liezhuan” 遊俠列傳 [Arrayed Traditions of the Roving Warriors] (SJ ch.124) contains

the pronouncement that, “If we sincerely compare the warriors of the hamlets and

130
Fan Ye also notes that “Later, Yun repented and wanted to stop [Cai Yong‟s death], but he was too late”
[允悔，欲止而不及] (HHS 60B.2006).

225
villages to men like Ji Ci and Yuan Xian131 in terms of actual authority and the power and

effect of their actions on their own times, then the former so far surpass the latter that

they can hardly be discussed on the same day” [誠使鄉曲之俠，予季次﹑原憲比權量

力，效功於當世，不同日而論矣].132 To do-nothing paragons of virtue, Sima Qian was

understood as preferring men of dubious moral standing who nonetheless acted boldly

and influentially: the assassins, the warriors, the wayward generals and talented retainers.

The Shiji even goes so far as to question the motives of recluses (along with other types

generally considered to be motivated by high moral principles, writing, “The hermits of

cliffs and caves establish a reputation for purity of conduct, but what is their ultimate

objective? Their objective is simply wealth” [隱居巖穴之士設為名高者安歸乎？歸於

富厚也].133

Ban Biao and Ban Gu found these remarks so objectionable that it became one of

their most serious criticisms against Sima Qian, that he (in Ban Biao‟s words)

“denigrated modest restraint and valued vulgar achievements” [賤守節而貴俗功]134 or

(as Ban Gu put it) “disparaged scholars in retirement and promoted heroic

131
Legendary disciples of Confucius who, as the text earlier explains,
were common village men: they studied texts, and cherished independence of action and the virtue
of the superior man. Their principles were not in harmony with their age, and their age in turn
merely ridiculed them. Therefore they lived in barren hovels with vine-woven doors, wearing
rough clothes and eating coarse food so meager that they never got their fill. [季次﹑原憲，閭巷
人也，讀書懷獨行君子之德，義不苟合當世，當世亦笑之．故季次﹑原憲終身空室蓬戶，
褐衣疏食不厭．] (SJ 124.3181)
My translation is inspired by, but altered from, Burton Watson‟s (Han II 409-410).
132
SJ 124.3183. The “Huozhi liezhuan” 貨殖列傳 [Arrayed Traditions of the Merchants] (SJ ch.129) goes
further in denigrating Yuan Xian (better known in the Lunyu as Zisi 子思): “Yuan Xian could not get even
enough chaff and husks to satisfy his hunger, and lived hidden away in a tiny lane. Zigong rode about with
a team of four horses attended by a mounted retinue…. Kongzi‟s fame being spread throughout the realm
was really altogether due to the assistance of Zigong” [原憲不厭糟穅 ，匿於窮巷．子貢結駟連
騎．．．夫使孔子名布揚於天下者，子貢先後之也] (SJ 129.3258).
133
SJ 129.3271.
134
HHS 40.1325.

226
scoundrels…honored position and profit, while heaping shame on poverty and humble

station” [退處士而進姦雄…崇勢利而羞賤貧].135

Huangfu Mi 皇甫謐 (215-282), in the preface to his Gaoshi zhuan 高士傳

[Traditions of lofty gentlemen] complained about the many recluses that Sima Qian

„omitted‟ from his history.136 He also included an anecdote which seems specifically

designed to show the error of Sima Qian‟s position. It takes the form of an exchange of

letters, purportedly between Sima Qian and a recluse named Zhi Jun 摯峻. Zhi Jun was

said to be from Chang‟an and was allegedly a good friend of Sima Qian‟s. After Zhi Jun

went away to hide himself at Mount Qian, Sima Qian wrote to persuade him to return and

take up office in (according to Huangfu Mi) the following words:

I, Qian, have heard that what the True Gentleman values is the Way. Of its three
essentials, the highest is to establish one‟s virtue. Next is to establish one‟s words.
And third is to establish one‟s deeds. I humbly consider that you, Boling, surpass
others in talent and ability, that your aspirations are high and lofty. You value
your person and so are pure as ice, flawless as jade. You do not become
entangled with responsibilities through petty actions. To be certain, your name is
already honored indeed! Yet, you have not yet completely fulfilled that from
which the highest essential proceeds. I pray that the honored master might
consider this for a moment. [遷聞君子所貴乎道者。三太，上立徳，其次立
言，其次立功。伏惟伯陵材能絶人，髙尚其志，以善厥身冰清玉潔，不以細
行荷累。其名固已貴矣。然未盡太上之所由也。願先生少致意焉。]137

I know of no Shiji scholar who takes this letter seriously as a coming from Sima Qian‟s

hand. It seems to have been concocted simply to set up Zhi Jun‟s letter, in which he

defends the practice of reclusion. Aat Vervoorn doubts not only the authenticity of Sima

Qian‟s letters but also the historical existence of Zhi Jun.138 Alan Berkowitz disagrees,

135
HS 62.2738. Whether these figures were actual or mere historical projections of a later fascination is of
course unclear and open to debate. See discussion below.
136
GSZ “Preface” 608.
137
GSZ 2.621-622.
138
Vervoorn 114.

227
suggesting that though Zhi Jun‟s story “may have been somewhat embellished…by all

indications he was more than a historiographical foil fabricated by Huangfu Mi or one of

Sima Qian‟s detractors,” and offers several other pieces of evidence and potential sources

for Zhi Jun‟s existence.139 Burton Watson in any case considered it Sima Qian‟s letter (as

well as the fu) to be “too fragmentary and suspect to merit consideration” in his study.140

The letter need not be authentic, however, for it to reveal something of how Sima

Qian‟s story was interpreted during the period. The meat of the anecdote, as far as this

study is concerned, comes after Zhi Jun‟s spirited reply.141 The anecdote continues:

Jun‟s defense of purity did not shift from this [stand]. Qian occupied the office of
Senior Archivist, and because he acted as a roving persuader for Li Ling,
descended to the punishment of rottenness, and thus to great shame and
humiliation. Jun, in accordance with his loftiness, never did take office and died
at [Mount] Qian. The people established a temple for him there, where they
offered and prayed to him for generations without end. [峻之守節不移如此。遷
居太史官，為李陵遊説，下腐刑，果以悔恡被辱。峻遂髙尚不仕，卒於岍。
岍人立祠，世奉祀之不絶。]142

This is a version of the view that Wang Chong and Ban Gu held, that Sima Qian‟s

misfortune was his own fault. But where those Han dynasty thinkers held that the

disaster was a result of fate or miscalculation, Huangfu Mi (or whoever framed the

exchange) ascribed it to a fundamental error in Sima Qian‟s attitude toward the political

realm.

139
Berkowitz 9, 81 n.73.
140
Watson, Ssu-ma Ch‟ien, 200 nt.2.
141
In his reply, as it appears in the Gaoshi zhuan, Zhi Jun defends reclusion with magnificent modesty,
concluding: “I want only to bow and bend, free from care, wandering at ease through my remaining age”
[徒欲偃仰從容，以遊餘齒耳] (GSZ 2.622).
142
Ibid.

228
“A Danger for One‟s Writing”: Wenxin Diaolong and Fan Ye

The Wenxin Diaolong sums up a kind of consensus view on the connection

between Sima Qian‟s life and the Shiji. Comments on, or influenced by, Sima Qian are

scattered throughout the text, but only one seems to have any bearing on what Liu Xie

might have thought about Sima Qian‟s life as it pertained to his work: “[Sima] Qian and

[Ban] Gu attained mastery,” wrote Liu Xie, “yet they have both been subject to criticism

by each subsequent generation. If one lets one‟s personal situation interfere with

uprightness, it is a danger for one‟s writing” [遷固通矣，而歷詆後世。若任情失正，

文其殆哉].143 The Wenxin diaolong does not enter any further into the question than

this.144

This one remark, however, is an interesting one. An example of the type of

criticism it alludes to can be found in Fan Ye‟s Hou Hanshu evaluation of Ban Biao and

Ban Gu:

[Ban] Biao and Gu criticize [Sima] Qian, considering that his judgements differ
from those of the sage. However, [Gu‟s] discussions and arguments often belittle
pure virtue unto death, condemn uprightness and directness, and do not present
sacrificing oneself for the perfection of benevolence as an act of excellence.
Therefore, he goes too far in making light of benevolence and morality, and
denigrating constancy in virtue. Gu injures Qian [in saying] that for all his
breadth of learning he was not able to have the wisdom to avoid the extreme
punishment [that befell him]. Yet [Gu] himself also ran afoul of a particularly
cruel death. He was wise enough to reach [his position], but not wise enough to
keep to it. Alas, how the ancients in making arguments suffer from the problem
of the eye and the eyelash.145 [彪、固譏遷，以為是非頗謬於聖人．然其論議
常排死節，否正直，而不敘殺身成仁之為美，則輕仁義，賤守節愈矣．固傷

143
WXDL 16.165.
144
It makes passing references to the “Letter in Reply to Ren An” and to the “Lament for Unappreciated
Scholars” but, aside from the one above-quoted remark, does not attempt to bring the interpretation of the
history together with the autobiographical content.
145
The longer form of the expression is 目不見睫. Similar to the saying that before you can remove the
mote from your neighbor‟s eye, you must first remove the beam from your own—in this case, your eye can
see others clearly but you are completely unable to see your own eyelashes.

229
 遷博物洽聞，不能以智免極刑；然亦身陷大戮，智及之而不能守之．嗚呼，
古人所以致論於目睫也！]146

Fan Ye does not deny that Sima Qian may have suffered from some of the faults that Ban

Gu assigns to him. He even writes of Ban Gu (perhaps implicitly in contrast to Sima

Qian) that “the way Gu narrates events is not incited to deceptiveness” [若固之序事，不

激詭]. But Fan Ye has begun to doubt that Ban Gu himself had any right to criticize.

Liu Xie displayed a similar tendency when he wrote that despite the penetrating

understanding of the two master historians, they continued to suffer criticism. The reason

is because of their personal involvement with the events they narrate.

The Wenxuan Context and the Expansion of the Sima Qian Romance

Given his monumental stature in later times, Sima Qian played a surprisingly

minor role in the Wenxuan 文選 [Selections of refined literature].147 The Wenxuan

explicitly omits historical narratives,148 but includes numerous (and some very lengthy)

prose pieces by Ban Gu.149 In contrast, Sima Qian has only one piece included in the

146
HHS 40B.1386.
147
The Wenxuan was an extremely influential anthology of prose and verse compiled by a group of
scholars working under the patronage of Xiao Tong 蕭統 (501-531), the heir-apparent of the Liang 梁
dynasty (502-557). For further background and partial translation, see David Knechtges, Wenxuan, or
Selections or Refined Literature vols.I-III.
148
As the “Preface” explains, histories were originally made “to praise right and censure wrong” [所以褒貶
是非] or “record and distinguish differences and similarities” [紀別入聲異同]. They were not first and
foremost literary works. Xiao Tong does, however, include “Judgments and Treatises with an intricate
verbal eloquence” [讚論之綜緝辭采] and “Self-Narrations and Evaluations interspersed with literary
splendor” [序述之錯比文華], considering that “their matter is the product of profound thought, and their
principles belong to the realm of literary elegance” [事出於沈思，義歸乎翰藻] (WX “Xu” 3; Knechtges,
Wenxuan I.90-91).
149
It is Ban Gu‟s “Rhapsody on the Two Capitals” 兩都賦 which begins the Wenxuan. Others of his works
are included in WX 14.635 ff., 45.2015 ff., WX 50.2226 ff., and WX 56.2406 ff. Particularly telling is that
the Wenxuan‟s section on historical narratives contains three different zan 贊 [evaluations] by Ban Gu and
one by Fan Ye (WX 50.2229 ff.), but none by Sima Qian.

230
anthology, his “Letter in Reply to Ren An.”150 One might feel that the now-beloved

historian has been given short shrift, especially in comparison to Ban Gu, and this may

have been no accident in light of the Wenxuan‟s royal sponsorship and the Shiji‟s

reputation as a “defamatory text” which was thought, at least by some, to have spitefully

criticized the ruling dynasty of its time. Still, the reanthologization of Sima Qian‟s

“Letter” almost certainly had an important positive effect on how educated readers would

come to read his autobiography, and ultimately the Shiji as well.

The Wenxuan context for the “Letter in Reply to Ren An” is quite different from

its Hanshu setting. In the Wenxuan, Sima Qian‟s letter is situated between a letter

purportedly written by Li Ling 李陵,151 and one by Sima Qian‟s grandson Yang Yun (楊

惲 d. 54 BCE).152 They are the first three letters in the Wenxuan section for the epistolary

genre.

The “Letter in Reply to Su Wu,” is only one of several extant letters supposedly

exchanged by Li Ling and Su Wu.153 It begins with an address contrasting Li Ling‟s own

unhappy circumstances with his friend‟s happier ones.154 It goes on to describe the

difficult conditions under which Li Ling is living—both because of his barbarian

surroundings and because of the execution of his family by the Han court. There then

follows a description of his fateful military campaign against the Xiongnu, ending in his

150
Wenxuan title: “Letter in Reply to Ren Shaoqing” 報任少卿書. Shaoqing was Ren An‟s courtesy name.
151
“Letter of Reply to Su Wu” 答蘇武書.
152
“Letter in Reply to Sun Huizong” 報孫會宗書.
153
K.P.K. Whitaker discusses the corpus in detail. See “Some Notes on the Authorship of the Lii Ling/Su
Wuu Letters, Part I.”
154
Li Ling was sent to do battle against the Xiongnu, was eventually defeated, and surrendered, incurring
the wrath of the Han Emperor Wu. Su Wu was sent to the Xiongnu as an envoy; he was taken prisoner but
refused to renounce his loyalty to the Han, and was released after nearly two decades of captivity, returning
to (as the letter has it) “a glorious reputation” and “an imperishable name.” Translations of this letter are
from Giles, 82-86.

231
capture, and a further defense of his decision not to commit suicide. The next section

supposedly responds to a passage in Su Wu‟s letter, an attempt to persuade Li Ling that

“the house of Han never fails to reward a deserving servant” [漢與功臣不薄].155 The

letter vehemently denies this, citing examples of others who had been mistreated by Han,

including his own grandfather Li Guang, and Su Wu himself, whose exceptional loyalty

(the letter argues) deserved more than the somewhat insignificant reward he received for

it. In closing, the letter explains that because of all this, Li Ling has renounced the Han

and decided never to return.

The authenticity of “Letter in Reply to Su Wu” (as well as of the other letters and

poems supposedly exchanged by the two men) has been questioned since the Tang. Liu

Zhiji 劉知幾 (661-721), for example, wrote, “When one surveys its style, one not would

classify it as having been written by a person of the Western Han. It is probably made by

someone from a later period, who falsely attributed it to Li Ling” [觀其文體，不類西漢

人，殆後來所為，假稱陵作也].156 The work‟s reputation received another serious

blow when the great Su Shi 蘇軾 (1036-1101) pronounced unequivocally that its

“speeches and expressions are frivolous and shallow” [詞句儇淺] and that “it is proper to

consider that it was the imitative work of a novice of the Qi or Liang dynasty. Certainly,

it was not written by a man of the Western Han” [正齊梁小兒所擬作，決非西漢文].157

Though the letter has continued to have its defenders,158 K.P.K. Whitaker‟s rigorously

155
WX 41.1851.
156
STTS 18.525.
157
From Su Shi‟s “Letter in Reply to Supervisory Officer Liu Mian” 答劉沔都曹書. Translated on Chung,
“A Study of the Shu,” 322.
158
Perhaps most recently, Zhang Peihang and Liu Jun, whose 1998 article focuses mostly on defending the
authenticity of the poems, but who also spend a page defending the letter. They do not, however, answer

232
researched and argued article, “Some Notes on the Authorship of the Lii Ling/Su Wuu

Letters” paints a fairly damning picture. One of the most vexing problems is that Sima

Qian‟s “Letter in Reply to Ren An” (which Li Ling could not have seen) seems to have

supplied the description of Li Ling‟s experience fighting the Xiongnu.

None of this would necessarily have been clear to Six Dynasties readers of the

Wenxuan, however. To them, the heart-rending expression of Li Ling‟s painful situation

could well have been influential in justifying Sima Qian‟s decision to defend him. If Li

Ling was made out to be a hero, then Sima Qian was one too, for courageously defending

him.

In contrast to Li Ling‟s and Sima Qian‟s cases, the inclusion of Yang Yun‟s

“Letter in Reply to Sun Huizong” in the Wenxuan probably did not improve Yang Yun‟s

reputation. Indeed, it may well have had the opposite effect. Yang Yun‟s Hanshu

biography (Hanshu 66) is an ambivalent and complex portrait.159 Yang Yun‟s letter,

anthologized there, is a key element of the plot: writing it marks a turning point in Yang‟s

career, the point beyond which there is no possible return from disgrace.160 In its second

appearance, it even serves as the final nail in his coffin and as such holds a certain

all the objections against the letter, and their efforts to defend it are more a testament to its desirable
qualities than to any renewed possibility that it might be authentic.
159
Yang Yun became involved in an exchange of (exaggerated) accusations with an enemy at court and
ended up being stripped of his rank. At that point, he went to live on his family land and began to make
money through grain speculation. His friend Sun Huizong wrote a letter to him reprimanding him for this
improper behavior, and the anthologized letter is Yang Yun‟s reply, rejecting his friend‟s well-meaning
advice and repaying it only with bitterness and sarcasm. See “Gongsun Liu Tian Wang Yang Cai Chen
Zheng zhuan” 公孫劉田王楊蔡陳鄭傳 (HS 66.2889-2897).
160
David Knechtges has called Yang Yun‟s letter a “masterful justification of the eremitic life” (Knechtges
1982, 43), and this judgment is closely echoed by Aat Vervoorn, who calls it “a masterly evocation of the
pleasures to be found in a [hermit‟s] life” (Vervoorn 94). However, Chinese readers have found it to be a
most uncomfortable document. “The resentment accumulated deep in his heart over a long period of
depression is like water from a burst dyke,” writes Zhang Weifang 張偉芳; and, “the acrimonious and
piercingly sharp sword-edge [of Yang‟s rhetoric] leaves Sun Huizong without any ground to stand on”
(“Bao Sun Huizong,” 9). Min Zeping 閔澤平 too complains of Yang‟s “haughty attitude” and writes,
“Yang Yun had talent but he was self-important about his abilities, and harshly intolerant” (“Handai
shuxinti,” 82).

233
fascination. But standing in the Wenxuan beside the much subtler and more repentant

letter of Sima Qian, Yang Yun‟s comes off as harsh and unpleasing, particularly in its

attack on the recipient, Yang Yun‟s “friend” Sun Huizong. Sima Qian, who lost not just

his rank but even his manhood, admits his mistakes and continues humbly to serve his

emperor while sublimating his resentment into the creation of a great historical work.

Owing to doubts about the authenticity of “Letter in Reply to Su Wu,” few

scholars have considered the three letters as a group. It is clear to anyone who reads them

together, however, that the one essential characteristic they have in common is a sense of

resentment. The immediate object of resentment is the political situation. As Whitaker

put it, the Li Ling and Sima Qian letters both “dwell on the same theme of injustice

meted out by the autocratic government to an honest but powerless individual.”161 She

might have added the Yang Yun letter as well, in which open denunciations of slanderers

reveal the writer‟s negative opinion about a dynasty whose emperor is so ready to believe

their slanderous remarks.

As discussed above, the Hanshu context for the “Letter in Reply to Ren An” was

an overall evaluation of Sima Qian‟s life and work. Ban Gu may have admired the

letter‟s expressive power, but blamed Sima Qian for his misfortune.162 Wenxuan readers,

with the example of the Han‟s fall to look back on, would be less inclined to blame

someone merely because he failed to preserve himself from harm. Furthermore, situated

(in the Wenxuan) between Li Ling, who turned his back on his dynasty forever, and the

impulsive, unrepentant Yang Yun, Sima Qian‟s position seems the least culpable.

Contextual arguments are of course not definitive. Nonetheless, the widespread

161
Whitaker “Lii Ling/Su Wuu,” 119.
162
Concluding, as mentioned above, that Sima Qian “failed to understand how to keep himself from harm.”

234
availability of the Wenxuan context for Sima Qian‟s letter seem to have gradually

changed the way people viewed the man and his life. While Ban Gu had cast Sima

Qian‟s movingly-expressed resentment as the regrets of a man who has been unwise, the

Wenxuan implicitly presents him, together with Li Ling and Yang Yun, as essentially

innocent men betrayed by their sovereign.

Putting together the Six Dynasties readings, sparse as they are, we see a gradual

alteration in attitude toward Sima Qian‟s works. Huangfu Mi, whose interest lay more

with remarkable men who chose not to engage with the world, presented an anecdote

arguing that Sima Qian‟s mistake lay, not in defamation or miscalculation, but in ever

trying to affect the political world to begin with. Fan Ye‟s and Liu Xie‟s remarks reveal

an assumption that Sima Qian was autobiographically motivated in writing the Shiji.

Their attitude towards this is slightly negative, however, or at best ambivalent. The re-

contextualization of the “Letter in Reply to Ren An” in the Wenxuan resulted in a more

sympathetic reading of Sima Qian‟s life. By extension, it led to a more sympathetic

attitude toward potential autobiographical influences on the Shiji. Even if that sympathy

did not necessarily extend to an endorsement of criticizing the emperor or disagreeing

with the Sage, it was an essential component of later developments, as the discussion

below will show.

AUTOBIOGRAPHICAL READINGS IN THE TANG

The Tang dynasty was the period in which the Shiji might be said to have come

into its own, if only as an offshoot of the immense Tang dynasty scholarly interest in the

235
Hanshu.163 The eighth century Shiji commentator Sima Zhen 司馬貞 explained its

previously lackluster popularity in the following comment:

[Sima Qian‟s] intent is difficult to ascertain in detail. Compared to Ban [Gu's

Han]shu, [the Shiji] is subtle and of ancient-style substance, thus the famous
worthies of the Han and Jin did not yet know to value it. Its goodness is like
when Marquis Wen of Wei heard old music and only feared he would doze off.
[其意難究詳矣．比於班書，微為古質，故漢晉名賢未知見重，所以魏文侯
聽古樂則唯恐臥，良有以也．]164

Sima Zhen proceeds to give a „geneaology‟ of Shiji commentators beginning with Xu

Guang 徐廣 (352-425) and Pei Yin. Sima Zhen owns that Pei Yin “had a rough

understanding of [Sima Qian's] subtle intention,” but adds that (despite the encyclopedic

thoroughness of his invaluable Shiji Jijie commentary) Pei Yin still “did not exhaust the

[possibilities for] discussion.” Sima Zhen‟s succeeding remarks clearly show that in his

time he still felt the lack of a good commentary edition of the Shiji and that this was why

he decided to produce one. Neither was he the only one to do so. In comparison to the

Hanshu, the Shiji may have been seen as relatively under-studied; in any case, the greater

part of readerly energy expended on the Shiji during the Tang took the form of

commentary. Another major issue for Tang readers of the Shiji was clearly its structural

and generic qualities (discussed in Part II above). Still, the men of the Tang have left

behind some texts which are of some interest to our investigation of Sima Qian‟s

tragedies and how they relate to the Shiji. For the most part, these continue the Six

Dynasties trend of using Sima Qian‟s tragedy as a focus for the problem of ruler-minister

relations.

163
For a description of this phenomenon, see McMullen 1988, 163 ff.
164
Shiji suoyin “Preface”.

236
Li Xian‟s Hou Hanshu Commentary

The first text I will consider is the Hou Hanshu commentary that took form under

the sponsorship of the Prince Li Xian 李賢 (654-684). He was the second son of Tang

Gaozong and Wu Zetian, and was also briefly the heir apparent of the realm. During his

time as heir apparent (in the mid and late 670s), he gathered together a group of scholars,

who compiled the Hou Hanshu commentary.165 Their work was later submitted to and

approved by Tang Gaozong. Their comment on the story of Cai Yong‟s death (discussed

above) provides one of the first concrete readings of the Shiji as autobiographically

motivated. They wrote:

Whenever historians recorded affairs, it was necessary to write of good and evil
deeds. We might say that in compiling the Shiji, it is just in those matters in
which the house of Han did wrong that Sima Qian invariably made defamations.
It is not only intended to mean the person of Emperor Wu, but also the matter of
Gaozu‟s being pleased with the steward and Emperor Wu‟s taxing capital and
problems with embezzling. Ban Gu brings it all together, saying, “When Sima
Qian compiled his text, he completed the words of a single family. When
punishment was visited upon him, he used subtle language to criticize and ridicule,
denigrating and deriding his own times. He was not a scholar with a sense of
rightness.” [凡史官記事，善惡必書．謂遷所著史記，但是漢家不善之事，皆
為謗也．非獨指武帝之身，即高祖善家令之言，武帝筭緡、榷酤之類是也．
班固集云：「司馬遷著書，成一家之言．至以身陷刑，故微文刺譏，貶損當
世，非誼士也．」]166

The first few lines of this passage are especially ambiguous. The implication of the first

statement is that all historians must record bad deeds, not just good ones. If that were the

case, then Sima Qian writing about the Han emperors‟ bad deeds would be not only

perfectly normal but even obligatory—if indeed they did make such mistakes. But, the

commentators add, only when Sima Qian recorded the Han rulers‟ mistakes did he

165
These scholars included Zhang Da‟an 張大安, Liu Neyan 劉訥言, Ge Xiyuan 格希元, and others (JTS
86.2831-32).
166
HHS 60B.2007.

237
somehow “make all his writings [about them] defamatory” (to translate the awkward

phrase somewhat more literally).

The commentators then give examples of what they considered defamatory (as

opposed to merely a record of bad deeds): first, Wudi zhi shen 武帝之身, which I have

rendered “the person of Emperor Wu.” It is not fully spelled out just what in the Shiji

this example was referring to. One supposes they might have been referring to the events

narrated in the “Feng Shan shu” 封禪書 [Treatise on the Feng and Shan] (Shiji ch.28).

The other two examples are more thoroughly spelled out. “The matter of Gaozu‟s

being pleased with the steward” refers to an incident in which the steward in Han

Gaozu‟s father‟s household suggested that it was improper for Gaozu to observe the

ordinary rites of a father toward his son (since his son was the emperor), and counseled

the father to behave submissively instead. Gaozu naturally protested and gave his father

an honorary title, but because Gaozu “was inwardly pleased with steward, he also

awarded the man five hundred catties of gold” [心善家令言，賜金五百斤].167 That this

is a defamatory story is clear even to the modern reader. The purpose of the anecdote,

from a historian‟s point of view, would have been to explain the honorary title awarded to

Gaozu‟s father. That Gaozu was also “inwardly pleased” could not have been a matter of

public record, and even though the reward to the steward might have been, in a sense it

was of negligible significance. Yet the inclusion of the story contributes to an

unflattering picture of Gaozu‟s character, built up elsewhere in the Shiji as well, seriously

undermining the dignity of the Han dynastic founder.

167
SJ 8.382.

238
 The other references are more difficult to identify. The phrase I have translated as

“capital tax”—literally taking a suan from strings (min)—seems to refer to a proposal

found in the “Ping zhun shu” 平準書 [Treatise on the Balanced Standard] (Shiji ch.30).

In order to obstruct merchants who were hoarding cash and profiting from it in various

ways, the great lords proposed that “the various merchants and those engaged in

secondary occupations of trade… should all be required to make a declaration of their

possessions and should be taxed at a rate of one suan on every two min of a thousand

cash” [諸賈人末作貰貸賣買…各以其物自占，率緡錢二千而一算].168 The Shiji jijie

quotes Li Fei 李 斐 (also echoed by the Shiji suoyin), as saying that “a min is the silk used

to string together cash; one string is a thousand cash, which yields twenty suan” [緡，絲

也，以貫錢也．一貫千錢，出二十算也]. In short, the proposal was to tax merchants

at a rate of fifty cash (one suan) from every two thousand they possessed, in effect, a two

and a half percent tax on capital.169 A number of other taxes were also instituted, on

boats and carts and craftsmen.

Probably the real sting of the proposal was that the penalty for false or incomplete

reports was confiscation and frontier service, an option which was heavily abused by

enforcers when the proposal went into effect, as the Shiji describes in colorful and

damning detail:

The emperor dispatched parties of assistants under the imperial secretary and
commandant of justice go to the various provinces and kingdoms and examine the
charges of concealed wealth. The wealth confiscated from the people as a result
was calculated in the billions of cash….Practically all the merchants and
tradesmen of middling means or better means were ruined. The people indulged

168
SJ 30.1430, trans. adapted from Watson, Han II, 72.
169
Burton Watson parenthetically supplies the information that a suan is equal to 120 cash, but does not
give his source. For convenience, I follow the Shiji commentators. In any case, the exact rate is not crucial
to my purpose here.

239
 in tasty food and fine clothing while they still had the opportunity, making no
effort to lay away any wealth for the future. [乃分遣御史廷尉正監分曹往，即
治郡國緡錢，得民財物以億計．．．於是商賈中家以上大率破，民偷甘食好
衣，不事畜藏之產業．]170

Passages of this sort were probably what Li Xian and his group meant in the reference to

Emperor Wu‟s taxing capital and embezzling. No doubt they had in mind, not merely the

tax policy and its unfortunate repercussions, but the entire tangle of policies, by turns

ineffectual and harsh, which Emperor Wu‟s economic advisors suggested in order to

control the currency and the acquisition of wealth. All of these are described minutely in

the “Ping zhun shu,” and while they might not constitute a direct attack on Emperor Wu

himself, such emphasis on the economic confusion that prevailed during his reign

certainly does not reflect well on him.

As for the final example raised by the Hou Hanshu commentators, which I have

translated as “problems with embezzling,” it is very murky indeed. The phrase, 榷酤,

usually refers to a government monopoly on the sale of alcohol. No reference to this

appears in the Shiji, although the government did have lucrative monopolies on salt and

iron. The only use of the phrase in the Shiji at all is by commentators discussing cases

which seem related to embezzling or other dishonest financial dealings.171 Very likely,

the general import is meant to be corruption among officials—another oblique criticism

of Emperor Wu—but the evidence is far from conclusive. Is it possible that Li Xian‟s

version of the Shiji was even more defamatory than the one we have today?

Li Xian‟s Hou Hanshu commentators (perhaps naturally, given their imperial

audience) sided with Ban Gu against Sima Qian on the matter of criticizing the Han. Yet

170
SJ 30.1435, trans. adapted from Watson, Han II, 77.
171
See SJ 59.1098-99 and SJ 120.3113.

240
their reasoning is curious. In mentioning “those matters in which the house of Han did

wrong” [漢家不善之事], they admit that the Han had done wrong. Thus, they are not

accusing Sima Qian of telling lies. What they seem to object to as “defamatory” 謗 is the

enthusiasm with which Sima Qian focused on such anecdotes, or the overall impression

they produce. In any case, the fact that Sima Qian had received harsh punishment at the

hands of Emperor Wu is presented (through the quotation from Ban Gu) as a key to their

interpretation.

Zhang Zhuo and an Emperor‟s „Mercy‟

I have shown so far that the Sima Qian/Li Ling tragedy was used in a variety of

ways. It often appears as an example of unwise action vis-à-vis the ruler or the state. Xu

Yanbo 徐彥伯 (d.714), for instance, mentioned Sima Qian in an essay on the importance

of correct speech (“Discussion of the Axis” 樞機論), as an example of the disasters

which befall those who speak unwisely.172 Bai Juyi 白居易 (772-846) would later

criticize Sima Qian outright for siding with Li Ling at all.173 Zhang Zhuo 張鷟 (ca.660-

740), however, made use of the Sima Qian story in a slightly different way, one which

shed more light on the details of Sima Qian‟s tragedy.

Zhang Zhuo received his jinshi in 679, and had a successful career. He was

several times in charge of examinations, and also at one time served as the District

Defender174 of Chang‟an. He held the rank of Censor (yushi 御史) in 695. His biography

in the Jiu Tangshu describes his words as humorous and facetious (詼諧), and perhaps

172
QTW 267.2718, JTS 94.3005.
173
QTW 677.6913-14.
174
Xianwei 县尉, see Hucker, Official Titles, 243.

241
this is what led to his troubles. At the beginning of the Kaiyuan era (sometime around

714), he was accused of defamation, namely, because “his words contained a great deal

of satirical criticism of the times” and that “he had spoken harshly against the Lingnan

project” [言鷟語多譏刺時，坐貶嶺南].175 It was at this time, one supposes, that Zhang

Zhuo wrote the “Chen qing biao” 陳情表 [Memorial expressing my feelings], which is of

interest here.

[These are] the words of your servant Zhuo, a piece of excrement who
deserves ten thousand deaths. Your servant, unworthy to be a courtier…
committed a crime which should bring about his death. I wish I could cut out my
own tongue and swallow the sound of my voice. I humbly await the sentence
which will turn my body to ashes and my bones to powder, submit willingly to the
executioner‟s axe. How could an insect lament its fate, a sparrow or mouse long
for its life? [And yet] this insignificant heart has that which has not been fully
[expressed]. [萬死糞土臣鷟言。。。罪應至死。自可鉗口吞聲。伏待刑書，
灰身粉骨，甘從斧鉞。豈可昆蟲惜命，雀鼠貪生？區區微心，有所未盡。]
All my life I was fond of study, and was quite fond of literary writing.
Though I have never been a literary genius, a request has come down to me for an
imperially commissioned „perusal draft‟. My recent poems, rhapsodies,
memorials, records, and other works might make up a volume. I have compiled a
draft for presentation, but the fair copy is not yet complete. I have held a position
of responsibility in this enlightened time, but I am about to suffer the extreme of
the law. I fear that Shiheng (=Lu Ji) will falter and the crane cry of the flowery
pavilion will be heard no more.176 Xi Kang looked back at reflection [of the sun]
and the music of Guangling was extinguished forever177… Formerly, Sima Qian
requested the “rotten punishment” in order to complete the Shiji. Emperor Wu of
the Han pitied his extreme earnestness, and compassionately allowed it. I beg and
pray that your majesty will fulfill that which in my heart I have longed for ten
thousand times, and grant your servant one hundred more days of life, that I may
collect records and make a fair copy for presentation for your court. [臣平生好
學，頗愛文章。雖不逮於詞人，濫流傳於視草。近來撰集詩賦表記等若干
卷。編集擬進，繕寫未周。負譴明時，方從極典。恐士衡止息，華亭之唳不
聞。嵇康顧影，廣陵之音永絕。。。昔司馬遷請就腐刑，以終史記。漢武帝

175
JTS 49.4023. This seems to be a reference to minister‟s Zhang Jiuling‟s 張九齡 project to build a road
that would open up the Lingnan area.
176
The reference is to a remark by Lu Ji 陸機; see JS 54.1480.
177
The story is related in the JS 49.1374, and in SSXY “Ya liang pian” 雅量篇 6.344. Roughly, Xi Kang
was about to be executed by the Sima clan. He looked at the image of the sun (顧視日影) and played the
music of Guangling on his qin, saying that since he had taught it to no one, it would, after his death, be cut
off forever.

242
 愍其至懇，矜而許之。伏願陛下遂臣萬請之心，覬臣百日之命，集錄繕寫，
奉進闕庭。]178

Like Sima Qian, Zhang Zhuo admitted his guilt but hoped to avoid execution long

enough to complete his literary work. The interesting thing about his presentation of the

Sima Qian story is that in his version, Sima Qian himself requested the punishment of

castration as an alternative to execution.

Neither the “Self-Narration” nor the “Letter” mentions such a request. The

“Letter” in particular is full of Sima Qian‟s bitterness at the humiliation of becoming a

eunuch.179 It is not immediately obvious that Sima Qian would have brought such a

punishment upon himself. Yet as Lü Xisheng 呂錫生 has argued, the punishment for

Sima Qian‟s crime of lèse majesté (wuwang zhushang 誣罔主上) should have been death;

castration was a punishment generally reserved for crimes of lasciviousness (yin 淫).180

The Hanshu, however, records an edict of Emperor Jing allowing those who have

committed a capital crime “but wish to receive the „punishment of rottenness‟ may be

permitted to do so” [欲腐者，許之].181 Lü Xisheng even cites another example, a

contemporary of Sima Qian named Zhang He 張賀, who was a retainer of the Wei heir

apparent. When the heir apparent‟s retainers were all executed after the Witchcraft

178
QTW 172.1749.
179
Sima Qian wrote there, for example, “That a man who has undergone [castration] is no longer fit to be
associated with is not the opinion of one age alone, but has been held since ancient times” [刑餘之人，無
所比數，非一世也，所從來遠矣] (HS 62.2727; trans. adapted from Watson, Ssu-ma Ch‟ien, 59). He
then gives a list of proper gentlemen avoiding the shame of the slightest association with eunuchs.
180
“Sima Qian gong xing xi yi,” 68.
181
HS 5.147.

243
Incident of 91 BCE,182 Zhang He requested and received the punishment of castration in

lieu of a death sentence.183

It is not clear whether Zhang Zhuo‟s reference reflects merely his own

understanding of the Sima Qian story or an understanding common in his time. In either

case, such an understanding makes good sense of the obsessive self-justification found in

both the “Letter” and the “Self-Narration.” If the terrible humiliation of castration was

Sima Qian‟s own choice, it would be that much more essential for him to make known to

the world the over-riding importance of the work for whose sake he endured it.

The political connotation of Zhang Zhuo‟s version of the story is also worth

noting. In the Han and Six Dynasties, the discourse surrounding Sima Qian‟s tragedy

often featured a wrathful Emperor Wu in an adversarial relationship with a resentful Sima

Qian. Here we have a merciful Emperor Wu granting Sima Qian the lesser punishment,

allowing him to live on and complete his work. The Cai Yong version of the story

(discussed above) features a similar understanding. But something has changed in the

intervening time, for Wang Yun called the Shiji a “defamatory text” and summarily

refused Cai Yong‟s request. In Zhang Zhuo‟s time, the sense of the Sima Qian‟s having

been inappropriately critical towards the ruler must have diminished somewhat, for why

otherwise would Zhang Zhuo have risked comparing it to his own literary work when

making a request for clemency?

Furthermore, Zhang Zhuo‟s request was successful. He was not executed but

rather exiled to Lingnan. Later the Minister of the Board of Punishments, Li Rizhi,

brought suit to complain that the punishment was too harsh, and Zhuo was allowed to

182
For a detailed description of the witchcraft incident, see Michael Loewe‟s Crisis and Conflict, 37-90.
183
HS 59.2651.

244
return from the frontier. He ended his days as in a minor post in the Transit

Authorization Bureau, and perhaps even completed his literary endeavors, though

ironically he produced no surviving work of any great significance.

Tang Commentators on Sima Qian‟s Bo Yi

Two of the three major commentaries on the Shiji—the Shiji suoyin 史記索隱

[Seeking hidden in the Shiji] by Sima Zhen and the Shiji zhengyi 史記正義 [Correct

meaning of the Shiji] by Zhang Shoujie—were written during the Tang dynasty. Despite

the promising title of the first (to those inclined toward autobiographical readings, such

readings must be among the Shiji‟s “hidden” meaning), and the thoroughness of both

commentaries, neither seems particularly concerned with using Sima Qian‟s tragedy to

interpret the Shiji.

There is one point in the Shiji where both Sima Zhen and Zhang Shoujie seem to

become uncharacteristically sensitive to the potential autobiographical significance of the

text. This is in the “Bo Yi liezhuan” 伯夷列傳 [Arrayed Traditions of Bo Yi] (Shiji

ch.60). In particular, there is a point where Sima Qian quotes the Lunyu, saying “The

gentleman hates not leaving behind a name when he is gone” [君子疾沒世而名不稱

焉].184 Sima Qian then gives two quotations from Jia Yi, “As a covetous man pursues

wealth, so an honorable gentleman pursues fame; as the proud will die for power, so the

commoner clings to life” [貪夫徇財，烈士徇名，夸者死權，衆庶馮生] and directly

afterwards, a quotation from the Yijing: “Those which shine with the same light are

mutually illuminating, and those who are of the same type seek each other out. Clouds

184
LY 15:20, trans. Lau, Analects, 135.

245
follow the dragon; wind follows the tiger. The sage arises and the ten thousand things are

manifest” [同明相照，同類相求．雲從龍，風從虎，聖人作而萬物覩].185 The

chapter goes on to describe how Bo Yi and Shu Qi were remembered because Confucius

praised them. It likens Yan Hui to a fly on the tail of a swift steed—he owes his

advancement to his teacher.

This passage, the final one of the chapter, prompted considerable comment from

both Sima Zhen and Zhang Shoujie. Sima Zhen writes (commenting on the Lunyu

quotation given above):

From this point on, although Sima Qian argues that Bo Yi had the Master to
glorify his name, that Yan Hui clung to the horse‟s tail and advanced, probably he
also wanted to subtly express how his own work was not completed, and that he
feared he would vanish from the world and his name would not be honored. Thus
he quotes Jiazi saying, “As a covetous man pursues wealth, so an honorable
gentleman pursues fame.” And again when he quotes [the Changes], saying,
“Those which shine with the same light are mutually illuminating, and those who
are of the same type seek each other out,” “Clouds follow the dragon; wind
follows the tiger”—he means to say that all things seek other things which are of
their own type. Thus the Honorable Senior Archivist is also saying of himself that
his conduct was modest and upright but his own age did not make use of him and
in the end he sank into condemnation and transgression. So he was of the same
type as Bo Yi and takes him as an analogy to put forth his own argument. [自此已
下，雖論伯夷得夫子而名彰，顏回附驥尾而行著，蓋亦欲微見己之著撰不
已，亦是疾沒世而名不稱焉，故引賈子「貪夫徇財，烈士徇名」是也．又引
「同明相照，同類相求」，「雲從龍，風從虎」者，言物各以類相求．故太
史公言己亦是操行廉直而不用於代，卒陷非罪，與伯夷相類，故寄此而發論
也．]186

Sima Zhen‟s comment is actually quite complex (not to say confusing) if examined

carefully. He seems to be making an analogy, but there are in fact two analogies. He

mentions Bo Yi and Yan Hui, both of whom owe their posthumous reputation to

185
SJ 61.2127, Zhouyi “Qian” 1.15. The first couplet is a paraphrase. In the Yijing it reads: “Those who
have the same tone resonate with one another; those which have the same qi seek each other out” [同聲相
應，同氣相求]. Still, the meaning is clearly quite similar.
186
SJ 61.2127.

246
Confucius. He then adds that because they were virtuous but not appreciated in their own

time, Sima Qian and Bo Yi were of the same type—i.e., in danger of being forgotten

unless rescued from oblivion by someone like Confucius. But Sima Zhen also says that

Sima Qian in writing this passage “wanted to subtly reveal how his own work of

ceaseless writing was also done out of fear that he would vanish from the world and his

name would not be honored.” Therefore in this context, and because this chapter of the

Shiji tells the story of Bo Yi, there is also an implied analogy between Sima Qian and

Confucius. In the Han, Confucius was honored perhaps first and foremost as the editor of

the Classics (particularly the Spring and Autumn Annals), remembered for his literary

endeavors just as Sima Qian hoped to be. Confucius honored Bo Yi as being without

resentment. Sima Qian also honored Bo Yi, but with the “Plucking Ferns” song,

disagrees with Confucius and recasts the ancient figure as being, in fact, full of poignant

resentment. Perhaps Sima Qian expected his readers to find this a more powerful and

affecting version of Bo Yi‟s story. If so, he was not entirely successful.

Bo Yi had Confucius; does he also need Sima Qian? Rather, it was Sima Qian

who needed Bo Yi. As Sima Zhen points out, Sima Qian merely used the chapter on Bo

Yi as an excuse to put forth a more general argument, relying on the fame of Bo Yi (as

well as that of Yan Hui) to give his argument force. Judging from the Suoyin comment

on this passage, Sima Qian would seem to occupy a position in between: in between the

clouds and the dragon, in between the wind and the tiger, in between the fly and the swift

steed. He hoped to be both the pure man of virtuous conduct and the one who confers

immortality upon such a man.

247
 Sima Zhen resolves this ambiguity in his next comment, which is attached to the

line about the sage arising and the ten thousand things falling into place:

[Sima Qian] also quotes the sentence, saying that when the sage arises and takes
his place, then the dispositions of the ten thousand things are all manifest.
Therefore he himself, in the present day, also is able to write a book that discusses
the comparative importance of the world‟s dispositions. [又引此句者，謂聖人
起而居位，則萬物之情皆得覩見，故己今日又得著書言世情之輕重也．]187

Sima Zhen here willingly made the comparison that Sima Qian himself did not quite dare

to make, placing Sima Qian in the position analogous to Confucius and suggesting an

analogy between their respective labors. Naturally Sima Zhen did not accord to Sima

Qian the same power as the sage is said to have, that of making things manifest. Still,

Sima Qian‟s work allows him at least to weigh events in the balance and make

judgements upon them.

Zhang Shoujie, though his comments on the Bo Yi passage are similar to Sima

Zhen‟s, tried to clarify the terms of the analogy. On the Lunyu quotation, he commented:

What the Princely Man hates is the fear that after he vanishes from the earth his
name will be annihilated and no longer honored. Those like Bo Yi, Shu Qi, and
Yan Hui established their reputation for pure conduct and later generations
honored them and told their story. The Honorable Senior Archivist wants
gradually to reveal the beauty of establishing a reputation through writing and
transmitting, as he himself did. [君子疾沒世後懼名堙滅而不稱，若夷﹑齊﹑
顏回絜行立名，後代稱述，亦太史公欲漸見己立名著述之美也．]188

Here, the analogy is slightly altered. It is not with respect to conduct that Sima Qian is

like Bo Yi. Rather, Sima Qian is like Bo Yi in the ability to establish a reputation (立名).

Sima Qian‟s means for doing this differs from Bo Yi‟s, in that Bo Yi‟s was a first order

virtue (pure conduct), while Sima Qian‟s is a second order virtue (celebrating those

187
SJ 61.2128.
188
SJ 61.2127.

248
whose conduct was pure) more closely resembling Confucius. In short, in Zhang

Shoujie‟s interpretation, Sima Qian is playing the roles of both Bo Yi and Confucius all

at the same time.

The Shiji zhengyi interpretation of the Jia Yi quotation is somewhat different from

Sima Zhen‟s. Regarding the first quotation, Zhang Shoujie writes, “The Honorable

Senior Archivist quotes Master Jia as a metaphor for the creation of the Shiji. „As a

covetous man pursues wealth, as an honorable gentleman pursues fame, as the proud will

die for power, as the commoner clings to life‟—thus did he make his Shiji” [太史公引賈

子譬作史記，若貪夫徇財，烈士徇名，夸者死權，衆庶馮生，乃成其史記].189

Here each of the four figures in the verse (the greedy man, the honorable gentleman, etc.)

seems to be an analogue for Sima Qian, and the thing they all have in common is their

passionate drive. In other words, Zhang Shoujie‟s comment tends to slightly de-

emphasize the special importance of the second line, and instead placed all four lines on

more or less equal terms.

Zhang Shoujie‟s comment on the last quotation, about the sage arising, arrives at

the same conclusion as Sima Zhen‟s, that Sima Qian is doing the work of the sage:

This [passage] shows real understanding. “The Sage has the virtue of nourishing
life; the ten thousand things have a disposition to be raised and cultivated. Thus
the [ten thousand things] respond to and resonate with the Sage.”190 This goes
together with “Those which shine with the same light are mutually illuminating”
above, words from the Zhou Yi figure of “Heaven.” The Honorable Senior
Archivist quotes these sorts of passages having to do with mutal resonance
because he wants to show the intention of his transmission and creation—to cause
the ten thousand things to become manifest. He lived just at a time five hundred
years after the death of Confucius,191 and thus wrote the Shiji to make manifest
189
SJ 61.2128.
190
A quotation from the Yijing commentary, though the present text reads “the ten thousand things have the
disposition to be nourished” [萬物有生養之情].
191
As modern scholars have pointed out, arithmetically, this is an approximation at best: Sima Qian lived
less than four hundred years after the death of Confucius.

249
 the ten thousand things. [此有識也．聖人有養生之德，萬物有長育之情，故
相感應也．此以上至「同明相照」是周易乾象辭也．太史公引此等相感者，
欲見述作之意，令萬物有覩也．孔子歿後五百歲而己當之，故作史記，使萬
物見覩之也．]192

Zhang Shoujie then goes on to quote passages from the “Self-Narration” about the five

hundred year sage cycle193 and about Sima Qian‟s understanding of the function of each

of the Classics.194 Like Sima Qian, he coyly stops short of calling the Shiji a seventh

Classic, but the implication is quite clear from his juxtaposition.

Sima Zhen‟s and Zhang Shoujie‟s autobiographical readings of the “Bo Yi

liezhuan” seem to depart somewhat from Sima Qian‟s tragedy, which has hitherto played

such a prominent role in autobiographical readings of the Shiji. Zhang Shoujie made no

mention of the tragedy at all here, and Sima Zhen merely alluded to it by saying that, as

was the case with Bo Yi, Sima Qian‟s “own age did not make use of him and in the end

he sank into condemnation and transgression.” Neither commentator, at least in

interpreting this passage, places much weight on the tragedy as having concrete influence

on what Sima Qian chose to write.

This negative result is actually quite revealing, however. The two commentators

are engaged in the exact type of autobiographical readings that would later become a

common way of linking Sima Qian‟s tragedy to his text. But they are scholars who

invested considerable time and effort into this text. Their own commentaries are like

flies on the tail of the Shiji, and they depend for their immortality on the Shiji‟s continued

propagation. Perhaps they wanted to paint the best possible picture of Sima Qian and his

work, a picture that elevates Sima Qian‟s imagined emotional state from unseemly

192
SJ 61.2128.
193
SJ 130.3296.
194
SJ 130.3297.

250
resentment to something approaching the sorrows of Bo Yi and Confucius. In short, in

order to achieve a truly sympathetic autobiographical reading in the Tang, it may have

been that Sima Qian‟s tragedy had to be despecified and subsumed under the more

general heading of „not meeting the time.‟ This is quite a different subject-ruler

relationship than most readers have imagined for Sima Qian (though perhaps not so

different than the one he imagined for himself): far from being tied to the petty faults of

his ruling dynasty, Sima Qian the immortalizer of virtuous men transcended not merely

his ruler but his entire era, almost ascending to the rarified realm of sagehood.

Liu Zhiji on the Li Ling Affair

Liu Zhiji was not terribly concerned with autobiographical readings of the Shiji.

He did, however, make an interesting observation regarding the texts from which much

of the Sima Qian biography is drawn.

It is clear from parts of the Shitong that Liu Zhiji knew as much as we do about

the details of the Li Ling affair, having read both the “Self-Narration” and the “Letter in

Reply to Ren An.” Liu observed, in the “Miscellaneous Sayings” chapter, that the

“Letter” spells out fully what the “Self-Narration” only hints at:

Sima Qian‟s “[Self]-Narrated Tradition” says that he had been the Senior
Archivist for seven years when “he met with the catastrophe of the Li Ling affair
and was hidden in darkness, bound in black ropes. He sighed deeply and said, „I
am to blame for this! My body is mutilated I can no longer be of use.‟”195 How
elliptical it is, him telling his story this way! From what he says—that “he met
with the catastrophe of the Li Ling affair and was hidden in darkness, bound in
black ropes”—it seems like he was captured along with Li Ling, and was
punished for that. And then again it seems like he was entangled with Ling, and
because of that was considered to have committed a crime against the state. In
this way, [Sima Qian] causes the reader to have a difficult time understanding the
details. We must rely on Ban Gu‟s having preserved the “Letter to Ren An,” as it
195
SJ 130.3300.

251
 narrates the entire matter for which Sima Qian was punished. Supposing we did
not have this record, would it have been at all possible to understand this matter?!
[司馬遷序傳云：為太史七年，而遭李陵之禍，幽於縲絏。乃喟然而歎曰：
是予之罪也，身虧不用矣。自叙如此，何其略哉！夫云遭李陵之禍，幽於縲
絏者，乍似同陵陷沒，以置於刑，又似為陵所間，獲罪於國。遂令讀者難得
而詳。頼班固載其與任安書，書中具述被刑所以。儻無此録，何以克明其事
者乎？]196

In the first section of this chapter, I explored the question of what readers who had

apparently not read the “Letter” thought had happened to Sima Qian. Ban Biao implied

that he was punished because of his failings as an historian, and Wei Hong went farther,

suggesting that the Li Ling affair was merely a pretext for Emperor Wu to punish Sima

Qian because Sima Qian‟s historical writings reflected poorly on the ruling house. It

should be noted that Wei Hong‟s version shows at least some awareness of what the Li

Ling affair was about, independent of the “Letter,” though if Yu Jiaxi‟s reading is correct,

Wei Hong had some of the details wrong. Finally, Wang Chong betrayed a fairly

thorough and detailed familiarity with the Li Ling affair, but lacked the sympathetic

attitude that might be expected from someone who had read the “Letter.” We can

probably conclude, then, that even without the “Letter” some knowledge of the Li Ling

affair would have survived. Still, Liu Zhiji‟s point is an important one: in the “Self-

Narration,” Sima Qian gives very little detail about his tragedy. Those who have

developed autobiographical readings of the Shiji have usually done so with separate

reference to the “Letter,” for the “Self-Narration” account of the tragedy is, as Liu Zhiji

observed, so spare as to be quite ambiguous.

196
STTS 16.460. Commentator Pu Qilong disagrees with Liu Zhiji “using the „Letter‟ to attack the „Self-
Narration‟” here, and adds—somewhat sardonically, I think—that if only Sima Qian had written that,
“ „After seven years he was accused of a crime because he argued for Li Ling,‟ then the matter would have
been made quite clear” [七年而以訟李陵獲罪，則由便明] (STTS 16.460).

252
 The lack of detail about the Li Ling affair is especially striking when contrasted

with the “Self-Narration” account of the other tragedy in Sima Qian‟s life, the death of

his father. If we consider the “Self-Narration” as a document intended to be a genesis

story for the Shiji, it might almost seem that Sima Qian wanted to place more emphasis

on his father‟s command as a motivating factor, rather than the Li Ling affair. And yet,

Sima Qian‟s work on the Shiji, from the narrative perspective of the “Self-Narration”,

does not seem directly connected to his father‟s command and subsequent death, while it

does seem directly connected with the Li Ling affair—so much so that readers, from Wei

Hong on down even to Liu Zhiji, consciously or unconsciously infer that the tragedy was

a result of the writing rather than a motivation for it.

Why was Sima Qian‟s “Self-Narration” so elliptical in describing the Li Ling

affair? Was it due to political sensitivity? to personal shame? or, as Pu Qilong suggests,

because “he feared it would be disrespectful toward the historical genre” [懼史體之褻

也]197? And again, why explain the affair in such detail in the “Letter”? Did he (as some

scholars have speculated) intend the “Letter” to survive for posterity and be read as a

supplement to his autobiography and interpretive key to the Shiji? Was that why he

omitted from the “Self-Narration” those details which he gave in the “Letter”? Given the

high estimation of Sima Qian to which we are accustomed, there is a temptation to read

all kinds of meaning into this discrepancy of detail.

Liu Zhiji, however, offered no particular interpretation. The word he used to

complain, lue 略 (elliptical), may carry some implication of carelessness,198 but overall

Liu Zhiji seems more concerned with pinning down the characteristics of a genre

197
STTS 16.460.
198
As, for example, in the common compound shulue 疏略.

253
(historical writing) than with explaining the vagaries of Sima Qian‟s tragedy.

Nonetheless, by complaining about the potential loss of Sima Qian‟s biographical details

(such as the Li Ling affair), Liu Zhiji showed that he considered them to be

historiographically important.

The Tang dynasty yields a few examples of autobiographical readings of the Shiji.

These are widely divergent in their sympathies. Li Xian‟s Hou Hanshu commentary still

endorsed Ban Gu‟s harshest criticisms, but Zhang Zhuo‟s use of Sima Qian‟s example in

begging for clemency suggests an overall increase in readers‟ tendencies to read Sima

Qian‟s story sympathetically. Sima Zhen‟s and Zhang Shoujie‟s comments on the Bo Yi

chapter even attempted to raise autobiographical interpretations of the Shiji out of the

entire “defamatory versus true” debate. They would place him instead in the company of

Confucius. Finally, Liu Zhiji‟s recognition of the “Letter‟s” importance, while not in

itself an autobiographical reading, affirmed the relevance of Sima Qian‟s story to Shiji

interpretation.

In the next chapter, I show how in the Song dynasty, the autobiographical reading

assumed something much more closely resembling the readings we are familiar with

today.

254
 Chapter 5

Personal or Political: Autobiographical Readings in the Song Dynasty

In the preceding chapters, I traced the development of autobiographical readings,

how readers connected Sima Qian‟s life-story with their interpretations of the Shiji. For

the most part, positions on this issue had remained stable since the Han dynasty. Those

who followed Ban Gu believed that Sima Qian‟s private resentment contaminated his

work to some extent, making it defamatory against the Han. Those who would defend

Sima Qian argued along two possible lines: first, they took up Yang Xiong‟s laconic

pronouncement that the Shiji was a “true record,” taking that to mean that any criticism of

the Han the Shiji contained was fully justified;1 second, if less frequently, they argued,

with Wei Hong, that such criticism preceded Sima Qian‟s castration and perhaps

contributed to it.

During the Song dynasty (960-1279), a shift occurred in how many readers saw

the relationship between Sima Qian‟s biography and the Shiji. The initial stages of this

shift seem to have been brought about by the circle of scholars and literati surrounding Su

Shi 蘇軾 (1037-1101).2 It seems likely that these figures developed new ways of reading

the Shiji as part of their overall response to the specific political circumstances in which

they found themselves during the Northern Song— in particular, the extreme factionalism

that arose around Wang Anshi‟s New Policies. Su Shi‟s students and friends saw their

hero persecuted by his enemies at court during the Crow Terrace Poetry Trial (Wutai

shian 烏臺詩案) of 1079 because he had written and spoken unwisely. Suddenly it

became not only easier to sympathize with Sima Qian but even to defend Sima Qian‟s

1
See discussion in chapter 6.
2
Though not, as I will show, by Su Shi himself.

255
“mistake” as an act of courage, and to reconceptualize the traces his experience might

have left on the Shiji as passionate outpourings of justified resentment.

The sensitivity to context I argued for in Part I is important to maintain if we are

to make sense of an entirely contradictory trend also found in Northern (as well as

Southern) Song writings on the Shiji, namely a tendency to bitterly condemn Sima Qian

for exactly the same expressions of resentment which other writers—or even the same

writers—had found sympathetic. This seeming contradiction can be explained by the

difference between private and public venues of expression. In a personal context it had

perhaps become easy to sympathize with Sima Qian, but in a political context, and

especially in the context of official historiography, Sima Qian‟s actions were not so

defensible.

After the Song‟s defeat by the Jin 金 (Jurchens) and the Song court‟s southward

relocation in 1127, the intellectual milieu was heavily influenced by the increasing

tendency of emperors and their highest ministers to restrict dissent. James T.C. Liu has

described the Southern Song political conditions as leading to “an inward turn” in

Chinese intellectual style.3 While he gives few details on how this inward turn

manifested in the area of historical criticism,4 I would argue that one result of Southern

Song political conditions was a reinterpretation of the Li Ling affair: whether due to

conscious historical analogism or unconscious projection, writers began to exonerate or

partially exonerate Sima Qian and instead blamed the Han Emperor Wu to an

3
This hypothesis is supported and elaborated throughout Liu, China Turning Inward (1988).
4
Liu, China Turning Inward 36. Liu‟s remarks on the area of historical criticism are brief and dismissive,
saying that in Southern Song “analyses of historical cases…were usually colored by subjective evaluation”
and had a tendency to “put…moral concerns ahead of historical facts” (China Turning Inward 24). I would
not dispute either of these characterizations, but consider that the way Southern Song figures reinterpreted
events like the Li Ling affair is more interesting than Liu implies, and repays closer study.

256
unprecedented extent. Scholars employed increasingly specific analyses of the Shiji in

service of their various agendas, now pointing to particular words and phrases as the

source of their judgements about the text. This raised the value of their comments in the

eyes of later readers. Consequently, their evaluations—though historically contingent

and circumstantial—acquired the coloring of profound truths about the Shiji.

NORTHERN SONG

In lieu of a strictly chronological treatment, I will begin with two poetic

evaluations of Sima Qian, one by Wang Anshi 王安石 (1021-1086) and one by Qin Guan

秦觀 (1049-1100). I will argue that the contrast between these two poems highlights the

transformation in how readers reacted to Sima Qian‟s story.

Two Poems on Sima Qian

Wang Anshi‟s poem on Sima Qian is an excellent example of readerly

ambivalence toward Sima Qian and his tragedy:

“Sima Qian” by Wang Anshi 王安石《司馬遷》

5

The peacock and phoenix wear patterned plumes; 孔鸞負文章，

they cannot bear to nest in a thorny hedge. 不忍留枳棘。
But alas he who underwent the saw and blade 嗟子刀鋸間，
stopped at leisure there for his sustenance. 悠然止而食。
He created a book for generations to come 成書與後世，
and his passionate frustration relied on its self-explanation. 憤悱聊自釋。
His appreciation was not limited to a single school; 領略非一家，
his lofty phrases might almost have been inspired by Heaven. 髙辭殆天得。
Though he was not as wise as Zhong Shan Fu, 雖微樊父明，
he did not fail in straightness of the eunuch Meng. 不失孟子直。
5
Wang Jing Gong shi zhu 王荆公詩注 [Commentary on the poems of Wang Jing Gong], SKQS.

257
 They who debase themselves from selfish greed— 彼欺以自私，
was [Qian] not worth ten, a hundred, of such people? 豈啻相十百。

The poem divides naturally into three quatrains. The first develops the metaphor of the

thorny hedge. The peacock and phoenix of the first line are noble birds, representing the

worthy and virtuous man.6 Such a person could not bear to remain in a dishonorable

situation. And yet as the second couplet shows, Sima Qian did just that. Wang Anshi, in

using the phrase “saw and blade” refers to the line in the “Letter in Reply to Ren An,”

where Sima Qian uses a similar phrase to refer to himself.7 The fourth line of the poem,

though less obvious, seems to refer to the fact that, after suffering castration, Sima Qian

was still content to serve Emperor Wu as Prefect of Palace Writers (中書令), which Ban

Gu described as “a respected and favored official position” [尊寵任職].8 The point is,

then, that Sima Qian was not on the level with men of the highest principle, given that he

was willing to accept such humiliation.9

The second quatrain focuses on the Shiji. First, it was a book given to future

generations. The word Wang uses, houshi 後世, prompted Southern Song commentator

Li Bi 李壁 (1157-1222) to suggest an allusion to the Cai Yong story (discussed in

chapter 4 above), where Wang Yun argues that by not killing Sima Qian, Emperor Wu

allowed him to transmit his defamatory text to future generations (houshi 後世). It does

6
The allusion is to a line from the Hanshu, spoken by Wang Huan 王渙, then-prefect of Henei, in praise of
the “virtuous official” Qiu Lan 仇覽, whose career Wang Huan helped to begin. Wang Huan had asked
Qiu Lan if he had the ambition to become a powerful man, literally, an eagle or hawk (yingzhan 鷹鸇). Qiu
Lan replied that he would rather be a phoenix, a bird of rare virtue. Wang Huan replied, “A thorny hedge is
not the place for a phoenix to nest; and how could this minor village be proper road for a great worthy?”
[枳棘非鸞鳳所棲，百里豈大賢之路？] (HHS 76.2480). Wang then helped Qiu get a place at the
Imperial College (taixue 太學).
7
I.e., “a remnant of saw and blade” 刀鋸之餘 (HS 62.2727).
8
HS 62.2725.
9
This part of Wang Anshi‟s evaluation is comparable to that in the Gaoshi zhuan (as discussed in chapter
4), though there the implicit judgement was that Sima Qian ought never to have served at all.

258
not seem obvious that Wang Anshi meant to introduce this kind of disapproving note into

his evaluation of the Shiji, for the rest of the quatrain is quite positive. The fenfei 憤悱 of

the next line alludes to a passage from the Lunyu, in which the Master says, “I will not

enlighten a student who is not passionate (fen). I will not hold forth for a student who is

not frustrated (fei)” [不憤不啟， 不悱不發].10 By attributing to Sima Qian qualities

which the Master pronounced desirable, indeed necessary for moral progress, Wang

Anshi seems generally approving of Sima Qian‟s efforts on the Shiji. Indeed, the phrase

houshi 後世 is innocuous enough that it might well not have been intended as a reference

to the Cai Yong story at all, but rather as a reference to Ban Gu‟s description of Sima

Qian‟s effort to “glorify his name for the eyes of later generations” [揚名後世],11 or even

to Sima Qian‟s own description of his work, that it was “awaiting the sages and

gentlemen of future generations” [俟後世聖人君子].12 (That Li Bi did read it as an

allusion to the Cai Yong story, however, shows the degree to which Wang Yun‟s

accusation—and, implicitly, autobiographical considerations resulting from Sima Qian‟s

relationship to his emperor—remained entangled with judgments on the Shiji.) With this

third couplet, Wang Anshi makes the point, taken up and much extended by writers in Su

Shi‟s circle, that the Shiji was a work of emotional expression.13 Wang does not,

however, dwell heavily on this but goes on to the work‟s philosophical aspects,

10
Lunyu VII:7.
11
“Dianyin”, WX 48.2158.
12
SJ 130.3320.
13
Li Bi‟s commentary here quotes from the “Self-Narration” (Shiji ch.130), and further he gives his own
paraphrase of the passage, apparently more in order to give a sense of Sima Qian‟s situation than as an
explanation for Wang‟s poem. The Shiji‟s role in its author‟s self-expression was clearly what caught the
commentator’s interest and attention, though the poet gives it only a single line.

259
apparently praising it for its breadth, its ability to select from various schools of thought,

and for the quality of its prose.

Having surveyed Sima Qian‟s achievement, Wang Anshi then alludes to Ban Gu‟s

judgment by comparing Sima Qian to the same two Shijing figures that Ban Gu did,

namely, Zhong Shan Fu and the eunuch Meng.14 Unlike Sima Qian‟s various defenders,

in the Song and before,15 Wang Anshi accepts rather than challenges Ban Gu‟s

conclusion. The final couplet, however, ends with a reasonably positive verdict on Sima

Qian. The allusion is to a quotation from the Mengzi, where Mengzi refutes the Mohist

doctrine of Xuzi—which would make all things equal—by arguing:

That things are unequal is part of their nature. Some are worth thrice or five times,
ten or a hundred times, even a thousand and ten thousand times, more than others.
If you reduce them to the same level, it will only bring confusion to the Empire.
[夫物之不齊，物之情也。或相倍蓰，或相什百，或相千萬。子比而同之，
是亂天下也。]16

In other words, many people seem to have done just what Sima Qian did, that is

compromise themselves and accept shame to save their own lives. To judge Sima Qian

as being the same as these, however, would be incorrect despite some superficial

similarity. As the discussion in Mengzi goes on to say, “If a roughly finished shoe sells

at the same price as a finely finished one, what merchant would produce the latter?” [巨

屨小屨同，賈人豈為之哉].17

Wang Anshi‟s poetic evaluation of Sima Qian does not attempt to downplay the

historian‟s shame: as far as Wang was concerned, Sima Qian did nest in the thorny hedge

and tolerated a level of humiliation that a true worthy would not accept. Nor does Wang

14
C.f. HS 62.2738, and the discussion in chapter 4 above
15
See especially the discussion of Qin Guan‟s “Sima Qian lun” 司馬遷論 below.
16
Mengzi 3A.4; trans. Lau, Mencius, 104.
17
Ibid.

260
try to deny Ban Gu‟s verdict that Sima Qian was courageous in his directness but not

very wise. On the other hand, admiring what Sima Qian achieved in his creation of the

Shiji, Wang Anshi seems to affirm that Sima Qian‟s gamble paid off in the end: Qian

accepted humiliation in exchange for a chance at immortal fame, and Wang Anshi

allowed that the compromise was justified.

Qin Guan‟s poem is an occasional piece, and seems less highly crafted than Wang

Anshi‟s. In Qin Guan‟s literary collection, it is prefaced by the explanation: “We were

dividing up rhymes and I got the character he (ravine)” [分韻得壑字]. The poem is

worth noting, though, for the striking difference it reveals between Qin Guan‟s attitude

toward Sima Qian and Wang Anshi‟s.

“Sima Qian” by Qin Guan 宋秦《司馬遷》

18

Zichang when young was unrestrained, 子長少不覊，

removed the wheelblock and traveled over hill and ravine. 發軔遍丘壑。
In his later years he suffered the Li Ling affair, 晚遭李陵禍，
and in passionate frustration entrusted his meaning to distant things.憤悱思逺託。
From his lofty phrases shine forth a pure and hidden light, 髙辭振幽光，
and his truthful brush punishes concealed evil. 直筆誅隱惡。
He galloped over records of several thousand years, 馳騁數千載，
stringing together the works of the hundred schools. 貫穿百家作。
Even today on the green bamboo strips, 至今青簡上，
the brilliance of his writing shines in crimson and gold. 文彩炳金艧。
In such lofty talent one ignores the little flaws-- 髙才忽小疵，
It would be difficult to measure him by ordinary standards. 難用常情度。
Compare him to the great bird Peng who flew out over the sea: 譬彼海運鵬，
Why would he turn to notice nets or arrows? 豈復顧繒繳。
A petty man like Ban Shupi (=Ban Biao), 區區班叔皮，
cannot easily criticize him for carelessness. 未易議疏略。

18
Quan Song Shi 全宋詩 1053.12065.

261
 The initial couplet of the poem alludes to Sima Qian‟s youthful travels.

Romanticizing this period in Sima Qian‟s life was, during the Northern Song, gradually

becoming an important part of the literary associations surrounding Sima Qian.19

Probably Qin Guan made use of this trope for two reasons. First, it seems the best

solution to the problem of how to use the difficult rhyme word he drew. Second, it

introduces the theme of Sima Qian‟s lack of restraint, which by the end of the poem will

grow to quite glorious proportions. The theme of motion, and particularly travel to

distant lands, continues metaphorically throughout the poem, from the “galloping” over

millennia of history to the sea crossing of the bird Peng.20 Beginning the poem with

Sima Qian‟s actual journey, however, gives these metaphors a real-world correspondence.

The second couplet introduces the Li Ling affair. Unlike Wang Anshi, Qin Guan

made no judgement on Sima Qian for his conduct in that situation. Qin Guan was not

concerned with whether Sima Qian was right in defending Li Ling. Rather, he made this

couplet parallel to the first one, implying that Sima Qian‟s youthful quality of being

“unrestrained” emerged again at the time of the Li Ling catastrophe, causing Sima Qian

to act as he did. Then to parallel the reference to the wheelblock in the second line, the

fourth line uses the same phrase as Wang Anshi, fenfei 憤悱, which I have translated in

both cases as “passionate frustration.” (More specifically, it is an emotional state in

which strong feelings are unable easily to find expression, like being simultaneously

obstructed and impelled forward.) Having cast away the wheelblock, the youthful Sima

Qian of the first couplet travels to distant lands. The older Sima Qian of the second

19
See for example Su Zhe‟s 蘇轍 (1039-1112) “Shang shumi Han taiwei shu”上樞密韓太尉書[Letter
Presented to Grand Marshall Han Qi], Su Zhe ji 22.381, discussed in chapter 2 above.
20
Zhuangzi 1.1.

262
couplet instead entrusts his meaning to “distant” things, surely a reference to the subtle

criticism which readers had long imputed to Sima Qian‟s creation.21 It is important that

Qin Guan uses the metaphor of traveling, rather than of seclusion, to describe this process.

In travel, one‟s own consciousness is juxtaposed with a vast new landscape; in writing

the Shiji, Sima Qian‟s passionate frustration was juxtaposed with the vastness of

historical events.

The next three couplets are devoted to praising what Sima Qian has achieved in

his writings. Like Wang Anshi, Qin Guan used the expression “lofty phrases” 髙辭,

presumably a conventional term of praise for good writing, but pairs it with “pure and

hidden light,”22 another reference to the allegorical content of the Shiji, but again a

positive one. What is hidden is not resentment or hatred, but something righteous and

pure, a shining light of virtue, in obscurity but clearly visible for those who know how to

perceive it. Thus the nature of lodged (yu 寓) or entrusted (tuo 託) words: they are

neither wholly obvious nor wholly hidden. The next line in the poem takes an interesting

stance on the matter of Sima Qian‟s brush. Qin Guan describes the brush as “straight,”

recalling Liu Zhiji‟s distinction (discussed in chapter 6 below) between the courage of

writing with a “straight” brush and the cowardice of writing with a “crooked” (qu 曲) one

(i.e., flattering those in power in order to protect oneself). In Qin Guan‟s poem, Sima

Qian uses his “straight brush” to criticize “concealed dislike” (or “concealed evil”). This

21
See chapter 4 above.
22
You 幽 is a particularly multivalent word and inconvenient of translation. The Shuowen glosses it merely
as yin 隱 (hidden), as I have translated it here, but its connotations are various and Qin Guan probably had
several in mind. One was surely “imprisoned,” for which the most apropos source is Sima Qian himself,
who described himself as having been “imprisoned and [tied] in black ropes” 幽於縲紲 (SJ 130.3300).
Other connotations include both “elegant” and “profound,” both of which Qin Guan might apply to the
Shiji, and finally—superficially—the connotation of darkness, in order to provide a pleasingly paradoxical
pairing with guang 光 (light).

263
is an allusion to Ban Gu‟s evaluation of Sima Qian, where the Shiji is praised for

recording events “without empty beautification and without concealed dislike” [不虛美，

不隱惡].23 This is a phrase that would be associated with the Shiji in the mind of any

educated reader at the time. But concealed dislike is just what readers (including Ban Gu

himself) would also impute to Sima Qian, claiming that in his resentment he satirized and

ridiculed his own generation and ruling house, producing a “defamatory text” (bangshu

謗書). In this line, then, Qin Guan unambiguously argues that though Sima Qian does

“entrust” his meaning to the events he writes about, he does not do so out of concealed

dislike, but quite to the contrary cuts through such unworthy sentiments and writes with a

straight brush. Here it is important to notice that you 幽 and yin 隱 are parallel in

meaning but opposite in connotation. The hidden (you 幽) light in Sima Qian‟s phrases is

a good thing, while the concealed (yin 隱) dislike was something he not only was not

guilty of, but himself actively condemned.

The final six lines develop a new theme, that true greatness need not be bothered

with petty rules or standards. To say that in a lofty talent one overlooks small flaws is to

brush aside the usual complaints about Sima Qian—his supposed preference for Huang

Lao or lack of consistency with the Classics, etc. Qin Guan refers to these as “little flaws”

(小疵). This is not to deny that they are flaws, but only to question their relevance to

judgements about the Shiji and its author. To strengthen his point, Qin Guan uses the

Zhuangzi‟s image of the great bird Peng. Recall that Wang Anshi had compared Sima

Qian with the phoenix and peacock and found his fastidiousness lacking—for Sima Qian

was willing to “feed in the thorny hedge.” Qin Guan‟s picture is very different. Not only

23
HS 62.2738.

264
does Sima Qian not compromise himself, but he is too great even to notice the trifling

rules (the nets and arrows) which lesser beings try to impose on him. Someone like Ban

Biao, Qin Guan argues, is too unimportant to have the right to criticize a writer as great

as Sima Qian. Qin Guan‟s choice of Ban Biao (as opposed to Ban Gu) was surely not

accidental. First, it correctly points out that Ban Biao was actually the originator of much

of the Hanshu evaluation on Sima Qian, and Ban Biao, though quite illustrious in his day,

did not leave behind the same kind of literary legacy as his son did.

Wang Anshi‟s portrait of Sima Qian was of a person who had been unwise and

then compromised his ideals to stay alive. Qin Guan‟s portrait was of a Sima Qian whose

actions and writings were all of a piece: his fundamental character was magnificent and

unrestrained, which explained both his actions and his writing. What we see in Qin

Guan‟s “Sima Qian” poem is a romanticization of Sima Qian‟s biography brought about

by an admiration for his writing. It is such a reading that would become extremely

common in later times, but Qin Guan‟s poem is a relatively early and fresh example.

Zhang Lei

Qin Guan‟s poem on Sima Qian may be considered representative of the new

reading of Sima Qian popularized by Su Shi‟s circle.24 In further considering this view of

the Shiji, we must face the awkward circumstance that Su Shi himself had no particular

fondness for the text.25 As Qin Guan‟s poem shows, however, Su Shi‟s students and

24
For my purposes, this term refers primarily to Su Shi‟s brother Su Zhe, his „six gentlemen‟ (liu junzi 六
君子), i.e., Huang Tingjian 黃庭堅 (1045-1105), Qin Guan 秦觀 (1049-1100), Chao Buzhi 晁補之 (1053-
1110), Zhang Lei 張耒 (1054-1114), Chen Shidao 陳師道 (1053-1101), and Li Zhi 李廌 (1059-1109).
25
The major piece of evidence for this contention is the statement, found in Chen Shidao‟s 陳師道 (1053-
1101) Houshan shihua 後山詩話 [Back Mountain poetry talk]: “Ouyang Yongshu [=Ouyang Xiu] is not
fond of Du Fu‟s poetry. Su Zizhan [=Su Shi] is not fond of Sima [Qian‟s] Shiji. Huang Luzhi [=Huang

265
friends had a lively interest in Sima Qian and produced a number of pieces relating his

work to his life. Within this circle, Zhang Lei 張耒 (1154-1114) was probably the one

who wrote about the Shiji most frequently.

In his youth Zhang Lei wrote in a letter to Zeng Gong 曾鞏 (1019-1083): “After

the Three Dynasties, I most enjoy reading the works of Honorable Senior Archivist and

of Han Tuizhi [=Han Yu]” [自三代以來，最喜讀太史公、韓退之之文].26 That same

letter elaborates Zhang Lei‟s theory of literary creation—that a person‟s moral qualities

are inevitably reflected in his writing. Thus Zhang describes the relationship between

Sima Qian‟s life and his writing:

Sima Qian has such heroism in the face of adversity as is rarely encountered.
From the time he was young, he traveled everywhere, meeting those who were
heroic and pure. His learning is best at discussing and analyzing the traces of
former ages, but due to his righteous air, he dared to speak and leaped to his own
disaster. This is why his writing is smooth and clear, simple but races forth with
great momentum. It is only that there was something suppressed in the
aspirations he pursued throughout his life. Thus even in the most insignificant [of
his writings], there is at times a sense of forceful emotion that can have no outlet.
[司馬遷竒邁慷慨，自其少時，周遊天下，交結豪傑。其學長於討論尋繹前
世之迹，負氣敢言以蹈於禍。故其文章踈蕩明白，簡樸而馳騁。惟其平生之
志有所鬱於中，故其餘章末句，時有感激而不洩者。]27

There is no question that in 1079 (when this letter was written) Zhang Lei had a positive

view of Sima Qian and even of the effect that Sima Qian‟s life had had on his writing.

Tingjian] and I always wonder and sigh, considering it is a very strange thing” [歐陽永叔不好杜詩， 蘇
子瞻不好司馬史記。 余每與黃魯直怪嘆， 以為異事] (in Biji xiaoshuo daguan 6.3660). Yu Zhanghua
俞樟華, in his article “Tang Song Ba Da Jia yu Shiji” 唐宋八大家與史記 [The Eight Prose Masters of the
Tang and Song, and the Shiji] argues against this assertion, seeming to consider it a slander against both the
Song poet and the Han historian. Listing numerous allusions and anecdotes, Yu proves that Su Shi was
deeply familiar with the Shiji and not averse to employing its style and phrases in his own writing. Perhaps
this kind of „defense‟ misses the point, however. The tone of Chen Shidao‟s anecdote suggests that a
personal and idiosyncratic preference was involved in both cases. Su Shi no doubt did know the Shiji well
(his father Su Xun, himself an important figure in Shiji studies, would have ensured that), but apparently it
simply did not appeal to him—its cast of thought did not resonate with his own. Part of his reason certainly
also involved some very specific political positions, as discussed below.
26
Zhang Lei ji 56.844.
27
Zhang Lei ji 56.844-845.

266
Like Qin Guan, Zhang Lei assigned an important role to Sima Qian‟s early travels, and

even in his mention of the Li Ling tragedy, he makes an implicit connection between the

fact that the historian “dared to speak and leaped to his own disaster” and the qualities of

“momentum” [馳騁] and “forceful emotion” [感激] found in his prose—qualities which

Zhang clearly considered admirable. As in Qin Guan‟s poem, we find here a

romanticization of Sima Qian‟s life, which is then connected with stylistic qualities as

well.28

Others of Zhang Lei‟s Shiji-related writings are not quite so easily interpreted. In

his “Discussion of Wei Bao and Peng Yue” [魏豹彭越論], Zhang Lei wrote:

I feel such sympathy when Sima Qian writes about Wei Bao and Peng Yue not
being ashamed of imprisonment and going to their punishment. He says, “They
had no ulterior motives, and their wisdom surpassed that of other people, and they
worried only that they would lose their lives. If they could but get the smallest
leverage, clouds could mass and dragons transform [i.e., they might well have met
a time of great success]. They longed for that which would let them achieve the
full scope of their ambitions, and that was why they allowed themselves to be
hidden away, imprisoned, and did not take their leave [i.e., commit suicide].”29
Alas! How his discussion gets to the heart of it! [予愛司馬遷論魏豹、彭越之不
耻囚虜以至刑戮也，曰：“彼無異故，智略絶人，獨患無身耳。得攝尺寸之
柄，其雲蒸龍變，欲有所會其度，以故幽囚而不辭。” 嗚呼！何其論之之
至也。]30

Later scholars would explicitly draw a connection between this Shiji passage and Sima

Qian‟s own situation,31 a connection that Zhang Lei may well also have perceived. Here,

however, Zhang Lei does not mention Sima Qian‟s life story. Perhaps Zhang Lei‟s

thoughts about the Shiji in this case were more personal than those of most readers

28
It is worth noting that the expressive theory underlying Zhang Lei‟s letter owes much to Han Yu‟s “Song
Meng Dongye xu” 送孟東野序, discussed in chapter 2 above.
29
A direct quotation from SJ 90.2595.
30
Zhang Lei ji, 40.653-655.
31
E.g., Dong Fen 董份 (dates unknown) in the late Ming (see Shiji pinglin 5.728).

267
discussed heretofore: when he read Shiji, Zhang Lei may not have been thinking of Sima

Qian‟s imprisonment, but of Su Shi‟s.

In summer of 1079, Zhang Lei‟s teacher and close friend, the famous poet Su Shi,

underwent the ordeal of the Crow Terrace Poetry Trial, and spent four months in prison.32

Su Shi was deeply shamed and felt that he had no hope at all of being acquitted, writing

of how during the time of his arrest and imprisonment he several times contemplated

suicide. Yet in the end Empress Dowager Gao herself helped plead his case to Emperor

Shenzong, and Su Shi was released.

It is not entirely clear when Zhang Lei‟s “Discussion of Wei Bao and Peng Yue”

was written, but it seems likely that the first section reflects Zhang Lei‟s attitude toward

the Crow Terrace Poetry Trial. The fact that Su Shi had suffered the shame of trial and

imprisonment did not change Zhang Lei‟s respect for him, and Zhang “worried only that

[Su] would lose his life.” Of course we cannot be certain that in writing this essay Zhang

Lei was thinking of Su Shi and his misfortune. But as long as the “Discussion” was

written after 1079, we can at least say that when Zhang wrote “How his discussion gets to

the heart of it!”, he had himself seen a person of extraordinary talent imprisoned and in

danger of death. Much more could be said about how Zhang Lei‟s reading of the Wei

Bao and Peng Yue chapter relates to the circumstances of his own time, but it is sufficient

for now to mention that Zhang Lei (like many other Song figures) often took an

32
See Ronald Egan‟s fascinating description in Word, Image, and Deed (esp. 48-53), upon which the short
summary below is largely based; Hartman, “Poetry and Politics” also gives an interesting analysis and
description together with many translations of primary sources from that incident.

268
instrumental approach in their readings of historical figures, Sima Qian included. They

adapted their judgements to suit an underlying purpose connected with their own time.33

From these and other pieces Zhang Lei wrote pertaining to the Shiji, we can see

that he respected Sima Qian and enjoyed reading his work. It is especially surprising

then to find Zhang Lei, in his “Discussion of Sima Qian,” specifically analyzing Sima

Qian‟s autobiographically-motivated preoccupations, and then attacking the historian for

them:

When Sima Qian made the “Traditions of Bo Yi,” he wrote: “If it was not
a just cause he did not speak out, yet he met with calamity and disaster.” 34 This
is only Qian speaking of himself. It is simply that he made an argument on Li
Ling‟s behalf, and Emperor Wu punished him. When he wrote of the affairs of
Guan [Zhong] and Yan [Ying],35 of Yanzi alone did he write, “If only Yanzi were
alive today, though I might only hold his whip for him [i.e., act as his driver]36, I
would do so with joy and admiration.” That [Si]ma Qian wrote these words is
probably because Yanzi extricated Yue Shifu from his bonds, and at the time that
[Sima] Qian was punished, the lords and ministers of Han did not speak up for
[Sima] Qian.37 Thus [Sima Qian] places special emphasis on Yanzi at that point.
[司馬遷作《伯夷傳》言“非公正不發憤而遇禍災”，此特遷自言為李陵辯而
武帝刑之耳。論管、晏之事，則于晏子獨曰：“使晏子而在，雖執鞭所忻慕
焉。”遷之為是言者，蓋晏子出越石父于縲紲而方遷被刑，漢之公卿無為遷
言，故于晏子致意焉。]
At the time of Li Ling‟s surrender, it was not known whether or not Ling
was still for the Han, but Qian alone was passionate and reckless, arguing for [Li
Ling] with all his might. To be like this comes near to folly, does it not? It is
different indeed from waiting until the [proper] time and only then speaking out,
settling the case with just a few words! How is it that [Qian] did not realize his
mistake and became confused about the rights and wrongs of the Way!? Going to
such extremes of resentment that people at the time did not rescue him from
catastrophe, and thinking obsessively about Yanzi—Qian is shallow indeed! Qian
is shallow indeed! [且方李陵之降，其為漢與否未可知，而遷獨激昻不顧出力
辯之如此，幾于愚乎！與夫時然後言，片言解紛者異矣。不知其失，而惑夫

33
See the discussion of historical didacticism and historical analogism in Ng and Wang, Mirroring the Past
151-155.
34
SJ 61.2125.
35
Shiji ch.62.
36
This expression, found in SJ 62.2137, is also used in the Lunyu to refer to potentially accepting a humble
position in the effort to do what is right (see Lunyu VII.12).
37
See the “Letter in Reply to Ren An” (HS 62.2730)

269
 道之是非，何哉？至怨時人之不援巳于禍，而拳拳于晏子，遷亦淺矣！遷亦
淺矣！]38

It seems that Zhang Lei was criticizing Sima Qian for allowing his own personal

circumstances to affect the way he portrayed history. Given the direct contradiction

between this passage and the one quoted above, the reader almost wonders if this might

not be written sarcastically—in praise of folly, as it were. But it is also possible that

Zhang Lei wrote this essay rather later, at a time in his life when his own circumstances

had changed considerably.39 In 1079, he looked upon the literary men of former times—

Qu Yuan, Sima Qian, Han Yu—as a sort of ideal, men whose lives and works were of a

piece, both being fine and noble. At that time, Zhang Lei, though a jinshi 進士

[presented scholar], was still very much an outsider. Because of his membership in Su

Shi‟s out-of-favor faction, he held offices far removed from the capital and was largely

excluded from the ranks of the powerful. However in the eighth year of Yuanfeng (1085),

Emperor Shenzong passed away. Emperor Zhezong was still very young, and the new

emperor‟s grandmother Empress Dowager Gao had considerable sway over what went on

at court. She was not a supporter of Wang Anshi‟s New Policies, and instead brought

Wang‟s rival Sima Guang into the limelight. Because of this, Su Shi and his disciples

finally had the opportunity to return to Kaifeng. The offices Zhang Lei eventually

obtained were involved with the compilation of the official history for the previous reign.

Like Han Yu (discussed in chapter 6 below), Zhang Lei found that this

involvement in official historiography brought him nothing but trouble. His position was

far from stable even during the generally favorable Yuanyou reign. Then, when the

38
Zhang Lei ji 41.664-665.
39
See Peter Bol, “Culture and the Way,” 344-419, which emphasizes the changes in Zhang Lei‟s thought
over the course of his life.

270
Empress Dowager passed away in 1093, Emperor Zhezong took the reigns of government

into his own hands. Displeased with the Yuanyou policies, the emperor tolerated or

perhaps even encouraged attacks on Su Shi and his friends. The following year, Deng

Runfu 鄧潤甫 (1027-1094) and Cai Bian 蔡卞 (1048-1117) submitted a memorial

referring to the “True Records of Shenzong”—which had been compiled by the Su Shi‟s

„Four Gentleman Scholars,‟ Zhang Lei among them—as a “defamatory history” (bangshi

謗史) which ought, they argued, to be completely rewritten.40

It seems possible that it was these or similar attacks against Zhang Lei and his

friends that led Zhang to reverse the verdict on Sima Qian, finding fault with what he had

formerly approved. Being compared to Sima Qian was certainly a term of praise by

Zhang Lei‟s time, at least in the realm of prose style. But, as we saw in the previous

chapter, there was a separate tradition of the Shiji being attacked as a “defamatory text.”

It was probably a similar charge of defamation that Zhang Lei was attempting to defend

against in his “Discussion of Sima Qian.” He could of course have chosen to defend

Sima Qian, as Pei Songzhi and others had done, by arguing that the historian had only

been writing the truth. But to do so, according to the personalizing logic of the “character

discussion” (renwulun 人物論) genre, would be to admit that his faction‟s own “True

Records” were, as his critics may well have claimed, highly critical of Emperor Shenzong

(just as Sima Qian had been toward Emperor Wu). Instead, Zhang Lei attacked Sima

Qian in such a way as to disclaim all similarity between Qian and himself. Zhang Lei

takes the moral high ground, inveighing against the Han historian first for acting rashly

40
See discussion in Shen Fen 湛芬, Zhang Lei, 444.

271
and then for allowing his selfish resentment to distort his work on the great undertaking

of writing a dynastic history.

There may be yet another turn of the screw. To accuse Sima Qian of folly and

shallowness might not have sounded strange in the Eastern Han (recall that Ban Gu had

called him “a gentleman without a sense of duty”). But by the Song, Sima Qian had been

much elevated. And what if Deng Runfu and Cai Bian were correct, and Zhang Lei and

his friends did embed subtly negative judgements in their version of Shenzong‟s “True

Records”? In choosing to (disingenuously?) attack Sima Qian for the same fault, Zhang

Lei would have been putting himself and his circle in very good company.

At this point let us lay aside the question of Zhang Lei‟s attitude toward Sima

Qian and consider instead the content of his autobiographical reading of the Shiji. Zhang

mentioned two chapters, the “Traditions of Bo Yi” (Shiji ch.61) and the “Traditions of

Guan and Yan” (Shiji ch.62), and even matched specific quotations from the Shiji with a

reading of Sima Qian‟s autobiography.

The first quotation, from the “Traditions of Bo Yi,” points to a passage in which

Sima Qian does not discuss Bo Yi and Shu Qi specifically, but the contrast between (on

the one hand) people “whose conduct does not follow what is right, and whose every act

violates taboos and prohibitions” [操行不軌，專犯忌諱] but yet “live all their lives in

leisure and happiness” [終身逸樂] and (on the other hand) people who “first carefully

choose the ground and only then step on it, who „speak only when it is time to speak,‟41

41
Lunyu XIV:13.

272
who „take no short-cuts‟42 and pour forth no passion for what is not upright and just” [擇

地而蹈之，時然後出言，行不由徑，非公正不發憤]. These latter, to the great

confusion of the Honorable Senior Archivist, “encounter misfortune and disaster in

numbers too great to be reckoned” [遇禍災者，不可勝數也].43

This passage is a sure invitation to the reader to think of examples. Sima Zhen‟s

Suoyin commentary remarks at this point, “This is talking about people like Long Feng,

Bi Gan, Qu Yuan, and Wu Xu” [謂龍逢﹑比干﹑屈平﹑伍胥之屬是也].44 Zhang Lei,

however, goes a step farther and adds Sima Qian to the list: “This is only Qian speaking

quite on purpose about himself: he made an argument on Li Ling‟s behalf, and Emperor

Wu punished him” [此特遷自言，為李陵辯而武帝刑之耳]. For Sima Qian to

complain against heaven on behalf of great figures of the past is one matter, but to do so

on his own behalf is quite another. True, Qu Yuan (according to the Shiji chapter that

came to be accepted as the story of his life) did this successfully in the Li Sao 離騷

[Encountering sorrow]. But Sima Qian did not have quite the same record of valiant

service to his country. What had he done, but attempt to defend the reputation of a

general who did, in the end, turn traitor? Thus Zhang Lei follows Ban Gu‟s

42
Lunyu VI:14.
43
This and the other lines quoted above are from SJ 61.2125.
44
SJ 61.2126. Longfeng (or Guan Longfeng, as he is called) was supposedly murdered by the cruel last
ruler of the Xia, and is frequently placed in parallel construction with Bi Gan, who had his heart torn out by
the cruel last ruler of the Shang. It may be that in writing this comment, Sima Zhen had in mind a passage
from the Xunzi “Yu zuo” 宥坐 chapter in which Zilu asks questions very like the ones Sima Qian asks in
the “Bo Yi” chapter, namely, “If Heaven rewards the good with good fortune, and requites those who are
not good with disaster, why is it that you, Master, who accumulates virtue and piles up rightness and holds
fast to all that is good, have practiced this way for so very long and yet always remain in obscurity?” [為善
者天報之以福，為不善者天報之以禍，今夫子累德積義懷美，行之日久矣，奚居之隱也] (Xunzi
28.647-648). Confucius answers that Bi Gan, Long Feng, and Wu Zixu are all examples of virtue not only
going unrewarded but actually meeting with catastrophe.

273
unsympathetic reading, pointing out that Sima Qian had acted rashly, had in fact poured

out his passion for a cause that was not necessarily upright or just.

The second passage for which Zhang Lei produces an autobiographical reading is

from the evaluation at the end of the “Traditions of Guanzi and Yanzi” [管晏列傳], Shiji

ch.62. That part of the evaluation in the Shiji reads,

Yanzi prostrated himself by the corpse of Duke Zhuang and wept for him. He
completed the rite and only then departed. How could anyone say of him that “he
saw what was right but did nothing, lacking courage”45? When it comes to
remonstrance and persuasion, to go against a ruler to his face [as Yanzi did] can
surely be described as, “In coming forth, completely loyal; in withdrawing,
repairing his mistakes”!46 If only Yanzi were alive today, though I might only
hold his whip for him [i.e., act as his driver]47, I would do so with joy and
admiration. [方晏子伏莊公尸哭之，成禮然後去，豈所謂「見義不為無勇」
者邪？至其諫說，犯君之顏，此所謂「進思盡忠，退思補過」者哉！假令晏
子而在，余雖為之執鞭，所忻慕焉。]48

There is no doubt that this evaluation expresses some kind of personal feeling. Sima

Zhen glossed the passage and concluded with the evaluation, “This is the extent to which

he was fond of the worthy and took pleasure in the good. How worthy he was, this fine

historian. He could be held up as a warning to all ministers of state” [其好賢樂善如此。

賢哉良史，可以示人臣之炯戒也].49 But both the original Shiji passage and Sima

Zhen‟s comment leave the specifics of Sima Qian‟s admiration open to question. The

ordinary reader of the “Guan Yan liezhuan” (SJ ch.62) evaluation would probably agree

with Sima Zhen that “what the Honorable Senior Archivist admires and longs for is the

45
A quotation from the Lunyu, II.24. Cui Zhu 崔杼 had assassinated Duke Zhuang and was planning to set
up his younger brother as ruler instead. In displaying ritual grief for Duke Zhuang, Yanzi was taking a
tremendous risk, and indeed, according to the Zuozhuan account, one of Cui Zhu‟s followers urged that
Yanzi be put to death, but Cui Zhu decided to spare him, see Zuo “Xiang” 25.2, CQZZ zhu 3.1097-1099.
46
Found in Zuozhuan Xuan 12, but also appears in the Xiaojing “Shi jun” 事君 chapter, 17.52.
47
This expression is used Lunyu to mean accepting a humble position in the effort to do what is right (c.f.,
Lunyu VII.12).
48
SJ 62.2136-37.
49
SJ 62.2137.

274
conduct of Pingzhong (=Yanzi)” [羨慕仰企平仲之行] (SJ 62.2137)—his mourning for

his ruler and his courageous remonstrance, and so on.

The reading Zhang Lei hinted at, however, is particularly sensitive to the potential

autobiographical subtext of the Yanzi chapter.50 Zhang singled out the story of Yue

Shifu as being the real explanation for the depth of emotion in Sima Qian‟s evaluation.

In order to fully understand this interpretation, we must first consider the Yue Shifu

narrative in the Yanzi chapter:

Yue Shifu was worthy, but was bound with black ropes [as a prisoner]. Yanzi
went out, and happened to encounter him on the road. He unharnessed his left
outside horse in order to ransom him and took him home. There, he went into his
own chamber without taking leave of the man. After some time, Yue Shifu
requested to break off relations. Yanzi was most surprised, and, straightening his
robes and cap, went to greet him, saying, “Though I, Ying, am not benevolent, I
did help you avoid trouble. How is it that you are so quick in seeking to break off
relations?” Shifu said, “It is not so. I have heard that a gentleman may suffer the
injustice of one who does not recognize his worth, but should be able to trust
those who do recognize his worth [to treat him properly]. When I was bound in
the black ropes, you did not know me at all. And yet, master, you already felt
some instinct about me and had me ransomed. This makes you one who
recognizes my worth. Yet if you recognize me and treat me without respect, it
were certainly better had I remained in the black ropes.” Yanzi thereupon invited
him to enter as an honored guest. [越石父賢，在縲紲中。晏子出，遭之塗，
解左驂贖之，載歸．弗謝，入閨。久之，越石父請絕。晏子戄然，攝衣冠謝
曰：「嬰雖不仁，免子於厄，何子求絕之速也？」石父曰：「不然。吾聞君
子詘於不知己而信於知己者。方吾在縲紲中，彼不知我也。夫子既已感寤而
贖我，是知己；知己而無禮，固不如在縲紲之中。」晏子於是延入為上
客。]51

This story exists in several other versions, including one in the Lüshi chunqiu 呂氏春秋

[Spring and Autumn of Master Lü], and another in the Yanzi chunqiu 晏子春秋 [Spring

50
It would surely be incorrect to attribute this way of reading the Shiji to Zhang Lei as his own personal
innovation. Much more likely, it was a strategy for reading that was shared in varying degrees by members
of his circle.
51
SJ 62.2135. It is worth noting that the Shiji “Traditions of Yanzi” is an extremely short chapter,
containing only a few anecdotes. The fact that Sima Qian selected this one for inclusion therefore becomes
all the more significant.

275
and Autumn of Yanzi]. Both versions are substantially similar to this one, except in the

opening lines. The Lüshi chunqiu first describes Yue Shifu as “a man dressed in animal

skin and carrying fodder on his back taking a rest by the side of the road” [反裘負芻息於

塗者].52 Later Yue Shifu himself states that “a man of Qi has made this person do hard

labor” [齊人累之]. The Yanzi chunqiu story is much the same, except it adds the detail

that the Yue Shifu wore a “worn-out cap” [弊冠] and has him say that he became a

servant because “I was not able to protect myself from cold and starvation” [不免凍餓之

切吾身].

It is quite likely that Sima Qian deliberately altered the opening lines of the

passage, and in a striking way: he has Yue Shifu “bound with black ropes” [在縲紲中],

with the clear implication that the man is not so much a slave as a prisoner accused of a

crime.53 As discussed above, Sima Qian also described himself as being “hidden away,

bound in the black ropes” [幽於縲紲] (SJ 130.3300) as a consequence of the Li Ling

affair. Zhang Lei pointed out that when Sima Qian was imprisoned, “the lords and

ministers of Han did nothing” to help him, or in Sima Qian‟s own words, “not one of my

friends would save me” [交遊莫救].54 Yet in this story, Yanzi without hesitation used

one of his own horses to ransom a man he did not even know. If Sima Qian had had a

Yanzi, suggested Zhang Lei, he would have been saved: that is why the historian betrays

such emotion in his evaluation at the end of the chapter.

52
LSCQ “Guan shi” 觀世 16.958, Knoblock/Riegel 2000, 279.
53
The word leixie 縲紲 appears in the Lunyu V:1 as a near variant 縲絏, where the Master says of Gong Yi
Chang that “though he is bound in black ropes, he is blameless” [雖在縲絏之中，非其罪也] (Lunyu V:1).
54
HS 62.2730.

276
Qin Guan

Zhang Lei was not the only one in his circle to produce this kind of specific

autobiographical interpretation of Shiji narratives. Qin Guan‟s “Discussion of Sima Qian”

shows many of the same characteristics, while (as we might expect from his poem,

discussed above) arguing for a much more sympathetic reading. Qin Guan begins his

essay by quoting the faults which Ban Gu imputed to Sima Qian. He then refutes Ban Gu

on each point. Leaving aside for now Qin Guan‟s interesting defense of Sima Qian‟s

promotion of Huang-Lao over the Six Classics, let us consider his comments on the

matter of Sima Qian‟s “disparaging gentlemen scholars who live in retirement and

speaking in favor of heroic scoundrels” and “honoring those who were skilled at making

a profit but heaping shame on those in poverty and low station.” In response to these

points, Qin Guan wrote:

Only an extremely low and stupid villager would do this! Who could say that a
person of Qian‟s great talent and breadth ever came to this? The way I see it, it is
not so. In truth, he put forth [his words] according to what he had experienced,
and spoke as he did just because he was provoked to do so. [非閭里至愚極陋之
人，不至是也，孰謂遷之高才博洽而至於是乎？以臣觀之不然，彼實有見而
發、有激而云耳．]55

The thesis of this part of the “Discussion” is itself a kind of autobiographical reading:

what Sima Qian appears to be doing (flying in the face of morality) is not what he is

actually doing, and in order to understand what he is actually doing, it is necessary to try

to understand Sima Qian‟s own experience, the circumstances which provoked [ji 激]

him to write as he did.

Qin Guan then gives his view of the historical situation which formed the

background for Sima Qian‟s authorship of the Shiji:

55
Huaihai ji 20.700.

277
 At that time, Emperor Wu of the Han was applying the law with great severity,
and was most urgent [in demanding] achievements and profit. If a high minister
spoke one word that did not accord with [the emperor‟s policies], he was
summarily arrested and sent to be executed. Those who were guilty and facing
punishment, if they had the means, could ransom themselves. Thus there were
those who helped to ransom officials [when they were accused]. Because of this,
courtiers made it their business to covertly form associations for the purpose of
avoiding harm, and it became a custom throughout the realm to steal money and
goods. [方漢武用法刻深，急於功利，大臣一言不合，輒下吏就誅；有罪當
刑，得以貨自贖，因而補官者有焉．於是，朝廷皆以偷合苟免為事，而天下
皆以竊資貨殖為風．]56

Like many in his circle, Qin Guan was particularly concerned with the problem of over-

zealous application of harsh laws. His own historical background certainly included

Wang Anshi‟s New Policies; his circle, most prominently Su Shi, took an oppositional

stance toward these policies which blighted their promising careers for quite some time.

One of Su Shi‟s poems describes having to watch people being whipped,57 and Su Shi

also wrote an essay attacking Sima Qian for what he considered excessively positive

portrayals of Shang Yang and Sang Hongyang (both proponents of harsh reforms in the

manner of Wang Anshi).58 Like Zhang Lei, Qin Guan had seen his teacher and friend

suffer in the Crow Terrace Poetry Trial, and the perils of running afoul of harsh laws

must have made a deep impression on him. These were the political influences that

formed the background for Qin Guan‟s description of Emperor Wu‟s reign as quoted

above: a situation in which the ill-considered harshness of the central government has any

number of bad effects on officials and even the entire realm.

Having set the stage, Qin Guan went on to produce an extremely explicit

autobiographical reading of this aspect of the Shiji in terms of Sima Qian‟s situation.

56
Ibid.
57
See the description in Egan, Word Image Deed, 41-43.
58
Dongpo zhilin 5.107.

278
 When Qian encountered the Li Ling affair, “[his] family lacked the
resources for bribes to ransom [him]. None of [his] acquaintances rescued him.
Of [his] retainers and intimates, none said a single word.”59 Therefore he suffered
the “rotten punishment” [i.e. castration]. His passionate sorrow and agitated
energy had no other outlet, and so he lodged it all in his writing. Thus in his
preface to the “Roving Warriors” he writes that “In antiquity, Emperor Shun was
caught in a [burning] granary and trapped in a well; Yi Yin carried tripods and
sacrificial stands; Fu Yue was a convict laborer at the cliffs of Fu; Lü Shang
encountered trouble at the Ji Ford; Yi Wu (=Guan Zhong) wore shackles; Baili
tended cattle; and Zhongni encountered difficulty at Chen and Cai.”60 Probably
this is [a reference to] Qian‟s own circumstances. [遷之遭李陵之禍也，家貨無
財賄自贖，交遊莫救，左右親近不為一言，以陷腐刑．其憤懣不平之氣無所
發泄，乃一切寓之於書．故其序游俠也，稱昔虞舜於井廩，伊尹負於鼎俎，
傅說匿於傅險，呂尚困於棘津，夷吾桎梏，百里飯牛，仲尼阨於陳蔡．蓋遷
自況也．]
He also wrote, “When gentlemen find themselves in desperate straits they
are able to entrust their lives to [these men]. Is this not what people mean when
they talk about the „worthy‟ and „heroic‟? In fact, in terms of power and
forcefulness, there is no comparison between the effect on their own time of these
warriors of the hamlets and villages and that of men like Jici and Yuan Xian.”61
He is probably saying that those in his own time who were known for their self-
cultivation and moral virtues were in fact all cowardly and avoided [any hint of
trouble] in order to protect themselves. There were none willing to exert
themselves in the cause of other people‟s difficulties. Thus, they were not even
the equal of commoners in earlier times. [又曰：「士窮窘而得委命，此豈非人
之所謂賢豪閒者邪？誠使鄉曲之俠，予季次﹑原憲比權量力，效功於當世，
不同日而論矣．」蓋言當世號為修行仁義者，皆畏避自保，莫肯急於人之難，
曾匹夫之不若也．]62

The first passage argues that in writing the “Traditions of the Roving Warriors,” Sima

Qian was thinking of his own situation, himself having suffered without friend or

protector the harshness of legal punishment. The second passage then uses this

understanding to read an underlying meaning into Sima Qian‟s reference to Ji Ci and

Yuan Xian. This reference should not, Qin Guan argues, be understood as pointing to

59
A near-exact quote from the “Letter in Reply to Ren An” (HS 62.2730).
60
An exact quote from the “Traditions of the Roving Warriors” preface (SJ 124.3182), with the exception
of the last phrase. In Shiji 124, the last line reads, “Confucius was threatened at Kuang, and between Chen
and Cai his supplies were cut off” [仲尼畏匡，菜色陳﹑蔡]. Qin Guan‟s version of that line is instead a
closer parallel with the version in “The Honorable Senior Archivist‟s Self-Narration” (SJ 130.3300).
61
SJ 124.3183.
62
Huaihai ji 20.700-701.

279
those actual figures. Rather it is an allusion to people in Sima Qian‟s own time who self-

righteously assumed a post of great virtue and cultivation but did not dare take the actions

necessary to bring substance to their reputations.

Just as the “Traditions of the Roving Warriors” has been taken as the main

referent of Ban Gu‟s criticism about gentleman scholars and heroic scoundrels, so the

“Traditions of the Merchants” (and, alternatively, the “Treatise on the Balanced

Standard”) is generally taken as the target of Ban Gu‟s other accusation, that Sima Qian

was guilty of equating profit with honor and poverty with shame. Here too Qin Guan

uses an autobiographical reading to defend Sima Qian against Ban Gu‟s charge:

When Sima Qian compiled the “Merchants,” he wrote about how “Qin Shihuang
ordered that Wuzhi Luo be [treated as] comparable in rank to an enfeoffed lord,
and allowed him to come to seasonal audiences at court together with the various
ministers”, and also about how “[Qin Shihuang also] proclaimed Qing, the widow
of Ba and Shu, to be a virtuous woman and treated her as an honored guest,
building for her the Nühuaiqing Terrace”63—probably he was using these to
satirize Emperor Wu. [其述貨殖也，稱秦皇令烏氏倮比封君，與列臣朝請，
以巴蜀寡婦清為正婦而客之，為築女懐清臺：葢以譏孝武也。]64

According to the Shiji, Wuzhi Luo was a person who raised domestic animals, but

through shrewd speculation and clever diplomacy managed to increase his wealth

significantly through dealings with the Rong barbarians. In short, he sold his stock and

bought presents for the Rong king, who was so pleased and flattered that he repaid Wuzhi

Luo with gifts many times the value of the original investment. The nouveau riche

63
Both of these quotations are passages from the “Traditions of the Merchants,” SJ 129.3260. The
characters of Wuzhi Luo‟s name differ in the received text: Qin Guan has 烏氏倮, while the Zhonghua
shuju edition of the Shiji has 烏氏劳. Pei Yin‟s Shiji jijie commentary quotes Wei Shao as saying that
Wuzhi is the name of a county (xian). Sima Zhen‟s Shiji suoyin adds that the character 氏 here should be
pronounced like 支, and also that 劳 is the man‟s name and should be pronounced 踝. The name of the
terrace built for the widow of Ba and Shu, Nühuaiqing, is clearly meant to be a pun on her personal name,
qing 清, which means pure and uncorrupted. The name of the terrace might therefore be translated
“Woman Treasuring Purity Terrace.”
64
Huaihai ji 20.701.

280
rancher was thereby granted by the First Qin Emperor a status far above the class he was

born into.

The widow of Ba and Shu was similarly wealthy, though of slightly older money:

the Shiji writes that her ancestors had acquired their capital through ownership of

valuable cinnabar mines, which they had parlayed into “inestimable amounts of wealth.”

Sima Qian writes that although Qing was “only a widow, she was able to carry on

business and used her wealth to buy protection for herself so that others could not

mistreat or impose upon her” [清，寡婦也，能守其業，用財自衛，不見侵犯].65 She

too managed to earn a degree of imperial recognition that certainly had more to do with

her political contributions than with her womanly virtues, despite the terms in which that

recognition was couched.

Qin Guan wrote elliptically that both of these figures were intended as a satire

against Emperor Wu, but he did not explain why. The answer, however, probably lies in

the “Treatise on the Balanced Standard,” which repeatedly emphasizes that Emperor

Wu‟s standing policy was to give out positions and titles in exchange for contributions to

the government. In both Emperor Wu‟s case and the earlier description of the First Qin

Emperor‟s dealings with his wealthiest subjects, what is problematic is the potential

confusion of moral worth with financial worth, a devaluation of the currency of virtue.

Sima Qian‟s actual stance toward the “balanced standard” policies was potentially

a matter of debate in the Song. As discussed above, some Tang commentators had

considered Sima Qian‟s narrative in that chapter to be a noteworthy example of the

Shiji‟s anti-Han defamation. However, a court debate recorded as taking place between

65
SJ 129.3260.

281
Sima Guang and Wang Anshi in 1068 shows that this interpretation of the “Treatise on

the Balanced Standard” may have remained ambiguous or open to question.

Discussing economic policy, Wang Anshi argued that “insufficient funds are the

result of not having someone who is good at managing money” [所以不足者， 由未得

善理財之人耳]. Sima Guang retorted that “A person who is good at managing money is

only someone with a head for squeezing out all the people‟s wealth. When the people are

impoverished they become thieves, and that is not good fortune for the state!” [善理財之

人，不過頭會箕斂以盡民財。民窮為盜， 非國之福]. Wang Anshi disagreed, using

the quotation from the Shiji to support his answer: “It is not so. Those who are good at

managing money „do not increase taxes but yet the state has sufficient funds‟” [不然，善

理財者，不加賦而國用足]. Sima Guang rebutted this in a long speech (also quoted and

approved by Su Shi)66:

The resources and goods and hundred kinds of things produced by heaven and
earth are limited to a certain number. If they are not possessed by the people, then
they are possessed by the officials. It can be compared to rainfall: if in summer
there are floods, in autumn there will be drought. Not increasing taxes but the
emperor having sufficient funds at his disposal [depends on] setting up laws for
seizing the people‟s profit, which is more harmful than raising taxes. These [i.e.,
the Shiji passage quoted by Wang Anshi above] are merely the words that Sang
Hongyang used to deceive the Han Emperor Wu. Scribe Qian [=Sima Qian]
recorded them only in order to reveal how uninsightful they were!” [天地所生財
貨百物，止有此數，不在民，則在官。譬如雨澤夏澇，則秋旱。不加賦而國
用足不過設法以隂奪民利。其害甚於加賦。此乃桑弘羊欺漢武帝之言， 史
遷書之， 以見其不明耳]. 67

The emperor ruled in favor of Sima Guang in this case, but the fact that this line from the

Shiji could be understood in two such diametrically opposite ways shows that Sima

66
In “Sima Qian er da zui” 司馬遷二大罪 [Sima Qian's Two Great Faults], Dongpo zhilin 5.107.
67
Zizhi tongjian houbian ch.76. C.f. further parallel passages on Songshi 336.10763-10764 and Lü
Zuqian‟s Dashiji jieti 大事記解題 [Explanation of Topics in the Chronicle of Great Events], 12.109.

282
Qian‟s portrayal of Emperor Wu‟s economic policy was deeply ambiguous. The

complexity goes far beyond just the issue of autobiographical reading that is the focus of

discussion here. Still, it is worth noting that even in a primarily economic debate, literary

interpretation is a key issue: Wang Anshi is portrayed as reading the Shiji with great

literalness, whereas Sima Guang goes beyond the surface reading to discover a critical

subtext.

To return to Qin Guan‟s essay however—it continues in a simpler and more

explicit vein, almost as if Qin had regained firmer ground with his interpretation:

[Sima Qian] also said, “The proverb states that „the son of a thousand-gold
household does not die in the marketplace.‟”68 These were not empty words!
Probably Qian himself was feeling the pain of having honed his purity and
tempered his conduct [like a fine blade], and yet merely due to being poor, he was
unable to escape a cruel punishment. To say because of this that he was
“disparaging gentlemen scholars who live in retirement and speaking in favor of
heroic scoundrels, honoring those who were skilled at making a profit but heaping
shame on those in poverty and low station”69—well, how could Qian not have had
special impetus for his words? In that, Ban Gu did not understand his real
meaning. [又云：諺曰「千金之子，不死於市。」非空言也。葢遷自傷砥節
礪行，特以貧故不免於刑戮也。以此言退處士而進姦雄，崇勢利而羞貧賤，
豈非有激而云哉？彼班固不逹其意。]

Qin Guan finished this part of his essay by accusing Ban Gu of the same kind of

literalism that Sima Guang disparaged in Wang Anshi. By emphasizing Sima Qian‟s

virtues, Qin Guan revealed a sympathy—and a tendency toward sympathetic reading—

that excuses Sima Qian (as Su Shi and perhaps Zhang Lei as well did not) for the

confusion his hidden meaning might potentially cause.

68
SJ 129.3256.
69
HS 62.2738.

283
Li Zhi

In Qin Guan‟s “Essay” we find another case of Northern Song politics having a

clear influence on the views of Sima Qian‟s life and how it should affect Shiji

interpretation. The final passage I will discuss here in connection with the Su Shi circle

comes from Li Zhi‟s 李廌(1059-1109) Shiyou tanji 師友談記 [A record of conversation

with teachers and friends]. As the “General Catalogue of the Siku quanshu” describes,

this text “records the conversations of Su Shi and Fan Zuyu, as well as Huang Tingjian,

Qin Guan, Chao Shuozhi, and Zhang Lei” (記蘇軾、范祖禹及黃庭堅、秦觀、晁說之、

張耒所談).70 As such, it can probably be taken as a good indicator of the kind of ideas

that this circle had in common or discussed and debated among themselves.

Unfortunately, the passage in question has apparently dropped out of the received text of

the Shiyou tanji. It survives from a quotation in Ma Duanlin‟s 馬端臨 (1245-1322)

Wenxian tongkao 文獻通考 [Comprehensive Investigation of Documents], where it is

said to be taken from Li Zhi‟s Shiyou dushu ji 師友讀書記 [A record of reading with

teachers and friends]. Whether this is another title for the same work or a different work

that has been lost, it seems quite likely that it would have featured the same cast of

characters, assuming the attribution is reliable.71

The passage is interesting enough to be worth considering despite this textual

uncertainty. It reads as follows:

When Sima Qian made the Shiji, it was mainly to satirize the many shortcomings
of Emperor Wu of the Han. Therefore he employed a far-reaching intention. The
discussions of Yang Xiong and Ban Gu do not get to the truth of the matter. The

70
SKQS “Zongmu”, 1609.
71
Yang Yanqi, in his compilation of comments on the Shiji, Lidai mingjia ping Shiji, does take it as reliable,
working from Zhang Yuanyi‟s 張元顗 supplemented Rongyuan congshu 榕園叢書 edition (SJYJJC 6.625).

284
 “Basic Annals of Qin Shihuang” is entirely a satire of Emperor Wu. One can
ascertain in the Shiji that where [Sima Qian‟s] intention is profound and far-
reaching, then his words become increasingly drawn out. When the events are
more numerous and scattered, then his words become increasingly simple. This is
the principle of the Odes and of the Spring and Autumn. [司馬遷作《史記》，
大抵譏漢武帝所短為多，故其用意遠，楊雄、班固之論不得實。《秦始皇本
紀》皆譏武帝也，可以推求《史記》，其意深遠，則其言愈緩，其事繁碎，
則其言愈簡，此《詩》、《春秋》之義也。]72

This is not the first time we have encountered the claim that the Shiji was meant as a

satire of Emperor Wu, for such a judgement is part and parcel with the “defamatory text”

(bangshu 謗書) theme that has run throughout Part III. However, up until the Song, most

of the bangshu interpretations had been unsympathetic, and most of the sympathetic

interpretations had downplayed any aspect of subtle or indirect criticism, emphasizing

instead that the Shiji was a “true record” and as such gave a clear and honest account of

the Han, including Emperor Wu, unprejudiced by Sima Qian‟s tragedy. This passage

does not make it entirely explicit whether Sima Qian‟s satirical purpose and methods are

being celebrated in a positive light (one wonders if the participants in the discussion

might not have had differing opinions on the matter), but the phenomenon is at least

noticed with interest.

The aspects of Yang Xiong‟s and Ban Gu‟s evaluations with which Li Zhi (et al.)

are disagreeing probably involve Yang Xiong‟s complaint that Sima Qian disagreed with

the sage and Ban Gu‟s “three faults of Sima Qian” (introduced in chapter 4 above). Most

earlier defenders of Sima Qian had merely suggested that it was easier to criticize a

history than it was to write one, or that Ban Gu himself had failed to live up to his own

72
WXTK 191.1621-1.

285
standard.73 Qin Guan, as shown above, criticized Ban Gu by saying that he missed the

point of Sima Qian‟s work. The participants in Li Zhi‟s discussion seem to subscribe to a

similar notion.

The next interesting feature of this passage is the specific reference to the “Basic

Annals of Qin Shihuang.” One of the more striking chapters of the Shiji, the “Annals of

Qin Shihuang” was also singled out by Emperor Ming, as described in Ban Gu‟s

narration in the “Dian Yin” (see chapter 4 above). In the Ming dynasty, Ling Zhilong‟s

凌稚隆 (dates unknown) Shiji pinglin 史記評林 [A forest of comments on the Shiji]

would to some extent pursue the direction of Li Zhi‟s analysis,74 but it is interesting to

find, as early as the Northern Song, an unequivocal statement that this chapter should not

be read at face-value and instead should be taken as a reflection on Emperor Wu.

The final lines of the comment claim that the Shiji‟s narrative style acts as an

indicator of whether the passage in question should or should not be read in terms of

covert satire. The claim, that the narration slows down (huan 緩) when Sima Qian wants

us to read with extra significance, and becomes very simple (jian 簡) when the matter is

not very important, seems potentially controversial. But regardless of how accurate this

claim might be, it is interesting to find it here. Several Ming authors would make similar

claims, but perhaps they did so in part owing to the ideas of the influential Su group.

Furthermore, the passage contains an explicit comparison with the Odes and the

Spring and Autumn in terms of method, a nod to Sima Qian‟s own ambitions in that

regard. We should remember that Sima Qian himself mentioned the writers of the Odes

73
Fan Ye‟s evaluation of Ban Gu (HHS 40B.1386), discussed in chapter 1 above, is representative of this
tendency.
74
See Stephen Durrant, “Ssu-ma Ch‟ien‟s Portrayal,” 28-50.

286
in his “Self-Narration” as part of the “suffering author” list, and in the corresponding list

in the “Letter,” he singles out the Spring and Autumn (alone from amidst Confucius‟

allegedly copious textual production) as being comparable to his own work on the Shiji.

In this respect, at least, someone had finally begun reading Sima Qian as the “Self-

Narration” and “Letter” seemed to want to present him.

Chao Gongwu

A generation or so later, the Southern Song bibliographer Chao Gongwu 晁公武

(ca.1105-1180) would sum up the more sympathetic autobiographical readings of the

Shiji by paraphrasing Qin Guan‟s “Discussion of Sima Qian.” In the entry on the Shiji in

his Junzhai dushu zhi, he began by quoting Ban Gu‟s three criticisms, and wrote:

Those in later generations who were fond of Qian considered [Ban Gu‟s]
argument to be incorrect. They said that Qian felt deeply the failings of his own
generation, and was resentful about what had happened to him, and therefore he
lodged [these feelings] in his writings. It was just that he was stirred up, and he
made these words. It is not that in his own heart he would truly consider them to
be true. [後世愛遷者以此論為不然，謂遷特感當世之所失，憤其身之所遭，
寓之於書，有所激而為此言耳，非其心所謂誠然也。]75

Chao Gongwu‟s evaluation is often cited as representative of the Song approach to

autobiographical subtext in the Shiji. Aspects of the text that moralists might find

objectionable are reinterpreted as allegorical or figurative (寓) expressions of legitimate

grievance. Sima Qian wrote the parts of the Shiji which Ban Gu found objectionable

because he was provoked to do so. Chao Gongwu, whose entry on Sima Qian

paraphrases extensively from Qin Guan, adds little that is new to this interpretation. He

does, however, express what seems to have been a known and recognized position on the

75
Junzhai dushu zhi 5.176.

287
issue, canonizing it as a valid way of reading the Shiji through inclusion in his

bibliographical description.

SOUTHERN SONG

As I have emphasized throughout this section, one of the important aspects of

autobiographical readings is how readers reacted to and interpreted the events described

in Sima Qian‟s autobiographical texts. Previous chapters have shown that Sima Qian‟s

actions in the Li Ling affair had initially been judged as foolhardy at best, or even

erroneous. In the Northern Song, writers like Qin Guan and Zhang Lei had re-envisioned

the Li Ling affair through a lens of sympathy or even empathy with Sima Qian. Certain

Southern Song writers put forth another view of the Li Ling affair, shifting the blame

squarely and specifically onto Emperor Wu. They emphasized that ruler‟s arbitrary and

overly emotional nature, giving new justification to the Shiji‟s critique of Emperor Wu

and of the Han as a whole. Again, this seems to be in part a response to the political

conditions of the time. Below, I will consider Southern Song reactions figures to the

newly sympathetic autobiographical reading of Shiji, produced or at least popularized by

Su Shi‟s circle.

Zhou Zizhi

Zhou Zizhi 周紫芝 (1082-1155) would have been in the generation just after Su

Shi‟s younger friends and students, but is the first figure we shall consider whose official

career was located squarely in the Southern Song. Already middle-aged, he received his

jinshi degree during the Shaoxing reign period (1131-1162). A comment in the “Self-

288
Narration” of his collected works specifies that he first took office when he was 61 years

of age (i.e., 1142 or 1143), and in 1147 he is referred to as holding the rank of Junior (you

右) Gentleman of Meritorious Achievement (digong lang 迪功郎) and revisor (shanding

guan 刪定官) under the Law Code Office (chiling suo 敕令所). Later he became a

Junior Compiler (bianxiu guan 編修官) in the Bureau of Military Affairs (shumi yuan 樞

密院), and eventually Vice Director (yuanwai lang 員外朗) of the Right Office (yousi 右

司).

A record in the Jiangnan tongzhi 江南通志 [Comprehensive record of Jiangnan]

recounts that in the beginning, the powerful chief councilor Qin Gui 秦檜 (1090-1155)76

was extremely fond of Zhou‟s poetry and was really quite generous toward him. Later,

however, one of Zhou‟s poems contained a line that offended Qin and in 1151, Zhou was

sent away from the capital and became the prefect of a military prefecture.77 Not long

afterwards, he retired to Mount Lu and died there. The offending lines suggest regret for

the Song Emperor‟s captured family members, as does a long memorial submitted much

earlier by Zhou and preserved in Xu Mengxing‟s 徐夢莘(1126-1207) Sanchao beimeng

huibian 三朝北盟會編 [Unified Compilation of Three Beimeng Courts] (ch.124). In that

memorial, which Xu dates to 1130, Zhou uses the captured Han dynasty envoy Su Wu 蘇

76
The all-powerful prime minister during that period. After a long stretch of frequent changes and joint
appointments for the office, Emperor Gaozong 高宗 (r.1127-1162) finally settled on the cautious,
politically astute, but vindictive Qin Gui, apparently because of his willingness to make peace with the Jin
Empire rather than insisting on a restoration of the original Song borders. For in-depth discussion of Song
Gaozong and his chief ministers, see Liu, China Turning Inward and Hsu, “Song Gaozong and his Chief
Councilors.”
77
Jiangnan tongzhi ch.167. The poem was “Imperially commissioned poem of sacrifice for delight in
spring, in celebration of peace” 恭和御製郊祀喜晴詩, found in Zhou‟s Taicang timi ji 太倉稊米集, ch. 37.

289
武 as a comparison and contrast for the Southern Song Emperor Gaozong‟s own behavior

toward his captured relatives.78

It is unclear when Zhou wrote his “Discussion of Sima Qian” [司馬遷論], but it

clearly reflects Zhou‟s experience with the painful north-south Song transition. The

shocking circumstance of Emperors Huizong and Qinzong being held in captivity by the

Jurchen, as well as many others who became high officials in Gaozong‟s court having

first suffered captivity and only later escaping, must have been part of what led Zhou

Zizhi to reexamine Li Ling‟s surrender and Sima Qian‟s reaction to it. Certainly his

“Discussion of Sima Qian” 司馬遷論 shows an awareness of the complications involved

with serving a nation at war.

Zhou began his text by quoting Fan Ye‟s evaluation of Ban Gu, including the

criticism that Ban Gu “excluded pure virtue unto death, and denied a place to rightness

and straightness.”79 As discussed above, this is part of a long-standing debate on the

relative merits and mistakes of Ban Gu and Sima Qian. Because Ban Gu‟s critique of

Sima Qian was so influential, defenders of Sima Qian at times used the strategy of

attacking Ban Gu merely to make room for a verdict on Sima Qian different from Ban

Gu‟s widely accepted one.

Zhou Zizhi then compared the historian‟s task to that of the ruler: the ruler

determines rewards and punishments in the world, whereas the historian determines them

within the framework of the historical record. The historian‟s responsibility, then, “is not

very different from that of the ruler; how could anyone deny the difficulty of it!” [其任常

78
Su Wu spent twenty years in captivity but was ultimately ransomed from the Xiongnu and brought home
by Han Emperor Zhao. The Southern Song Emperor Gaozong never succeeded in forcing or negotiating
the release of his predecessors, arguably because he lacked a whole-hearted commitment to doing so.
79
HHS 40B.1386.

290
與人主相為重輕，顧不難哉].80 Given the profound effect the historian could

potentially have on the future, Zhou Zizhi wrote, it is all the more important not to make

errors such as the ones Fan Ye ascribes to Ban Gu. If such errors are allowed to stand,

“the number of people who know to throw all their effort into doing good will be ever

fewer” [人知勉於為善益，寡矣].81

At this point, Zhou Zhizhi begins to write about Li Ling. It is a provocative

transition, for Li Ling was known as a general who failed to die for his ruler. Here,

however, he is described in heroic terms similar to the ones Sima Qian used—that Li

Ling‟s troops were outnumbered, that they fought dozens of skirmishes, and even used up

all their arrows. Zhou Zhizhi concluded: “Although you may say that his troops were

defeated and his merit was not established, that he brought shame upon his person and his

reputation was ruined, if we consider his intention from the beginning, was it ever lacking

in loyalty or courage?” [雖曰兵敗而功不立，身辱而名巳隳，究其初心，豈不忠且勇

哉？].82 In Zhou Zizhi‟s analysis here, “pure virtue unto death” does not require that one

actually die, only that one is loyal and courageous to the point of being willing to die. In

his view, we should judge Li Ling on the intentions he displayed, and also should

consider the difficulty of his circumstances.

Still remaining close to the narrative in Sima Qian‟s letter, Zhou Zizhi described

the situation back at the Han court: Emperor Wu‟s rage, Li Ling‟s lack of defenders in

court, and Sima Qian‟s courageous action:

Historian Qian forcefully opposed the arguments of everyone else, saying that
Ling had fought with a passion that was utterly selfless, offering himself for his

80
Taicang timi ji 太倉稊米集 [Collected rice shoots in the great granary] 45.1B.
81
Ibid.
82
Ibid.

291
 country‟s sake, and that even the famous generals of antiquity would not have
been able to surpass him by much. But Emperor Wu suspected that [Sima] Qian
wanted to undermine the Sutrishna general,83 that he was acting as a „roving
persuader‟ on Ling‟s behalf, and thus inflicted on him the disaster of the silkworm
chamber. Alas! [Sima] Qian could be called worthy indeed! [史遷力奪羣議，以
謂陵奮不顧身以狥國家，雖古名將無以逺過。而武帝疑遷欲沮貳師，為陵遊
說，遂罹蠶室之禍。噫！遷亦可謂賢矣哉！]84

Here we see Sima Qian‟s action on Li Ling‟s behalf described in unproblematically

laudatory terms. It is tempting to suppose that the Southern Song‟s military and political

situation during Zhou Zizhi‟s lifetime contributed to his insight on the problematic results

of military defeat. Invariably, the disappointed ruler seeks a scapegoat, and conditions

are ripe for the most unpleasant kinds of intrigue and back-stabbing. The swift

succession of chief ministers in Song Gaozong‟s court shows that this was a particularly

sensitive point in the early Southern Song.

Zhou Zizhi then mentioned Emperor Wu‟s quickness to resort to punishment,

calling it an inborn feature of his character. In Zhou‟s portrait of Emperor Wu, we might

see reflections of the mercurial and vengeful Song Gaozong: Zhou‟s own observation of

politics would suggest that to stand up to such an emperor was a righteous, courageous,

and sometimes necessary act. He adds:

It was fortunate indeed that the emperor only responded to Qian‟s arguing for
Ling by punishing him thus [with castration]! Suppose his anger had been
insatiable—how could we know that he would not have responded by killing
[Qian]? It is true that [Qian] was very nearly a gentleman of pure virtue unto
death, and yet Gu criticized him for being unable to “have clear principles and
protect his person.” Does this make any sense? [遷之議陵，帝從而刑之，幸也。

83
I.e., Li Guangli 李廣利 (d.ca.88 BCE). The Emperor honored him with the special title of “Sutrishna
General” (Ershi jiangjun 貳師將軍) because while on expedition to Ferghana (Dayuan 大宛) he succeeded
in conquering the city of Sutrishna. He was also the eldest brother of Emperor Wu‟s favorite consort,
Madame Li.
84
Taicang timi ji 45.2A.

292
 使其怒而不巳，安知其不從而殺之乎！是亦幾於死節之士，而固方且譏其不
能明哲保身。此何理也？]85

Zhou Zizhi‟s intention in beginning with Fan Ye‟s criticism of Ban Gu now becomes

clear. Sima Qian showed himself willing to die for his principles, and very nearly did.

But instead of praising Sima Qian‟s pure virtue (almost) unto death, Ban Gu criticized

him for lack of judgement.86

Zhou Zizhi wrote that the princely man‟s way of having clear principles and

protecting his person is “to be wise enough to anticipate problems, and knowledgeable

enough to protect himself”[ 其智足以慮患，識足以周身].87 In short, the princely man

knows enough to “avoid being like a person walking blindfold, blundering into the net of

the law” [不至冥行以觸罪罟而巳]88—coming to harm without intention or anticipation.

Zhou contrasts this kind of behavior with that of someone who out of caution and desire

to “protect his wife and child” [保妻子] says nothing, even when the ruler is falling into

error. Finally, Zhou returns to the original Shijing context of the line, which was applied

to Zhong Shan Fu. “Suppose that Zhong Shan Fu had remained silent, sitting by while

looking upon the errors of his ruler, and that that was how he practiced the way of

„having clear principles and protecting his person‟—how could it have been said that „if

the ruler erred in his duty, Zhong Shan Fu made up for it‟?” [使仲山甫以緘黙不言，坐

視人主之過為明哲保身之道，則衮職有闕，仲山甫補之果何謂哉].89 Again, we can

85
Ibid.
86
As discussed in chapter 4 above, this criticism appeared in Ban Gu‟s evaluation in Hanshu 62, and refers
to the Shijing ode “Sheng min”: “Intelligent is he and wise,/Protecting his own person” [既明且哲，以保
其身]. Zhou Zizhi‟s slightly condensed version (明哲保身) does not appear in the actual Hanshu
evaluation but is more convenient of reference.
87
Taicang timi ji 45.2A.
88
Ibid.
89
Ibid., 45.2B.

293
discern here the influence of Zhou‟s own political milieu, a criticism of those too

cautious to remonstrate with the emperor, pandering to his self-interested desire for peace

rather than espousing the more righteous (if risky) cause of retaking the north.

Zhou Zizhi then approaches the problem of culpability in the Li Ling affair from

another direction, responding to the objections of an anonymous interlocutor. This

„devil‟s advocate‟ argues as follows:

[Someone] said: Regarding his battle with the Shanyu, Ling should have died for
his country and should not have surrendered. Therefore, surrendering was the
crime that Ling committed. Regarding Li Ling‟s surrender, Qian should have
fairly admitted his error and should not have argued for him. Arguing for him
therefore was the crime that Qian committed. Regarding [Si]ma Qian‟s arguing,
the Han [emperor] should have tolerated it and should not have punished him.
Punishing [Qian] therefore was the mistake that the Han made. Han‟s blaming
Qian was not without error. But Qian‟s seeking to rescue Ling could not avoid
failing every time. And so Gu‟s reproach of Qian was correct. [曰：單于之戰，
陵當死國而不當降，降則陵之罪也。李陵之降，遷當直其過而不當辨，辨則
遷之罪也。馬遷之辨，漢當容之而不當刑，刑則漢之過也。漢之責遷不為無
罪。遷之救陵未免於屢敗，固之所以責遷者是也。]90

Zhou Zizhi‟s rebuttal relies heavily on the interpretation of events found in the “Letter in

Reply to Ren An,” from which he quotes copiously. The unknown interlocutor has

suggested that, although the Han emperor was to blame, neither Li Ling nor Sima Qian

were innocent either. If Li Ling did wrong, then Sima Qian did wrong too in defending

him, so to clear Sima Qian‟s name, Zhou must also clear Li Ling‟s. He therefore adopts

Sima Qian‟s argument, that Li Ling allowed himself to be captured alive because “there

was still hope of [aiding] the Han” [猶冀得當以報漢也]. He adds that “Qian arguing for

Ling in the way that he did might indeed be called understanding [Ling‟s] inmost heart!

How did he know that [the court] would not accept his words, but on the contrary

90
Ibid., 45.2B-3A.

294
suspected that he was engaging in „roving persuasion‟?” [遷之論陵如此，可謂得其心

矣！奈何不納其言，而反疑以遊説乎？].91

It is highly unlikely that Zhou Zizhi had any new information or independent

corroboration about the events in the Li Ling affair. He had come to a verdict different

from (for example) Bai Juyi because he lived in a different time. On the one hand,

accepting compromise was unavoidable in Zhou‟s political world; on the other hand, the

possibility of questioning decisions made by the emperor or his surrogate seemed to be

fast disappearing.92 Scholars of today might argue that by merely offering Sima Qian‟s

own version of events, Zhou Zizhi was adding nothing new. It is easy to make a case that

Sima Qian had reason to believe Li Ling‟s actions and his own to be defensible; the

question is whether Sima Qian‟s interpretation of events was correct. However, what is

significant about Zhou‟s version of events is his willingness to give Sima Qian‟s political

interpretation a completely sympathetic reading. Zhou defends Sima Qian against Ban

Gu because of his own political climate: he served at a time when frequent, unpredictable

shifts in the winds of policy ensured that it was not only hot-headed fools who got into

trouble for saying the wrong thing. A further example of this is Zhou‟s explanation of

why Li Ling failed to return to the Han, always a thorny point for those who would

defend the general‟s loyal intentions and Sima Qian‟s interpretation of them:

Ling did not return to the Han after his defeat was because he knew the
Han would certainly kill him. If they killed him, then it would have been a
useless death. It seemed better to use his defeat in order to achieve something for
the Han, which he still had hope of doing. One cannot say that Ling lacked skill
in making the most of his death!93 When Ling heard that Qian had spoken on his

91
Ibid., 45.3A.
92
See Liu, China Turning Inward, 18-19 on the increasingly autocratic and absolutist tendencies of the
Southern Song court.
93
Surely a reference to the famous line in Sima Qian‟s “Letter in Reply to Ren An”: “It is certain that a
person has but one death, which can be as weighty as Mount Tai or as light as a goose feather—it is how in

295
 behalf and barely avoided death, he knew that the Han would certainly kill him.
It is for this reason that when he was summoned [back] he did not go. After he
did not go, [the Han] responded by killing his mother and wife. At this point,
Ling‟s hope in [the Han] was at an end, and though they should send a hundred
envoys with ten thousand strategies to summon him, would he have gone back?
Speaking from this point of view, it is clear that the error was on the part of the
Han and not on the part of Qian. [且陵敗而不歸漢，知漢之必殺巳也，殺之則
無益於死，不若因敗以立功於漢，猶有望焉。陵不可謂不善於處死者也。遷
為陵言而幾不免死，陵之聞之，知漢之必殺巳也，是以招之而不至。及其不
至，又從而殺其母、妻。陵之望於是絶矣，雖遣百使萬方而招之，其有至哉？
由是言之，過在於漢而不在遷明矣。]
Suppose that Qian‟s words actually had their intended effect in his own
time. Then the Han would not be known for turning its back on a loyal servant
[Ling], nor would it have committed the crime of punishing the wife and mother
of a loyal servant, nor would it have made the mistake of ignoring a remonstration.
In one stroke it would have attained three good results. This is the reason why
Qian spoke so many times on Ling‟s behalf. [借使遷言果效於當時，漢無負忠
臣之名，無報忠臣母、妻之罪，無拒諌不納之失，一舉而三善隨之。此遷所
以反覆為陵言也。]94

The extent to which Zhou‟s argument is convincing to us is the extent to which we

believe Sima Qian was right about Li Ling. Was Sima Qian‟s rhetoric about Li Ling

wanting to requite the Han really a knowledgeable interpretation, or was it ex post facto

rationalization? All arguments about the Li Ling affair ultimately depend on whether one

chooses to believe Sima Qian‟s version of events, and this essay makes it clear that the

verdict was not unanimous. In the early Southern Song, when the issues of loyalty and

surrender were painfully close to the hearts of everyone, from emperors to generals and

even perhaps to ordinary citizens, there was space to reconsider the complexity of the Li

Ling affair. I would argue, however, that Zhou Zizhi‟s reading of Sima Qian and Li Ling

reflects a developing tendency to exonerate Sima Qian and instead blame Emperor Wu,

one uses it that the difference lies” [人固有一死，死有重於泰山，或輕於鴻毛，用之所趨異也] (HS

62.2732).
94
Taicang timi ji 45.3A-3B. The careful reader will note an oddity in the last line: Zhou wrote that Sima
Qian spoke “many times” (反覆) on Li Ling‟s behalf. No other version of the events suggests this.
Possibly it is a mis-reading based on there being two versions of the story (in the “Letter” and Hanshu
ch.54)?

296
and that this tendency was originally influenced by the Southern Song political situation.

The next two readings I will discuss, by Wang Guanguo 王觀國 (fl. mid-12th c.) and Lü

Zuqian 呂祖謙 (1137-1181), each continue this trend.

Wang Guanguo

Wang Guanguo‟s dates are unclear, but like Zhou Zizhi he was active during the

Shaoxing reign period (1131-1162) of Song Gaozong, especially in the 1140s. He does

not appear in the Songshi 宋史 [History of the Song] and is known primarily for his

collection of scholarly studies, the Xuelin 學林 [Forest of learning], which the Siku

quanshu general catalogue describes in the following terms: “There were only a few

among the Southern Song ru who paid attention to evidential scholarship, and someone

like Guanguo could be called truly outstanding and extraordinary” [南宋諸儒講考證者

不過數家，若觀國者亦可謂卓然特出矣].95 The brief, small-scale studies in Xuelin

concern themselves primarily with philological problems: the meanings or pronunciations

of individual characters or how a phrase should be understood. Often these small facts

can have larger implications, however, as in the study which concerns us here.

The Xuelin note entitled “Fermented Wine and Seizing of Ranks” [酎酒奪爵]

initially seems to have little to do with Sima Qian. It concerns an incident not mentioned

at all in the Shiji, but which appears in the Hanshu “Basic Annals of Emperor Wu.” The

text, which Wang Guanguo quotes, reads:

95
SKQS “Zongmu”, 1582. The term I have translated here as „evidential scholarship‟ (kaozheng xue 考證
學) would eventually represent an intellectual tradition with much to say about the Shiji. Though the bulk
of these kaozheng scholars‟ work is beyond the scope of this study, some discussion can be found in Part
III below.

297
 In the fifth year of Yuanding (112 B.C.E.)…in the ninth month, various marquises
were accused of not having presented, for the sacrificial [offering of] the eighth
month fermented wine in the [imperial] ancestral temples, the [amounts] of real
gold specified by the law. [Noble titles] were taken away from a hundred and six
persons. [元鼎五年。。。九月，列侯坐獻黃金酎祭宗廟不如法，奪爵者百六
人。] 96

Wang also quotes Hanshu commentator Fu Qian‟s partial explanation of the incident:

“This was because the eighth month was the time for presenting wine for sacrifice at the

ancestral temple, but [Emperor Wu] caused the various lords each to present gold to

assist the sacrifice” [因八月獻酎祭宗廟時，使諸侯各獻金来助祭也]. This does

something to explain why they would have been expected to present gold in the first

place, but does not explain the mass demotions. As Wang comments, “Since it was not a

great offense against the Way, [the punishment] should not have extended to having their

rank taken away and losing their fiefs” [自非大惡不道，則不至于奪爵失侯]. 97

Wang Guanguo then offers a hypothesis about the solution to the puzzle, saying

that, “now the crime of paying less „fermented wine‟ gold than specified by the law was

not typically punished by loss of one‟s title, but Emperor Wu seized upon the law [as an

excuse]” [今酎金不如法，辠不至于奪爵，而武帝削奪之典]. Emperor Wu‟s reasons

for doing so can only be understood from the narrative in the Hanshu “Treatise on

Foodstuffs and Commodities,” which gives a more complete story:

The Southern Yue rebelled, and the Western Qiang invaded the border....The
Prime Minister of Qi, Bu Shi, submitted a memorial requesting that he and his
sons be allowed to die [fighting] the Southern Yue. The Son of Heaven sent
down an edict praising him and rewarding him with the promotion to marquis
within-the-passes, as well as with 47 jin of gold, and ten qing of fields. [This
edict] was publicized throughout the empire, but no others responded to it.
Among the various lords, of which there were several hundred, none sought to
join the military forces. When it came time for the drinking of the eighth-month

96
HS 6.187; translation based on Dubs II.80.
97
Xuelin 3.84.

298
 liquor, the Privy Treasurer inspected their gold and more than a hundred were
sentenced because their eighth-month gold fell short. [南粤反，西羌侵邊。齊
相卜式上書願父子死南粤。天子下詔褒揚，賜爵闗内侯，黄金四十斤，田十
頃。布告天下，天下莫應。列侯以百數，皆莫求從軍。至飲酎，少府省金，
而列侯坐酎金失侯者百餘人。]98

The part of this passage which is interesting for my purpose is Wang Guanguo‟s

analysis of this incident in terms of Emperor Wu‟s psychology. After some summary and

explanation of the empire‟s military and economic situation, Wang adds that due to this

lack of response

Emperor Wu grew angry, and thereupon used [the nobility‟s] failure to pay in full
the specified „fermented wine‟ tribute as an excuse to seize the titles of more than
a hundred lords. This is something unprecedented in Han law. It was only
because of his personal anger that [Emperor Wu] exercised his power this way.
Thus, although the penalties regarding “fermented wine” gold were light,
Emperor Wu seized upon the law [and punished the transgressors] with great
severity. [武帝因此發怒，乃以酎金不如法而列侯奪爵者百餘人。此在漢法
未之有，特以私怒而加威，故酎金之辠雖輕，而削奪之典特重。]99

Wang Guanguo draws very clearly here the contrast between what the law prescribes and

the power exercised by the emperor to alter the penalties, in this case making them more

severe. It is interesting that he does not criticize this practice outright, adding, “It would

be permissible to exercise one‟s authority like this on a single occasion, for after all, it

was not a law that was written in stone” [權一時之宜可也，非不刋之法也].100

At this point, however, he changes track and gives another example of what he

considers to be Emperor Wu‟s arbitrary exercise of authority in contrast to what the law

would prescribe:

In antiquity, those who recommended someone as being worthy but the

recommended person later proving unfit, the recommender was deprived of his
title and position, no more. Of the five punishments of antiquity, the “palace

98
HS 24B.1173.
99
Xuelin 3.85.
100
Ibid.

299
 punishment” was reserved for those who committed crimes of a sexual nature.
When Sima Qian was accused of having recommended Li Ling, and was sent
down to the silkworm chamber, his crime and his punishment seem not to fit at all.
[古之舉賢不當者，削爵黜位而已，古之五刑，犯淫者待之以宫刑。司馬遷
坐舉李陵降匈奴而下蠶室，其辠與刑，頗不從類。]101

In the last chapter, we already saw this problem and its possible solution hinted at

by Zhang Zhuo 張鷟 (ca.660-740). Clearly Wang Guanguo did not arrive at the same

solution as modern scholars like Qian Mu and Lü Xisheng, that Sima Qian himself must

have requested the punishment of castration.102 Instead Wang quotes Wei Hong‟s

account, in particular, the line which previously I translated “[Sima Qian] was sitting in

attendance and put in a good word for Li Ling.”103 Wang‟s interpretation hinges on the

way he reads the character ju 舉, which he understands as “to recommend for office.” It

seems likely that Wei Hong, in the first century, used ju 舉 to mean something like

“elevate” as in Lunyu, “Elevate the straight and set them over the crooked” [舉直錯諸

枉],104 a broader sense that could certainly include official recommendation (as in the

common phrase 舉賢良 “to recommend the good and the worthy”) but would not be

limited to it and could also refer to something like simple moral praise. In order to make

his point, however, Wang Guanguo interpreted Wei Hong‟s story to mean that it was

Sima Qian who recommended Li Ling for office. Wang does not seem to find this or any

101
Ibid.
102
Qian Mu, “Taishigong kao shi,” 26-27; Lü Xisheng, “Sima Qian gongxing xiyi,” 68. See chapter 4
above for a more detailed discussion.
103
The full passage, as preserved in Pei Yin‟s Shiji jijie commentary, is discussed in chapter 4 above:
“When Sima Qian made the „Basic Annals of Emperor Jing,‟ he expended great passion in discussing
[Emperor Jing‟s] shortcomings and Emperor Wu‟s excesses. Emperor Wu was infuriated. He destroyed
and cast away [those annals]. Later, [Sima Qian] was sitting in attendance and put in a good word for Li
Ling. Ling surrendered to the Xiongnu. Because of this Qian was sent to the Silkworm Chamber.
Resentful words were spoken, and [Qian] was sent to jail where he died.” [司馬遷作景帝本紀，極言其
短及武帝過，武帝怒而削去之。後坐舉李陵，陵降匈奴，故下遷蠶室。有怨言，下獄死。] (SJ
130.3321.)
104
Lunyu XII:22.

300
other aspect of the Wei Hong anecdote problematic. Instead he uses it as further

evidence that Emperor Wu bent the laws because of his private feelings:

From this [i.e., Wei Hong‟s anecdote] we see that when Sima Qian was tried for
recommending Li Ling and was sent down to the Silkworm Chamber, it was
really a punishment that resulted from Emperor Wu‟s private anger. That is why
the punishment and the crime seem not to fit at all. [由此觀之，則司馬遷坐舉
李陵而下蠶室，實武帝私忿之刑，故罪與刑所以不從類也。]105

It is unclear why a careful scholar like Wang Guanguo saw no contradiction (as

Yu Jiaxi did106) between his reading of Wei Hong‟s account and the version of events

given in the “Letter.” There is a slight possibility that in fact there is no contradiction—

after all, the “Letter” was written ostensibly in part as a discussion of recommending

worthies, and Sima Qian may not have explicitly said he recommended Li Ling merely

because it was obvious to his reader (not us, of course, but Ren An). Another possibility

is that Wang Guanguo made no special study of the matter, and made a careless but

superficially plausible error. Still, it is interesting that the Wei Hong account could still

have been as believable in the Southern Song, for it shows the continuance of a particular

understanding of the Shiji‟s critical content: that Sima Qian‟s criticism of Emperor Wu

was present even before the Li Ling affair and was not strictly a result of it.

In any case, the point Wang Guanguo makes both in the fermented wine case and

in Sima Qian‟s is that Emperor Wu‟s capricious exercise of power, guided only by his

own private anger, led to troubling irregularities in the functioning of the political system.

In Wang‟s eyes, the emperor was not acting as an emperor should. In such a case, Sima

105
Xuelin 3.85.
106
In his “Taishigong wangpian kao,” Yu writes, “If we investigate it in the Hanshu, Qian‟s offense was
sitting in council and coming to the aid of Li Ling, nothing more. It was never that he recommended Li
Ling for his post as General, and there was never any matter of his dying in prison.” [考之漢書，遷之得
罪，坐救李陵耳，未嘗舉以為將，亦無下獄死之事。] Clearly Yu Jiaxi‟s reading of the character ju 舉
was influenced by Wang Guanguo‟s.

301
Qian‟s criticism was not defamatory writing (bangshu 謗書) motivated by private

resentment, but rather a selfless attempt to set the record straight, and perhaps even to

remonstrate.

Lü Zuqian

Lü Zuqian‟s reading of the Li Ling affair is not quite so readily accessible as the

others‟, for it appears in a genre of writing whose conventions were very different from

the poems and prose pieces by Song figures considered above. The major

historiographical activity in the Song dynasty took the form of rewriting and continuing

earlier histories. Many of these rewritings add very little that is new, but a few may be

mined for indications of their compilers‟ changing judgements on historical events. This

is the case with Lü Zuqian‟s rewriting of the Li Ling affair in his unfinished historical

text, the Da Shi Ji 大事記 [Record of major events].

In order to understand Lü Zuqian‟s work, it is first useful to establish a context for

it by considering other versions in the genre. Sima Guang‟s version of the Li Ling affair

in one of the most famous rewritings of history, the Zizhi tongjian 資治通鑒

[Comprehensive mirror to aid in governing], is very close to the Hanshu account, only

slightly abbreviated. Zhu Xi‟s abridgement of Sima Guang‟s work, the Zizhi tongjian

gangmu 資治通鑒綱目 [Outline of the comprehensive mirror to aid in governing],

includes the Hanshu/Sima Guang version as commentary under a heading which presents

Li Ling‟s failure in the most stark and negative possible terms: “In the summer of the

second year, Li Guangli was sent to lead troops and strike at the Xiongnu. Another

302
general, Li Ling, fought, was defeated, surrendered, and was taken prisoner” [二年夏，

遣李廣利將兵擊匈奴，别將李陵戰敗降虜].107

By contrast with both of these, Lü Zuqian‟s version analyzes the reasons behind

the Li Ling affair from a much broader perspective. First let us consider how the Hanshu

account explains the events surrounding Li Ling‟s defeat: “In the beginning, the emperor

sent out the Sutrishna General [Li Guangli], and only sent Li Ling to provide support

troops. It happened that Ling engaged directly with the Shanyu, but the Sutrishna

General‟s achievement was very minor” [初，上遣貳師將軍出，財令陵為助兵，及陵

與單于相値，而貳師功少].108 The discomfort produced by this rivalry between

generals, at least according to Ban Gu, was part of why Emperor Wu reacted so badly to

Sima Qian‟s defense of Li Ling. Lü Zuqian took up this same theme in his analysis of

the heading “The Senior-Director Archivist Sima Qian is sent down to the Silkworm

Chamber” [大史令司馬遷下蠶室]. Lü first quoted the Hanshu account mentioned

above109 and then wrote:

Emperor Wu‟s favored generals were Wei Qing and Huo Qubing. Those like Li
Guangli, though their reputation was flourishing and they were brave heroes, he
looked down on as if they were mere dung and dirt. The more people like Li
Guang and his descendants were neglected and constrained, the more they were
admired by brave heroes. The split into two factions was already well established
[by this time]. If Emperor Wu had been able to put the public good above his
own selfishness, and let go of his collected resentment, then how would this kind
of schism develop between the old and the new [generals]? The emperor being
unable to engage in this sort of self-examination, he inevitably glared at those
below him, cherishing his grudge and storing up his resentment, so that it was
only a matter of time before it poured out. Sima Qian‟s words happened to touch

107
ZZTJ gangmu 5A.28B.
108
HS 54.2456.
109
With an interesting difference: the Hanshu version has it that “the emperor believed that Qian was
treacherously deceiving him out of a desire to slander the Sutrishna general” [上以遷誣罔，欲沮貳師]
(HS 54.2456). Lü‟s version omits the two characters wuwang 誣罔 (treacherously deceiving), which
actually alters the connotation considerably, making Sima Qian‟s purported crime seem far less sever.

303
 upon a sensitive spot,110 and so it made sense that he was punished. [武帝寵將衛
青、霍去病，李廣利之屬，名位雖盛而豪傑，賤之如糞土。如李廣父子愈擯
抑，而豪傑愈宗之，分為兩黨乆矣。茍武帝以公滅私，消彌衆憤，則安得有
新舊彼此之隙哉。帝既不能自反，則必疾視其下，懐怒蓄憤，有待而發。司
馬遷之言適觸其機，宜其不免也。]111

An interesting feature of Lü Zuqian‟s analysis is that although it does criticize Emperor

Wu, the criticism is not that he was excessively harsh in punishing Sima Qian, but rather

that he set up the situation in which Li Ling was defeated. In part, this is a feature of the

genre in which Lü Zuqian is writing: because he is not focused primarily on Sima Qian

(as, for example, Zhou Zizhi is), he takes a broader view, presenting an analysis of the

situation in which Sima Qian‟s role was in fact only a minor one. As in Zhou‟s essay,

however, we can see here an image of the Southern Song tragedy reflected back upon the

Han. The Song Emperor Gaozong, who was very nearly toppled in a military coup at the

beginning of his reign, never trusted his generals and frequently seemed intent on

passively or even actively weakening them.112

As for Lü‟s reaction to Sima Qian‟s fate, it sketches out an interesting position in

which Sima Qian‟s action comes out morally correct, and yet Lü expresses no sympathy

or sense that injustice was done.

At this time, the various ministers were all blaming Ling, but Qian said, “Ling
getting involved with this matter was one piece of misfortune. All those ministers,
whose [only concern is to] keep themselves safe and protect their wives and
children, take advantage of this to plot against him.” Probably it was out of
loyalty to Emperor Wu [that Qian said this], and he simply did not have leisure to
consider the piled-up resentment. Because of his loyalty he committed this
offense, [but] having already done his duty as a minister, what more could he
seek!? Nonetheless, [Qian] resentfully sought to blame those around him for not
saying a word [in his favor]. He did not consider that he had previously
110
Literally, “hit its pivot.”
111
Dashiji jieti 12.133B.
112
The circumstances and results of the coup, known in Chinese as “Miao Liu zhi bian” 苗劉之變 [the
Miao-Liu affair], are described in detail in Yeong-Huei Hsu‟s “Song Gaozong and his Chief Councillors,”
64-74.

304
 condemned those around him as being people who were [only concerned] to keep
themselves safe and protect their wives and children. Was it not misguided for
him still to be pettily hoping that they would devote themselves [to helping him]?
This is a problem of being inadequately learned. [當是時羣臣皆罪陵，遷乃言
陵今舉事，一不幸，全軀保妻子之臣隨而媒孽其短。蓋為武帝忠，計不暇顧
衆怨耳。以忠獲罪，既得為臣之義，餘何求哉。反憤然追咎左右親近不為一
言，抑不思左右親近，則遷前日詆以為全軀保妻子者也。猶區區望其致力，
不亦惑乎。此學問不足之病也。]113

Lü Zuqian‟s rhetoric in this case is clever; nonetheless when Sima Qian wrote in the

letter “among those surrounding the Emperor no one said a word for me” [左右親近不為

壹言],114 it does not seem likely that he entertained hopes about the same group of people

he had been criticizing. More likely there was a factional situation in which Sima Qian‟s

group failed to come to his defense as he might reasonably have expected them to do.

That might explain why the above quoted phrase is paired with another, that “not one of

my friends would save me” [交遊莫救].115 Whether or not we accept Lü‟s analysis,

however, it does show a Song tendency to make nuanced judgments on historical figures

like Li Ling and Sima Qian.

Lü Zuqian‟s surviving writings do not give a full picture of the degree of

importance Lü placed on the Shiji.116 I have been able to locate only one other passage in

which Lü seems to comment directly on what he thought Sima Qian‟s purpose and

intention was. Commenting on a memorial of the Han Emperor Wen, Lü wrote:

The Shiji “Annals of Emperor Wen” records many edicts, but when it comes to
the “Annals of Emperor Jing” then none of them are recorded. Probably he
considered them unworthy of being recorded; how subtle was his aim! [《史記·

113
Ibid.
114
HS 62.2730.
115
Ibid.
116
According to Zhu Xi, “Bogong [=Lü Zuqian] and Ziyue [=Lü Zuqian‟s brother Zujian] took the learning
of the Honorable Senior Archivist as their revered model, and believed that no Han classicist came up to
him. I once argued bitterly with them [over this]…” [伯恭 子約 宗 太史 公 之學 ， 以為 非漢 儒 所及 ，
某嘗 痛 與之 辨 …] (ZZYL 122.2951).

305
 文帝紀》多載詔書，至《景帝紀》則皆不載，蓋以為不足載也，其旨微
矣！]117

The above passage actually introduces Lü‟s most famous argument regarding the Shiji,

namely that most of the so-called ten “missing” chapters—including the “Annals of

Emperor Jing”—were never missing at all. I mention it briefly here, however, because it

shows that Lü, himself the author of a private historical work, appreciated and admired

“subtle” techniques for praise or blame that a historian like Sima Qian seemed to have

had at his disposal.

It would be misleading to suggest that all Southern Song writers were increasingly

sympathetic towards Sima Qian. I mentioned above that Zhu Xi, a towering figure in

Song intellectual history, recast the history of Li Ling‟s defeat in such a way as to

completely dismiss the heroism Sima Qian ascribes to the captured general. He also had

quite a number of critical things to say about Sima Qian,118 but ignored any form of

autobiographical reading. The closest he comes to discussing such a reading is his

complaint that the “Traditions of Bo Yi” 伯夷列傳 “is resentful words from beginning to

end, completely slandering Bo Yi!” [一傳中首尾皆是怨辭，盡說壞了伯夷！ ].119

Yet Zhu refuses to comment on any of the possible motivations behind such resentment.

This in itself indicates that the autobiographical reading must have commanded

considerable sympathy during Zhu Xi‟s day, and Zhu Xi perhaps considered that his

critique would be more compelling if he turned his back such questions as why the

“Traditions of Bo Yi” might be full of resentment.

117
Dashiji jieti 10.111A.
118
Discussed at length in chapter 6 below.
119
ZZYL 122.2952.

306
 Later reactions against those sympathetic to Sima Qian were less restrained. A

particularly revealing contrast to Qin Guan‟s sympathetic autobiographical reading and to

Lü Zuqian‟s implicit approval of Sima Qian‟s techniques of criticism is that of Wang

Ruoxu 王若虚 (1174-1243). Wang Ruoxu was actually an official of the Jin 金 (Jurchen)

dynasty, but clearly he was thoroughly familiar with Song scholarship. He was known

for his literary ability and scholarship and served in a variety of posts. Among other

duties, he was charged with the compilation of Jin Emperor Xuanzong‟s “True Records.”

Wang himself was no stranger to politically dangerous situations: his biography in the Jin

shi 金史 [History of the Jin] records that during the troubled final years of the Jin dynasty,

he was summoned by the faction of the turncoat Cui Li 崔立, who had surrendered

Kaifeng (together with the Jin royal family) to the invading Mongol forces. Cui Li‟s

confederate Cui Yi 崔奕 ordered Wang Ruoxu to make a commemorative stele for Cui

Li‟s “achievements.” Wang Ruoxu said that death would be preferable but added that he

would compose such a stele if Cui Yi would tell him what achievements there were that

would be worth writing about. Ultimately, Wang managed to evade any major

responsibility for composing the stele, which in any case was never erected. After the

fall of the Jin in 1234, he went home to his native place to quietly live out his days. He

died during a trip to Mount Tai in 1243.120

Wang Ruoxu wrote extensively on the Shiji in his Shiji bianhuo 史記辨惑

[Refuting errors regarding the Shiji], which occupies chapters 9-19 of his Hunan yilao ji

120
Jinshi 金史 126.2737-2738.

307
滹南遺老集.121 A particularly striking feature of his work on the Shiji is that his

evaluation is unrelievedly negative. Not only does Wang take Sima Qian to task for a

variety of factual and doctrinal errors, but even goes out of his way to refute

conventionally accepted praise of Sima Qian as a great prose stylist.122 Amongst his

comments, we find one which echoes the argument found Zhang Lei‟s “Discussion of

Sima Qian,” but is utterly lacking in any potentially sympathetic double meaning:

Ban Gu ridiculed [Sima] Qian‟s discussion of the roving warriors and his
narration of the merchants, and the world has pronounced him correct. Yet Qin
Shaoyou [=Qin Guan] disputed it, considering that when Qian suffered the rotten
punishment, his family was poor and he was unable to ransom himself, while
none of his acquaintances would come to his rescue, thus he poured forth his
resentment and so on. This really does capture [Sima Qian‟s] basic intention.
However, a reliable history will be a model for ten thousand generations, and is
not written for one‟s self. How is it then that [Qian] poured forth his private
resentment!? [班固譏遷論游俠述貨殖之非，世稱其當，而秦少游辨之，以為
遷被腐刑，家貧不能自贖，而交游莫救，故發憤而云。此誠得其本意，然信
史將為法於萬世，非一己之書也，豈所以發其私憤者哉。 ]123

Wang Ruoxu, an upright and steadfast professional historian, denies the appropriateness

of any personal feeling in the compilation of official history. Underlying his criticism is

the interesting statement that history can serve as a model for ten thousand generations.

It echoes and contradicts a statement from Su Xun‟s “Discussion of History,” where Su

121
The Shiji receives by far the most extensive treatment of any single work within the Hunan yilao ji, with
the next longest being the Lunyu, with a mere five chapters. Apparently the Shiji was especially replete
with errors that needed refuting.
122
For example, Wang criticized a well-loved passage about General Li Guang, writing, “Li Guang „saw a
stone in the grass. Believing it was a tiger, he shot it. He hit the stone and the arrow point sank in. He
went to look at it, and saw that it was a stone. Therefore he went back and shot at it again, but in the end
was never again able to sink an arrow into the stone‟ (SJ 109.2871-2872). In total, [Sima Qian] uses three
more „stone‟ characters than he needs to. He ought to have written, „He believed he saw a tiger and shot it,
and the arrow sank it. When he realized it was a stone, he therefore went and shot at it again but in the end
was unable to get it to go in.‟ Or writing, „He once saw that there was a tiger in the grass and shot it, so the
arrow sank in, then saw that it was a stone‟ would also do.” [李廣“見草中石，以為虎而射之，中石沒
鏃，視之石也，因復更射，終不能復入石矣”。凡多三“石”字，當云“以為虎而射之，沒鏃，既
知其石，因復更射，終不能入”。或云“嘗見草中有虎，射之，沒鏃，視之石也”亦可] (Hu nan
yi lao ji 15.97).
123
Hu nan yi lao ji 19.117.

308
said, “a history is not a constant rule for ten thousand generations” [史非萬世之常法,

emphasis added].124 Su Xun had argued that the Classics are like “compass, square, level,

or marking line—things that determine a vessel” [夫規矩準繩所以制器], while the

histories are like the vessels. “The histories wait for the Classics in order to be made

correct. But without the histories, the Classics would be obscure” [史待經而正，不得史

則經晦].125

Wang Ruoxu‟s point is that the vessel can also be a constant model, if it is made

true and reliable according to the measuring tools of the Classics. But—unlike Qin Guan,

who placed considerable value on private literary artistry—Wang completely denied the

value of a vessel made according to individual idiosyncrasy.

It is worth noting that Wang used the notion of Sima Qian‟s resentment in the

more conservative, unsympathetic way, to argue that Sima Qian was not merely a good

historian writing a true record. In another passage, one that responds to Lü Zuqian‟s

above remark about the difference between the Emperor Wen and Emperor Jing annals,

Wang argued even more forcefully that the interference of Sima Qian‟s private

resentment impoverished, rather than enriched, his history.

The Dashi ji [says]: “The Shiji „Annals of Emperor Wen‟ records many edicts, but
when it comes to the „Annals of Emperor Jing‟ then none of them are recorded.
Probably he considered them unworthy of being recorded; how subtle was his
aim!” I would say that an historical text is a „true record‟, and edicts and imperial
announcements are the momentous affairs of any era. Even if what an emperor
puts into effect is not able to match his words, how can one allow [those words] to
vanish away? This results from Qian‟s private resentment, but Master Lü deeply
approved it. [《大事記》：“《史記·文帝紀》多載詔書，至《景帝紀》則皆
不載，蓋以為不足載也，其旨微矣！”予謂史書實錄也，詔誥一時之大事，

124
Jiayou ji 9.229.
125
Jiayou ji 9.230.

309
 縱使帝之所行不能副其言，豈容悉沒之乎？此自遷之私憤，而呂氏深取
之。]126

Again, Wang Ruoxu is an exceptional case. He expressed no interest what the proper

verdict on the Han emperors might be, but merely argued for a historiographical principle,

that an emperor‟s edicts should be recorded in his “Basic Annals,” even regardless of

how effectively they were put into practice, let alone what the historian‟s opinion of him

was. In this, Wang Ruoxu‟s argument sounds almost modern, advocating that primary

sources should be preserved intact so that later generations can make their own decisions

about them.

As with Zhang Lei‟s “Discussion of Sima Qian” analyzed above, one other

notable thing about Wang Ruoxu‟s critique here is the expansion of specific aspects of

the Shiji now accepted as “subtle” (微) criticism or autobiographically motivated hidden

messages. Even though Lü and Wang differ in their judgements, they now agree that the

“Annals of Emperor Jing” has joined the list of chapters that should be considered in this

way.

Huang Zhen

The last figure I will discuss is Huang Zhen 黃震 (1213-1280), who lived at the

very end of the Southern Song. Huang Zhen got his jinshi degree in 1256. His official

biography in the Songshi 宋史 reads almost like a hagiography, listing his many daring

efforts to relieve the common people from suffering and from the deceptions and

depredations of the wealthy. His long official career included a stint as District Defender

126
Hu nan yi lao ji 19.116-117.

310
of Wu Province, as well as political setbacks due to his bold remonstrations and slanders

from his enemies. He also participated in the compilation of official histories, in

particular the True Records of Emperors Ningzong and Lizong. He is known for the

motto, “Read nothing that was not written by a sage; write no prose or poetry unless it

does some good” [非聖人之書不觀，無益之詩文不作]. Intellectually, his roots lay

with Ye Shi‟s utilitarianism (功利之學), but he founded of his own school of thought

known as the Dongfa school (東發學派). He is considered to be a Song martyr, someone

who committed suicide rather than continuing to live under Mongol rule.127 His studies

of the Shiji (and indeed of many other texts as well), have been preserved in the

Huangshi richao 黃氏日抄 [Daily copying of Master Huang], and were apparently

carried out through the process of copying from the text, with each chapter followed by a

record of Huang‟s thoughts and reactions.

I will begin by briefly considering Huang‟s overall evaluation of Sima Qian,

which actually appears in the section devoted to the Hanshu. Huang writes,

[Sima] Qian, with his dauntless and independent spirit, was guiltless yet received
humiliation. He was stirred to literary composition, a powerful vision of a
thousand [years] of antiquity. Ah, magnificent indeed! It is a pity that he had not
heard of the Way... [遷以邁往不群之氣，無辜受辱，激為文章，雄視千古。
嗚呼亦壯矣！惜乎其未聞道也]128

Following Zhang Lei, Huang Zhen suggested a link between Sima Qian‟s fundamental

character—his qualities of rushing forward (“dauntless” 邁往) and of not merely doing

what others do (“independent” 不群) caused him to take a brave and unpopular stand on

127
Although Huang Zhen‟s Songshi biography does not mention the manner of his death, the Song Yuan
xue an 宋元学案 [Case studies of Song and Yuan learning] that he “starved to death at Bao Zhuang” [餓于
寶幢] (86.2885).
128
Huangshi richao 47.13a.

311
the Li Ling affair.129 Importantly, Huang Zhen emphasized that Sima Qian was

“guiltless”: he did not make a mistake in speaking up for Li Ling, but was punished

anyway. And, using the same word as Zhang Lei, Qin Guan, and others, Huang wrote

that Sima Qian was “stirred” (激) to create the Shiji. Huang expressed admiration for

Sima Qian‟s ability to survey the vastness of antiquity, but then expresses his reservation

as well, criticizing Sima Qian for not having heard of the way. The rest of the passage

goes on to explain this remark, complaining that in writing about the Spring and Autumn

period, Sima Qian included just what the Master decided to discard. There is no

indication that Huang Zhen wanted to extend his criticism to Sima Qian‟s portrayal of the

Han (as Ban Gu, for example, had done).

Indeed, Huang Zhen followed the Southern Song trend I have pointed out, in

being thoroughly critical of Emperor Wu and his endeavors. His response to the

“Honorable Senior Archivist‟s Self-Narration” was again fairly sympathetic and admiring

toward Sima Qian, yet seems to take him to task for not being critical enough:

Tan had a son, Qian, who was able to use his literary prose to carry forth his
family‟s traditions and glorify their name in later generations. Indeed, this could
be called skillfully carrying on a person‟s [i.e., Sima Tan‟s] aspirations. However,
when Tan was near death, and weeping entrusted [the task to Qian], it was only
because he was full of regret and resentment at not going along to conduct the
Feng sacrifice at Mouth Tai, and Qian narrated this. Yet how could Qian not
have known the falseness of the Feng and Shan? [談生遷，能以文章世其家，揚
名後世，亦可謂善繼人之志者矣。然談垂死，涕泣之囑，唯以不得從封泰山
為恨，而遷述之。豈遷亦不知封禪之為非耶？]130

It is worth again emphasizing the warm admiration for Sima Qian and his family task

which Huang Zhen expressed. But this is an odd comment in a variety of ways. Huang

129
Recall in the “Letter to Zeng Gong,” Zhang Lei had described Sima Qian‟s tragedy as being because
“due to his righteous air, he dared to speak and leaped to his own disaster” [負氣敢言以蹈於禍] (Zhang
Lei ji, 56.844).
130
Huangshi richao 46.74B-75A.

312
Zhen would have known from the “Treatise on the Feng and Shan” 封禪書 (Shiji ch.28)

that Qian did take quite a dim view of the sacrifices Emperor Wu conducted. In short,

there is reason to believe that Qian did concur that they were without any true basis in

antiquity. Huang seems to be suggesting that Sima Qian ought to have mentioned this in

his narrative at this point. (Another possible meaning is that Qian should have mentioned

it to his father and perhaps prevented his death.) Does Huang Zhen have a valid point?

In considering this question, we must remember the differences between the age in which

Sima Qian lived and the time of Huang Zhen.

Various chapters of the Shiji suggest that the ru and the fangshi (as well as other

contenders who might be called “Legalist”) were engaged in an ongoing struggle for the

approval of the emperor. In Sima Qian‟s time, it was a live issue, an open contest that

had not been settled. The Feng sacrifice, supposedly performed by ancients, was a

skirmish that could have been won by any group. As it turned out, it was the fangshi who

carried the day. We can guess that Sima Tan was no friend of the fangshi, but nor was he

precisely a ru. It is impossible to guess whether his presence at the Feng sacrifice would

have had an effect on the proceedings, but clearly his exclusion meant that whatever

group he represented was not even in the running.

Unlike Sima Qian, Huang Zhen wrote at a time when hope for his dynasty was

fast running out. To him, the important issue was not the worldly advantage gained by

whoever won the right to direct the sacrifice, but whether the ancient sage kings had ever

performed it to begin with.131 It short, it was a question of legitimacy rather than missed

131
On this question, Huang Zhen agrees with other Southern Song figures that the evidence is
unconvincing. See, for example, opinions by Ye Shi and Wang Yinglin, as recorded in Yang Yanqi 楊燕
起, Shiji jiping 史記集評 (on SJYJJC 6.368-369).

313
opportunity. Huang Zhen would have Sima Qian pronounce outright that the Feng

sacrifice was a fake and a fraud, as it seemed to him that it was. We should not exclude

Huang Zhen‟s view merely because it was disconnected from the circumstances of Sima

Qian‟s time. His unconditional condemnation of the Feng, which seems to reflect a

consensus opinion in his time, was a precursor to scholars like Fang Bao 方苞 (1668-

1741), who read a great deal of complexity into Sima Qian‟s portrayal of the Feng, using

it to argue for ever subtler autobiographical interpretations of the Shiji.

Though there is not much depth in the passages by Huang Zhen discussed above,

he did record a more interesting comment under the heading of the “Traditions of the

Harsh Officials” 酷吏列傳, Shiji ch.122. Here, Huang analyzes in detail the subtle

strategies employed by Sima Qian, not just in his portrayal of the harsh officials, but in a

number of different chapters. He points to specific juxtapositions which show Sima

Qian‟s deliberate intention to highlight Emperor Wu‟s errors.

Huang unexpectedly began his reading, not with the harsh officials as we might

expect, but with the way the Xiongnu are treated in the Shiji:

Regarding Emperor Wu‟s military affairs, the Honorable Senior Archivist

precedes them with [a narrative about] the marriage treaties of [Emperors] Wen
and Jing, and how the Xiongnu trusted the Han. After that, he discusses the two
generals going out to the border in successive years, and also invariably follows it
with the Xiongnu crossing the border and how many people they killed and
captured. [太史公於武帝征伐事，先之以文景和親，匈奴信漢，然後論兩將
軍連年出塞，又必隨之以匈奴入塞，殺略若干。]132

Describing the marriage treaties, which kept peace between the Han and the Xiongnu, is a

way of drawing a contrast between foreign relations under Emperors Wen and Jing, and

the more aggressive policy which Emperor Wu pursued. The two generals were Wei

132
Huangshi richao 46.67B-68A.

314
Qing and Huo Qubing, and according to Huang Zhen, Sima Qian deliberately juxtaposed

the record of their campaigns and the casualties inflicted by Xiongnu retaliation. The

implicit conclusion would be that Emperor Wu‟s policy (of attempting to force the

Xiongnu into subjugation instead of maintaining the long-standing treaties of his

predecessors) was unwise and only caused the people suffering.

Huang Zhen then moved to the “Harsh Officials” chapter. He again remarked

upon the structure of the narration, and explained the connection with the Xiongnu:

Now in this „Traditions of the Harsh Officials,‟ [Sima Qian] begins by [saying
that] because “the officials governed honestly,”133 “the people were simple and
afraid to commit crimes.”134 After that, he discusses how the ten harsh officials
take control of affairs each in succession, and also inevitably follows by [saying]
how “the people increasingly flouted the law, and how thieves and brigands
multiplied and rose up.”135 That being so, the „transformations‟ in the Xiongnu,
and in the thieves and brigands, were brought about by the emperor‟s exhausting
the troops and instituting harsh penalties. How can displays of force and
punishments not continually increase?! [於今《酷吏傳》，先之以吏治烝烝，
民朴畏罪，然後論十酷吏更迭用事，又必隨之以民益犯法，盜賊滋起。然則
匈奴盜賊之變，此帝窮兵酷罰致之，威刑豈徒無益而已哉！]

In Huang Zhen‟s analysis, the Shiji account is designed to show that Emperor Wu was

personally to blame for both the domestic and foreign difficulties that occurred during his

reign: his tendency toward displays of force and punishments actually led to potential

enemies (both the Xiongnu and domestic troublemakers) having less respect for the

government, requiring him to continually increase his commitment in both cases.

133
A direct quote from the Shiji, SJ 122.3131.
134
SJ 122.3133. This statement actually occurs in the section devoted to the first “harsh official”, Zhi Du,
explaining the state of the people during the time he was in power.
135
SJ 122.3151. Although Huang Zhen quotes this phrase directly (so I have translated 必 as “inevitably”
rather than “invariably”), there are a number of other situations in which an account of a harsh official‟s
career is followed by a brief description of the resultant bad effects on the people. See, for example, the
statements that under Zhang Tang‟s rule “the common people did not find peace in their existence” [百姓
不安其生] (SJ 122.3140); that under Yi Zong “throughout the province, people trembled even when it was
not cold, and cunning, treacherous people assisted officials in governing” [郡中不寒而栗，猾民佐吏為治]
(SJ 122.3146). The conditions quoted above are said to have arisen because everyone in authority began
imitating the governing strategies of Wang Wenshu.

315
 Huang Zhen then shifts to a third area of Emperor Wu‟s behavior, namely, his

actions in the religious sphere:

When it comes to having a hundred different manifestations of worshipping and

sacrificing, each time it is followed with words about the seeming phenomenon of
a cosmic resonance. As for seeking immortality but being without a method, then
each time it is followed with words about how in the end they could not get it.
The way [Sima] Qian‟s subtle writing shows his intention is always like this, and
Emperor Wu‟s being without the Way is glaringly obvious indeed. [至於禱祠百
出，則各隨之以若有符應之言；於求仙無方，則各隨之以終不可得之言。遷
之微文見意，往往如此，而武帝之無道昭昭矣。]136

Huang Zhen is here referring to the “Treatise on the Feng and Shan Sacrifices” 封禪書

(Shiji ch.28). In that chapter, there are two distinct spheres of activity, one which has to

do with “legitimate” rituals, and another which involves Emperor Wu‟s fruitless

obsession with immortality. It may not always be easy for the modern reader to

distinguish them, but as Huang Zhen points out, the “Feng and Shan” chapter seems laid

out to provide intentional contrast, making the ritualists (the Honorable Senior Archivist

mentioned several times as being among them) succeed and the magicians (fangshi 方士)

fail time and time again. Yet the emperor never ceased being taken in by the magicians,

time after time.

Scholars like Qin Guan and Zhang Lei in the Northern Song dynasty changed the

way later readers would see Sima Qian‟s tragedy and its relationship with the Shiji. For

pre-Song readers, the terms of the debate were still heavily influenced by Ban Gu‟s

critique, and had focused primarily on whether Sima Qian was merely telling the truth or

whether he was defaming his emperor. In the Song, under the influence of Su Shi‟s circle,

new considerations arose. Sima Qian‟s life-story was romanticized and linked to his

136
Huangshi richao 46.68a.

316
prose style, while the potentially political satire contained in the Shiji was psychologized

and seen as emotionally linked to the Li Ling tragedy. For Su Shi‟s group, at least, Sima

Qian began to be viewed not just as a historian but also as an artist. The fact that these

Northern Song figures had themselves run afoul of political persecution undoubtedly

influenced and motivated their interpretations of Sima Qian.

In the Southern Song, Sima Qian‟s tragedy was placed in a quite different context.

The dynastic crisis (defeat by the Jurchens and the subsequent southward relocation) had

led to the reign of Song Gaozong, a highly autocratic emperor motivated more by

realpolitik than by moral considerations. With the elevation of Qin Gui to the post of

preeminent councilor, expressions of disagreement with official policy were discouraged

or even forbidden. In this context, Zhou Zizhi, Wang Guanguo, and Lü Zuqian all re-

analyzed the Li Ling affair as an historical event, reversing the tendency of Sima Qian‟s

critics to portray Sima Qian as biased and resentful for personal reasons. Each instead

portrayed Sima Qian as essentially blameless or even praiseworthy. Meanwhile, it was

Emperor Wu whom they criticized for exactly the same fault: allowing his personal

feelings to interfere with the proper business of government. The autocratic government

of Song Gaozong and his surrogates surely had some influence on these interpretations.

Despite this tendency to shift blame from the historian to the emperor, moralists

like Zhu Xi and Wang Ruoxu continued to follow the line of Ban Gu‟s criticism,

condemning Sima Qian for his failure to accord with the Classics and the Sage, for his

idiosyncratic choices as regards form, and for allowing personal resentment to interfere

with his task. I have ended with Huang Zhen because on the one hand he serves as an

accurate summary of Song dynasty views (sympathy and admiration for Sima Qian,

317
combined with a slight tendency to criticize his failure to accord with Classical morality,

and condemnation of Emperor Wu); on the other hand, with his focus on specific

techniques and multi-chapter interpretations, he foreshadows the late Ming style of Shiji

interpretation.

318
 Chapter 6

A ‘True Record’: The Rhetoric of Reliability

There are a number of things we could mean by reliability in the historical context.

For a Western historian, one thing that might come to mind is Leopold von Ranke‟s

dictum that a historian should write about the past wie es eigentlich gewesen (“how it

actually was”). This has been taken to mean something like objectively, factually, or

scientifically, though even for Ranke, the notion had some component of theological

commitment: Ranke seems to have believed “that history provides the locus where God is

witnessed,” that “the finger of God” is discernible in history, and that the task of the

historian is to “decipher its divinely guaranteed coherence.”1

Charles Beard, criticizing this approach to history, attempted to tease out the

problematic assumptions it involved, including the notion “that history… has existed as

an object or series of objects outside the mind of the historian” and “that the historian can

face and know this object or series of objects and can describe it as it objectively

existed…divest[ing] himself of all taint of religious, political, philosophical, social, sex,

economic, moral, and aesthetic interests.” Furthermore, and more interesting as a

comparison to the Chinese case, Beard suggested that there was also an assumption “that

the multitudinous events of history as actuality had some structural organization through

inner (perhaps causal) relations, which the impartial historian can grasp by inquiry and

observation and accurately reproduce or describe in written history.”2 As Elizabeth Clark

describes, this “objectivist creed” among twentieth century historians led to “a quest for

scientific history”, a “cult of Research.” In the words of Hayden White, “the idea was to

1
Elizabeth Clark, History, Theory, Text, 202 n.46-47.
2
From a 1935 essay, “The Noble Dream” cited on Clark, 14.

319
let the explanation emerge naturally from the documents themselves, and then figure its

meaning in story form.”3

There are a number of potential analogues in the early Chinese historical context;

certain parts of the Shiji are especially suggestive. The “Self-Narration,” for example,

contains a quotation attributed to Confucius and associated with the Spring and Autumn:

“I would record it as empty words, but that is not as profound, incisive, clear, or

enlightening as showing it in actions and events” [我欲載之空言，不如見之於行事之

深切著明也].4 The it which is the object of this recording or showing can be tied both to

“history as actuality” and a profound “structural organization” that it is understood to

possess. Could it be that for Sima Qian, as for Ranke, there was an underlying moral

truth that would emerge naturally from a careful examination of available documents?

Again, the Shiji itself seems to suggest that that is so. At times Sima Qian looks

very much like a proto-scientific historian. As Wai-yee Li has pointed out, with regard to

certain passages in the Shiji:

For the first time in the Chinese tradition, the historian informs us about how he
chooses his sources, visits historical sites, confronts representations of historical
figures, and tells of his encounters or personal associations with historical figures
or their descendants or associates.... He also tells how he personally participated
in key historical events.5

For each of these aspects, Li cites key passages from the Shiji which seem to display

these sorts of social-scientific tendencies. The question, though, is: does a selection like

Li‟s say more about modern preoccupations than it does about the Shiji? Readers of the

present time, responding to the strong scientific bias of a modern era, want badly to get
3
Cited in Clark, 15.
4
SJ 130.3298. This saying is not found in the Lunyu, but rather only in Dong Zhongshu‟s 董仲舒 (179-
104 BCE) Chunqiu fanlu 春秋繁露 [Abundant dew of the Spring and Autumn]. See Han Zhaoqi, Shiji
jianzheng 9.6358, n.10.
5
“The Idea of Authority”, 377.

320
some grasp on the ancient Chinese past as “an object or series of objects” that

“objectively existed” and can be described as such. This longing finds resonance in

passages from the Shiji which look most closely scientific.

It is perhaps more true to the Shiji (as a textual object independent of our desires),

however to look at what it was and how it was perceived to be before it came into our

grasping hands. In short, I will not decide what scientific history looks like and then go

looking for what in the Shiji might correspond to that. Instead of studying the Shiji‟s

most apparently scientific passages, I examine the discourse of reliability within the

Chinese context, especially as it relates to the Shiji. I look for the kinds of arguments and

terms early readers actually applied to the Shiji when they were concerned with questions

of reliability. One of the most important of these is the term shilu 實錄, whose historical

development I discuss in detail below.

I then turn to early readings of the Shiji by Chu Shaosun, Huan Tan, and Wang

Chong. To them, the “trueness” of the Shiji‟s record was not yet authoritatively

established, leaving considerable room for argument and interpretation. In the Six

Dynasties, the focus of debate was on the issue of whether or not the Shiji was a

defamatory text. Calling the Shiji a true record enabled its defenders to justify not only

Sima Qian‟s honest criticism, but potentially their own as well. In the Tang, the problem

is not so much what constitutes a “true record” but the practical political difficulty of

producing one. The trueness of a record was most at issue when that record was

potentially displeasing to those in power. Finally, I discuss two issues regarding the Song

view of the Shiji‟s reliability: I begin with Su Xun‟s “Discussion of History,” which

321
elevated the Shiji and Hanshu almost to the status of Classics. I then turn to Zhedong

school, which hoped to use the Shiji and other historical texts as criteria for pragmatic

truth—a practical guide to what government should look like. Zhu Xi, whose first loyalty

was firmly with the Classics, in reacting against this threatening trend by attacking the

Shiji and the notion of histories generally with particular vigor.

ON THE TERM „TRUE RECORD‟

In chapter 4, I introduced the fundamental disagreement regarding Shiji

interpretation. On the one hand, Sima Qian was accused of contradicting the Classics and

the Sage, as well as slandering his dynasty, because of his personal feelings. On the other

hand, he was praised a “good archivist” 良史 who succeeded in creating a “true record”

shilu 實錄. Ban Gu, one of the most influential interpreters of the Shiji, seems to have

held both opinions at once. This created a problem both for later readers of the Shiji and,

on a deeper level, for those who hope to understand what it meant for something to be a

“true record” shilu 實錄. In order to clarify this issue and open a general discussion of

the traditional Chinese rhetoric of reliability as it relates to authorship of historical texts, I

will begin by considering in depth the problem of the term “true record” shilu 實錄.

The term shilu in a sense encapsulates and represents the multivalent nature of the

entire issue of reliability in historical texts. It was first used by Yang Xiong 揚雄 (53

BCE-18 CE), whose few comments on the Shiji became a permanent part of the text‟s

reputation. As the brief and archaic terms which Yang used to describe the Shiji were

repeated over the centuries, later writers frequently read their own ideas into them. The

322
following comment Yang made about the Shiji is the one which bears most directly on

the issue of reliability:

Someone asked, “The Offices of Zhou?” [I] said, “They establish procedures.”
“Master Zuo?” [I] said, “It appraises and evaluates.” “Senior Archivist Qian?” [I]
said, “A true record.” [或問「周官」。曰：「立事。」「左氏」。曰：「品
藻。」「太史遷」。曰：「實錄。」]6

In chapter 4 above, I discussed what this passage tells us about how the Shiji was

contextualized during Yang Xiong‟s time. Here I want to focus on the meaning of the

term shilu (實錄) itself.

Yang Xiong himself offered no further clarification of the term, nor did he use it

elsewhere in his works. However, Ban Gu in his evaluation at the end of the “Arrayed

Traditions of Sima Qian” 司馬遷列傳 (HS ch.62) offered what can be understood as a

gloss on, or definition of, the term shilu:

Since Liu Xiang and Yang Xiong were extremely knowledgeable about the
multitude of books, they all praised Qian as a having the talent of a good archivist,
and testified that he was excellent at narrating events and their causes, that he
made arguments without being flowery, that he was substantial but not unpolished,
that the writing was direct and the events relevant, that he did not emptily beautify
nor covertly vilify, and therefore they called [his work] a „true record.‟ [然自劉
向﹑楊雄博極群書，皆稱遷有良史之材，服其善序事理，辯而不華，質而不
俚，其文直，其事核，不虛美，不隱惡，故謂之實錄。]7

Clearly Ban Gu was aware of Yang Xiong‟s judgment on the Shiji (as seen in the Fayan).

Ban‟s expansion of that judgment does not refer only to Yang Xiong‟s opinion however.

Instead it should probably be considered a synthesis of different views, including not only

those of Liu Xiang and Liu Xin,8 but also Ban Gu‟s father Ban Biao, and probably others

whom Ban Gu did not specify. Still, from a historical point of view, this Hanshu
6
FY 10.413.
7
HS 62.2738.
8
It seems likely that whatever Liu Xiang wrote about Sima Qian, to which Ban Gu here alludes, would
have appeared in the lost bibliographic catalogue he compiled, the “Separate Records” 別錄.

323
comment on the phrase shilu played a deeply influential role in determining its

subsequent meaning, and is worth analyzing in detail. The characterization cited above

and summed up as amounting to a „true record‟ has three main aspects. The first is that

Sima Qian‟s work allows his readers to understand shili 事理, what I have translated

above as “events and their causes.” It is perhaps more literally rendered “the patterns

within the events”,9 and suggests that the Shiji goes beyond a mere factual rendering and

offers an understanding of the patterns behind history, a deeper truth. Ban Gu‟s

contemporary, Xu Shen 許慎 (1st c. CE), understood shi 實 to derive from wealth, strings

of cowry shells (guan 貫) under one‟s roof.10 Extensionally, there is a sense of

“richness” in a historical account that comprehends the underlying pattern as well as the

surface facts.

The second aspect of Ban Gu‟s definition is stylistic. Three different descriptive

terms in the passage above can be understood as referring to an essential characteristic of

Sima Qian‟s prose: it is “not flowery” 不華, it is “substantial” 質, and it is “concise” 核

in its narration of events. It might seem odd to find style being used as part of a criterion

for “truth”, but it should be kept in mind that semantic categories are divided up

differently in classical Chinese than they are in English. One meaning of shi 實—the

word I have been translating as “true” or “veritable”—is “the fruit or seed or a plant”, in

explicit contrast to its flower. Thus the description “not flowery” 不華 is almost a

paraphrase for “true” 實—if the opposite of “true” can be understood not as “false”, but

as insubstantial, peripheral, not yet fully matured. Similarly, the word I translate as

9
The earliest meaning of li 理 is the pattern of veins in jade, and the profound philosophical significance
that the word gradually acquired can be seen as being extensional from this meaning.
10
Shuowen jiezi 7B.150.

324
“concise” 核 literally means “kernel” or “nucleus,” very close in meaning to the sense of

shi 實 that refers to the core or seed. Regarding “substantial” 質, Kongzi referred to

“native substance” zhi 質 as a necessary complement to “acquired refinement” wen 文.11

This substance/form dichotomy is a close parallel to the contrast of shi 實 and hua 華. In

short, the aesthetic quality associated with truth is not beauty but mature, concise

substantiality.12

The final aspect of Ban Gu‟s “gloss” is associated with a historian‟s personal

moral sense, and was the most historically important. It includes the assertion that Sima

Qian‟s “writing is direct” [其文直]13 and that “he does not emptily beautify nor covertly

vilify” (不虛美，不隱惡). Commentators on the Fayan select out merely this last line

for their own gloss on Yang Xiong‟s use of shilu.14 Emptiness, xu 虛 (like floweriness) is

an antonym for shi 實, which in contrast represents that which is full, rich, substantial.

Empty beautification is like floweriness but with a more sinister purpose—flattery at the

expense of truth. Thus Sima Qian‟s record is a true one because it does not hand out

undeserved praise. The second part of the description is far more problematic, both as

part of a definition for shi 實 and as a description of the Shiji. As discussed in chapter 4

above, Ban Gu himself accused Sima Qian of using “subtle writing and piercing satire” to

11
Lunyu VI:18, trans. Lau, Analects, 83.
12
Yang Xiong‟s use of the term shilu has been explored in detail by Kai Vogelsang in “Historical
Judgement” 153ff.; his discussion has greatly informed my own.
13
The quality of directness, whose opposite is deviousness (qu 曲), I place in this third category rather than
the previous one because this is the way it would come to be understood. See discussion below.
14
This applies not only to modern commentators. See discussion of Wang Su 王肅 (below).

325
“denigrate and detract from his own generation” [微文刺譏，貶損當世].15 Certainly

this sounds like another way of describing “covert vilification.”

Rather than continuing to delve into these conundrums and contradictions, I

believe it is more useful to turn instead to subsequent interpretations of the phrase. One,

that of Wang Su as portrayed in the Sanguo zhi, is essentially a paraphrase and

condensation of Ban Gu‟s evaluation cited above.16 Also as discussed in chapter 4,

Sanguo zhi commentator Pei Songzhi used similar terms to defend Sima Qian against

charges of defamation, though he does not use the exact phrase shilu: “Qian did not

conceal (yin 隱) the faults of Filial Wu, and instead wrote a direct (zhi 直) account of his

affairs” [但遷為不隱孝武之失，直書其事耳].17

Another example, which might shed more light on the problem, is also drawn

from Pei Songzhi‟s commentary, though in this case he was evaluating Chen Shou‟s

historical practice as opposed to Sima Qian‟s. In narrating the dramatic battles at the

beginning of the Three Kingdoms period, the Sanguo zhi at one point states that: “at this

time, the Lord [Cao Cao] had less than ten thousand troops, and two or three out of every

ten were wounded” [時公兵不滿萬，傷者十二三].18 Pei Songzhi objected, writing:

I, Songzhi, consider that when [Emperor] Wu of Wei [i.e., Cao Cao] first raised
troops they were already a massive five thousand. Afterward, there were a great
number of battles and of victories. Those that were lost could only have been two

15
“Dian yin” WX 48.2158. Of course, the context of the “Dian yin” was very different, and this difference
must not be ignored.
16
This passage will be discussed below. The relevant portion of Wang Su‟s speech is as follows: “Sima
Qian recorded events without emptily beautifying or covertly vilifying. Liu Xiang and Yang Xiong both
admitted that he was very good at narrating events and had the talent of a good archivist, calling [his work]
a true record.” [司馬遷記事，不虛美，不隱惡．劉向﹑揚雄服其善敘事，有良史之才，謂之實錄]
(SGZ 13.418). It may be that the nature of the rearrangement carries some significance, but any attempt to
tease it out would likely be too tenuous.
17
SGZ 6.180.
18
SGZ 1.19-20.

326
 or three out of every ten.19 However, when [the Emperor] defeated the Yellow
Turbans, he received the surrender of more than three hundred thousand soldiers.
The number which his army absorbed cannot be recorded precisely. Thus,
although he fought battles and had losses and casualties, the actual size of his
army could not have been this small…. The person who transmitted the record
wanted to use the small number of troops to make it seem most wondrous. This is
not a true record of it.20 [臣松之以為魏武初起兵，已有衆五千，自後百戰百
勝，敗者十二三而已矣．但一破黃巾，受降卒三十餘萬，餘所吞并，不可悉
紀；雖征戰損傷，未應如此之少也。。。將記述者欲以少見奇，非其實錄
也．]21

Pei Songzhi‟s criticism focuses on two major aspects. The first is the inaccuracy of the

number reported. The second is the motive behind it: by exaggerating the odds stacked

against Cao Cao, the narrative makes his victory appear more amazing than it really

was—an example of “empty beautification,” or more generally, politically motivated

distortion of facts.22

Other illustrations of what it meant not to be a shilu 實錄 allow us to explore the

range of that critique. Some examples suggest less explicitly the same kind of politically

motivated distortions Pei Songzhi complained about. For example, in the “self-narration”

of the Songshu, Shen Yue 沈約 (441-513) criticized previously-compiled histories of the

Liu Song 劉宋 dynasty (420-479) for a variety of faults, including the fact that “events

which belonged to [the compilers‟] own times contained much that was not a true record”

19
It is not entirely clear whether Pei Songzhi means 20-30 percent of the battles or of the troops, but given
that the figure echoes the one in the Sanguo zhi, one might suppose the latter.
20
Or depending on how one interprets qi 其: “This is not his true recording” (where “it” refers to the event
in question).
21
SGZ 1.20.
22
An interesting Western parallel to this situation can be found in accounts of the Battle of Agincourt
(October, 1415). As historian Anne Curry has noted, a key perception about the Battle of Agincourt is that
it was a “victory of the few against the many” (Agincourt, 10). Yet Curry cautions that even modern
commentators fall prey to the desire for a good story just as Chen Shou did. Like Pei Songzhi, Curry
considers that the requirements of a „true record‟ involve careful reasoning about actual numbers involved.
In the case of Agincourt, “since the armies of both [France and England] were paid, we can ascertain their
size, composition and even names of soldiers, thereby liberating ourselves from over-reliance on narrative
accounts” (Agincourt, 13). Certainly in the case of Agincourt, the numbers suggest that it was not as
unequal a contest as the narrative accounts would have us believe.

327
[事屬當時，多非實錄].23 Shen Yue delicately allows readers to draw their own

conclusions about why historians‟ records of their own time might not be trustworthy, but

political self-interest does seem a likely explanation.

A more involved example can be found in the story of Wu Jun‟s 吳均 (469-520)

ill-fated project of compiling a Qi Chunqiu 齊春秋 (Spring and Autumn annals of the Qi)

for the recently deposed Southern Qi dynasty (479-502). Both the Liangshu and the

Nanshi mention this, but the the Nanshi gives greater detail:

Since [Wu] Jun had made for himself a reputation as a historian, he wanted to
compile a Qishu [History of the Qi]. He sought to borrow the imperial Diaries of
Activity and Repose and the posthumous Records of Conduct for the various
ministers. Emperor Wu did not permit it. Thereupon, [Jun] privately compiled a
Qi Chunqiu and submitted it to the throne. His text stated that the Emperor24 had
assisted Emperor Ming [r.494-498] of Qi in his succession to the throne. The
Emperor hated his “true record”, considering his writing to be untrue. He sent the
Secretariat Drafter Liu Zhilin to interrogate him on several dozen items, and it
turned out that he was incoherent and was unable to answer [satisfactorily]. By
imperial edict [Wu Jun] was ordered to hand it over to the provincial authorities,
who burned it, and [Wu Jun] was relieved of office. [均將著史以自名，欲撰齊
書，求借齊起居注及群臣行狀，武帝不許，遂私撰齊春秋奏之．書稱帝為齊
明帝佐命，帝惡其實錄，以其書不實，使中書舍人劉之遴詰問數十條，竟支
離無對．敕付省焚之，坐免職．]25

In the days of the Qi Emperor Ming, Emperor Wu of the Liang [r.502-549]—the Liang

dynastic founder—had been merely Xiao Yan 蕭衍, a distant cousin of the Qi ruling

family. He would have personally witnessed the Emperor Ming‟s succession.

Presumably it would have been accurate and defensible to say that he “assisted” in the

succession, at least in some sense. But this is to emphasize the Liang Emperor Wu‟s

treachery in later turning against heirs of a ruler he had previously supported. The reason

23
Songshu 100.2467.
24
That is, Emperor Wu of the Liang.
25
NS 72.1781. C.f. Liangshu 49.698-99, which closely parallels the passage quoted above but omits the
detail about Wu Jun‟s historiographic treatment of Emperor Ming.

328
for Emperor Wu‟s opprobrium is therefore no mystery, and it is unsurprising that Wu Jun

was incoherent when questioned about his portrayal.26 The extent to which Wu Jun‟s

comments were factually true is unclear; however, the political motivation behind them—

a jab at the legitimacy of the Liang and at Emperor Wu‟s moral character—were certainly

a case of deliberate vilification.

In this understanding of what was not a shilu 實錄, we can see a reflection of the

conflict, introduced in chapter 4 and also discussed at length below, over the Shiji‟s

portrayal of the Han. Did Sima Qian‟s resentment lead him to distort the facts? Or was

he merely supplying an accurate portrait? Could he have used the true facts to produce a

distorted portrait? Contemporary theories about history tend to rule out a “view from

nowhere,”27 suggesting that all accounts are in some sense distorted by the historian‟s

perspective. Chinese historians, however, had not given up hope as to the existence of a

proper objective standard which could provide the basis for a common perspective that

everyone could share and agree on. Assisted by this underlying assumption, the ideal of

shilu 實錄 remained in place, even though it was not always clear how to reach it or

judge it.

The standard histories for the Six Dynasties period, mostly compiled in the Sui

and early Tang28 using pre-existing materials (and in some case pre-existing

compilations), show that the term shilu 實錄 developed two other quite distinct meanings

26
It is interesting to note that the Nan Qi shu we have today was compiled under the leadership of Xiao
Zixian 蕭子顯 (587-537), also a grandson of Emperor Gao of the Qi. Xiao Zixian was fifteen years old at
the official founding of the Liang.
27
The works of Hayden White, to name one example. For a relatively optimistic discussion about attaining
a “view from nowhere”, at least in some contexts, see Thomas Nagel‟s 1989 book of that title.
28
Of the official histories for the Six Dynasties, the Songshu 宋書 and Nan Qi shu 南齊書 were compiled
during the Liang 梁 (502-557) dynasty, and the Weishu 魏書 was compiled under the auspices of the
Northern Qi 北齊 (550-577) dynasty; the rest were completed in the Sui or Tang.

329
during the Six Dynasties period. First, it began to be used as the term for a genre of

official record-keeping. Dennis Twitchett, who argues that the “Veritable Records”

(shilu 實錄) as a full-fledged genre “were in fact a Tang innovation”29 nonetheless lists

several Six Dynasties works which contained shilu 實錄 in their titles, including the

Dunhuang shilu 敦煌實錄 [Veritable records from Dunhuang] by Liu Bing 劉昞,30 a

Liang Huangdi shilu 梁皇帝實錄 [Veritable records of the Liang emperor] by Zhou

Xingsi 周興嗣 (d.521), and another work of the same title by Xie Wu. As Twitchett

discusses in great detail, a “Veritable Record” was a kind of first draft dynastic history,

an intermediate stage between the Qi ju zhu 起居注 [Diaries of activity and repose]

recorded by court diarists, and the larger project of the dynastic history as a whole, which

was generally compiled after the fall of the dynasty. He also emphasizes that “the

compilation of a Veritable Record was not simple a routine re-writing of the official

record but was in every case a major and deliberate political act.”31

The fact that the term shilu 實錄, which was first and foremost associated with

Sima Qian‟s Shiji, became the name of an official (and controversial) genre shows the

official acceptance and/or co-opting of works like the Shiji—and of what was seen, at

least by some, as the Shiji‟s underlying motivations. It was desirable and instructive to

have a compilation of this nature. However, as Twitchett describes, emperors began

more and more to feel the need to control what was written about them.32

29
Twitchett, Official History, 119.
30
Or perhaps Liu Jing 劉景; see Twitchett, Official History, 119 nt.1.
31
Twitchett, Official History, 120.
32
Ibid., 119-159.

330
 A second context in which we find the term shilu 實錄 being used during the

same time was anything but official. In the realm of personal historiography, the term

shilu 實錄 was applied (as a term of praise) to a description of a person that was

particularly vivid and revealing. One interesting example of this usage applied to Tao

Qian‟s autobiographical statement, as recorded in the Jinshu:

[Tao Qian] once wrote the “Story of Master Five Willows” as a conceit for
himself. It said: [嘗著五柳先生傳以自況曰：]

No one knows where the master came from, nor was his surname or
courtesy name known for certain. Near his studio there were five willow
trees, and thus he took that for his style. He was withdrawn and quiet, not
much given to speech. He did not long for glory or profit. He was fond of
reading books, but did not pursue any elaborate exegesis. Whenever he
had an insight, he would be so happy that he forgot to eat. By nature he
had a special liking for drink, but his household was impoverished and he
could not always get it. His relatives and closest friends knew it was like
this, and would sometimes buy strong drink and invite him over.
Whatever he started drinking, he was sure to drink it the dregs. After a
time, he was certain to become drunk. When, being drunk, he had to retire,
he never hesitated about expressing it. He lived in a tiny hovel which was
barely furnished; he went unsheltered from wind or sun, and wore a short
robe that was ragged and much mended. His bowl and gourd often went
empty, but he was at peace with himself. He was always writing little
pieces to amuse himself, and also in some sense to express his aspirations.
He could dismiss success and failure from his mind, and did so until the
natural end of his days. [先生不知何許人，不詳姓字，宅邊有五柳樹，
因以為號焉．閑靜少言，不慕榮利．好讀書，不求甚解，每有會意，
欣然忘食．性嗜酒，而家貧不能恒得．親舊知其如此，或置酒招之，
造飲必盡，期在必醉，既醉而退，曾不吝情．環堵蕭然，不蔽風日，
短褐穿結，簞瓢屢空，晏如也．常著文章自娛，頗示己志，忘懷得
失，以此自終．]33

[Tao Qian‟s] self-narration being like this, people of the time considered it a true
record. [其自序如此，時人謂之實錄．]34

33
See also Yuan Xingpei‟s annotated version of “The Story of Master Five Willows” in Tao Yuanming ji
jianzhu, 502-507.
34
JS 94.2460-2461.

331
“The Story of Master Five Willows” adopts an explicitly fictional pose in its first

sentence, one which disavows any specific information about the master‟s identity. On

the other hand, it quickly becomes clear that Tao Qian‟s description of Master Five

Willows is a true portrait of an ideal, an expression of the author‟s own unconventional

aspirations. The relevant question for the current study is, when the historian35

commented that people of the time considered the piece a true record, in what sense was

the term being understood?

There are several possibilities, none mutually exclusive. First, it could mean that

despite the fictional pose, Tao Qian‟s contemporaries recognized that Tao was writing

about himself: his portrait of Master Five Willows was true of him. Second, it could

mean that people saw in this portrait a deeper truth—about happiness, life in the world,

and so forth. The record was not only a true portrait of its author but true in its

implication that being happy does not require material possessions or worldly success.

Third, it could simply be a truth claim on the part of the historian, who in introducing

“The Story of Master Five Willows” stated that it was autobiographical, but then had

somehow to explain the fictional pose. Of course the story says that the man‟s name is

unknown, the compiler implies, but everyone at the time knew that the piece was really

autobiographical and that is why I am using it as such. Finally, in a more skeptical vein,

it is important to consider that Tao Qian‟s portrait of Master Five Willows was more true

to the image Tao Qian wanted to project than to the actual reality of his life. In this

connection, we might consider the difference between zikuang 自況—something like a

35
Probably Fang Xuanling 房玄齡 (579-648), lead compiler of the Jinshu, although it is quite possible that
he was copying earlier material.

332
comparison of oneself to another—and zixu 自序, which I have translated as “self-

narration.”

I have considered this case so closely because the slippage between

autobiographical self and fictionalized other is a key aspect of Shiji interpretation.

Though I have found little evidence that the Shiji was read this way in the Six Dynasties

or Tang periods, it seems important that the term shilu had begun to develop this meaning

when applied to other texts. There being a continued close association between the Shiji

and the term shilu, gradual transformation in the meaning of the term potentially had an

effect on its meaning with regard to the Shiji.

Another interesting example of a seemingly personal shilu 實錄 is found in the

biography of Bian Bin 卞彬 (d.500) in the Nan Qi shu 南齊書. Bian Bin, who held a

variety of official posts in the Liu-Song dynasty (420-479) and then the Southern Qi

(479-502), was described as being “singular in talent and conduct” [才操不羣]. The

historian also claimed that “his writings contained a great deal of satirical meaning” [文

多指刺].36 Among other examples, the historian included the preface to Bian Bin‟s (now

lost) “Rhapsody on Fleas and Lice” 《蚤虱賦》:

I live in poverty, wearing cotton clothes for ten years without having new ones
remade. The coarse fabric of a single robe is all one needs to survive. It provided
for me in both times of cold and of heat, for I had not another with which to
exchange it. I was one with a sickly constitution, very earnest in all my
undertakings, and took my repose in a nest of waste cotton, being unable to ease
myself.37 My basic nature is entirely lazy. I am especially lazy in taking care of
my skin; nor am I meticulous in bathing and scrubbing. I miss the regular time
for washing my hair. My four limbs are shaggy and disheveled.38 Furthermore, I

36
NQS 52.982.
37
The phrase I have translated as “ease myself” is equally interpretable as “explain myself,” a double
meaning whose presence is clearly deliberate.
38
The binome seems to apply generally to feathers or fur.

333
 am stinking and filthy. Therefore, in the ruffled tassels of my reed mat, the fleas
and lice move about in great numbers. The overflow of itching spreads and
sprawls; they are continually intruding upon my flesh. I root about in my clothes,
hoping to seize or capture them, but I cannot get a hand on them. Lice have a
proverb which says, “Born at dawn, have grandsons by evening.” Furthermore,
lice like mine do not have to worry about [the dangers] of the bath and are
untroubled by the sorrow of mourning for one another. They gather and feast in
an ancient jacket and underclothes of rotting cotton, clothing which is never
changed. Since it is impossible [for me] to scoop out the biters, they become
unrestrained, slow, and lazy. They no longer make preparations for attacks or
incursions, but merely multiply generation upon generation, and have been doing
so for thirty-five years. [余居貧，布衣十年不制．一袍之縕，有生所託，資其
寒暑，無與易之．為人多病，起居甚真，縈寢敗絮，不能自釋．兼攝性懈
惰，嬾事皮膚，澡刷不謹，澣沐失時，四體寉寉，加以臭穢，故葦席蓬纓之
間，蚤虱猥流．淫癢渭濩，無時恕肉，探揣擭撮，日不替手．虱有諺言，朝
生暮孫．若吾之虱者，無湯沐之慮，絕相弔之憂，宴聚乎久襟爛布之裳，服
無改換，搯齧不能加，脫略緩嬾，復不懃於捕討，孫孫息息，三十五歲
焉．]39

The historian adds that “these brief words were entirely a true record” [其略言皆實錄

也].40 Again, we have to ask, in what sense is this a true record? It would clearly be too

simple-minded to suggest that the historian considers it a true confession of Bian Bin‟s

questionable hygiene practices. Given the prevalence of satire in other samples of Bian

Bin‟s writing, it is quite reasonable to suppose that the lice are meant to be understood

symbolically. Certainly Li Yanshou 李延壽 (fl.650) read it this way. In compiling the

Nanshi biography of Bian Bin, he also included this passage, but introduced it by

commenting, “[Bian Bin] was dissatisfied with his official career, and so made

rhapsodies on fleas and lice, snails, toads, and so forth. All of these contain a great deal

of reproach and denunciation” [仕既不遂，乃著蚤蝨、蝸蟲、蝦蟆等賦，皆大有指

斥].41

39
NQS 52.892-893. Cf. NS 72.1676.
40
Ibid.
41
NS 72.1767.

334
 How does this affect our understanding of the truth contained in a “true record”?

We must conclude that a symbolic or allegorical truth could also fall within its province.

Furthermore, this usage does not match Ban Gu‟s gloss—“not covertly vilifying”—since

covertly vilifying is exactly what Bian Bin is doing. Again, there is no indication that

this sense of the term shilu was yet being directly applied to the Shiji, but, as discussed

above, the changed understanding of this term would gradually begin to affect the

understanding of the Shiji as well.

Liu Zhiji 劉知幾, reflecting upon the whole of the historical tradition up to his

day, used shilu primarily in the sense of factual correctness. For example, he complained

about the treatment of place names:

When the author [of a history] makes for a person a „Traditions‟ chapter, if he
records where they are from, he will invariably use the ancient name for that place,
applying it to the present [as if the name were still in use]. If one wants to seek
the true record, is it not indeed difficult!? [作者為人立傳，每云某所人也，其
地皆取舊號，施之於今。欲求實録，不亦難乎]42

The last two phrases represent the most common context in which Liu Zhiji uses the

word shilu 實錄. It might be slightly ambiguous whether it is the writer of the history

who finds it difficult to create a true record, or whether it is the reader who is frustrated in

his efforts to gain access to the true information.43 The historian‟s motive in using

antiquated place names—perhaps a matter of tradition, or simply convenience?—is not

entirely clear, but Liu Zhiji‟s objection is. If place names cannot be matched to actual

42
STTS 19.144.
43
Because the verb is “to seek” 求, and because Liu Zhiji writes more from a critic‟s perspective than that
of an author, I would be more inclined toward the latter interpretation. Either is quite possible, however.

335
locations, then the record becomes detached from any reality the reader has access to. It

is then no longer a “true” (shi 實) record.

In the paired chapters of the Shitong on “The Direct Brush” 直筆 and “The

Crooked Brush” 曲筆, Liu Zhiji discussed a more serious source of difficulty as regards

true records. He began by citing examples of historians who suffered or died because

their commitment to producing a true record conflicted with the wishes of those in power.

He concluded, “These are just sufficient to prove that the road of life has many narrow

and difficult passes, and to make us aware how difficult a true record is to come by” [足

以驗世途之多隘，知實録之難遇耳].44 A true record is difficult to come by because

only a few sterling personalities have ever been willing to risk life and limb in order to

produce one. In “The Crooked Brush,” Liu went on to describe in more detail the reason

most histories are not honest:

Supposing an official historian gets to decide on his own what should be loved
and hated, and that [judgements of] loftiness or inferiority exist in his mind [i.e.,
are subjectively determined by him]. Going in [to court] he does not fear lord or
law. Retiring, he behaves shamelessly in his own house. If you want to look to
him for a true record, is it not indeed difficult [to obtain]!? Alas, this is something
that those in charge of the state would do well to correct or reform. [令史臣得愛
憎由已，髙下在心，進不憚於公憲，退無媿於私室，欲求實録，不亦難乎？
嗚呼！此亦有國家者所宜懲革也。]45

Liu Zhiji was here describing the easy path for an official historian, one which would not

expose him to the disasters suffered by the heroic figures in “The Direct Brush” 直筆

chapter (Sima Qian among them). It might not seem obvious that allowing a historian

freedom to make his own judgements would be a bad thing. We have to understand,

however, what sort of freedom Liu Zhiji was talking about: he does not mean freedom

44
STTS 24.193.
45
STTS 25.199.

336
from compulsion and punishment, but rather freedom from an objective standard—

freedom to cave in to political pressure and write only what those in power want

written.46 What I here term an „objective standard‟ might bear little relation to scientific

truth in the modern Western sense, but clearly has some of the same of the same

properties.

A final statement by Liu Zhiji, from the chapter on “Narrating Events” 叙事,

illustrates the full moral import Liu gave to the term shilu 實錄. He wrote: “Keeping

one‟s will focused on preserving a true record—this is the means whereby good and evil

are made distinct from one another” (志存實録，此美惡所以爲異也).47

Liu Zhiji used the term shilu 實録 in much the same way Pei Songzhi did, to

mean both factual accuracy and a deeper moral integrity and courage in the history-

writing process. It is startling, then, to see that in the Northern Song dynasty, Sima

Guang clearly had a very different understanding of the term.

Among his many accomplishments, Sima Guang collaborated on a commentary to

Yang Xiong‟s Fayan. In his note on the passage where Yang Xiong first used the term,

Sima Guang departed from earlier understandings (such as Liu Zongyuan‟s), and

interpreted Yang Xiong‟s use of shilu 實錄 to mean that Sima Qian‟s work was “a true

record and nothing more” [實錄而已] (emphasis added).48 This gloss effectively

contrasted the meaning of factual accuracy from overall message of deeper moral truth.

In light of Yang Xiong‟s other remarks about Sima Qian, Sima Guang‟s understanding
46
It should further be noted that Liu Zhiji closely echoes the first line of the above citation in the following
chapter, where he writes: “Fondness or dislike proceed from individual difference…. Probably the three
kings suffered defamation... because things have a constant standard but the mirror has no fixed image” [愛
憎由其各異…葢三王之受謗也…則物有恒準而鑒無定識] (STTS 26.204).
47
STTS 22.167.
48
Yangzi fayan 7.20A.

337
may be closest to what Yang Xiong originally meant. To say that the Shiji was nothing

more than a true record was also a way of hinting that Sima Guang, as a historian himself,

had higher ambitions.

Having explored in some detail the changing range of meaning behind a key

general term used to evaluate reliability, I turn now to some of the ways in which the

issue of reliability played out in early readings of the Shiji.

SEEKING A STANDARD

Chu Shaosun was both an early reader of the Shiji and a much-reviled contributor

to its content. One of his most interesting contributions, appended to Shiji ch.13, is a

dialogue between himself and his fellow student Zhang Chang‟an. The text has been

analyzed it detail by Dorothee Schaab-Hanke,49 so I will not discuss it extensively. It is

worth noting as a response to one of the Shiji‟s more troubling features—troubling, at

least, from the perspective of those who would see it as a true record: namely, its internal

contradictions.

At the beginning of the “Yin benji” 殷本紀 [Basic annals of Yin, SJ ch.3], the

text states that Xie 契, the supposed ancestor of the Yin (Shang) dynasty, was produced

by miraculous conception: his mother Jian Di was out walking with two other women

when she “saw a dark bird laying an egg. Jian Di took the egg and swallowed it, and

because of that became pregnant and gave birth to Xie” [見玄鳥墮其卵，簡狄取吞之，

因孕生契].50 Meanwhile, the “Sandai shibiao” 三代世表 [Table by Generations of the

49
See her article, “Did Chu Shaosun Contribute to a Tradition of the Scribe?” OE 44 (2003), no.4: 11-26.
50
SJ 3.91. This reading is supported in, and probably derived from, the Shijing “Xuan niao” 玄鳥 ode
(Mao #303).

338
Three Dynasties, SJ ch.13] states, “Gao Xin fathered Xie” [高辛生 ].51 If Xie was

conceived as the result of his mother swallowing a dark bird‟s egg, how could Gao Xin

be his father? Similarly, the beginning of the “Zhou benji” 周本紀 [Basic annals of Zhou,

SJ ch.4] states that the Zhou ancestor Hou Ji was conceived when his mother

Jiang Yuan went out into the wilds and saw the footprint of a giant. Her heart was
rejoicingly gladdened, and she strongly desired to step into it. When she stepped
into it, her body experienced a contraction as with pregnancy, and after the time-
span had passed, she gave birth to a boy. [姜原出野，見巨人跡，心忻然說，
欲踐之，踐之而身動如孕者，居期而生子。]52

Yet again, the “Sandai shibiao” makes him the son of Gao Xin.53

For modern scholars, such contradictions are easily understood by reference to the

features of comparative mythology, such as the miraculous births of heroes, and so forth.

The question, though, is what did Sima Qian and his early readers make of them? The

Shiji makes no comment on its own contradictions, but, as mentioned above, Chu

Shaosun‟s addition addresses the question directly:

Zhang Fuzi [=Zhang Chang‟an] asked Master Chu, “The Odes say that Jie and
Houji were both born without having been fathered. Now regarding these cases,
the various traditions and records all say that they had fathers, and that their
fathers were both descendants of the Yellow Emperor. Is not this using the
Shi[jing] to contradict the [Spring and] Autumn [Annals]?” [張夫子問褚先生
曰：詩言契、后稷皆無父而生。今案諸傳記咸言有父，父皆黃帝子也，得無
與詩謬秋？]54

Chu Shaosun, like Zhang Chang‟an, was a scholar in the Lu Shijing tradition. Thus, as

Schaab-Hanke has pointed out, “The question Zhang raises toward Chu is whether or not

51
SJ 13.489. The archaic character that appears here is accepted as being an unproblematic variant of 契.
52
SJ 4.111. Again, this story is found in the Shijing, “Sheng min” 生民 ode (Mao #245).
53
SJ 13.489.
54
SJ 13.504.

339
by considering this doctrine [of common descent from the Yellow Emperor]…he might

deviate from the orthodox line as an exegete of the Book of Songs.”55

Chu Shaosun is able, however, to resolve the problem at least in some sense. I

translate only the first part of his response:

Master Chu said, “It is not so. When the Shi say that Xie was born from an egg
and Hou Ji from a person‟s footprint, their intention was just to show that these
two had the true essence of Heaven‟s Mandate. Even ghosts and spirits cannot be
born of themselves, but need people in order to arise. How much more [unlikely
for a person] to be born without a father? One said they had fathers, another said
they did not have fathers. Those who believed transmitted their belief, and those
who doubted transmitted their doubt. Therefore there are two different stories
about it. [不然。詩言契生於卵，后稷人跡者，欲見其有天命精誠之意耳。
鬼神不能自成，須人而生，柰何無父而生乎！一言有父，一言無父，信以傳
信，疑以傳疑，故兩言 之。]56

The first part of Chu‟s logic is less than perfectly clear, but Schaab-Hanke‟s

understanding seems a good explication, namely that “every human being necessarily

must have a father and that if tradition emphasizes the case of a supernatural birth, then

this would merely be a symbol meaning that the ruler whose birth is thus mystified has

received a mandate by Heaven to rule.”57 This line of reasoning in fact has something in

common with the explanation a scholar today might propose, that in this obviously

mythologized dynastic pre-history, two forms of legitimation are simultaneously being

brought to into play: the genealogical and the supernatural.

The latter part of the above passage alludes to an important historiographical

principle, which might be said to have its roots in a cluster of sayings by Confucius: “A

gentleman, with regard to what he does not know, probably leaves it blank” [君子於其所

55
Schaab-Hanke, “Chu Shaosun,” 13.
56
SJ 13.505.
57
Schaab-Hanke, “Chu Shaosun,” 13.

340
不知蓋闕如也]58; “Listen much and set aside what is doubtful” [多聞闕疑]59; and “I am

old enough to have seen scribes who left a blank when they did not know a word” [吾猶

及史之闕文也].60 It also echoes the statement from the Shiji itself, in the preface to the

same “Table”, which reads, “There is some [information] but also many lacunae and

things that cannot be recorded. Thus, where there are doubts one transmits the doubt—

this is probably the cautious thing to do” [或頗有，然多闕，不可錄。故疑則傳疑，蓋

其慎也].61

From the Lunyu sayings to the Shiji, there is already a certain degree of

development—the Lunyu seems to advocate complete omission in doubtful cases: leaving

a blank, saying nothing. The Shiji, on the other hand, is beginning to suggest something

more like piecing together some kind of account from what information there is—

certainly that is what the “Sandai shibiao” seems to be—while remaining ready to admit

one‟s lack of certainty. Chu Shaosun‟s version goes even farther, that where there are

multiple versions both should be preserved (and potentially reconciled). This certainly

seems to be what the Shiji does with the variant traditions about Xie and Hou Ji. This

may seem to bring Sima Qian closer to the cautious methodology of today‟s historian, but

did it also make the Shiji more a true record according to the standards of its own time?

58
Lunyu XIII: 3.
59
Lunyu II: 18.
60
Lunyu XV:26. Although this last may seem less clearly related, a remark by the Eastern Han
commentator Bao Xian 包咸 (dates unknown) shows that, whatever its original import, the quotation came
to be read as relating to this issue: “The good scribes of antiquity, with regard to the writing of characters,
would leave a blank when in doubt, waiting until it was known” [古之良史於書字，有疑則闕之，以待知
者] (Lunyu zhushu 15.140). The doubt involved, though apparently only related to the form of the
character, is readily extendable—following from the expansion of what it meant to be a “good scribe”—to
more profound matters, such as choice of words.
61
SJ 13.487.

341
Or, as was certainly the case with Chu‟s addition, was „truth‟ perhaps being pressed into

service for causes of political legitimation?

Chu Shaosun‟s comments showed a tendency to try to reconcile contradictions, or

at least show that they were operating on different levels of meaning. I turn now to a

comment by Huan Tan, which addresses a very different aspect of reliability, namely

how easily narrative plausibility can create the impression of truth. The incident in

question involves Emperor Gaozu of the Han and his conflict with a combined Xiongnu

and rebel force in 200 BCE. As Timoteus Pokora points out, this incident is narrated

three separate times in the Shiji. In the “Gaozu benji” 高祖本紀 [Basic Annals of Gaozu]

(SJ ch.8), it is presented as follows:

In the seventh year (200 BCE) the Xiongnu attacked Xin, the king of Hann, at
Mayi. Xin joined with them in plotting a revolt in Taiyuan. His generals, Manqiu
Chen of Baitu and Wang Huang, set up Zhao Li as king of Zhao in revolt [against
the emperor]. Gaozu in person led a force to attack them, but he encountered
such severe cold that two or three out of every ten of his soldiers lost their fingers
from frostbite. At last he reached Pingcheng, where the Xiongnu surrounded him.
After seven days of siege they finally withdrew. [七年，匈奴攻韓王信馬邑，
信因與謀反太原。白土曼丘臣﹑王黃立故趙將趙利為王以反，高祖自往擊
之。會天寒，士卒墮指者什二三，遂至平城。匈奴圍我平城，七日而後罷
去。]62

The account in the “Xiongnu liezhuan” 匈奴列傳 [Arrayed Traditions of the Xiongnu, SJ

ch.110] is similar but much more detailed, naming the Xiongnu leader as Maodun, and

giving specifics about troop movements as well as particulars of the Han army‟s suffering

during the siege. It then explains the escape as follows:

Emperor Gaozu sent an envoy in secret to Maodun‟s consort, presenting her with
generous gifts, whereupon she spoke to Maodun, saying, “Why should the rulers
of these two nations make such trouble for each other? Even if you gained
62
SJ 8.384-385, trans. adapted from Watson, Han I, 78.

342
 possession of the Han lands, you could never occupy them. The ruler of the Han
has his guardian deities as well as you. I beg you to consider the matter well!”
[高帝乃使使閒厚遺閼氏，閼氏乃謂冒頓曰：「兩主不相困。今得漢地，而
單于終非能居之也。且漢王亦有神，單于察之。」]63

This timely persuasion, together with suspicions Maodun had begun to have about his

rebel allies, had the hoped-for effect: the chapter then describes how Maodun allows the

Han emperor to slip away.

This brings us to a third narration of the siege at Pingcheng. We find it in “Chen

Chengxiang shijia” 陳承相世家 [The Hereditary Household of Prime Minister Chen] (SJ

56), where the eponymous character (naturally) plays a more significant role:

The following year Chen Ping in his capacity as colonel of the guard
accompanied the emperor in an attack upon the rebellious King Xin of Hann in
Dai. When they finally reached Pingcheng they were surrounded by the Xiongnu
and for seven days were unable to obtain food. Emperor Gaozu, following an
ingenious plan suggested by Chen Ping, sent an envoy to the consort of the
Shanyu. The siege was finally raised and Gaozu managed to escape. The exact
plan used was secret so that nowadays no one knows just what it was. [其明年，
以護軍中尉從攻反者韓王信於代。卒至平城，為匈奴所圍，七日不得食。高
帝用陳平奇計，使單于閼氏，圍以得開。高帝既出，其計祕，世莫得聞。]64

The narration in the Xiongnu chapter implies that the gifts alone are enough to win the

consort‟s help. But in Chen Ping‟s chapter, the strategy by which the consort‟s help was

won remains explicitly mysterious.

It is here that Pei Yin quotes from Huan Tan‟s Xinlun. It is rather long, but I

quote it in extenso because of what it reveals about its author‟s approach towards history.

Huan Tan‟s Xinlun [contains the following passage]:

Someone said, “Of Chen Ping‟s raising of the siege of Emperor Gao it is
said that “the exact plan used was secret so that nowadays no one knows just what
it was.”65 The job was done in a very adroit and excellent manner, so it remained

63
Trans. Watson, Han II, 138, SJ 110.2894.
64
SJ 56.2057; trans. Watson, Han I, 78.
65
A direct quote from the Shiji passage cited above.

343
 hidden and was not spread about. Could you, perhaps, reflect upon and
comprehend this matter?
I answered, “To the contrary, the plan was mean and low, clumsy and bad.
That is why it was kept secret and not disclosed.
“When Emperor Gao was besieged for seven days, Chen Ping went to
plead with the consort of the Shanyu. When she subsequently spoke to the
Shanyu, he released the Han Emperor. From this we know the method Chen Ping
used to persuade her.
“Chen Ping must have told her: „Han women are excellent and beautiful.
Their appearance is beyond compare in the entire world. Being besieged, [the
emperor has sent messengers to rush back and return with [women] whom he
intends to offer to the Shanyu. When the Shanyu sees the women, he will
certainly be deeply fond of them, and if he is fond of them them, you, his consort,
will become increasingly estranged from him. The best thing to do is to let the
Han escape before [the messengers] arrive [with the women], because once the
Han are freed, they will not bring their women.‟
“The consort and [the Shanyu‟s] women had jealous natures and they
certainly would have despised [the Han women]; therefore, they spurred the
Shanyu to let the Han go. This explanation is simple and to the point. Once the
ruse proved effective, [Chen Ping] wished to present it as something mysterious,
and therefore he kept it secret so it would not leak out.”
When Liu Zijun [=Xin] heard my words, he immediately praised them and
expressed his agreement.
桓譚新論：「或云：『陳平為高帝解平城之圍，則言其事祕，世莫得而聞
也。此以工妙踔善，故藏隱不傳焉。子能權知斯事否？』吾應之曰：『此策
乃反薄陋拙惡，故隱而不泄。高帝見圍七日，而陳平往說閼氏，閼氏言於單
于而出之，以是知其所用說之事矣。彼陳平必言漢有好麗美女，為道其容貌
天下無有，今困急，已馳使歸迎取，欲進與單于，單于見此人必大好愛之，
愛之則閼氏日以遠疏，不如及其未到，令漢得脫去，去，亦不持女來矣。閼
氏婦女，有妒媔之性，必憎惡而事去之。此說簡而要，及得其用，則欲使神
怪，故隱匿不泄也。』劉子駿聞吾言，乃立稱善焉。」66

A first thing to notice is the use of the direct quote from the Shiji.67 Though this

particular discussion is a literary conceit, it does show that debates about historical events

in Huan Tan‟s time took place with the Shiji text as part of the background information.

Another thing to add is that the Shiji was certainly not seen as an infallible authority, for

66
SJ Jijie 56.2057-58; trans. Pokora, Hsin-lun, 165-166.
67
Of course, since the Shiji is not explicitly named, it is possible that the quotation came from a third
source which the Shiji also quoted. In this case, however, it seems unlikely. The context of the quotation
in the Shiji seems much more like the historian explaining a lack of sources than actually using his sources.

344
where Sima Qian says of the plan that “no one knows just what it was,” Huan Tan insists

that it can be deduced from the information available to Sima Qian.

One important question to ask as regards Shiji reading practices is, was Huan Tan

aware of the other Shiji narrations of this incident, especially the one in Shiji ch.110?68

Unfortunately, there is no way to answer for sure. The Shiji ch.110 narrative suggests

that the Shanyu‟s consort merely allowed herself to be bribed: certainly a “mean and low”

plan, but not the one Huan Tan puts forth. That would argue against Huan Tan having

Shiji ch.110 in mind. On the other hand, if Huan Tan put the Shiji ch.110 narrative

together with than in Shiji ch.56, he might have concluded that though Shiji ch.110 tells

part of the story, part of the story must also be concealed, for otherwise the “mystery” of

Shji ch.56 would be easily “solved” by the narrative in Shiji ch.110. On the whole, I

consider the second alternative to be more plausible.69 Gifts, of course, would sweeten

the deal. But the narrative in Shiji ch.110 leaves doubts in the mind of a thoughtful

reader (especially one trained in the clever strategies and speeches of early Chinese

narrative prose)—would a consort be persuaded by material goods alone to advise her

husband against his own best interests?

Huan Tan‟s narrative bridges this gap nicely. There is no evidence that he had

any historical sources for it. Even Sima Qian a hundred years earlier lacked them, and

did not venture to invent anything. Huan Tan‟s method instead claims to be deductive:

“From this we know the method,” he confidently proclaims, referring to the

circumstances of the narrative, and then proceeds to supply a speech for Chen Ping.

68
Grant Hardy believes that an important part of the Shiji‟s construction is to train the reader to be see
events from different points of view (see Bronze and Bamboo, 61-85). It would be interesting to know if
real reading practices support this hypothesis.
69
David Honey‟s doubts about the text of Shiji ch.110 notwithstanding: see “The Han-shu, Manuscript
Evidence, and the Textual Criticism of the Shih-chi.”

345
“Chen Ping must have told her…” We know that in other places Sima Qian must also

have availed himself of this type of historical interpolation.70 Some scholars would see it

as evidence that fiction thrived under the auspices of history, but I believe that to the

historians themselves, Sima Qian as much as Huan Tan, such incidents were considered

to be deduced rather than invented. The historian begins with certain assumptions—in

this case, that the Xiongnu women are jealous, and insecure about the beauty of the Han

women—and with the circumstances of the narrative as given. The resulting narrative is

not, in their eyes, how it could have been, but in fact how it must have been, given these

assumptions and circumstances. Hence, Liu Xin is said to agree immediately: given their

shared assumptions, Huan Tan‟s solution must be correct.

This comment of Huan Tan shows the Shiji as a sort of background against which

new stories could be embroidered. This is something like the approach to the text taken

by Chu Shaosun, but Chu did not (as far as we know) fill details into existing Shiji

narratives, but rather grafted whole narratives on to Shiji chapters. The behavior of both

readers gives a sense that the text had a certain fluidity. It was not yet a canonical

account of history, not assumed to be complete; instead it was considered still open to

additions and changes.

I turn now to Wang Chong, who was using the Shiji rather than adding to it. In

particularly, Wang Chong frequently mentioned the Shiji as an arbiter of the reliability of

information. An important example is a detail in the story of Jing Ke and Prince Dan. At

the end of his narration of the Jing Ke story, Sima Qian wrote, “When people in our

70
The most striking example is the conversation among Zhao Gao, Li Si, and Hu Hai directly after the
death of the First Qin Emperor at Shaqiu (SJ ch.87). For a full discussion of doubts about this passage, see
Derk Bodde, First Unifier , 91-95.

346
generation speak of Jing Ke and claim that at the command of Heir-Apparent Dan „the

heavens rained grain and horses grew horns,‟ they are really going too far” [世言荊軻，

其稱太子丹之命，「天雨粟，馬生角」也，太過].71 Wang Chong, arguing against

the idea of fictitious influences—such as that a person‟s behavior could influence Heaven

to produce anomalies as described above—quotes these lines,72 and adds “The Honorable

Senior Archivist was a man who wrote down the true events of the Han era. When he

writes „idle talk,‟ it comes close to meaning „untrue‟” [太史公書漢世實事之人，而云

「虛言」，近非實也].73

The events of the Jing Ke story of course took place before the Han era, so why

should Sima Qian‟s recording of Han events have any bearing on the question? I believe

the statement is elliptical, but that it is actually making an argument: 1) Sima Qian‟s

recording of Han history—which can be independently verified—is reliable. 2)

Therefore it is better to trust Sima Qian‟s opinions on pre-Han history (which are less

easy to verify) than to trust the “empty talk” of others. Another possibility is that the Jing

Ke story, which took place in 227 BCE,74 was might possibly have been considered

recent enough to be part of the Han era.

Wang Chong discusses a similar subject in “Bian dong” 變動 [Change and

motion] (LH ch.43), where he is particularly concerned to refute a story that when Zou

71
SJ 86.2538.
72
Wang Chong‟s version, though directly attributed to Sima Qian, contains some variants from today‟s
Shiji text: “The Taishigong says, „People say of Prince Dan that he induced Heaven to rain grain and make
horses grow horns. All this is most likely idle talk‟” [太史公曰：「世稱太子丹之令天雨粟，馬生角，
大抵皆虛言也。」] (trans. Forke II.177, LH 19.236). It is likely that Wang Chong was quoting from
memory, but (given the excellence of his memory) also possible that the version of Shiji that he saw was
not quite the one we have today. One would think that Wang Chong would have preferred “gross error” to
“idle talk” and would be unlikely to have changed the former to the latter in his recollection.
73
LH 19.236.
74
According to SJ 15.755.

347
Yan was imprisoned, he implored Heaven to cause frost to fall. Among many other

arguments against this unlikely miracle, Wang Chong adds:

The imprisonment of Zou Yan resembles the adventures of Fan Sui and Zhang Yi.
Why does Sima Qian omit to mention this? Since it is not mentioned in Zou
Yan‟s biography that during his imprisonment he caused the frost to fall, it must
be an invention, a random statement like the story of Prince Dan, who is believed
to have ordered the sun to return to the meridian and heaven to rain grain. [鄒衍
見拘，睢、儀之比也，且子長何諱不言？案衍列傳，不言見拘而使霜降。偽
書遊言，猶太子丹使日再中、天雨粟也。]75

Here we see that Wang Chong wants to extend Sima Qian‟s dismissal of the Prince Dan

stories to Zou Yan and even to a more general case, interpreting the comment in Shiji

ch.86 to mean that Sima Qian did not believe in such phenomena at all, not just that he

doubted the specific instance of it.

Another case where Wang Chong cites the Shiji as an authority is in his chapter

“Tan tian” 談天 [A discussion of the sky]. In the elaborate process of debunking some

myths of his time regarding terrestrial and celestial geography, Wang Chong quotes the

skeptical evaluation at the end of “Dayuan liezhuan” 大苑列傳 [Arrayed Traditions of

Dayuan] (Shiji ch.123), which expresses doubts about various mythological accounts of

the far west. Wang Chong remarks that when Sima Qian writes, “„I dare not express

myself‟ [he] means that there is no truth in [it]” [夫弗敢言者，謂之虛也]. Wang Chong

goes on to add, “In the opinion of the Honorable Senior Archivist, the reports of the

„Mountain Classic‟ and the „Chronicle of Yu‟ are inventions” [案太史公之言，山經、

禹紀，虛妄之言].76 Wang Chong did not automatically accept the Shiji‟s verdict.

75
LH 43.659, trans. adapted from Forke I:115-116.
76
LH 31.476, trans. adapted from Forke I:254.

348
Instead he brought it together with other authorities, and weighed the evidence on either

side. However, he did end up affirming the correctness of Shiji‟s argument.

Wang Chong never hesitated to question authority, and though in some cases cites

the Shiji an authority, he does so only when convenient. When he disagreed with the

Honorable Senior Archivist‟s opinion in some certain case, he had no compunctions

about criticizing him. In “Dao Xu” 道虛 [Daoist untruths], he wrote:

The Honorable Senior Archivist in his evaluation on the “Five Emperors” also
says that, having performed the Feng and Shan sacrifices, Huangdi disappeared as
an immortal, and that his followers paid their respects to his garments and cap,
and afterwards buried them. I say that this is not true. [太史公記誄五帝，亦
云：黃帝封禪已，仙去，臣朝其衣冠。因葬埋之。曰：此虛言也。]77

Here we probably have a case of quoting from memory gone slightly wrong. The

anecdote in question is nowhere to be found in the “Wu di ben ji” 五帝本紀 (Basic

Annals of the Five Emperors) chapter, but rather in the “Fengshan shu” 封禪書 (Treatise

on the Feng and Shan), and even then it is not framed as being in Sima Qian‟s own voice

but rather is the explanation given by a follower of Emperor Wu, an ad hoc answer to the

emperor‟s rather tricky question.78 (Again it is possible that Wang Chong‟s memory

served him well and it is our transmitted text that has failed, but in this case it seems less

likely.79)

77
LH 24.314, trans. adapted from Forke I:332.
78
According to the “Fengshan shu” 封禪書 (SJ ch.28) , the discussion took place in 110 BCE, after the
emperor had sacrificed at the grave of the Yellow Emperor:
The emperor said, “I have heard it said that the Yellow Emperor did not die. Yet now we find his
grave here—how could that be?” Someone replied, “After the Yellow Emperor had been
transformed into an immortal and had ascended to heaven, his ministers made a grave here for his
robes and hat!” [上曰：「吾聞黃帝不死，今有冢，何也？」或對曰：「黃帝已僊上天，臣
葬其衣冠。」] (SJ 28.1396).
79
It is easy to imagine Wang Chong misremembering a story having to do with the Yellow Emperor as
being in the chapter on the earliest material, the “Wu di benji” 五帝本紀 (SJ ch.1). On the other hand, it is
difficult to imagine Sima Qian including such a bizarre and almost comical story in the “Wu di benji”—
where, he writes, he has “selected those words which seem most excellent and elegant” 擇其言尤雅者 (SJ

349
 Though the above anecdote shows that Wang Chong did not hesitate to question

the Shiji‟s account, he did at times defend it when the question was in doubt. The

argument in “Ziran” 自然 [On spontaneity] remarks on the Honorable Senior Archivist‟s

skepticism, but ends by offering a justification for the Shiji‟s account, which even its own

author doubted. The Shiji account records the story of Zhang Liang‟s meeting with an

old man, who gives Zhang Liang a wonderfully efficacious text and then disappears, but

adds that they will meet again and he will be in the form of a yellow stone. And indeed,

as the Shiji has it, Zhang Liang does find the yellow stone just where and when the old

man predicted. Thus the narrative; but in the evaluation, the Honorable Senior Archivist

expresses doubt about it: “Most learned people agree that there are no such things as

ghosts and spirits, though they concede the existence of [certain abnormal] things. What,

I wonder, are we to make of the old man whom Zhang Liang met and the book he gave

him? It is a thing to be wondered at” [學者多言無鬼神，然言有物。至如留侯所見老

父予書，亦可怪矣].80

Wang Chong argued for a theory of spontaneity that explains such bizarre

phenomena without resorting to the idea that heaven responds to man.

The spontaneity of these processes seems dubious and is difficult to understand.

Externally there seemed to be activity, but as a matter of fact, there was
spontaneity internally. Thus the Honorable Senior Archivist recording the story
of the yellow stone has his doubts and cannot find the truth…. The transmission
of the book by the „yellow stone‟ was a sign of the rise of the Han dynasty. That
the supernatural fluid becomes a ghost, and that the ghost is shaped like a man, is
spontaneous, and not the work of anybody. [自然之化，固疑難知，外若有
為，內實自然。是以太史公紀黃石事，疑而不能實也。黃石授書，亦漢且興
之象也。妖氣為鬼，鬼象人形，自然之道，非或為之也。]81

1.46)—and then it somehow being later moved to the “Fengshan shu” with a whole separate anecdotal
context invented for it.
80
SJ 55.2049.
81
LH 54.779, trans. adapted from Forke I:96.

350
The way this argument is set up displays Wang Chong‟s trust in the Shiji as a reliable

narrative, even as he sets himself up as superior to Sima Qian, arguing that Sima Qian did

not understand the phenomenon he was recording.

Wang Chong again used an example from the “Liu hou shijia” 留侯世家 [The

hereditary household of the Lord of Liu] in “Shi zhi” 實知 [Real knowledge]. This one is

remarkable in that it places Sima Qian and Confucius in parallel:

Confucius seeing an animal named it rhinopithecus, and the Senior Archivist had
the idea that Zhang Liang looked like a woman. Confucius had never before seen
a rhinopithecus, but when it arrived he could give it its name. The Senior
Archivist belonged to another age than Zhang Liang, but his eyes beheld his shape.
If the people at large had heard of this, they would have looked upon both as
divine beings who were prescient. However Confucius could name the
rhinopithecus because he had heard the songs of the people of Zhao, and the
Senior Archivist knew Zhang Liang from a picture which he had seen in the
Emperor‟s memorial hall. They kept secret what they had seen and concealed
their knowledge, using their mental powers in a way that was profound and secret.
[若孔子之見獸，名之曰狌狌；太史公之見張良，似婦人之形矣。案孔子未
嘗見狌狌，至乌能能名之；太史公與張良異世，而目見其形。使眾人聞此
言，則謂神而先知。然而孔子名狌狌，聞昭人之歌；太史公之見張良，觀宣
室之畫也。陰見默識，用思深祕。]82

Minor though the “knowledge” discussed here may be, this argument illustrates two

important things about Wang Chong‟s view of the Shiji. First, Wang Chong believed that

Sima Qian had a rational basis for his writings. Contrary to Wang Chong‟s final

comment, there is no great secret here: the picture of Zhang Liang is actually mentioned

in the Shiji.83 But presumably what Wang Chong meant was that in cases where the

82
LH 78:1079-1080, trans. adapted from Forke II:123.
83
SJ 55.2049. No source survives for the anecdote about Confucius, though as Lunheng commentator
Huang Hui points out, it is of a type similar to a fragment from the no longer extant Hanshi 韓詩 [Han
Odes tradition] quoted in the Guangyun 廣韻 [Comprehensive rhyme dictionary] (LH 78:1079). There are
also two typologically similar stories told in the Shiji “Kongzi shijia” 孔子世家 [Hereditary household of
Confucius] (SJ 47.1912-1913). Both involve Confucius identifying excavated bones, demonstrating his
amazing erudition.

351
Honorable Senior Archivist did not specify the evidence for his account, we should

assume that he nonetheless possessed such evidence. The second conclusion we might

draw from this passage is that Wang Chong considered it appropriate to treat the Sage

and the historian in parallel, which is to say, Confucius here appears no more divine than

Sima Qian, pace later believers in the mythos of the Supreme Sage.

Wang Chong was a peculiarity in his time. Having formulated his own ideas

about what was true of the world, he used whatever means possible to support them. Nor

did he hesitate to discard anything that could not be placed within his philosophical

framework, however venerable its cultural provenance. He was not an analytic

philosopher in the modern sense of the term, but he shared some of the same attitudes.

When he wrote that “The Honorable Senior Archivist was a man who wrote down the

true events (實事) of the Han era” it seems like a significant pronouncement, if still

somewhat formless.

A much more typical Han dynasty response to the issue of the Shiji‟s reliability is

that displayed by Yang Xiong, Ban Biao, and Ban Gu: the impulse to use the Classics and

the words of Confucius as a standard against which to measure the reliability of the

Shiji—the Shiji was only reliable insofar as it agreed with these more authoritative

sources. Qiao Zhou 譙周 (ca. 200-270) seems to have similar standards:

Qiao Zhou considered that when Sima Qian‟s Shiji wrote of the period of Zhou,
Qin, and before, he at some points adopted vulgar sayings and words of the
hundred schools, and did not exclusively rely on the correct Classics for his
evidence. This is why [Qiao] Zhou made the Gushi kao in twenty-five pian,
relying on the old canonical texts throughout, in order to rectify [Sima] Qian‟s
errors. [譙周以司馬遷史記書周秦以上，或採俗語百家之言，不專據正經，
周於是作古史考二十五篇，皆憑舊典，以糾遷之謬誤．]84

84
JS 82.2141.

352
Since the Gushi kao survives only in fragments, it is difficult to say for certain whether or

not this is an accurate characterization.85 Something worth noting, however, is the

intermediate position we see the Shiji occupying: it is somewhere directly in between the

Classics—an almost universally accepted standard of truth and reliability—and the vulgar

sayings and „Hundred Schools‟ texts, which were by Qiao Zhou‟s time a metonym for the

opposite, the very essence of unreliability.

AGAINST „DEFAMATORY TEXT‟ READINGS

In chapter 4, above, I explored the roots of autobiographical readings of the Shiji,

showing that by and large autobiographical readings were used negatively, in the context

of criticizing Sima Qian for his authorial role. In the section which follows, I discuss the

types of readings put forth by Sima Qian‟s defenders in order to counter these

condemnations. These defenses almost invariably involved claiming that Sima Qian was

only telling the truth.

An anecdote, found in Chen Shou‟s 陳壽 (233-297) San Guo Zhi 三國志, seems

to be a direct rebuttal of Ban Gu‟s position that Sima Qian was in the wrong and “not a

gentleman with a sense of duty.” It occurs in a chapter devoted to the various memorials

and official words of the scholar Wang Su 王肅 (195-256).86 The particular context is

85
J. Michael Farmer, who has painstakingly translated and analyzed all the surviving comments, has come
up with a number of interesting conclusions about the nature of the Gushi kao. See his recent monograph
The Talent of Shu.
86
It is interesting to note that Wang Su must have been an expert on matters historical, or perhaps even
specifically on the Shiji, for his name is mentioned close to two hundred times in the Three Scholars
Commentary on the Shiji.

353
that the emperor87 has begun a discussion with Wang Su concerning those who are so

bold as to criticize their rulers.

The Emperor once asked, “In the time of Emperor Huan of the Han, the
Prefect of Baima, Li Yun, submitted a memorial saying, „An emperor is one who
is attentive.88 Yet this emperor makes no effort to be attentive.‟ At that point,
how was it that Li Yun managed not to be put to death!?” Su replied, “It is only
that in choosing his words he failed to measure what would be pleasing or
unpleasing to his audience. From the beginning, his basic intent was entirely in
making a whole-hearted effort, his thoughts concentrated on helping his country.
Furthermore, the might of an emperor surpasses that of a thunderclap. Killing one
common fellow is not different from killing an ant or a cricket. By being
magnanimous and forgiving [Li Yun], [the emperor] was able to show his
willingness to tolerate cutting words, spreading [his reputation for] gracious
generosity throughout the realm. Therefore I humbly consider that killing him
would not necessarily have been the right thing to do.” [帝嘗問曰：「漢桓帝
時，白馬令李雲上書言：『帝者，諦也．是帝欲不 諦．』當何得不死？」
肅對曰：「但為言失逆順之節．原其本意，皆欲盡心，念存補國．且帝者之
威，過於雷霆，殺一匹夫，無異螻蟻．寬而宥之，可以示容受切言，廣德宇
於天下．故臣以為殺之未必為是也．」]
[The emperor] then asked, “Sima Qian, because he suffered punishment,
harbored in his breast a hidden condemnation (qie 切). When he compiled his
Archivist‟s Records, he denounced and criticized Emperor Wu. It makes one
want to gnash [切] one‟s teeth!” [Su] replied, “Sima Qian recorded events
without empty beautification and without hidden dislike. Liu Xiang and Yang
Xiong both admitted that he was very good at narrating events and had the talent
of a good historian, calling [his work] a true record. When Emperor Wu heard
that [Sima Qian] was compiling the Archivist‟s Records, he took up [his father]
the Filial Jing‟s basic annals and those of himself, and read them. Thereupon he
flew into a rage, breaking [the strips] into pieces and throwing them down. Even
today those two annals have only titles but no texts. Later, when the Li Ling
affair occurred, [Emperor Wu] had Qian sent to the Silkworm Chamber. This
would suggest that the hidden grudge was on the part of Emperor Wu, and did not
lie with Archivist Qian. [帝又問：「 馬遷以受刑之故，內懷隱切，著史記非
貶孝武，令人切齒．」對曰：「司馬遷記事，不虛美，不隱惡．劉向﹑揚雄
服其善敘事，有良史之才，謂之實錄．漢武帝聞其述史記，取孝景及己本紀

87
Wang Su served under both the Wei Emperor Wen (Cao Pi 曹丕, r.220-226) and his successor Emperor
Ming (Cao Rui 曹叡, r. 226-239). Here the placement of the anecdote suggests that the emperor in
question is Emperor Wen.
88
In Chinese, this is a paronomastic gloss, with the words for “emperor” and “attentive” being both near
homophones and graphically similar. In the Hou Hanshu account of Li Yun‟s remonstrance, this phrase is
attributed to Confucius by way of the Spring and Autumn Annals tradition; see HHS 57.1852 and HHS
57.1853 n.6.

354
 覽之，於是大怒，削而投之．於今此兩紀有錄無書．後遭李陵事，遂下遷蠶
室．此為隱切在孝武，而不在於史遷也．」]89

Here Sima Qian is used as an example within the context of remonstration. The fact that

he appears in parallel with Li Yun is important: Li Yun‟s blunt and rather ill-considered

words might well have earned him a death sentence, but their very bluntness and

ineptitude was a kind of argument for their truth (and beyond that, their correctness of

intention).90 In this one might find echoes of Sima Qian‟s inept defense of Li Ling. But

certain instabilities in the Sima Qian story introduce another level of complication.

While Emperor Ming follows what we might call the Hanshu version (as outlined in the

“Letter”) of Sima Qian‟s story, Wang Su counters it with a version of Wei Hong‟s

account (discussed in chapter 4 above).91

Wang Su cleverly juxtaposed the story of Emperor Wu‟s wrath with two other

well-known pronouncements on the Shiji, those of Liu Xiang and Yang Xiong: these

eminent figures pronounced that Sima Qian was a good historian, and the Shiji was a true

record. This reframing of the “destruction story” results in a crucial shift of sympathies.

Where Wei Hong had accused Sima Qian of emphasizing, even exaggerating [Emperor

Jing‟s] shortcomings and Emperor Wu‟s mistakes, Wang Su implied that Sima Qian‟s

89
SGZ 13.418.
90
Li Yun‟s own story is clearly problematic, for at least two accounts in the Hou Hanshu (including his
own biography) say that as a consequence of his blunt remonstrance “he died in prison” [下獄死] (HHS
7.307, HHS 57.1852). It is only in the biography of Chen Fan 陳蕃 that Li Yun is said to escape that fate:
“Fan submitted a memorial and saved Yun, whereupon he was excused and returned to his rustic village”
[蕃上書救雲，坐免歸田里] (HHS 66.2161). Of course the present anecdote only works if we accept the
latter (and to my mind less likely) interpretation.
91
It is quite possible that calling it “Wei Hong‟s account” is not strictly accurate. As I proposed above,
Wei Hong may have had some perfectly respectable source for his version, and the emperor in this
anecdote may have gotten the story from the same source. It is simply impossible to say for certain, since
no other version but Wei Hong‟s exists with reliable attribution. It might be worth noting that the story
does appear in nearly identical form in the Xijing zaji (XJZJ 6.139), but the authorship of this text is
problematic to say the least. See Nienhauser, “An Interpretation of the Literary and Historical Aspects of
the Hsi-ching tsa-chi (Miscellanies of the Western Capital).”

355
criticism was perfectly justified. Therefore, in Wang Su‟s eyes at least, the powerful

Emperor Wu had proven himself inferior to the much-weaker Emperor Huan: Emperor

Wu had shown himself unwilling to “tolerate cutting words” and unable to “spread [his

reputation for] gracious generosity throughout the realm.” Even more significantly,

Wang Su rejects Ban Gu‟s accusation that the Shiji contains “subtle writing and piercing

satire,” choosing instead to follow the “Wei Hong” tradition that it was so openly critical

that the emperor could not keep himself from destroying the chapters he read.

Shiji scholarship has largely ignored anecdotes like the one cited above because

their version of Sima Qian‟s story (the Wei Hong version) is contradicted by the “Letter.”

Yet Wang Su‟s use of the story shows that it was still very much in circulation even after

the “Letter” had appeared to contradict it. Indeed, if we consider things from Wang Su‟s

perspective, the “Letter” does not entirely rule out the Wei Hong version. Sima Qian

wrote that he was punished because he had not made himself clearly understood in

speaking about Li Ling, and hinted that the complicated rivalries among generals‟

factions within the court had also played a part. That does not mean Emperor Wu could

not have had another, prior, grudge.92

The interesting point about this story for my purposes is the way Wang Su used

the “true record” reading of the Shiji to neutralize certain aspects of the autobiographical

reading. In Wang Su‟s version, Sima Qian produced his true but offensive portrait of the

mid-Western Han long before the Li Ling affair—before, that is, he had any particular

reason to be resentful. This redeems the (conveniently lost) annals of Emperors Jing and

92
A reader today would surely agree with Yu Jiaxi that the Wei Hong version sounds far-fetched and does
not seem to ring true to Sima Qian‟s time, or as Yu Jiaxi rather contemptuously pronounces, “There is no
need for any deep investigation in order to know it is false.” After all, why would a powerful emperor like
Filial Wu need to wait for an excuse to execute one of his minor officials? But not all emperors could
dispose of their officials at will, and perhaps the fiction rang more true in other times.

356
Wu from charges of having been distorted by Sima Qian‟s rancor. By extension, the rest

of the Shiji might also be redeemed as well.

In chapter 4 above, I already discussed the death of Cai Yong, and Sima Qian‟s

incidental involvement with the narrative. I want to revisit that story here, to consider

what reactions to that story tell us about “true record” readings and their opposition to

autobiographically motivated ones.

Fan Ye‟s version of Cai Yong‟s death was not the earliest one. At least two other

versions existed in latter Han histories now no longer extant. One, in Zhang Fan‟s 張璠

Hou Han Ji 後漢紀 [Annals of the latter Han], is much abbreviated and contains no trace

of Wang Yun‟s reference to the Shiji. Pei Songzhi 裴松之 (372-451), in his commentary

to the San Guo Zhi, quotes another version (fairly similar to Fan Ye‟s, see note 94 below)

by Xie Cheng 謝承. Pei then follows it with an interesting critique:

I, Songzhi, believe that, although Cai Yong was personally employed by Zhuo, he
could not possibly have felt himself to be one of Zhuo‟s faction. Could he have
not known that Zhuo was treacherous and a villain, or [not been aware of] the
harm that [Zhuo] did to the realm? Hearing of [Zhuo‟s] death, it stands to reason
that he would not have sighed or regretted it. Or let us grant for a moment that it
was so, he would not have talked back at Yun‟s council meeting. This is merely
Xie Cheng‟s unfounded recording. Archivist Qian‟s annals and traditions are
extensive and are a marvelous achievement of this world. Yet [Xie Cheng] relates
that Wang Yun said Filial Wu should have killed Qian earlier [to prevent him
from writing it]. These are not the words of a knowledgeable person. It is only
that Qian did not conceal the faults of Filial Wu, and instead wrote a direct
account of his affairs. In what sense is this calumny? Wang Yun, in his loyalty
and uprightness, could be called one who could reflect on himself without error.
He could not have been afraid of calumny, and yet he wanted to kill Yong. In
discussing whether or not Yong should die, how could he worry about [Yong‟s]
calumniating him, and unjustly kill a good man!? These are all false accusations,
and furthermore extremely incomprehensible. [臣松之以為蔡邕雖為卓所親
任，情必不黨．寧不知卓之姦凶，為天下所毒，聞其死亡，理無歎惜．縱復
令然，不應反言于王允之坐．斯殆謝承之妄記也．史遷紀傳，博有奇功于
世，而云王允謂孝武應早殺遷，此非識者之言．但遷為不隱孝武之失，直書
其事耳，何謗之有乎？王允之忠正，可謂內省不疚者矣，既無懼于謗，且欲

357
 殺邕，當論邕應死與不，豈可慮其謗己而枉戮善人哉！此皆誣罔不通之甚
者。]93

In the translation above, I have used “calumny” instead of “defamation” as a translation

for bang 謗, because Pei Songzhi was (perhaps polemically) insisting on using the term

to mean something that is expressly untrue. For Pei Songzhi, an honest account—the sort

that he believes Sima Qian to have written—cannot be a bangshu. Furthermore, Pei

argued, it does not even make sense for Wang Yun to use such reasoning to justify killing

Cai Yong. Casting the whole episode in doubt, he cites Zhang Fan‟s account as being the

preferable version.94

It should not surprise us that Pei Songzhi should be sympathetic to Sima Qian.

His son, Pei Yin 裴駰 (fl.438) would later write the first surviving commentary on the

Shiji, showing that the family had a tradition of historiographical expertise. But the

debate over whether or not the Shiji was a bangshu 謗書 (as well as over just what

constituted a bangshu) was clearly a lively one in this period. It is worth pausing a

moment to consider this term, which in the Six Dynasties at least appeared as almost an

antonym to “true record.”

The word bangshu 謗書 actually appears in the Shiji itself, in what would become

by far the most ubiquitous historical allusion involving the term:

Lord Wen of Wei ordered Yue Yang to lead an attack on Zhongshan. After three
years he conquered it. But when he returned and his merit was being discussed,
93
SGZ 6.180.
94
Note that Fan Ye‟s account actually addresses all of Pei Songzhi‟s concerns, keeping most of the original
anecdote intact, but including other details and scenes. For example, Pei Songzhi claims that Yun would
not have acted that way. Fan Ye has him first arrest Cai Yong in a fit of rage and then, after he has made
the speech comparing Cai Yong to Sima Qian, Fan Ye uses Ma Midi to suggest that indeed, Wang Yun‟s
reasoning was faulty. Also, in Xie Cheng‟s version, Cai Yong makes a speech in his own defense. Pei
Songzhi complains that Cai Yong would not have been so stupid as to speak that way and indeed Fan Ye
cuts Cai Yong‟s speech down to a succinct paraphrase. Finally, unlike Xie Cheng‟s version (as far as we
know), Fan Ye included the detail from Zhang Fan, that Wang Yun regretted his harshness.

358
 Lord Wen showed him a box full of letters criticizing him (bangshu), and Yue
Yang, bowing his head twice, said, “The victory was due to no merit of mine, but
to the might of our lord and ruler!” [魏文侯令樂羊將而攻中山，三年而拔之．
樂羊返而論功，文侯示之謗書一篋．樂羊再拜稽首曰：此非臣之功也，主君
之力也．]95

In the Shiji, this story is one of a string of historical examples mentioned by Gan Mou,

who is trying to gain some assurance that the king of Qin will continue to support an

attack on the state of Han and not be swayed by criticisms of Gan Mou‟s rivals at court.

In context, the nature of the bangshu (here, critical or defamatory letters) is not

necessarily libelous—the historian does not record whether the criticism they contained

was true or false. The effect they produced, however, was to cheat Yue Yang of his

victory, a fate that Gan Mou wants to avoid.

The phrase bangshu yikui 謗書一篋 (“a box of defamatory letters”) and the

closely related bangshu mankui 謗書滿篋 or bangshu yingkui 謗書盈篋 (“defamatory

letters enough to fill a box”) eventually became set-phrases frequently used by the ruler

or other people in high-position, often in the context of complaining about a lack of

support for their actions.96 Whether the criticism is deserved or not, calling something

bang 謗 implies that the criticism it contained was not constructive or well-intended;

bang 謗 is criticism meant to punish rather than to correct.

Is the Shiji a defamatory text in this sense? It certainly seemed so to some readers,

as we will see, even if they did not say it in so many words. Other readers, like Pei

Songzhi, seem to have judged that the “true record” of a “good historian” could not be a

95
SJ 71.2312; trans. adapted from Watson, Qin, 104. Note that there is an extremely close parallel to this
anecdote in the Zhanguoce, see Qin 2, 149-150.
96
E.g., JS 42.1213; Songshi 332.10675 and 402.12192; Mingshi 259.6710.

359
bangshu—a useless criticism—because a history‟s utility lay in its very truthfulness. The

only useless criticism was a false one.

THE DANGERS OF THE STRAIGHT BRUSH

Tang dynasty discussions of truth and reliability in history seem, probably not by

accident, to focus on the political dangers of writing true history. For Tang historians,

what constituted historical truth seems to have been a somewhat lesser worry. The bigger

problem was facing the potential consequences that might result from being honest about

the shortcomings of those in power.

Liu Zhiji and the „Straight Brush‟

Liu Zhiji was a contemporary of Shiji commentator Sima Zhen but had quite a

different perspective on Sima Qian, as we will see. The two do not seem to have been

allies in academic circles, taking, for example, opposing positions in a court debate over

the adoption of commentaries to the Laozi and other works.97 Dissatisfied with the way

in which official histories were being written in his day, Liu Zhiji eventually resigned his

post and devoted himself to the creation of the Shitong, a monumental work of historical

criticism.98 The Shitong has much to say about the Shiji (especially about its relationship

to the Classics and its generic characteristics), but curiously little to say about the Sima

Qian tragedy. No doubt this reticence was deliberate—Liu Zhiji‟s aim was to focus

attention on the histories themselves rather than getting entangled in ad hominem

97
See XTS 132.4522 and McMullen, State and Scholarship in T‟ang China, 85.
98
For a more detailed description of Liu Zhiji‟s career leading up to the composition of the Shitong as well
as his specific complaints, see McMullen, State and Scholarship, 177-178. See also Michael Quirin, Liu
Zhiji und das Chun Qiu.

360
arguments about their authors. Nonetheless, he does make an important contribution to

the picture we have been constructing, of how the issue of reliability—here referred to as

“the straight brush”—influenced the perceived connection between the Shiji and Sima

Qian‟s tragedy.

In the chapter “The Straight Brush,” Liu Zhiji discusses the historian‟s dilemma

of writing the truth or taking refuge in indirection:

People despise that which is bent and twisted, and [that kind of writing] is the
Way of the petty man. People value that which is upright and straight, and [that
kind of writing] is the Virtue of the princely man. And yet the [people of the]
world have a great tendency toward bentness and they discard what is upright.
They do not tread in the footsteps of the princely man, and instead follow the
petty man‟s path. Why is this? There is a saying that goes: “Straight as a string,
the road to death; crooked as a hook, you wind up a marquis.” Thus, one would
rather follow obediently along preserving one‟s good fortune, instead of going
against the grain and being harmed thereby…. If one writes directly about the
affairs of a villainous minister or rebellious son, an unrestrained prince or an
incorrigible ruler, if one does not cover up these blemishes, then the traces of this
pollution reflect on the entire dynasty and it will bear that bad name for thousands
of years. When one puts it this way, ah, how frightening it is! [若邪曲者，人之
所賤，而小人之道也；正直者，人之所貴，而君子之徳也。然世多趨邪而棄
正，不踐君子之跡，而行由小人者，何哉？語曰：直如弦，死道邊；曲如
鈎，反封侯。故能順從以保吉，不違忤以受害也。。。其有賊臣逆子，淫君
亂主，茍直書其事，不掩其瑕，則穢跡彰於一朝，惡名被於千載。言之若
是，吁可畏乎。]99

Liu Zhiji makes it clear, here and elsewhere, that a historian should strive to write with a

„straight brush‟, and produce a true record. Yet Liu‟s own involvement in the

compilation of official history must have amply demonstrated to him the difficulty of

that proposition. As McMullen describes, Liu Zhiji‟s complaints about his task included

the fact that “material, including court diaries, was not available, confidentiality was

farcical, and no single policy was upheld for the historian compilers to follow.”100

99
STTS 24.192.
100
McMullen, State and Scholarship, 178.

361
 Liu Zhiji concludes, regarding the „straight brush‟ that “those who did it in a time

when one could do it got along fine, but those who did it in a time when one could not do

it had an evil fate befall them” [夫為於可為之時則從，為於不可為之時則凶].101 Thus

Dong Hu, who wrote that Zhao Dun assassinated his lord,102 survived together with his

reputation for honest recording. And yet,

When it came to the scribes of Qi who wrote that Cui assassinated his lord,103 or
to [Si]ma Qian who narrated the mistakes of Han, to Wei Shao who upheld justice
in the court of Wu,104 or Cui Hao offending against taboos in the state of Wei105—
some had their bodies mutilated by the axe and were ridiculed in their time,106
others see their work go to fill in walls or cellars, not to be heard of in later ages.
[至若齊史之書崔弑，馬遷之述漢非，韋昭仗正於吳朝，崔浩犯諱於魏國，
或身膏斧鉞，取笑於當時，或書填坑窖，無聞於後代。]107

Significantly, the Qing commentator of the Shitong, Pu Qilong, annotated the reference to

Sima Qian with the anecdote from the Hou Hanshu story of Cai Yong‟s death (discussed

above), for that is the place in the canonical histories which most directly resembles Liu

101
STTS 24.192.
102
The case of Dong Hu, found in Zuo Zhuan Xuan 2 is a peculiar paradigm case for honest history-writing
in the Chinese historiographical tradition. See Zuozhuan Xuan 2 (CQZZ zhu, 662-663). Duke Ling of Jin
was a cruel and irresponsible ruler. He attempted to kill his chief minister, Zhao Dun, which resulted in
Zhao Dun going into hiding. Later, in 606 BCE, Zhao Dun‟s brother killed Duke Ling. Yet the Senior
Archivist, Dong Hu, wrote, “Zhao Dun assassinated his lord.” Zhao Dun was blamed for the murder
because he had not crossed the border out of Jin at the time when it occurred. Confucius comments on this
anecdote that “Dong Hu was a good archivist of antiquity. His method of recording concealed nothing.”
103
Zuo Zhuan Xiang 25 (CQZZ zhu, 1099). Cui Shu of Qi had killed his ruler. The Senior Archivist
recorded it thus, and for that Cui killed him. Two of his younger brothers persisted in that interpretation of
events and were also killed for it. A third brother did the same, but was spared.
104
Pu Qilong summarizes the San Guo Zhi narrative as follows:
When Sun Hao wanted assumed the throne in Wu, he wanted to make an annal for his father Sun
He. Yao [i.e., Wei Shao] persisted in saying that since Sun He did not ascend to the position of
Emperor, it was fitting that his [history] be called a „Traditions‟ and not an „Annals‟, and there
were more than one that agreed with him. Hao‟s dislike and rage increased until in the end he
executed Yao. [孫皓即位，欲為父和本紀，曜執以和不登帝位，宜名為傳，如是者非一。皓
積嫌憤，遂誅曜。] (STTS 4.40; c.f., SGZ Wu 65.1462-1464)
105
The anecdote is found in the Beishi, as Pu Qilong notes. Apparently some of Cui Hao‟s associates
wanted to put up a stone with an inscription praising Cui Hao‟s talents. Rivals from another faction
accused them of wanting to set Cui Hao up in place of the emperor. The emperor flew into a rage and had
Hao executed. (See Beishi 21.789.)
106
Perhaps a reference to the passage in Sima Qian‟s “Letter in Reply to Ren An,” where he complained
that both he and Li Ling were “a sight for all the world to scorn” [重為天下觀笑] (HS 62.2730; Watson,
Ssu-ma Ch‟ien, 62).
107
STTS 24.192-193.

362
Zhiji‟s reference. Yet as with earlier references of this type (beginning with Ban Biao),

Liu Zhiji‟s use of the Sima Qian story blurs the distinction between the Li Ling affair and

Sima Qian‟s written criticism of the Han. In juxtaposition with the other figures, it would

seem that Sima Qian criticized the Han and was therefore ridiculed in his time. But when

Sima Qian complained that Li Ling and himself were ridiculed in their own time, history-

writing was in no sense a cause. It was due (respectively) to Li Ling‟s surrender and to

Sima Qian‟s own reckless defense of that act.

Han Yu and the Veritable Record of Shunzong

The Xin Tangshu 新唐書 [New History of the Tang] records a conversation

between Tang Emperor Wenzong (r.826-840) and one of his ministers, Zheng Tan 鄭覃

(d.842).108 In it, the emperor suggests a comparison between Han Yu 韓愈 (768-824)

and Sima Qian:

The emperor would often say something like, “The [record of] events [in the reign
of] Shunzong [r.805] is not detailed or veritable. How could the historical
minister Han Yu wrong a ruler [who was] his contemporary? Formerly, Sima
Qian of the Han wrote a letter to Ren An, and his phrases are full of hatred and
resentment. That is why the basic annals of Emperor Wu very often miss the
truth.” Tan replied, “In his middle years, Emperor Wu raised a great number of
troops and sent them to the frontier, causing waste and weariness for the people,
and emptying the granaries and store-houses. What Qian transmitted were not
exaggerated words.” Li Shi said, “What Tan has submitted uses Emperor Wu to
remonstrate with your majesty, whom he hopes will in the end attain the most
flourishing virtue.” The emperor said, “Truly it is so. Everything has its
beginning and rarely can one overcome the end.” [帝每言：順宗事不詳實，史
臣韓愈豈當時屈人邪？昔漢司馬遷與任安書，辭多怨懟，故武帝本紀多失
實．覃曰：武帝中年大發兵事邊，生人耗瘁，府庫殫竭，遷所述非過言．李
石曰：覃所陳，因武帝以諫，欲陛下終究盛德．帝曰：誠然，靡不有初，鮮
克有終．]109

108
Note that this incident is absent from the Jiu Tangshu.
109
XTS 165.5067.

363
This anecdote has several curious features, but let us first consider the background of the

comparison.

Han Yu, literary giant and controversial minister, received a commission in 813 to

write the Shilu 實錄 [Veritable Records] of Emperor Shunzong‟s brief reign. He does

not seem to have been wholly pleased with the assignment. In a letter responding to Liu

Ke‟s110 encouragement and congratulation, Han Yu detailed his objections to the job.

First, he wrote, he was not (as his correspondent must have suggested) engaging in praise

and blame in the high tradition of Confucius, for, he wrote,

the Spring and Autumn already has complete in it the entirety of the great model
by which historians praise and blame. Later writers engage in proving it with the
traces of events and veritable records, so good and evil appear of themselves.
However, this [making veritable records] is still something that this shallow and
indolent person could not dare engage in; how much less praise and blame? [愚
以為凡史氏褒貶大法，春秋已備之矣。後之作者在據事跡實録，則善惡自
見。然此尚非淺陋偷惰者所能就，况褒貶邪？]111

Furthermore, Han Yu (like Liu Zhiji, discussed above) was acutely aware that

historiographers had an alarming propensity to encounter disaster:

Confucius was a sage. He made the Spring and Autumn. [Yet] he encountered
shame at Lu, Wei, Chen, Song, Qi, and Chu. In the end he did not meet his time
and died. Of the Senior Archivists of Qi several brothers were exterminated. Zuo
Qiuming recorded historical events from the time of the Spring and Autumn and
lost his sight. Sima Qian made the Shiji and was sentenced to punishment. Ban
Gu starved to death….Now, those who make histories—if they did not encounter
some human catastrophe, then Heaven punished them. How can one not be
terrified [of this], and lightly accept [the task]? [孔子聖人，作春秋，辱於魯衛
陳宋齊楚，卒不遇而死；齊太史氏兄弟幾盡，左丘明紀春秋時事以失明；司

110
Liu Ke‟s courtesy name was Xiren 希仁. He would become a jinshi in 820 with the help and patronage
of Bai Juyi. E.G. Pulleyblank has suggested that in writing this letter, “Liu, whose greatest ambition was to
be an historian, hoped to be recognized as a kindred spirit by the most noted ku-wen writer of the day”
(Pulleyblank, 150). For a thorough discussion of Liu Ke‟s historical traces, see Pulleyblank‟s article, “Liu
K‟o, A Forgotten Rival of Han Yü.”
111
Han Changli 5.473.

364
 馬遷作史記，刑誅； 班固瘐死。。。夫為史者，不有人禍，則有天刑，豈
可不畏懼而輕為之哉！112

Again we see Sima Qian‟s punishment tied more to his history-making than to his

involvement with the Li Ling affair. But Han Yu solves the problem of anachronism by

hinting that the very enterprise of history-making seems to be cursed, given the degree to

which its practicitioners are dogged with misfortune. There need be no actual causal

connection between Sima Qian‟s history-making and his misfortune (as there was in the

case of the Senior Archivists of Qi, see above). It was as if Heaven itself caused his and

the others to suffer catastrophe.

Bernard Solomon has characterized this letter “more as a passing mood rather

than a serious statement of [Han Yu‟s] views.”113 Other less than earnest-sounding

claims Han Yu makes in the letter (such as that the Bureau of History job is a sinecure he

received because those in power felt sorry for him, or that the number of worthy rulers

and minister in the Tang is so great that compiling a history would be impossible anyway)

seem to support this idea. David McMullen, in contrast, interpreted the letter as an

expression of profound pessimism about the interference of political interests with the

historian‟s task:

He believed that the political self-interest of [the History Bureau‟s] directors

vitiated the material of official history to an extent that made it almost worthless.
“In extreme cases,” he wrote, “the fondnesses or hatreds of the factions adhered to
will be so different that good or evil deeds may be constructed out of nothing.”114

Certainly Liu Zongyuan took the issue seriously enough, when Han Yu‟s letter to Liu Ke

happened to fall into his hands. His “Letter to Han Yu Discussing Officials in the Bureau

112
Han Changli 5.473-474.
113
Solomon, Veritable Record, xv.
114
McMullen, State and Scholarship, 349 n.160, translation of Han Changli 5.476.

365
of Historiography” (與韓愈論史官書) lambasts Han Yu on every point,115 and was

probably the reason why Han Yu‟s letter to Liu Ke was excluded from his literary

collection proper and relegated to the waiji 外集 [outer collection].116 E.G. Pulleyblank

points out that Liu Zongyuan‟s own motives in this matter were far from disinterested—

as a former member of Shunzong‟s faction, he “had reason to feel sensitive about the way

in which the history of the reign of Shunzong was written.”117

Despite Han Yu‟s misgivings about the historian‟s task, he did eventually submit

a draft of the “Veritable Records of Shunzong” to the throne. As Charles Hartman writes,

“It is probably impossible to determine how the present text of the Veritable Record of

Emperor Shun-tsung relates to the manuscript that Han Yü submitted to the throne in the

summer of 815,”118 but it seems clear that, “the original text related in some detail the

eunuch role in the ascension of Emperor Hsien-tsung.”119 Hartman adds that “the

eunuchs criticized this text at the time of its submission and instigated for a revision until

this was accomplished about 830.”120 One wonders if the eunuchs‟ dissatisfaction with

115
Among Liu Zongyuan‟s complaints against Han Yu‟s letter, he includes the ahistoricity of attributing
Confucius‟ and Sima Qian‟s misfortunes to their historical endeavors: “That Confucius encountered
difficulty in Lu, Wei, Chen, Song, Cai, Qi, and Chu was because he lived in a dark time, and the feudal
lords were not able to put his ideas into practice….Even if he had not written the Spring and Autumn, he
would still have died without having met his proper time” [孔子之困於魯、衛、陳、宋、蔡、齊、楚
者，其時暗，諸侯不能行也。。。雖不作春秋，孔子猶不遇而死也] (LZYJ 21.808). As for Sima Qian,
he “struck at the emotions of the Emperor”[觸天子喜怒] (namely, by speaking up for the out-of-favor Li
Ling and—as it was thought at the time—secretly criticizing the Ershi general, of whom Emperor Wu was
fond).
116
As Song commentator Fan Rulin 樊汝林 wrote, “Li Han said himself that when he was collecting up
posthumous papers he left nothing out. Yet he put this piece outside of the literary collection proper. Why
throw it away just because Liu Zongyuan once disputed and criticized it?” [李漢自謂收拾遺文無所失
墜，乃逸此篇于正集之外，豈以其甞為子厚所辨駁而遂棄歟？] (Han Changli 5.474). Li Han (jinshi
813) was Han Yu‟s student and son-in-law. As the above quotation implies, after Han Yu‟s death, Li Han
was the one responsible for collecting and organizing his literary collection.
117
“Liu K‟o,” 151.
118
Bernard Solomon attempted to do just that in the introduction to his translation of the present text (see
Solomon, Veritable Record, xvi-xxii).
119
Hartman, T‟ang Search for Unity, 78.
120
Ibid.

366
Han Yu‟s work played a role in Emperor Wenzong‟s criticism of it, quoted above. As

Bernard Solomon points out, the Jiu Tangshu “Traditions of Lu Sui” states that “When

Han Yu compiled the „Veritable Records of Shunzong,‟ he spoke quite frankly of palace

affairs. The eunuchs disliked it and had frequently criticized its inaccuracies in front of

the emperor; in reign after reign edicts were issued ordering its revision” [初，韓愈撰順

宗實錄，說禁中事頗切直，內官惡之，往往於上前言其不實，累朝有詔改修].121

Thus, possibly influenced by the eunuch‟s displeasure, the emperor was said to have

complained (in the Xin Tangshu anecdote above) about the “Veritable Records of

Shunzong” to Zheng Tan, using Sima Qian as a metaphor.

By referring to Sima Qian‟s resentment of Emperor Wu, the Emperor Wenzong of

the anecdote is implying that, perhaps because Emperor Xianzong sent Han Yu into exile

over the Buddha relic incident, 122 Han Yu resented him. This resentment then caused

Han Yu to write the “Veritable Records” of Xianzong‟s father in such a ways as to

suggest that Xianzong had come to power through less than fully legitimate channels.

This is surely a distortion; during the time he was compiling the “Veritable Records of

Shunzong,” Han Yu seemed to have been very optimistic about the new reign—and of

course the Buddha relic incident had not yet occurred. However, one can see how

superficially apt the comparison might seem.

Zheng Tan, showing the brilliance one often manages to display in one‟s own

standard history biography, turns the comparison around, suggesting that Sima Qian did

not make mistakes in writing of Emperor Wu‟s reign, that in fact Sima Qian‟s criticism of

121
JTS 159.4193; trans. Solomon, Veritable Record, xvii, slightly modified.
122
For details of this incident, which occurred in 819, see Hartman, Han Yü and the T‟ang Search for Unity,
84-85.

367
his emperor was justified, and that Emperor Wu had exhausted the realm with frontier

wars. One wonders if some oblique allusion to Wenzong‟s grandfather Xianzong‟s

campaigns against renegade military governors (fanzhen 藩镇) was intended. The upshot

of these campaigns was a temporary stabilization of the realm, but also an unfortunate

increase in eunuch power which by Wenzong‟s reign had seriously weakened the throne.

Even in the late (or, depending on the source for this anecdote, post) Tang period,

we see the Shiji figuring in conflicts between ruler and minister, just as in the Six

dynasties and before. An emperor (here Wenzong) attacks Sima Qian for daring to

criticize his ruler; a wise minister-figure defends Sima Qian, claiming that the criticism of

Emperor Wu was true and justified.

SONG DYNASTY DEVELOPMENTS

In the following section, I focus on two Song dynasty developments in the

discourse surrounding the Shiji‟s reliability. The first is a highly influential essay by Su

Xun. The second is the Southern Song debate between Zhu Xu and thinkers of the

Zhedong school.

Su Xun‟s “Discussion of History”

Su Xun begins by arguing that the writing of histories and Classics was in both

case motivated by the same concern about petty people, and thus their “principle is the

same” [史与經皆憂小人而作，其義一也].123 Furthermore, they (and all writing) have

four aspects: events, words, the Way, and a rule. But the forms of histories and Classics

123
Jiayou ji 9.227.

368
differ in that in histories, events and words are most prominent, whereas in the Classics,

the Way and the rule are. He then goes on to develop the ways in which the histories and

the Classics are different but mutually dependent for their proper functioning.

If the Classics were not accompanied by history, there would be nothing to serve
as evidence for their praise and blame. If the histories were not accompanied by
the Classics, there would be nothing with which to measure their importance. A
Classic is not the true record of a single age, and a history is not a constant rule
for ten thousand generations. In their form, they do not imitate each other, but in
their use and substance, they complement each other. [經不得史無以證其褒
貶，史不得經無以酌其輕重；經非一代之實錄，史非万世之常法，体不相
沿，而用實相資焉。]124

There was nothing new in saying that the histories needed the Classics as a kind of moral

standard against which to measure and correct their judgments. But it was rather bold to

say that the Classics needed the histories as well.

The argument that Classics needed histories is not entirely unprecedented, of

course. What is generally considered the first narrative history, the Zuozhuan, was

widely believed to have arisen as a commentary on the laconic Spring and Autumn

Annals, written in order to clarify such of the Masters‟ oral teachings as had become

obscure and confused through disagreements between different disciple lineages.125 But

rarely are a Classic and its commentary considered to have the kind of separate but equal

124
Ibid., 9.229.
125
The Shiji story of the Zuozhuan‟s creation, which is the only one that has been passed down, is as
follows:
The seventy disciples received the meanings of the [Spring and Autumn Annals] tradition [from
Confucius]. Since there were patterned words that satirized and ridiculed, praised and obscured, and
impugned, they could not make them plain by writing them out. The Lu gentleman Zuo Qiuming
feared that the various disciples, differing in their biases, would be content with their own opinions and
lose what was genuine. Therefore, taking Confucius‟ scribal records as his basis, he put in order all
their words and completed the Chunqiu of Master Zuo. 七十子之徒口受其傳指，為有所刺譏裦諱
挹損之文辭不可以書見也。魯君子左丘明懼弟子人人異端，各安其意，失其真，故因孔子史記
具論其語，成左氏春秋。(SJ 14.509-510; trans. Schaberg, A Patterned Past, 318).
For an excellent discussion of the Zuozhuan‟s authorship, see David Schaberg‟s A Patterned Past, 315-324.

369
relationship that Su Xun seems to be describing. Particularly radical was Su Xun‟s

argument, pushing the point that the Classics were not true records.

Sometimes the Classics use false death announcements in their recording, or for
reasons of concealment and taboo do not record something. There are many cases
like this, but it is all only for the purpose of transformative education. Therefore I
say, “A Classic is not the true record of a single age.” [經或從偽赴而書，或隱諱
而不書，若此者眾，皆适于教而已。吾故曰：經非一代之實錄。]126

Here we see that Su Xun was employing a factual (rather than moral) notion of „true

record.‟ This comes closest to Sima Guang‟s gloss mentioned in the discussion of the

term „true record‟ above, namely, a true record and nothing more.

Su Xun goes on to employ the metaphor of a vessel (the histories) and the

measuring tools (the Classics) used to make the vessel: neither is useful without the other:

Now a compass, a square, a level, and a marking-line are things that determine a
vessel. The vessel is what waits and is made correct. However, if there is no
vessel, then the compass will have nothing on which to impose its roundness. The
square will have nothing on which to employ its squareness. The level will have
nothing on which to make manifest its evenness. And the marking-line will have
nothing on which to arrange its straightness. The histories wait for the Classics in
order to be made correct. But without the histories, the Classics would be obscure.
Therefore I say, “Their forms are not connected, but the pragmatic value [of the
one] and the real essence [of the other] assist each other.” [夫規矩准繩所以制
器，器所待而正者也。然而不得器則規無所效其圓，矩無所用其方，准無所
施其平，繩無所措其直。史待經而正，不得史則經晦。吾故曰：体不相沿，
而用實相資焉。]127

Even more extreme, Su Xun even proposes that the best of the histories—Sima Qian‟s

Shiji and Ban Gu‟s Hanshu—are not merely true records, but also (like the Classics)

“both have the Way and model” [亦兼道與法而有之]. Thus, “at times they measure up

to Zhongni‟s inherited intentions” [時得仲尼遺意焉].128 Su Xun then goes on to show

126
Jiayou ji, 9.229-230.
127
Jiayou ji, 9.230.
128
Ibid., 9.232.

370
subtle patterns of meaning-making within both works which make them seem to function

in ways similar to the Classics, especially the Spring and Autumn.

The growth of this notion, that historical works (and especially the Shiji) could

potentially challenge the Classics, blossomed into what might almost be called „open

conflict‟ during the Southern Song.

The Shiji in the Zhedong School‟s Conflict with Zhu Xi

In the Southern Song, a new development also occurred which would have a

profound impact on Sima Qian‟s reputation as a „reliable‟ historian, at least in the short

term. This was the role played by the Shiji in the conflict between the Zhedong school

and Zhu Xi. I will argue that although certain thinkers of the Zhedong school did place a

serious emphasis on the Shiji and other historical texts, Zhu Xi deliberately exaggerated

this emphasis for rhetorical purposes of his own. The degree to which he felt threatened

by the Zhedong thinkers, especially after the death of Lü Zuqian, led him to attack the

Shiji (and particularly its reliability) much more violently than he might otherwise have

been inclined to do. The prestige of Zhu Xi and the eventual adoption of his opinions as

state orthodoxy in turn caused the Shiji‟s reputation to suffer a considerable decline.

Zhu Xi was of course a towering figure in the Southern Song intellectual scene.

His philosophy, known as Daoxue 道學 or Lixue 理學 (and in English as Neo-

Confucianism) is of such complexity that even a basic outline is far beyond the scope of

this study. One important feature of his methodology, however, is particularly relevant

here: he advocated for primary study the new sub-canon that he developed, immortalized

in Chinese tradition thereafter as the Four Books 四書. These included the Lunyu, the

371
Mengzi, the “Great Learning” [Daxue 大學], and the “Doctrine of the Mean” [Zhongyong

中庸]. In general, the course of study Zhu recommended placed heavy emphasis on the

Classics, and relegated the histories to a distant secondary role.

In this, he differed very self-consciously from the thinkers associated with the

Zhedong school.129 Just what the Zhedong school did advocate is not especially easy to

pin down. They did not seem to have the kind of coherent and unified program of study

that Zhu Xi had, or if they did it has not been preserved in any convenient, detailed form.

One of their better known principles, however, is “learning for the sake of results” [事功

之學], which Hoyt Tillman has understood (at least in the case of Chen Liang) as

“utilitarian Confucianism.” An important aspect of this principle was to place a greater

emphasis on the study of history.

This is not to say that the Zhedong thinkers forsook the Classics entirely. Ye Shi

葉適 (1150-1223), a major representative of the school, expressed the vision of

complementarity very well when he wrote:

The writing of a historian takes events as the warp and the model as the weft. The
writings of sagely wisdom [i.e., the Classics] take principle as the warp and words
as the weft. One should base [one‟s learning] on the sagely wisdom, and
supplement it with history. [以事為經，以法為緯，史氏之文也。以理為經，
以言為緯，聖哲之文也。本之聖哲而叅之史。]130

129
The strand of thought now known as the Zhedong school [浙東學派] was a complex association of
various thinkers generally from the Zhejiang area, each in fact classified under their own geographically
named “schools.” These include Lü Zuqian and his brother Lü Zujian 呂祖儉 (d.1100), representatives of
the Jinhua 金華 School; Xue Jixuan 薛季宣 (1134-1173), Chen Fuliang 陳傅良 (1137-1203), and Ye Shi
葉適 (1150-1223), representatives of the Yongjia 永嘉 School; and Chen Liang 陳亮 (1143-1194) as
representative of the Yongkang 永康 School. For a detailed study of the Zhedong School, see He
Bingsong‟s 何炳松 Zhedong xuepai suyuan 浙東學派溯源 [Origins of the Zhedong school]; see also Hoyt
Tillman, Utilitarian Confucianism.
130
Shuixin wenji 水心文集, “Preface,” 2B.

372
Ye Shi‟s version of the complementary roles played by Classics and histories recalls Su

Xun‟s “Discussion of History,” which proposed a similar scheme. Perhaps Ye Shi

represents a conservative strand of the school‟s thought however.

A more threatening challenge to those like Zhu Xi—who would uphold the

primacy and orthodoxy of the Classics—was an account of early intellectual history such

as the one proposed by Chen Fuliang 陳傅良 (1137-1203)131:

After the capture of the unicorn, Mengzi and Xunzi revered Kongzi and
clarified the main points of the rites and duties. When these two masters died, the
hundred thinkers increasingly disordered what was true. Old ru like Fu Qiubo132
and Fu Sheng, here and there in the interval between Qin and Chu, carried the
Classics and privately protected them. Yet their strength was not adequate to
develop or elaborate on earlier thought. After the Han had passed its first sixty or
seventy years, Great Officer Dong [Zhongshu] first investigated the great
enterprise. Tian He, Kong Anguo, Dai Sheng, Dai De, and Master Mao all
appeared. Each had that which they compiled, but still they were not able to
harmonize the various writings into one, or to cut out what did not harmonize in
order to preserve the harmony. This was Senior Archivist Tan‟s intention.
However, the discussion of the Six Schools still reveres Laozi and looks down on
the ru. [獲麟以後，孟荀推尊孔氏，明禮義之統紀，二子死百氏益亂真，老
儒如浮邱伯、伏生之徒，區區於秦楚之際抱經自[守]而其力不足以發揮前
緒。至漢六七十年間，董大夫始究大業，田何、孔安國、戴聖、戴徳、毛莫
並出，各有[所]著而又未[能]合群書爲一，削其不合以存其合者，太史談有
意矣。然六家之論猶崇老抑儒。]
Qian completed the learning of his family, taking all the essence of the
hundred schools and judging it according to the Six Arts. For the Changes, he
based [his understanding] on Tian He; for the Spring and Autumn, he based it on
Dong Zhongshu; for the Revered Documents he based it on Kong Anguo; for the
rites, he based it on [King Xia]of Hejian. The only regrettable thing is that he had
not seen the Mao Odes. In general, he merges together the various strands [of
thought], gathering them into a single bundle, and rejecting miscellaneous
discussions. From the Annals of the Five Emperors on down, it is profoundly
well-founded. After Xun Qing, there is only this book. [遷卒家學，乃盡百家之

131
Chen Fuliang, with his friend and associate Xue Jixuan 薛季宣 (1134-1173), founded the so called
Yongjia 永嘉 School. He was recommended for the position of Overseer in the National University
(taixuelu 太學錄) in 1176, and also served as Prefect (知軍) of Guiyang Prefecture in Hunan starting in
1188. He resigned over political disagreements in 1193-94, and never served in office again, despite
receiving imperial summons to do so.
132
A disciple of Xunzi‟s, probably lived until the time of Han Emperor Wen, but was not employed at court
due to the then-current emphasis on Huang Lao.

373
 精而斷以六藝。易本田何，春秋本董仲舒，尚書本孔安國，禮本河間，獨恨
不見毛氏詩耳。盖其融液九流，萃為一篇，罷黜雜論，自五帝紀以下盛有依
據。荀卿之後，僅見此書爾。]133

Chen hints, without saying so explicitly, that the intellectual division of the Warring

States period did serious harm to the integrity of the Classics and to the Confucian

intellectual transmission they ought to contain. The Han dynasty classicists (Dong

Zhongshu, Tian He, Kong Anguo, Dai Sheng, Dai De, and the Mao brothers) went some

way toward repairing the damage, but they were not able to bring Classical learning

together and create a fully harmonized philosophical synthesis. In Chen Fuliang‟s view,

it was Sima Tan who formed the intention of doing so, and Sima Qian who actually

succeeded, at least in a preliminary way.134 He thus creates a direct lineage from Kongzi,

through Mengzi and especially Xunzi, down to Sima Qian.

In the same piece, Chen Fuliang also gave his own response to Ban Gu‟s criticism

of Sima Qian‟s faults, and the historical fate of the Shiji, writing:

Unfortunately, Ban Gu saw that he could not surpass [Sima Qian], and falsely
[claimed] that [Qian] had serious errors. Later scholars followed [his view], until
eventually it became a fixed belief. For fifteen hundred years, this book [i.e., the
Shiji] has been in obscurity. It is truly up to our group to open its own eyes, and
not be confused by successive arguments. [惜自班固看渠不過，妄有瑕摘，後
生沿習，遂成牢談。千五百年之間，此書湮晦。正頼吾黨自開隻眼，不惑於
紛紛之論。]135

It is probably anachronistic to suggest that Ban Gu believed he could not surpass Sima

Qian. Ban Gu himself, and most readers for several centuries, did consider Ban Gu to be

the superior writer and historian. But Chen Fuliang‟s analysis does perhaps have a grain

of truth in it. Because so much of the Hanshu was borrowed from the Shiji, Ban Gu had
133
Zhizhai wenji, 35.9A.
134
Because he followed a variant Odes tradition, often identified with the Lu school, rather than the Mao
tradition that ultimately won out, his understanding of the Odes was not seen as satisfactory by orthodox
Classicists.
135
Zhizhai wenji, 35.9A.

374
to find some fault with his predecessor in order to justify his own labors. Chen Fuliang‟s

suggestion was mainly that Ban Gu‟s criticism was taken too seriously by later readers.

It is also true that Chen seemed to propose a re-evaluation of the Shiji, at least within his

own circle.

Another Zhedong figure, Chen Liang 陳亮 (1143-1194),136 also denied Ban Gu‟s

criticism of the Shiji and placed Sima Qian in the company of the most serious Western

Han thinkers:

The book of Master Jia [=Jia Yi], the three policies of Dong Zhongshu, Sima
Zichang‟s [Sima Qian] record of the successive dynasties, and Liu Gengsheng‟s
[=Liu Xiang] transmitting of the Five Phases—they were all useful to the
[government] of their times, and were not contrary to the Sage. Certainly they
were already different from the various masters [of the Warring States]. [賈生之
一書，仲舒之三䇿，司馬子長之記歴代，劉更生之傳五行，其切於世用而不
悖於聖人，固巳或異諸子矣。]137

It is worth noting that this list was put together by Liu Zongyuan (see above). Liu,

however, seemed to have been placing greater emphasis on prose style. When Chen

Liang praised Shiji, he did not see it merely as a stylistic model, but as an important

political tool.

In the foregoing discussion, I have tended to emphasize the Zhedong interest in

the Shiji. In fact, this is surely something of an exaggeration. It is true that Chen Fuliang

seemed to have fairly high praise for the Shiji. However, the above letter, from which I

quoted extensively, was written to one Jia Duanlao 賈端老 (dates unknown), apparently a

student of Chen Fuliang‟s, whose name has survived only because more famous people

136
Chen Liang was a Southern Song poet and thinker. He was born in Yongkang 永康 (modern Zhejiang
province, Jinhua county). In the late 1170s he attempted the examinations several times but did not
succeed. He was a friend of the great lyric poet Xin Qiji 辛棄疾 (1140-1207). For more on Chen Liang
and his thought, see Tillman, Utilitarian Confucianism.
137
Longchuan ji, 9.8A-B.

375
addressed a few pieces to him. He can hardly be considered an influential thinker of his

day. Nor does Chen Liang‟s remark about Sima Qian not being contrary to the Sage

probably deserve the weight my discussion seems to assign to it: it was made in the

context of a piece whose explicit purpose is to argue that “Yang Xiong‟s thought

surpasses that of the various masters” [揚雄度越諸子] (this is in fact the title of the

essay). Chen Liang‟s purpose in mentioning the Western Han figures is largely to point

out an upward trend, of which he considers Yang Xiong to be the culmination.

Any investigation of the Zhedong school‟s views on the Shiji must confront the

curious fact that the richest source of information seems to be Zhu Xi. “The scholars of

Zhejiang promote and revere the Shiji,” he said. “They consider that „putting Huang Lao

first and the Six Classics last‟ was just from the learning of Senior Historian Tan. As for

Qian, he entirely revered Kongzi” [浙間學者推尊史記，以為先黃老，後六經，此自

是太史談之學。若遷則皆宗孔氏].138 Elsewhere, Zhu commented regretfully:

“Bogong [=Lü Zuqian] and Ziyue [=Lü Zujian] revered the study of the Lord Grand

Scribe [=Sima Qian], and believed there was not a Han dynasty classicist who could

compare with him” [伯恭子約宗太史公之學，以為非漢儒所及].139 Most extreme,

Zhu Xi ranted that:

Bogong incites and encourages people to look to the Zuozhuan and [Sima] Qian‟s
Shi[ji], which allows Ziyue [=Lü Zujian] and various others to elevate Sima Qian
without any sense of proportion, just as if [he] were about the same as Kongzi!”
[伯恭動勸人看左傳遷史，令子約諸人睺得司馬遷不知大小，恰比孔子相
似！]140

138
ZZYL 122.2956.
139
Ibid., 122.2951.
140
Ibid.

376
One can guess that the “various others” are meant to include Chen Fuliang and perhaps

Chen Liang as well.

Under these circumstances, the modern historian might be tempted to follow He

Bingsong, who suggested that Zhu Xi‟s remarks, however disapproving, represent a true

portrayal of the Zhedong school‟s interest in the Shiji:

After the Southern Song, the historical study of the Zhedong school flourished
greatly. At that time it came to the point that a Daoxue scholar reviled Zhe
learning by saying that they knew Shi Qian and did not know there was a Kongzi.
One can see from this that its flourishing was at one point quite extreme. [自南宋
以後，浙東史學大興。當時道學家至詬浙學為知有史遷而不知有孔子。其盛
極一時之情形，即此可見。]141

The problem with this interpretation is that if the Zhedong school really thought as highly

of the Shiji as Zhu Xi‟s words suggest, one would expect to find more evidence for that in

their own writings.

I argue for a different interpretation, one which depends on a historically

contextualized understanding of the Shiji‟s reputation during the Song. For various

reasons which will be explored below, scholars today are inclined to see the Shiji as the

“timeless masterpiece” and “monument of Chinese culture.” Serious thinkers in the Song

dynasty did not tend to hold that view. Instead they saw the Shiji as a well-written and

useful source of information but also potentially untrustworthy, riddled with mistakes,

and dogged by its reputation for being a slanderous and anti-Confucian book. One might

admire its style, but hesitate to subscribe to all of its views. To be sure, Chen Fuliang

was proposing a revised perspective on the Shiji. But it was not Chen Fuliang, or any of

the Zhedong thinkers, who made Sima Qian the mascot of Zhedong thought. It was Zhu

Xi himself.

141
Zhedong xuepai, 4.

377
 Zhu Xi must have found the “utilitarian” tendencies of the Zhedong school deeply

threatening. In the Zhuzi yulei, he juxtaposes it with the Jiangxi focus on chan (zen)

Buddhism:

The learning of Jiangxi is nothing more than zen, while what the learning of
Zhe[dong] specializes in material gain. As regards the study of chan, the scholars
who come after will try to get a feel for it and [soon discover] that there is nothing
to get a feel for, and of themselves will turn back. But as for material gain,
scholars can practice it, and can even see some results. So this idea [i.e., Zhedong
utilitarianism] is something one really has to worry about. [江西之學只是禪，
浙學卻專是功利．禪學後來學者摸索一上，無可摸索，自會轉去．若功利，
則學者習之，便可見效，此意甚可憂！]142

These theories of material gain (utilitarianism) were associated with the Zhedong

school‟s increased emphasis on history. Thus Zhu Xi used his rhetorical powers to attack

the study of history as a whole:

Our master said: “Looking at history is like looking at people hitting each other.
What is so great about watching people hitting each other? Chen Tongfu‟s
[=Liang] entire life has been ruined by history.” [先生說：「看史只如看人相
打，相打有甚好看處？陳同甫一生被史壞了．」]143

It should be noted that Zhu Xi‟s overall intellectual production does not support the idea

that he had a sustained hostility toward historical study. He himself compiled the Zizhi

tongjian gangmu 資治通鑑綱目 [Outline of the comprehensive mirror for government],

which would come to be officially considered, at least for a time, the best history after

that of Confucius.

Instead, as seen in the quotation about Chen Liang above, Zhu Xi considered that

the study of history was an inappropriate place to start, and as something that could be

ruinous to those who lacked a thorough grounding in the Classics (in particular, the Four

Books). He elaborated on this in criticizing Lü Zuqian‟s emphasis on the study of history:

142
ZZYL 123.2967.
143
ZZYL 123.2965.

378
 It is really impossible to understand this idea of his. If someone is seeking for
what is constant rather than what is particular, one does not dare encourage that
scholar to look to history. One also does not dare encourage that scholar to look
to the Classics. One does not even dare instruct him to look to the [Lun]yu and the
Meng[zi], but rather orders him to look to the “Great Learning.” [他此意便是不
可曉．某尋常非特不敢勸學者看史，亦不敢勸學者看經．只語孟亦不敢便教
他看，且令看大學．]144

It really seems as if Zhu Xi‟s opposition to history, and in particular to the Shiji, became

more extreme in reaction to his quarrel with the Zhedong school. Regarding the Lü

brothers‟ particular fondness for Sima Qian, Zhu added,

I once argued bitterly with them [about this]. Ziyou [=Su Zhe] in his Gushi said of
[Si]ma Qian, “He is shallow and vulgar, and not really learned. He is careless, and
gullible.” These two sentences are superlative in pinpointing [Si]ma Qian‟s
failings, and Bogong [Lü Zuqian] disliked them in the extreme. [某嘗痛與之辨．
子由古史言馬遷「淺陋而不學，疏略而輕信」．此二句最中馬遷之失，伯恭
極惡之．]145

Elsewhere, Zhu went further, specifically attacking Sima Qian.

[Sima] Qian‟s education included theories of benevolence and duty, and theories
of trickery and force, the employment of political strategy, and the employment of
material gain. However, his fundamental ideas only had to do with political
strategy and material gain. Kongzi said of Bo Yi: He sought benevolence and got
benevolence; what further was there to resent? But this chapter [in the Shiji] is
resentful words from beginning to end, completely slandering Bo Yi! Zi You
[=Su Zhe] in Gushi cuts all of this out, and instead uses only Kongzi‟s words to
make [Bo Yi and Shu Qi‟s] “Traditions.” How could one possibly say that Zi
You was the one who did wrong, while [Si]ma Qian was in the right? It is such a
pity that Zi Yue has died. This argument [between us], even to his death, did not
get clarified. Sages and worthies handed down their instructions to us by means of
the Six Classics, which shine like a colorful painting. There is no theory in them
that is counter to benevolence, duty, the Way, or virtue. Now to seek for
principles, going not to the Six Classics but rather to the careless and shallow Zi
Zhang [=Sima Qian], is the extreme of wrong-headedness! [遷之學，也說仁
義，也說詐力，也用權謀，也用功利，然其本意劔只在於權謀功利．孔子說
伯夷「求仁得仁，又何怨」！他一傳中首尾皆是怨辭，盡說壞了伯夷！子由
古史皆刪去之，盡用孔子之語作傳，豈可以子由為非，馬遷為是？可惜子約

144
ZZYL 122.2951.
145
ZZYL 122.2951.

379
 死了，此論至死不曾明！聖賢以六經垂訓，炳若丹青，無非仁義道德之說．
今求義理不於六經，而反取疏略淺陋之子長，亦惑之甚矣！]146

Again, it is important to realize that these arguments against Sima Qian tend to

appear only in the context of Zhu Xi‟s debate with the Zhedong school. If one sets aside

these debates, one might be surprised to find literally dozens of positive or neutral

references to the Shiji in other parts of the Zhuzi yulei. In an argument with Lü Zuqian,

Zhu Xi exclaimed in exasperation when the other brought in a quotation from the Shiji:

“How is Sima Qian adequate evidence!?” [司馬遷何足證]. Yet he himself frequently

used the Shiji as evidence when explaining things to his own students.147

Zhu Xi gave a reasonable explanation of his position in his correspondence with

Lü Zujian. Though Lü‟s letter has been lost, it was clearly a pastiche of arguments in

favor of the Shiji, for Zhu Xi‟s letter consists largely of refutations and criticisms.

However, at the end he wrote:

[Sima Qian‟s] text has several tens of thousands of characters; how could he be
without advantages? However, in discussing his greater aim, if Archivist Qian
were reincarnated and had to face the two sayings of Master Su, I fear that he
would not be able to explain and acquit himself. Now if one refuses to discuss
[Qian‟s] shortcomings, but ever emphasizes his good points, if one maintains that
he in no way deviated from the sage‟s intention, if one promotes and reveres him,
even coming to the point of comparing him favorably to the Six Classics, if upon
hearing there is someone who is arguing about his faults, one lets anger show in
both words and countenance, clenching one‟s fists and rolling up one‟s sleeves, as
if springing up vigorously and breaking the other fellow in two—then would I not
look upon such a person as having considerable shortcomings? [其書數十萬言，
亦豈無好處？但論其大旨，則蘇氏兩語，恐史遷復生不能自解免也。今乃諱
其所短，暴其所長，以為無一不合聖人之意，推尊崇奬，至與六經比隆，聞
有議其失者，則浡然見於詞色，奮拳攘臂，欲起而折之，一何所見之低矮
耶！]148

146
ZZYL 122.2952.
147
See, for example, ZZYL 11.195-196.
148
“Letter in Reply to Lü Ziyue” 答吕子約, Zhu Xi ji 朱熹集 2334.

380
Zhu Xi‟s opposition to the Shiji became more extreme because he felt that the Zhedong

thinkers were too extreme in their promotion of it. It seems clear that he also at times

over-emphasized their enthusiasm for the Shiji, by making such remarks as that they

“elevate Sima Qian without any sense of proportion, just as if [he] were about the same

as Kongzi.”

Like most critics who favored the Hanshu over the Shiji, Zhu Xi upheld Ban Gu‟s

criticism of Sima Qian. However, his real objections to the Zhedong approach went

deeper than that. In a more thoughtful moment, he gave perhaps the clearest

differentiation of his own concerns and theirs:

Now you must first rectify the beginning of the road, clearly distinguish
the separation between self and other…That way, your contemplation naturally
leads to comprehension; your knowledge naturally leads to clarity, and your
practice and conduct will naturally be correct. As the days accumulate and the
months pile up, you will gradually mature, and gradually attain naturalness. If
your view cannot fully penetrate, your road has gone astray at the beginning.
Then, although you may read a great number of books, and put the most diligent
daily effort into your writing, in the end you will not succeed in doing [important]
things. [今須先正路頭，明辨為己為人之別，直見得透…則思慮自通，知識
自明，踐履自正．積日累月，漸漸熟，漸漸自然．若見不透，路頭錯了，則
讀書雖多，為文日工，終做事不得．]
For comparison, look at our friends from the Zhe[dong school]. Some say
of themselves that they are able to fully comprehend the Zuozhuan, and some say
of themselves that they are able to fully comprehend the Shiji. They take Kongzi
and stand him against a wall, and yet they take the heterogeneous writings of
Master Zuo and Sima Qian and study them intensively, holding them in great
esteem. They say, here are the causes of flourishing and decline, here are the
roots of success and failure. [比見浙間朋友，或自謂能通左傳，或自謂能通史
記；將孔子置在一壁，劔將左氏司馬遷駁雜之文鑽研推尊，謂這箇是盛衰之
由，這箇是成敗之端．]
If you look at it from a different perspective, is it not your most crucial
responsibility to work on yourself? Your own self has so very very many aspects
that you have no choice but to attend to and understand, and so very very many
flaws that have not yet been weeded out. If you still come around talking about
all this flourishing and decline, rising and perishing, the well-governed and the
chaotic—this is no more than cheating yourself. [反而思之，干你身己甚事？你

381
 身己有多多少少底事合當理會，有多多少少底病未曾去，劔來說甚盛衰興亡
治亂，這箇直是自欺！]149

Zhu Xi here emphasizes an inward turn, while he characterizes the Zhedong thinkers as

being too closely focused on the outside world. It is important to note here that Zhu is

specifically talking about the “beginning of the road.” There might be nothing wrong

with studying history, but the Zhedong thinkers, in his way of thinking, put the cart

before the horse.

I have discussed Zhu Xi‟s response to the Zhedong thinkers in considerable detail

for the following reason: his debate with them seems to have had considerable influence

on his view of historical study, which in turn influenced the entire Chinese intellectual

milieu for centuries. Yet the context of Zhu Xi‟s Shiji critique was not evaluated or taken

into account by succeeding generations. Instead Zhu Xi‟s most biting rhetorical attacks

on the Shiji were taken at face value.

Furthermore, Zhu Xi eclipsed his rivals even (ironically) as a historian.

According to Benjamin Elman, in the context of the Ming examination system, Sima

Guang‟s 司馬光 (1019-1086) revival of chronologically organized histories with his

Zizhi tongjian 資治通鑑 [Comprehensive mirror for government]—and even more Zhu

Xi‟s 朱熹 (1130-1200) Zizhi tongjian gangmu 資治通鑑綱目 [Outline of the

comprehensive mirror for government]—cast a shadow on topically organized histories

like the Shiji and Hanshu. These histories were criticized for a variety of failings, but

first and foremost for departing from the Classics in genre and message.150 What we see

149
ZZYL 114.2757.
150
Elman, “History in Policy Questions.”

382
reflected in Ming examination questions and answers is not the most creative thinking on

the Shiji but rather restatements of an old debate. Still, the official line on the Shiji was

an important backdrop against which later developments must in turn be seen.

383
 PART III

TEXTUAL PROBLEMS AND SHIJI AUTHORSHIP

In the foregoing discussion, I have by and large avoided challenging traditional

assumptions about the integrity of the Shiji text and the authenticity of Sima Qian’s

“Letter in Reply to Ren An.” In this section, however, I examine various debates

surrounding these textual issues. In general, I do not attempt to weigh in on such debates.

Instead, I analyze the positions taken and the evidence employed, arguing that these

technical discussions contribute to Shiji interpretation and to the construction (or

deconstruction) of Sima Qian as an author figure.

How confident can we be that any given part of the Shiji was written by the

historical Sima Qian? The evidence we have to work with is ambiguous and sometimes

contradictory. There are reasons to doubt—and also to accept—Sima Qian’s authorship

of various passages; yet very little can be taken as conclusive. In Chapter 7, I attempt to

put this entire debate in historical context. The issues I address will be familiar: the

contributions of Sima Tan and Chu Shaosun; the ten lost chapters and post-Sima Qian

contributors to the Shiji, named and unnamed; and the idea of a drastically damaged Shiji

text. Here I am not concerned with the (in)authenticity of any particular part of the Shiji.

Instead, I am interested in the construction of Sima Qian as revealed through these

readers’ assumptions. I also consider how these textual debates influenced the

conception of the Sima Qian author function.

In chapter 8, I turn to one of the most crucial texts in Shiji interpretation, the

“Letter in Reply to Ren An.” Western scholars have tended to question the authenticity

of this “Letter,” while Chinese scholars are deeply pre-occupied with the complex issue

384
of the “Letter’s” dating. I discuss both types of debate. I also contrast the different

versions of the “Letter,” and analyze parallels between the “Letter” and other texts.

Ultimately, I conclude that one of the major issues at stake in all these debates over the

“Letter” is the problematic nature of its intended meaning and audience.

It is difficult to imagine Shiji interpretation without the “Letter” and the “Self-

Narration”—without the authorial figure of Sima Qian. It is not that without Sima Qian

Shiji interpretation would be impossible. Rather, the figure of Sima Qian (whatever

qualities the reader might impute to that figure) is there to be drawn upon. It is difficult

to see around him. Intended or not, the effect of a textually problematized Shiji (and

“Letter”) is potentially to do just that. How might our interpretation of the Shiji change if

its connection to the Sima Qian we know from the “Letter” were seriously undermined?

385
 Chapter 7

Multiple Authors, Damaged Text

This section explores three authorship issues that go beyond Sima Qian: the

problem of Sima Tan‟s possible authorship, the contributions of Chu Shaosun, and the

idea that the Shiji was badly damaged in the course of its transmission. All three have

bearing on how Sima Qian‟s authorial role has been conceptualized by readers.

Furthermore, these questions remain matters of debate even among scholars today.

In my discussion, I generally avoid making arguments about the authenticity or

authorship of any particular part of the Shiji. Rather I am concerned to analyze the

attitude toward authorship issues that emerges through the course of the debate, both

historically and in the present day.

THE SIMA TAN PROBLEM

No discussion of the Shiji‟s authorship would be complete without a consideration

of „the Sima Tan 司馬談 problem.‟ Édouard Chavannes was reasonably sanguine about

identifying Sima Tan‟s and Sima Qian‟s respective contributions.1 Chavannes suggests

two relatively simple criteria. First, “Daoism in the Shiji can only be blamed on Sima

Tan; it ought to have been him alone, not Sima Qian, that Ban Biao and Huan Tan

incriminated.”2 Second, “though the accusation of Daoism that has been brought against

Sima Qian is unfounded, a more deserved reproach against him is that he wrote a satirical

1
See Mémoires historiques I.xlvii-lxi.
2
“Taoïsme dans les Mémoires historiques, il faut n‟en accuser que Se-ma T‟an ; c‟est lui seul, et non Se-
ma Ts‟ien, que Pan Piao et Hoan T‟an auraient dû incriminer” (Mémoires historiques I.li).

386
book.”3 Developments in Shiji studies since Chavannes‟ time have shown us that neither

of his criteria is wholly satisfactory or easy to apply. Western Shiji scholars have by now

largely disavowed the project of trying to separate Sima Tan‟s and Sima Qian‟s

respective contributions.4 Some Chinese scholars have been more optimistic about the

issue,5 but the attempt to sort out the father‟s work from the son‟s runs aground on the

lack of reliable evidence. Zhang Dake aptly summed up the situation when he wrote:

There is no doubt that by the time Sima Tan lay dying, he was already leaving
behind a certain number of chapters or sections [of what would become the
Shiji].... However, Sima Tan‟s compositions have already been recast by Sima
Qian. In practice, it is impossible to search for chapters that were the result of
Sima Tan‟s historical composition, or to cut [his writings] apart from the overall
system of the Shiji. [司馬談臨終時，已留下若干篇章，這是沒有疑義
的。。。但司馬談之作，已為司馬遷所重新鎔鑄，考求司馬談作史的完整篇
目，割裂《史記》體系，則為事實所不容。]6

The fact that there are few or no reliable criteria for separating Sima Tan‟s work from

Sima Qian‟s has certainly not stopped scholars from trying. Even writing back in 1985,

Zhang Dake found it necessary to create a multi-page chart in order to summarize the

complex and mutually contradictory arguments that had been advanced. Nor has Zhang‟s

pessimistic conclusion laid the question to rest. 7

3
“Si l‟accusation de taoïsme portée contre Se-ma Ts‟ien est sans fondement, on lui a fait un reproche plus
mérité quand on a dit qu‟il avait écrit un livre satirique” (Mémoires historiques I.lii). Chavannes follows
this observation with a discussion of Shiji passages that place Emperor Wu in a negative light, a convenient
and concise example of the type of autobiographical reading whose development was the subject of
chapters 4-5 above.
4
Li Wai-yee can be taken as typical, when she writes “It is impossible to verify exactly which sections of
the [Shiji] were written by Sima Tan” (“Authority,” 371-372 n.49). Grant Hardy also tables the problem,
writing only that Sima Tan “had some, rather unclear, role in the creation of the Shiji” (Bronze and Bamboo,
xiii). There are exceptions however. One exception is E. Bruce Brooks, who in 2005 held a workshop on
this question. See http://www.umass.edu/wsp/conferences/wswg/21/index.html.
5
See, for example, Li Changzhi 李長之, Sima Qian zhi ren’ge yu fengge , 24-37.
6
Shiji yanjiu, 54.
7
See, for example, Li Fuyan‟s 李福燕 2004 article, “Jianlun Sima Tan dui Sima Qian de yingxiang” 簡論
司馬談對司馬遷的影響 [Brief discussion of Sima Tan‟s influence on Sima Qian].

387
 There is an entire range of attitudes toward Sima Tan. At one extreme, there is

the belief that his contributions were either minimal or so freely edited (“recast”) by Sima

Qian that whatever he did is more in the nature of a source than a collaborative effort. At

the other extreme, Sima Tan is presented as an overlooked genius whose fame has been

unjustly eclipsed by his son, and who may in fact have been responsible for a very

significant portion of the Shiji. No doubt there are elements of truth in both positions. In

the following sections, I will trace the development and variation of traditional readers‟

attitudes toward the Sima Tan problem. Where possible, I will show why and how Sima

Tan‟s contribution mattered to them (or failed to). My study will not attempt to advance

any particular theory of Sima Tan‟s authorship. Instead, I use the Sima Tan problem to

show something about authorship generally, how assumptions about it may shape a text.

Sima Tan’s Authorship in the Shiji

It makes sense to begin by briefly reviewing the Shiji‟s own most direct statement

on the matter. First, two passages from the narrative of Sima Tan‟s death-bed

conversation with his son suggest that Sima Tan‟s work on the Shiji may not have been

extensive or significant:

When I am dead, you must become a senior archivist. When you have become a
senior archivist, do not forget what I had intended to discourse upon and
make manifest. [余死，汝必為太史；為太史，無忘吾所欲論著矣。]8

Since the capture of the unicorn, it has been more than four hundred years.
Various lords [reigned] concurrently, and [as a result] archival recording was
abandoned or cut off. Now the Han has arisen, and all within the four seas are
united. Enlightened rulers, worthy lords, loyal ministers, gentlemen who would
die for rightness—I have been senior archivist and yet I have not discoursed
upon nor recorded them. What I fear most deeply is that the archival writings

8
SJ 130.3295, emphasis added.

388
 of the realm will be cast aside. You must remember this always! [自獲麟以來四
百有餘歲，而諸侯相兼，史記放絕。今漢興，海內一統，明主賢君忠臣死義
之士，余為太史而弗論載，廢天下之史文，余甚懼焉，汝其念哉！]9

On the other hand, we also find two further passages, one from the end of the same scene

(part of what is reported as Sima Qian‟s response) and one from later on the “Self-

Narration” (purportedly describing Sima Qian‟s commitment to fulfilling his father‟s

command). These two passages suggest that Sima Tan did do work on the Shiji prior to

his death:

Your small son is not intelligent, but he begs to discourse upon all the stories of
old that his ancestor has put in order. I will not dare to fail in this! [小子不
敏，請悉論先人所次舊聞，弗敢闕。]10

There is no greater crime [in this world] than to let fall that of which one’s
ancestor has spoken. [墮先人所言，罪莫大焉。]11

Finally, a single line describes Sima Qian‟s work on the project prior to the Li Ling affair:

“Thereupon [I, Sima Qian] discoursed upon and put in order its patterned words” [於

是論次其文].12

The first thing to say about all these passages is that caution is necessary in

deciding how much weight to give them. Stephen Durrant, for example, has argued that

“Sima Tan‟s deathbed admonitions… closely resemble the dramatic speeches delivered at

crucial moments thoughout the pages of [the Shiji].” Furthermore, given that the “Self-

Narration” was probably one of the last parts of the Shiji to be written, “Sima Qian‟s

memory [of his father‟s death] is inevitably molded by his own intervening experience,

and that experience demands that he provide the strongest conceivable justification for

9
Ibid., emphasis added.
10
Ibid., emphasis added.
11
SJ 130.3299, emphasis added.
12
SJ 130.3300, emphasis added.

389
being alive and speaking out at all.”13 Since Sima Qian “remember his own past and

reinterprets that past, like all human beings, through a haze of subsequent events… the

voice of Sima Qian‟s father, dying near Luoyang, unlike the voice of his philosophical

treatise, is at least partially Sima Qian‟s own voice.”14

If Sima Qian had wanted to specify which (if any) parts of the Shiji were written

by his father and which by himself, he could easily have done so. After all, there is no

ambiguity—at least in the text as we have it today!—regarding Sima Tan‟s authorship of

the essay “Essentials of the Six Schools,”15 nor for that matter about Jia Yi‟s authorship

of “Discourse on the Faults of Qin.”16 Given that there are no such indicators in the rest

of the Shiji, we are left with three possibilities.

1. Sima Tan wrote none of the Shiji, but only the “Essentials of the Six Schools”
essay.
2. Sima Qian (or some later editor) intentionally downplayed or hid the distinction
between Sima Tan‟s work on the text and Sima Qian‟s own.
3. The Shiji originally contained indicators of the Sima Qian‟s versus Sima Tan‟s
authorship, but they were accidentally lost or erased sometime during the process
of textual transmission.

The passages I cited above suggest the first or the second possibility: the phrase “what I

had intended to discourse upon and make manifest” [吾所欲論著] is actually quite literal

in specifying that the actions remained, for Sima Tan at least, in a counter-factual future.

The second passage is more difficult. Regarding “enlightened rulers, worthy

lords, loyal ministers, gentlemen who would die for rightness,” Sima Tan purportedly

said, “I have not discoursed upon nor recorded them” [弗論載]. Depending on context,

the passage leaves open the possibility that Sima Tan had completed the pre-Han portions

13
Cloudy Mirror, 8.
14
Ibid., 10.
15
SJ 130.3288 ff.
16
SJ 6.276 ff.

390
of the Shiji but not the Han portions. True, what directly precedes the “enlightened

rulers” etc. is a description of the Han unification. However, I would argue that the

persons named are to be seen as not only referring to persons of that sort in the Han era,

but in fact to all the people of that category after the “capture of the unicorn” (i.e., the end

of the Chunqiu). Furthermore, in today‟s Shiji, there is no indication that Sima Tan

envisioned the family project as encompassing the Spring and Autumn period, much less

high antiquity. He (reportedly) wanted his son to “continue the Chunqiu” [繼春秋];17

there is not a word about supplementing or rewriting it.18

In discussing the next passage, it is important to consider the specific verbs

involved. In the first passage, the words „Sima Tan‟ used were lun 論 (discourse upon)

and zhu 著 (make manifest), which I have understood as two separate—or at least

separable—actions. In the second passage, it was lun 論 and zai 載 (record), again

potentially separable. Both sets of verbs refer to things that Sima Tan ostensibly did not

do. In the third passage, what the Sima Qian of the narrative requests to do is again lun

論, but what he proposes to discourse upon is “the stories of old that his ancestor has put

in order” [先人所次舊聞]. The word xianren 先人 seems to be used by the Sima Qian

persona mostly to refer to Sima Tan, though it is possible to interpret it (as Burton

Watson has19) to mean the Sima family more generally.20 The verb used here, this time

17
SJ 130.3296.
18
For other arguments against a Qin-Han Shiji authored by Sima Qian and a pre-Qin Shiji authored by
Sima Tan, see Zhang Dake, Shiji yanjiu, 63.
19
In Watson‟s rendering: “the reports of antiquity which have come down from our ancestors” (Ssu-ma
Ch’ien, 50).
20
It is clear from context that xianren 先人 refers to Sima Tan in SJ 130.3299 (“I have heard my father
say…” 余聞之先人). In SJ 130.3319, it is clear that the same word refers to more distant ancestors
(“When I think of how my ancestors were in charge of these affairs…” 余維先人嘗掌斯事). The case on
SJ 130.3296 is ambiguous (where the xianren discusses the five hundred year sage cycle), with Sima Zhen

391
apparently referring to something that Sima Tan has done, is ci 次 (to order, arrange).

This would imply that Sima Tan had already done some arranging of old material,21 but

perhaps had not yet formally “discoursed upon” it.22

The fourth passage uses still another verb, yan 言 (to speak). It is possible to

understand this as an oral transmission of certain historical material from Sima Tan to

Sima Qian (possibly extending further back up the family line as well, owing to the

ambiguity of xianren). More likely, however, given the context, the thing that Sima Tan

had purportedly talked about was the project itself. The dialogue in question, after all,

has already been given in the “Self-Narration.”

Finally, the narrative states that Sima Qian lun 論 (discoursed upon) and ci 次

(ordered, arranged) “its patterned words” [其文]. I take “its” (qi 其) to refer to the Shiji.

As for wen 文, its meaning could and probably should be understood to encompass a

whole range, from carefully-written prose to structural organization to cultural principles.

We should not be troubled by the fact that Sima Tan is already said to have “arranged the

stories of old,” and Sima Qian further arranges it. A text as massive and heterogeneous

as the Shiji would clearly require a great deal of arranging and rearranging. What is more

interesting and perhaps significant is the emphasis on lun. Lun is what Sima Tan

reportedly said he had wanted to do but did not; lun is what Sima Qian is portrayed as

asking to do; and lun is what Sima Qian is reported to have done. If these passages

glossing it as “a worthy of the former age” 先代賢人 and Zhang Shoujie glossing it as referring to Sima
Tan.
21
Frequently linked to the tantalizing “writings of the stone chamber and metal casket” [石室金匱之書]
Sima Qian claimed to have consulted (SJ 130.3296).
22
See also Li Changzhi‟s discussion, Sima Qian zhi ren’ge yu fengge, 133.

392
reveal anything about Sima Qian‟s authorial claim as opposed to his father‟s, I argue it

has to do with the activity of lun.

Unfortunately, the character lun 論 is itself ambiguous. It could refer to two

different words: lùn, to debate, discuss, or evaluate; 23 or lún, to put in order or select.24

By convention, the term has been taken in the first sense and associated with so-called

discursive (as opposed to narrative) passages of the Shiji, especially the so-called

taishigong yue 太史公曰 (the Honorable Senior Archivist said) comments. This may in

part be because the term lunci 論次 actually appears elsewhere in the Shiji, tellingly

referring to Kongzi‟s work on the Odes and Documents.25 Did Kongzi, in Sima Qian‟s

view, merely select and arrange the Odes and Documents, or did he discuss and evaluate

them as well? In fact, the question of what lun means is not solved by this second use of

it, but only made more consequential. One thing is clear, however. Whatever activity is

being referred to by lun, the implication is that that activity is a profoundly important part

of the text‟s coming into being.

Early Readers and Sima Tan

A few traditional readers of the Shiji seem to have been confused even about the

one aspect of Sima Tan‟s oeuvre that seems clear to us today: his authorship of the

“Essentials of the Six Schools.” We find in Yang Xiong‟s Fayan the following dialogue:

Someone inquired [about the following]: “Sima Zichang has a saying that
the Five Classics cannot match the Laozi for conciseness, for „generations of

23
As in the modern Chinese compounds yilun 議論 (to debate or criticize) or pinglun 評論 (to discuss or
evaluate).
24
A variant of graphically related 倫 (order, ranking) or 掄 (to choose, select), most commonly associated
with the title of the best-known compilation of Kongzi sayings, the Lunyu 論語.
25
SJ 121.3115.

393
 scholars could not master their study, nor could a man in his whole lifetime
thoroughly comprehend all their rules.‟”26
[Yang Xiong] said, “If this be so, then the Duke of Zhou is benighted, and
Confucius an outlaw. The scholars of antiquity tilled the land to feed themselves,
but in three years could master one [Classic]. As for the scholars of today, it is
not only that they go in for flowery adornment, but follow that by further
embroidering their frills. How could it have anything to do with Laozi one way or
the other?”
Someone asked, “Then the explanations of scholars could be abridged?”
He said, “They could be abridged and still explain the subject.”
或問：「司馬子長有言，曰五經不如老子之約也，當年不能極其變，
終身不能究其業．」
曰：「若是，則周公惑，孔子賊．古者之學耕且養，三年通一．今之
學也，非獨為之華藻也，又從而繡其鞶帨，惡在老不老也？」
或曰：「學者之說可約邪？」
曰：「可約解科．」27

The anonymous interlocuter remarks that “Sima Zichang has a saying, that the Five

Classics cannot match the Laozi for conciseness” [emphasis added]. No such saying can

be found in the Shiji today. The two lines that follow, however, do come Sima Tan‟s

essay “the Essentials of the Six Schools.” Sima Qian, though ostensibly only the

transmitter of this essay, is given full credit (or blame) for the sentiments expressed

therein. Commentators were at pains to correct this. Wang Rongbao 汪榮寶 (1878-1933)

pointed out that this is Tan‟s essay and that Tan was a known Huang-Lao adherent.

Wang supplied various pieces of evidence to show that Sima Qian‟s own intellectual

predilections were quite otherwise. Wang also quoted Hu Yujin‟s 胡玉縉 (1859-1940)

argument that, based on the structure of Shiji, Sima Qian himself preferred ru to Huang-

Lao.28 Both seem to consider it a misattribution on the part of Yang Xiong and his

anonymous interlocutor, a slanderous mistake. But was it?

26
The passage in quotation marks is from the Shiji “Self-Narration,” for which I use the translation of
Burton Watson, Ssu-ma Ch’ien, 45.
27
FY 7.222.
28
Ibid.

394
 Sima Qian is certainly not held to be responsible for every shred of writing he

quotes in the Shiji. No one criticizes him as actually holding the opinions found in Li

Si‟s sycophantic memorials to the Second Qin Emperor, for example. But perhaps the

material found in the “Self-Narration” had a special status, or perhaps the fact that Sima

Qian claimed he was “establishing the words of one family” (成一家之言)29 made him

fully responsible for the literary products of that family. As remarked above, real interest

in Sima Tan‟s contribution to the Shiji is a relatively recent phenomenon. We might

conclude from this “misattribution” that even as early as Yang Xiong, the awareness of

Sima Tan as Shiji co-author had all but disappeared. Or perhaps we should conclude that

by anthologizing his father‟s essay in his own “Self-Narration,” Sima Qian was being

seen as taking responsibility for his father‟s views in addition to his own. The third, and

most unsettling, possibility is that the “Six Schools” essay was not, in Yang Xiong‟s time,

marked off as clearly as it is today.

I have already mentioned in chapter 3 Liu Xie‟s implication that Sima Tan was

the originator of the intention behind the Shiji: “Senior Archivist Tan grasped the

bamboo slips as his hereditary [duty]. Zichang [=Sima Qian] carried on [Sima Tan‟s]

intention, examining and putting in sequence the accomplishments of emperors” [太史

談，世惟執簡；子長繼志，甄序帝勣].30 The sense that the Shiji was a collaborative

effort between Sima Tan and Sima Qian was certainly present to Six Dynasties and Tang

readers.31 Indeed, the Sui Shu “Treatise on Classics and Records” seems to assign

primary responsibility to Sima Tan:

29
SJ 130.3319.
30
WXDL yizheng 16.573.
31
Zhang Dake also points this out; see Shiji yanjiu, 65.

395
 The office of Honorable Senior Archivist was first set up in the time of the Han
Emperor Wu, and he commanded Sima Tan to serve in this capacity, and thus he
performed this task. At that time, the written registers of the realm were all first
submitted to the Senior Archivist, and, next, to the prime minister. „Left behind‟
writings, ancient matters: there was nothing that did not finally arrive [at his
office]. Tan thereupon relied on the Zuoshi, Guoyu, Shiben, Chu-Han Chunqiu,
connected them with later matters, and it became the words of a single jia.32
When Tan died, his son Qian in turn became Senior Archival Director, inheriting
and completing [his father‟s] intention. Beginning with the Yellow Emperor and
ending in the Fiery Han, it brought together twelve benji, ten biao, eight shu,
thirty shijia, and seventy liezhuan. This is what is known as the Shiji. [漢武帝
時，始置太史公，命司馬談為之，以掌其職。時天下計書，皆先上太史，副
上丞相，遺文古事，靡不畢臻。談乃據左氏、國語、世本、 戰國策、楚漢
春秋，接其後事，成一家之言。談卒，其子遷又為太史令，嗣成其志。上自
黃帝，訖于炎漢，合十二本紀、十表、八書、三十世家、七十列傳，謂之史
記。]33

In Wei Zheng‟s view, the work of selecting both ancient and contemporary texts was

essentially all completed by Sima Tan: it was Tan who “established the words of a single

jia” [成一家之言], however we understand that phrase. Qian‟s role was to continue what

was primarily his father‟s project. By implication, Wei ascribes both the formal structure

and the choice of starting and ending points to Sima Qian however: these are not

described until after Tan‟s death and Qian‟s work have been mentioned.

As for Tang readers, Sima Zhen‟s statement about the authorship is curious. He

wrote, in the very beginning of his “Preface”: “The Shiji was transmitted by the Han

Senior Archivist Sima Qian, father and son” [史記者，漢太史司馬遷父子之所述也].

Supposing we did not know from other sources, this attribution would be ambiguous.

Would a more natural reading not be Sima Qian and his son? Why not “Honorable

Senior Archivists, father and son” [太史公父子]? Clearly it was important to Sima Zhen

that the Shiji‟s author be named, not merely referred to by title. But why did he not

32
It is unclear in this context whether jia means specialist, family, or school [of thought].
33
SuiS 33.959, compiled by Wei Zheng 魏徵 (580-643) et al.

396
employ the more conventional “Senior Archivist Sima Tan, father and son” [太史司馬談

父子]? The most likely answer seems to be that Sima Zhen wanted to emphasize Sima

Qian‟s contribution.

Liu Zhiji, on the other hand, tended to give Sima Tan slightly less credit:

In the era of Filial Wu, the Honorable Senior Archivist Sima Tan desired to
weave together the ancient and modern, tying it up into a single history. His idea
was not yet realized when he died. His son Qian thereupon transmitted the
intention his father left behind, selecting from Zuozhuan and Guoyu… [孝武之
世，太史公司馬談欲錯綜古今，勒成一史，其意未就而卒。子遷乃述父遺
志，採《左傳》、《國語》。]34

The passage ends in a way very similar to Wei Zheng‟s description above, with a list of

sources and a description of the Shiji‟s structure. To Wei Zheng, Sima Tan was the one

who selected from the various sources, but Liu Zhiji implied that it was Sima Qian.

In short, readers of the Shiji up through the Tang acknowledged that the Shiji was

a collaboration between Sima Tan and Sima Qian. However, they disagreed as to the

extent of the collaboration, what part, if any, Sima Tan played beyond being the

originator of the project. They did not, however, seem to acknowledge that there was a

disagreement. In the Northern Song dynasty, Su Xun 蘇洵 (1009-1066) pointed out the

difficulty as part of his critique of the Shiji:

The “Self-Narration” says, “Tan became the Lord Grand Scribe.” It also says,
“The Lord Grand Scribe encountered the Li Ling disaster.” This is making no
difference between his own title and his father‟s. An earlier Ru said that, to the
contrary, [Ban] Gu suppressed [his father Ban] Biao‟s reputation, and was unlike
[Sima] Qian who yielded to and honored [Sima] Tan.35 But I do not know if the
one [Sima] Qian calls the Lord Grand Scribe in the annals, the charts, the treatises,
the hereditary households, and the memoirs, is actually his father? Or is it himself?
These are [Sima] Qian‟s flaws. [其《自序》曰：談為太史公。又曰：太史公

34
STTS 12.337.
35
Possibly xian ru 先儒 should be understood in the plural (earlier Ru generally). However, the origin of
this criticism of Ban Gu appears to have been Fu Xuan (see chapter 1 above).

397
 遭李陵之禍。是與父無異稱也。先儒反謂固沒彪之名，不若遷讓美于談。吾
不知遷于紀、于表、于書、于世家、于列傳所謂太史公者，果其父耶抑其身
耶？此遷之失也。]36

Su Xun‟s observation seems to dove-tail well with the second of the hypotheses proposed

above, namely, that the distinction between Sima Tan‟s and Sima Qian‟s work on the

Shiji seems to be intentionally downplayed. Yet as Liu Zhiji and/or other earlier Ru have

observed, this was for the most part not done at Sima Tan‟s expense.

It is perhaps worthwhile to split the problem of Sima Tan‟s authorship into

several different aspects. First, there is the question of whose idea it was to create the

Shiji. Second, who did the selection of material from existing sources and of editing

those selections to create more or less new narratives based on them? Third, who wrote

the discursive (taishigong yue 太史公曰) sections? And fourth, who devised the sub-

genres of the Shiji and its overall structure? There seems little doubt that Sima Tan is

responsible for the first, and Sima Qian for the last.37 It is the second and third that are a

matter of contention.

Selection and editing of material might be seen as roughly corresponding to three

of the verbs discussed above: zhu 著, zai 載, and ci 次. Taking the “Self-Narration” at

face value, it seems as if Sima Tan and Sima Qian both engaged in these tasks. Therefore,

as Zhang Dake argues:

In the end, it was Sima Qian who completed the Shiji. Whatever Sima Tan wrote
or transmitted was, as far as Sima Qian was concerned, something that he selected
from, as he did from the Zuozhuan and Guoyu, or something he abridged, as he
did with the Shiben and Zhanguoce. Within Sima Qian‟s own final draft of the
Shiji, he would have cut and pruned, or wholly recast [Sima Tan‟s work]. [《史
記》最後完成於司馬遷之手，父談的著述，對於司馬遷來說就是同采擇《左

36
Jiayou ji 9.238.
37
Fang Bao does claim that Sima Tan invented all the sub-genres, but Zhang Dake doubts this claim. See
Shiji yanjiu, 59.

398
 傳》、《國語》，刪《世本》、《戰國策》一樣，剪裁鎔鑄在自己定稿的
《史記》之中。]38

This aspect of the Sima Tan problem, then, truly is reducible to the more general problem

of Sima Qian‟s use of sources.

In some sense, more is at stake with the discursive (taishigong yue 太史公曰)

passages. As mentioned above, these passages might be taken to correspond with what in

the “Self-Narration” is described as lun 論. If that is the case, then the “Self-Narration”

seems to claim these as Sima Qian‟s. Yet various readers, apparently beginning in the

Qing and continuing down to today, have been dissatisfied with that conclusion and

sought for signs of Sima Tan within the pages of the Shiji. Fang Bao is an interesting

example of this tendency. In two different short essays,39 Fang Bao claims that certain

instances of the title “Honorable Senior Archivist” (太史公) should be understood as

referring to Sima Tan.

When Qian made prefaces for the ten tables, only the “Twelve Lords” (Shiji
ch.12), “Six States” (Shiji ch.15), “Qin-Chu Transition” (Shiji ch.16), and “Lords
from Hui to Jing” (Shiji ch.19) say, “The Honorable Senior Archivist read...”
This refers to things that [Qian‟s] father wanted to discourse upon. This is [also]
why in the “Gaozu‟s Meritorious Followers” (Shiji ch.18) it says, “I read” in order
to differentiate it. [遷序十表，惟《十二諸侯》、《六國》、《秦楚之際》、
《惠景間侯》者稱「太史公讀」，謂其父所欲論著也；故《高祖功臣》稱
「余讀」以別之。]40

When the Shiji generational tables say, “The Honorable Senior Archivist read…”,
[Sima Qian] was referring to his father. Thus, when he is referring to himself, he
says, “I read...” in order to differentiate it. In the other Treatises and Traditions,
when something in the beginning or middle of a chapter is marked with “The
Honorable Senior Archivist says...”, this is merely Chu Shaosun‟s mistaken
[addition or change]. Thus, if one in general removed those four characters from

38
Ibid., 65.
39
I.e., “After Copying the Shiji‟s Ten Tables” [書史記十表後] and “After Copying Again the Honorable
Senior Archivist‟s Self-Narration” [又書太史公自序後].
40
Fang Bao ji 2.48.

399
 within of these chapters,41 the writing still connects up quite correctly. [史記世表
曰：「太史公讀」者，謂其父也；故於己所稱，曰「余讀」以別之。其他
書、傳篇首及中間標以「太史公曰」則禇少孫之妄耳；故凡篇中去此四字，
文正相續。]42

Fang Bao‟s hypothesis at first seems quite arbitrary and unsupported. It is important,

though, to consider the context of the two above-quoted passages: they belong to the

“after copying” genre. Explicit evidence either for or against Fang Bao‟s claim was not

available. He offered instead an argument implicitly based on the expertise and

familiarity gained through the practice of hand-copying a text. The implication is that

Fang Bao had acquired special sensitivity to nuances of style or structure, a sensitivity

that the more casual reader would lack.

In the early twentieth century, Li Changzhi made a much more systematic attempt

to differentiate Sima Tan‟s work from Sima Qian‟s. Based on three principles, he

identified eight chapters that he thought came from Sima Tan‟s hand.43 First, he

compared Shiji chapters with what he took to be the philosophical slant of the “Six

Schools” essay. Second, he assigned to Sima Tan eyewitness accounts which he took to

be too early for Sima Qian to have actually experienced them. And finally, he considered

that if a chapter failed to taboo the character tan 談, then it must have been written by

41
I understand 篇中 to refer to all parts of the chapter excluding the final comments, but there is a slight
ambiguity here.
42
Fang Bao ji 2.60.
43
I.e., the “Xiao Jing benji” 孝景本紀 (ch.11), “Lü shu” 律書 (ch.25), “Jin shijia” 晉世家 (ch.39), “Laozi
Han Fei liezhuan” 老子韓非列傳 (ch.63), “Cike liezhuan” 刺客列傳(ch.86), “Li Si liezhuan” 李斯列傳
(ch.87), “Li sheng Lu Jia liezhuan” 酈生陸賈列傳 (ch.97), and “Rizhe liezhuan” 日者列傳 (ch.127).

400
Sima Tan.44 Of these criteria, Zhang Dake has dismissed the first as “completely

untenable” [根本就不能成立],45 pointing out that

The intention to model the Shiji after the Chunqiu was set by Sima Tan…so who
can say that Sima Tan did not revere Confucianism? The “Lament for Scholars
who did not Meet their Time” [a poem attributed to Sima Qian] is heavily colored
with Daoism, so who can say that Sima Qian did not revere the Dao? There
might have been some differences between Sima Tan‟s and Sima Qian‟s thought,
but basically they were in agreement… bringing together and selecting the best
from Ru, Mo, names, laws, Dao, and yin-yang—all of the hundred schools. [《史
記》效《春秋》，這一宗旨為司馬談所定。。。誰說司馬談不尊儒？《悲士
不遇賦》充滿道家色彩，誰說司馬遷不尊道？談、遷父子思想有差異，但基
本一致。。。兼采儒、墨、名、法、道陰陽之長。]46

The third, too, is problematic: as Zhang Dake noted,47 the observance of taboos in

Western Han writings was sporadic.48 Furthermore, it is even more likely that later

editors would fail to observe taboos and „correct‟ some instances of taboo in the text

(while failing to correct others).

The second criterion, however, requires more background. As Zhang Dake has

pointed out, many of these supposed problems with dating (which cause some scholars to

assign chapters to Sima Tan) result from the theory that Sima Qian was actually born in

135 BCE, not 145 BCE as is traditionally believed.

44
Zhang Dake discusses a longer list of criteria based on a broader sampling of scholarship; see Shiji yanjiu,
59-60. I limit myself to Li Changzhi‟s criteria here because they are the most widely accepted and
representative arguments made with regard to Sima Tan‟s possible authorial role.
45
Zhang Dake‟s reasonable stance on this question does not by any means enjoy universal acceptance
from Western scholars. See, for example, a footnote in Wai-yee Li‟s “Idea of Authority” where she argues
that “the sympathies expressed in [ch.86] are more characteristic of Ssu-ma Ch‟ien than of his Taoist
father” (372 n.49).
46
Shiji yanjiu, 63.
47
He cites Hu Shi‟s 胡適 Xi Han ren lin wen bu hui kao 西漢人臨文不諱考 [Investigation into Western
Han People‟s Failures to Observe Taboos in Writing] and Chen Yuan‟s 陳垣 Shi hui ju li 史諱舉例
[Examples of Taboos in Historical Works].
48
Shiji yanjiu, 61. Note that Dirk Bodde also discussed the implication of the Tan taboo in China’s First
Unifier, 101-111.

401
 I will not here go into detail regarding the technical aspects of the argument over

when Sima Qian was born, since this has been done by others.49 Instead, I want to take a

broader perspective on this issue. The disagreement seems to have arisen already with

the Tang commentators. Sima Zhen‟s Shiji suoyin implies that Sima Qian was born in

135 BCE, while Zhang Shoujie‟s Shiji zhengyi corrects it to 145 BCE.50 Wang Guowei

and Liang Qichao both accepted Zhang Shoujie‟s version for a variety of textual

reasons.51 This has not prevented scholars from continuing to promote the 135 BCE

theory. The interesting question to ask about this tendency is not whether or not it is

well-founded, but rather, what do its proponents gain from it?

One of the most significant ramifications, as it turns out, is that if Sima Qian were

born in 135 BCE, Sima Tan‟s role in the creation of the Shiji would be much greater and

better defined. For various reasons, certain scholars began to find it desirable to see Sima

Tan as having played a greater authorial role in the creation of the Shiji. But what is the

difference between Sima Tan‟s authorship and Sima Qian‟s?

There are two distinct Sima Tans, as Stephen Durrant has pointed out.52 There is

Sima Tan the author of the “Six Schools” essay, and then there is the Sima Tan who was

said to have died of resentment and who charged his son with the completion of what

would become the Shiji. The difference between Sima Qian and that second Sima Tan

does not seem very significant, at least as far as Shiji interpretation is concerned. Casting

the first Sima Tan as a major author of the Shiji, however, could potentially have

profound implications for our understanding of how Han dynasty intellectual history in

49
For an brief review of the dating argument, see Zhang Dake, Sima Qian pingzhuan, 19-21.
50
For both comments, see SJ 130.3296. (Sima Zhen‟s comment is found in nt.1 and Zhang Shoujie‟s in
n.4.)
51
See Zheng Hesheng, Sima Qian nianpu, 9.
52
Cloudy Mirror, 6.

402
general is portrayed in the Shiji. Unfortunately, the argument is both circular and

speculative: chapters are assigned to Sima Tan based on their supposed Daoist bent, and

then that same Daoist bent is contrasted with Sima Qian‟s intellectual orientation. In fact,

we know very little about Sima Qian‟s own attitude toward Daoism, and there is no solid

argument to be made about it based on (or in support of) Sima Tan‟s authorship.

Interest in Sima Tan‟s authorship reveals an underlying desire on the part of some

readers to create interpretive distance between the Shiji and the Sima Qian author-

function. This tendency is also evinced, though with different characteristics, in the role

played by Chu Shaosun.

THE ROLE OF CHU SHAOSUN

If Sima Tan‟s authorial role in the Shiji was by and large seen in terms of a

differing ideological/philosophical orientation, Chu Shaosun‟s role was almost

exclusively that of scapegoat. As Burton Watson has written, “Chu took upon himself

the thankless task of making additions and continuations to the Shiji, [which undertaking]

has won him nothing but condemnation from all later commentators.”53 The traditional

low opinion of Chu has actually changed somewhat in the past century. Yu Jiaxi pointed

out that, contrary to the traditional opinion, Chu Shaosun was not such a bad stylist.54

Timoteus Pokora considered Chu to be the “third author of the [Shiji]”, and had planned

to write a book on the subject (unfortunately never completed). Dorothee Schaab-

Hanke‟s article, “Did Chu Shaosun Contribute to a Tradition of the Scribe?” presents Chu

53
Ssu-ma Ch’ien, 226.
54
“Taishigong wangpian kao,” 69.

403
as someone who adopted and elaborated on “ideas or even a kind of doctrine” from what

she calls the “tradition of the scribe” as found in the Shiji.55

In my discussion, I first give a brief introduction to Chu Shaosun and his “signed”

additions to the Shiji. I then examine what I call “the Chu Shaosun author-function,”

analyzing the role Chu Shaosun has played in Shiji authorship issues.

Chu Shaosun’s Life and Works

According to the Hanshu, Chu Shaosun56 studied the Odes (Shi 詩) with Wang

Shi 王式 (Western Han, dates unknown). Wang was at one point a tutor of Liu He 劉賀

(d.59 BCE), the Prince of Changyi 昌邑王 and also short-lived successor to Emperor

Zhao. When Liu He was deposed in favor of Emperor Xuan57 (r.73-49 BCE), Wang Shi

narrowly avoided a death sentence in the resulting purge of Liu He‟s followers. Facing

accusations of immorality by association, Wang Shi argued that he had in fact attempted

to remonstrate with his royal charge.58 Sometime after these events (which took place in

74-73 BCE), Chu Shaosun and others went to study with Wang Shi, though there is a

suggestion that their “studies” may not have amounted to much. The Hanshu account

writes, “When they asked about some chapters of the [Poetry] Classic, [Wang] Shi

excused himself, saying, „What I have heard from my teacher is all just this. Take it upon
55
“Tradition of the Scribe,” 11. Schaab-Hanke provides a valuable close reading the Chu Shaosun material,
using it to argue that Chu Shaosun‟s “reading” of the Shiji promotes an underlying political message,
namely that “the Han dynasty might not have received Heaven‟s mandate” (23). Schaab-Hanke argues
that this message was also intended by the Shiji authors.
56
I have neglected to provide dates for Chu Shaosun because his dates are in no way certain. Timoteus
Pokora estimates 104 BCE-30 BCE (“Narrator of Stories,” 430), while Yu Zhanghua, Shiji yanjiu jia, gives
his dates as 71 BCE-20 BCE (SJYJJC 13.397). Yi Ping‟s examination of events and posthumous names in
Chu‟s comments suggests that he perhaps ceased his additions to the Shiji around 47 BCE (“Chu Shaosun,”
163-170).
57
With, incidentally, the timorous consent of Sima Qian‟s son-in-law, Yang Chang 楊敞 (d. 73 BCE), and
the more active enthusiasm of Sima Qian‟s daughter (see HS 66.2889).
58
HS 88.3610.

404
yourselves to add embellishments.‟ He was not willing to teach them any more” [問經數

篇，式謝曰：「聞之於師俱是矣，自潤色之．」不肯復授].59

Nonetheless, when Chu received the post of Erudite Disciple, he and his fellow

student Tang Changbin 唐長賓 behaved with perfect command of their demeanor and

etiquette during their court appearance. They then commended Wang Shi as their teacher,

with the result that the old man was summoned to court. Unfortunately, the standing

Erudites humiliated Wang over a disagreement in Odes interpretation and he returned

home, where he ended his days. Meanwhile, Chu went on to become an Erudite on the

strength of his Spring and Autumn studies, and also founded one of three branches of the

Lu Odes tradition active at that time. But though the branches founded by his fellow

students thrived in different ways, nothing more is heard of Chu Shaosun except for what

we can infer from his comments on the Shiji.

What exactly is Chu‟s Shiji material, and what can it tell us? Chu‟s signed

contributions to the Shiji are appended to ten different chapters:

1. “Sandai shibiao” 三代世表 (SJ 13)

2. “Jianyuan yilai houzhe nianbiao” 建元以來侯者年表 (SJ 20)
3. “Chen She shijia” 陳涉世家 (SJ 48)
4. “Waiqi shijia” 外戚世家 (SJ 49)
5. “Liang Xiao Wang shijia” 梁孝王世家 (SJ 58)
6. “San Wang shijia” 三王世家 (SJ 60)
7. “Tian Shu liezhuan” 田叔列傳 (SJ 104)
8. “Kuaiji liezhuan” 滑稽列傳 (SJ 126)
9. “Rizhe liezhuan” 日者列傳 (SJ 127)
10. “Guice liezhuan” 龜策列傳 (SJ 128)

59
HS 88.3610. My translation is informed by, but differs from, that of Pokora in “Narrator of Stories,” 406,
which translates the entire section on Wang Shi, as well as that on Shaosun‟s great uncle Chu Da 禇大. For
the sense of Wang Shi‟s difficult remark, I have followed Yan Shigu‟s explanation (HS 88.3611).

405
Later commentators attributed to him a number of other chapters as well, which will be

discussed below.

Chu Shaosun’s Claims to Authority and Access

Shiji scholars have been intrigued by the question of Chu Shaosun‟s access to the

Shiji text. Which version did he have, and how much of today‟s Shiji did it contain?

How closely did it resemble today‟s Shiji? Chu Shaosun‟s few remarks on the matter

cannot fully answer these questions, but they are worth considering. Together with his

claims about access to the Shiji text, Chu also makes certain claims about his own

worthiness as a successor to Sima Qian. These claims are almost universally disregarded,

but they too are worth reviewing.

The first thing to note is that in eight of the ten chapters he supplemented, Chu

refers to his service as a palace gentleman (lang 郎).60 Three of these references describe,

in slightly varying terms, the means by which Chu was fortunate enough to attain this

position: “due to my aptitude in classical learning” [得以文學],61 or again, “due to my

canonical mastery” [得以經術].62 The third reference is more elaborate and seems to

outline his entire academic career: “Because of my penetration of canonical mastery, I

was charged with the work of an Erudite. I mastered the Spring and Autumn, and

because I was ranked first, I became a palace gentleman. I came and went in the palace

60
These references can be found on the following pages: SJ 13.507, 49.1981, 58.2089, 60.2114, 104.2779,
126.3203, 127.3221, 128.3225.
61
SJ 60.2114.
62
SJ 126.3203.

406
for more than ten years” [臣以通經術，受業博士，治春秋，以高第為郎，幸得宿

衛，出入宮殿中十有餘年].63

These references to Chu‟s service in the palace, and the abilities that led to it,

seem to be making a claim for Chu‟s intellectual authority, his credibility as an erudite

writer. The other five references describe people with whom he spoke in order to obtain

or confirm the information he presents in his comments, all prefaced by the reference to

his position as a marker of time: “When I was a palace gentleman…” [臣為郎時]. These

frequently repeated references suggest that Chu‟s position as a palace gentleman was the

major source of his authority for his work on the Shiji.

A second thing to note, perhaps more important, is Chu Shaosun‟s relationship to

the Shiji as seen through these comments. Did he have access to an imperial copy of the

Shiji? And was that copy complete during his time? Scholars such as Yu Jiaxi 余嘉錫,

Zhang Dake 張大可, and Yang Haizheng64 have suggested that due to his position as a

palace gentleman, Chu did have access to the imperial text. Meanwhile, Pokora argues

that although Chu “surely had access to the Palace…there is nothing to prove that he

could enter even the forbidden parts” and read carefully controlled texts like the Shiji.65

Yi Ping argues similarly, concluding, first, that Chu lacked access to the palace copy, and

second, given that there was no sign of a relationship between the modestly positioned

Chu and the much more exalted Yang Yun, Chu probably did not have access to Yang

63
SJ 128.3225.
64
See Zhang, Shiji yanjiu,192; Yang, Han Tang, 11. Yu makes no explicit statement to this effect, but the
assumption is implicit in his argument that Chu‟s additions were attached directly to the imperial edition.
65
“Narrator of Stories,” 426. Although whether or not the text really was carefully controlled may be open
to question: see my discussion in chapter 1 above.

407
Yun‟s family edition either. Instead, Yi Ping argues, Chu would have relied on scattered

copies circulating “among the people.”66

The strongest evidence for Chu having access to the palace copy is juxtapositional.

In his appendix to Shiji chapter 60, Chu wrote, “As I was fortune enough to have become

a serving gentlemen due to my aptitude in classical learning. I was fond of reading and

surveying the Arrayed Traditions of the Honorable Senior Archivist” [臣幸得以文學為

侍郎，好覽觀太史公之列傳].67 And again in chapter 128, Chu writes: “I came and

went in the palace for more than ten years, and was fond of the „Traditions of the

Honorable Senior Archivist‟” [出入宮殿中十有餘年，竊好太史公傳].68 These

references, to “Taishigong zhi liezhuan” 太史公之列傳 [Arrayed Tradition(s) of the

Honorable Senior Archivist] and to the “Taishigong zhuan” 太史公傳 [Tradition(s) of

the Honorable Senior Archivist], are sometimes been taken to refer to the whole of the

Shiji, but as Yi Ping has pointed out, in both chapters 60 and 128, the reference is

followed by a quotation, and both these quotations (with variants) are today to be found

in Shiji 130, the “Taishigong zixu” 太史公自序 [Honorable Senior Archivist‟s Own

Postface]. It is possible (though not necessary) that in mentioning the “Traditions of the

Honorable Senior Archivist,” Chu is referring only to the “Postface.”69

Of course we know that Chu did have access to other chapters of the Shiji. His

continuation of Shiji 20 begins with the words, “The Honorable Senior Archivist‟s

66
“Yang Yun,” 34.
67
SJ 60.2114.
68
SJ 128.3225.
69
Yi Ping (“Yang Yun,” 34-35) makes much of the fact that Chu‟s quotations do not exactly match today‟s
Shiji, arguing that this proves Chu‟s version was a vulgar popular edition. However the differences are not
as stark as he makes them out to be. In one case (SJ 60.2114, cf. SJ 130.3312), there is some difference of
opinion as to where the quote begins and ends (see note below), and the other is an extremely minor textual
variant (SJ 128.3225, cf. SJ 130.3318).

408
recording of events ends with the affairs of Emperor Wu,”70 certainly implying that Chu

had seen the chart to which he explicitly appended his own comment and continuation.

Of the six stories he wrote concerning ironical critics in Shiji 126, he clearly stated, “I put

them in as an additional appendix to the above part which is three stories of the

Honorable Senior Archivist” [以附益上方太史公之三章].71 It is therefore fair to

conclude that he had seen that chapter as well.

More fascinating to scholars are the chapters which Chu claimed to have searched

for but was unable to find, namely chapters 60 and 128. For the former, Chu writes, “The

Tradition [i.e., SJ 130] says, „The Hereditary Household of the Three Kings, their

patterned words can be surveyed.‟ I looked for this Hereditary Household, but in the end

was unable to obtain it” [傳中稱三王世家文辭可觀，求其世家終不能得].72 Does this

mean that it was missing from the palace copy already by Chu‟s time? In other words,

where did Chu look for it? Chu‟s comment on chapter 128 adds another layer to the

mystery, for there he writes, “I have gone back and forth within Chang‟an seeking the

„Guice liezhuan,‟ but I was unable to obtain it” [臣往來長安中，求龜策列傳不能得].73

70
Yi Ping points out that in today‟s Shiji, the last four entries in Emperor Wu‟s reign are preceded by the
notation, “On the right [i.e., the foregoing] is the Honorable Senior Archivist‟s own table” 右太史公本表
(SJ 20.1058). He suggests that Yang Yun was therefore responsible for the last four entries (“Yang Yun,”
36-38), but it is only a speculation.
71
SJ 126.3203.
72
SJ 60.2114. Scholars have pointed out that at least as far as the transmitted Shiji is concerned, this is a
misquotation. (The Shiji reads: 三子之王，文辭可觀 [SJ 130.3312].) Yi Ping even uses this to suggest
that Chu‟s version of Shiji 130 was a corrupt “popular” edition rather than the palace edition (see
discussion above). Initially, this would seem a plausible explanation, especially since the logic of Chu‟s
comment would then be: 1) the Tradition says that the chapter can be surveyed but 2) I was unable to
survey it. However, if Chu did have the entirety of the “table of contents” found in today‟s Shiji 130, he
should have noticed the formal anomaly—in all cases, the entry ends with the title of the chapter rather than
starting with it. To me it seems much more likely that the phrase should be punctuated 傳中稱三王世家
「文辭可觀」，求其世家終不能得 [The Tradition says of the Hereditary Household of the Three Kings,
“their patterned words can be surveyed.” I looked for this Hereditary Household, but in the end was unable
to obtain it].
73
SJ 128.3226.

409
Those who argue that Chu read the imperial copy might interpret this to mean that, not

finding the chapter in the archives, Chu went to seek it among acquaintances, or in the

marketplace. Those who argue that Chu did not read the imperial copy might take this as

evidence that Chang‟an was where Chu also sought Shiji 60, and all the other chapters he

did succeed in finding (with the possible exception of Shiji 130, which does seem to have

been available to him at the palace).

To further complicate the story, it is here, with chapters 60 and 128 (as well as

chapter 127 for reasons that will soon become clear), that Chu Shaosun‟s additions to the

Shiji become entangled with the tortuous arguments regarding the so-called ten missing

chapters of the Shiji.74 Ban Gu comments in Hanshu 62 that “ten chapters are missing,

having titles but no texts” [十篇缺，有錄無書].75 To the great regret of later scholars,

he did not specify which ones they were. In the third century, Zhang Yan 張晏 provided

a list of these chapters, which is quoted in Pei Yin‟s 裴駰 (fl.438) Shiji jijie 史記集解

commentary.76 The three chapters mentioned above appear on Zhang Yan‟s list, and

furthermore, Zhang even writes that:

In the reigns of Emperors Yuan and Cheng, Master Chu filled in the gaps and
made the “Annals of Emperor Wu”, the “Hereditary Household of the Three
Princes”, the “Traditions of the Tortoise-shell and Divining Straws”, and the
“Traditions of the Diviners of Lucky and Unlucky Days”. His words and phrases
are superficial and inferior, and he goes against Qian‟s original intentions. [元成
之閒，褚先生補闕，作武帝紀，三王世家，龜策﹑日者列傳 ，言辭鄙陋，
非遷本意也．]77

74
The most extensive study of the ten missing chapters is Yu Jiaxi‟s “Taishigong wangpian kao” 太史公亡
篇考 [Investigation of the missing chapters of the Honorable Senior Archivist].
75
HS 62.2724.
76
SJ 130.3321. The list is also reproduced on HS 62.2724-25.
77
SJ 130.3321.

410
Had Zhang Yan explained where he got his information, we might possess the solution to

a great puzzle in the history of the Shiji, or at least be closer to solving it. As it is, Zhang

Yan‟s statement is cast into doubt by the unpalatable fact that all four chapters mentioned

above contain comments attributed to the Honorable Senior Archivist.

For now I will leave aside the “Annals of the Present Emperor” (Shiji ch.12)

which are extracted from the relevant section of the “Treatise on the Feng and Shan”

(Shiji ch.28), but show no specific evidence that Chu was the one who did the extracting.

Of the remaining chapters, Chu could not find chapter 60 or chapter 128, as mentioned

above. In each case, he discusses the sources he used to compile a replacement.78

However, his discussions all appear after discussions attributed to the Honorable Senior

Archivist. For Shiji ch.127, the “Traditions of the Diviners of Lucky and Unlucky Days,”

Chu does not mention being unable to find the text, and even makes a reference79 that

suggests he was familiar with the text we now have, or something very like it (this neither

proves nor rules out the hypothesis that Chu wrote that text).

We are left with a problem. In Zhang Yan‟s time, the chapters were obviously

missing and/or attributable to Chu Shaosun. Liu Zhiji 劉知幾 (661-721), however, still

attributed chapters 127 and 128 to Sima Qian.80 Furthermore, by Lü Zuqian‟s 呂祖謙

(1137-1181) time, the Shiji was more or less in the state we find it today: Lü Zuqian

made the argument that none of the chapters were missing but were merely left

incomplete or in rough draft form by Sima Qian. Today, opinions are divided, according

to whether or not the scholar in question wants these chapters to be part of Sima Qian‟s

78
SJ 60.2114, 128.3226.
79
SJ 127.3221.
80
See Yu Jiaxi, “Wangpian,” 74.

411
oeuvre. The challenge for those who want to believe in their authenticity is to explain

when and how they “re-emerged”—to prove they were not simply forged (or written

from the “Honorable Senior Archivist‟s” point of view in homage to Sima Qian, as Yu

Jiaxi‟s more generous reading would have it81)—and to explain the considerable

discrepancy between the chapters as written and the small prefaces ostensibly describing

them in chapter 130.82 The challenge for those who want to believe that Chu Shaosun

supplied the chapters (and necessarily did so before Ban Gu‟s time) is to explain why

Ban Gu made no mention of Chu‟s work on the Shiji, seeming to know him only as a

scholar of the Lu 魯 school of Shijing 詩經 [Classic of odes] interpretation.83

If Chu Shaosun wrote supplements to the Shiji, he must have written supplements

to some particular version. In various places, as mentioned above, he wrote that he was

appending his comments to Sima Qian‟s chapter. Those who believe that Chu had access

to the edition in the imperial archives must then assert that Chu, without any known

official directive, nonetheless had the authority to tamper with the text there—adding,

moreover, texts that fully express his self-confessed fondness for decidedly non-

canonical authors.84 Those who believe that Chu did not have access to the imperial

edition, but was working from some copies of scattered chapters that he was able to

obtain, must then explain how Chu‟s comments made their way into the official version

of the Shiji.

81
“Wangpian,” 82-83.
82
Yu Jiaxi, in his detailed argument that no part of these chapters comes from Sima Qian‟s hand, made full
use of this discrepancy. In the case of Shiji ch.128, he used textual evidence to suggest that the
“Taishigong yue” section must also have been written after Chu Shaosun‟s time, and that it displays
familiarity with Chu Shaosun‟s supplement (“Wangpian,” 77-79).
83
It is not until Zhang Yan that we find specific mention of Chu Shaosun‟s supplements (see HS 62.2725).
84
SJ 126.3203.

412
 One complicating factor, not to be missed, is the fact that, without acknowledging

Chu Shaosun‟s work, Ban Gu apparently made use of it in three places.85 Yu Jiaxi has

reviewed various theories of how Chu Shaosun‟s version came to be attached to the

Shiji.86 One of these, also adopted by Timoteus Pokora,87 is that the history of the Chu

family was written (it is quoted by Pei Yin) and became available around the time of

Zhang Yan—for Zhang had much more information about Chu Shaosun than was

evinced by Ban Gu, despite the fact that Ban Gu lived much nearer Chu Shaosun‟s time.

Perhaps it was the increased attention on the Chu family at that time that only then led to

Chu Shaosun‟s additions actually being included in the Shiji.

Chu Shaosun as Author-Function

I will not here analyze in detail Chu Shaosun‟s contributions to the Shiji. Instead,

I will consider what I call “the Chu Shaosun author-function”: the way Chu as an author

was perceived by readers of the Shiji. There seems to have been a tendency for Chu‟s

perceived contribution to increase over time, though the change was not the result of any

new evidence. Rather, I will argue, it came about because of changing attitudes towards

Sima Qian’s authorship, and the perceived characteristics thereof.

In the previous section, I already mentioned Zhang Yan‟s statement (quoted by

Pei Yin) that Chu Shaosun had filled in four of the ten missing chapters.88 Pei Yin refers

to this comment again at the beginning of Shiji ch.12:

85
According to Yu Jiaxi‟s note, the Hanshu “Traditions of Emperor Wu‟s Five Sons” includes material
from Chu‟s “Hereditary Household of Three Kings”; the Hanshu “Traditions of Wei Qing” makes use of
Chu‟s material on the Lord of Pingyang; and the Hanshu “Traditions of Imperial Relatives” makes use of
Chu‟s material on the Prince of Xiucheng. See “Wangpian,” 85.
86
“Wangpian,” 104-108.
87
“Narrator of Stories,” 430.
88
SJ 130.3321.

413
 The Honorable Senior Archivist‟s “Self-Narration” says “I made the basic annals
of the present emperor.” Also, in recounting events [Honorable Senior Archivist]
invariably says, “the present emperor” or “the present son of heaven.” Wherever
one sees the words “The Filial Emperor Wu”, it has been fixed by some later
person. Zhang Yan said, “The annals of [Emperor] Wu were a supplement
created by Master Chu. Master Chu‟s personal name was Shaosun, and he was an
Erudite of the Han.” [太史公自序曰「作今上本紀」，又其述事皆云「今
上」，「今天子」，或有言「孝武帝」者，悉後人所定也．張晏曰：「武
紀，褚先生補作也．褚先生名少孫，漢博士也．」]89

This passage raises two issues. First, Pei Yin proposes that the use of “the Filial Emperor

Wu”—as opposed to “the present emperor” or “the present son of heaven”—should be

considered an indication that Sima Qian was not the author. While Sima Qian, despite

his mutilation, might have outlived Emperor Wu, the “Self-Narration” does seem to have

been written during Emperor Wu‟s reign (for the reason Pei Yin outlined above). Since

“Wu” was a posthumous name not determined until after the emperor‟s death, any

mention of it would have to be an interpolation of some kind, though on what scale is

impossible to determine.

The second point to be made concerns Chu Shaosun‟s involvement with the

twelfth chapter of the Shiji. Of the ten „missing‟ chapters of the Shiji, chapter 12 is the

only one which is truly missing.90 The current chapter 12 begins with a brief introduction

to Emperor Wu‟s background: that he was a middle son of Emperor Jing, but became the

heir when the previous heir was set aside. The passage ends with the line, “From the

time the filial Emperor Wu first assumed the position [of ruler], he paid special reverence

to sacrifices for ghosts and divinities” [孝武皇帝初即位，尤敬鬼神之祀].91 This same

89
SJ 12.451.
90
Martin Kern has argued that the “Treatise on Music” (Shiji ch.24) should also be considered to have been
entirely lost, since the current chapter is either wholly derivative or (in the few apparently original sections)
historically problematic (see Kern, “A Note on the Authenticity and Ideology of Shih-chi 24”). Certainly
the chapter is at best a borderline case, as discussed below.
91
SJ 12.451.

414
line is echoed in the “Treatise on the Feng and Shan” (Shiji 28), except that there it is

written as “From the time the current son of heaven first assumed the position [of

ruler]…” [今天子初即位],92 and from that point on, the two chapters are nearly identical.

There is no explicit indication in the text that it was Chu Shaosun who was responsible

for the current state of chapter 12, but no one seems to have doubted Zhang Yan‟s

assertion that this was so.

The use of the “Feng and Shan” chapter to substitute for the official annals of

Emperor Wu‟s reign has been largely condemned by scholars. Tang commentator Sima

Zhen, for example, complained:

Chu Shaosun in supplementing the Shiji, [should have] gathered together Emperor
Wu‟s affairs in chronological form. In fact, though, he merely selected from the
“Treatise on the Feng and Shan” to fill it in. Truly his talent was meager. [褚先
生補史記，合集武帝事以編年，今止取封禪書補之，信其才之薄也．]93

Later readers also thoroughly condemned the chapter wholesale.94 The only exceptions I

have been able to locate are Mao Kun 茅坤 (1512-1601) and Wu Jiansi 吳見思 (17th c.),

who appear to take the substitution as the work of Sima Qian himself.95 They imply that

Sima Qian‟s tragedy made him too cautious to give the full account of the reigning

emperor.

Aside chapter 12, which Zhang Yan explicitly attributed to Chu Shaosun, the

other major unsigned additions that came to be associated with Chu were events which

92
SJ 28.1384, emphasis added. The same kind of substitution is also found in one other place in the two
chapters:
“When the current ruler assumed the position… [及今上即位] (SJ 28.1384) versus “When the Filial Wu
assumed the position…” [及武帝即位] (SJ 12.453). There are also places, however, where the substitution
has not been made. See for example SJ 28.1403-04 versus SJ 12.485.
93
SJ 12.451.
94
See SJPL 12.247ff.
95
See SJPL 12.248 and Shiji lunwen, 17, respectively.

415
seemed too late to be plausibly attributable to Sima Qian. Again, it is in Sima Zhen‟s

commentary where we find mention of this:

Wei Ling96 said, “Chu Yi‟s family tradition [has] a Chu Shaosun, the grand-
nephew of Chu Da, prime minister of Liang…. [Chu Shaosun] was honored with
the title of „Master‟, and continued the Writings of the Honorable Senior
Archivist.” Ruan Xiaoxu97 also considered this to be correct. [韋稜云「褚顗家
傳褚少孫，梁相褚大弟之孫。。。宣帝代為博士，寓居于沛，事大儒王式，
號為『先生』，續太史公書」。阮孝緒亦以為然也。]98

What does it mean that Chu Shaosun supposedly “continued” the Shiji? It could mean

that he was merely one of the numerous Han figures known to have engaged in the same

activity. On the other hand, it might also be taken to mean that anything in the Shiji that

Sima Qian could not have written99 was authored by Chu Shaosun.

Pei Yin, who probably predated both scholars mentioned by Sima Zhen in the

above passage, understood the Shiji to contain continuations by various undetermined

hands. He wrote tersely in his commentary to the heading “The first year of the Taishi

reign period” [太始元年]100 in the last of the Shiji tables:

Ban Gu said, “Sima Qian‟s record of events went down to the Tianhan [reign
period].”101 Everything after that was continuations by later people. [班固云：
「司馬遷記事訖于天漢」，自此已後，後人所續。]102

96
Author of a Hanshu xuxun 漢書續訓 in three juan, according to the Suishu bibliographic treatise (SuiS
33.953). His dates are not known, but he was the son of Wei Rui 韋叡 (442-520). The latter has a chapter
in the Liangshu 梁書 (12.220ff.) to which the traditions of three of his sons, including Wang Ling, are also
appended.
97
Ruan Xiaoxu 阮孝緒 (479-536) was a Southern Dynasties (Nanchao 南朝) bibliographer and recluse
(see Liangshu 51.739ff.).
98
SJ 12.451.
99
This in itself is a controversial type of judgement given the uncertainty about the year of Sima Qian‟s
death.
100
I.e., 96 BCE. See SJ 22.1142.
101
I.e., 100-97 BCE. The quotation is found on HS 62.2737. There it reads “down to the great Han” [訖于
大漢], though the editors of the Zhonghua shuju Shiji edition have emended this to Tianhan 天漢, based
perhaps in part on this quotation, but also, explicitly on the justification by Yang Shuda 楊樹達 (1885-1956)
that “great Han” is “meaningless/insignificant” [無義] (see HS 62.2739). There are several ways in which
a compound in Chinese can be wuyi 無義. First, it could appear nowhere else as a compound and seem
incomprehensible in the location where it does appear. By that criterion, there is nothing wrong with “great

416
Sima Zhen‟s comment on the same item, some two to three centuries later, added more

information:

Pei Yin considered that everything after Tianhan was continuations by later
people. It was Master Chu who filled it in. In the records of later historians, it is
also no different. Thus there is no need to discuss it at present. [裴駰以為自天
漢已後，後人所續，即褚先生所補也。後史所記，又無異呼，故今不討論
也。]103

The comment on later historians is slightly ambiguous, and could mean one of two things:

either that later accounts of the Han period, such as Ban Gu‟s, did not disagree with the

supplements to the Shiji; or, that the practice with later histories was much the same

(namely, that when the primary author of a history laid down his brush, a secondary

author might pick it up again to add a few pertinent-seeming notes). In either case, Pei

Yin‟s anonymous “later person or persons” has suddenly become Chu Shaosun.

Sima Zhen did not merely adopt this theory in general, but also seemed to have

tried to apply it in particular cases. For example, after the Honorable Senior Archivist‟s

remarks in Shiji chapter 96, the narrative continues:

In the time of Filial Wu, there were extremely many [successive] chief ministers,
and so they are not recorded. [Here] I write nothing of their conduct or actions,
their appearances, or stratagems, [but] merely make a chronology of their service
from the Zhenghe period [92-89 BCE]. [孝武時丞相多甚，不記，莫錄其行起
居狀略，且紀征和以來。]104

Han.” It can be found in numerous places even in the Hanshu itself (i.e., HS 22.1075, 29.1694, 57.2601,
64A.2784, 87A.3539, 87B.3568, 94A.3780, 100A.4228, 100B.4267), and has a perfectly reasonable
interpretation in this context, namely, that Sima Qian‟s record came down to the Han dynasty. Yang Shuda
was presumably aware of this, and considered the text to be wuyi 無義 because it is a wholly uninteresting
(i.e., insignificant) statement about the Shiji. It fails to give us the information we very much want it to
give: namely, the point at which Sima Qian actually stopped his portion of the Shiji account. Given how
close tian 天 and da 大 are graphically, it would be reasonable to suppose that the original version (or
versions) could have had either one. Both Pei Yin and Yang Shuda chose to read tian 天 because it is the
more significant reading, but there is no evidence that it is the better reading.
102
SJ 22.1142.
103
Ibid.
104
SJ 96.2686.

417
The account then continues down to Kuang Heng 匡衡, who succeeded Wei Xuancheng

韋玄成 as chief minister upon the latter‟s death in 36 BCE.105 This certainly had to have

been written after the death of Emperor Wu. But did it have to have been written by Chu

Shaosun? Sima Zhen also commented:

From Ju Qianqiu106 on down, they are all recorded by Master Chu et al. However,
the traditions of these chief ministers are all abbreviated and vague, and it is only
in the Hanshu that there is a complete account. [自車千秋已下，皆褚先生等所
記，然丞相傳都省略，漢書則備。]107

Here Sima Zhen did not claim outright that the continuation was the work of Chu

Shaosun but wrote “Chu Shaosun et al.” [褚先生等].

Sima Zhen also cites Zhang Yan concerning the authorship of a highly

problematic passage near the end of the “Traditions of the Xiongnu”:

Zhang Yan said, “From Shanyu Hulugu on down, it is all recorded by Liu Xiang
and Master Chu. Ban Biao also (re-)wrote and (re-)ordered it. This is why the
Hanshu „Traditions of the Xiongnu‟ has two chapters [instead of just one].” [張晏
云：「自狐鹿姑單于已下，皆劉向、褚先生所錄，班彪又撰而次之，所以漢
書匈奴傳有上下兩卷。」]108

It is worth noting that even Zhang Yan was not here making any definite claims about

Chu Shaosun‟s authorship of the passage. Nor did Sima Zhen offer an opinion of his

own. Did Zhang Yan mean to imply that it was a collaboration between Chu Shaosun

and Liu Xiang, or was he merely making a guess as to the probable authorship? To me

the latter seems more likely.

105
This is followed by a problematic “Honorable Senior Archivist says” passage, SJ 96.2689.
106
The first chief minister mentioned after the above-quoted comment.
107
SJ 96.2686.
108
SJ 110.2919. For further discussion of the textual problems in the “Traditions of the Xiongnu” chapter,
see Honey, “Textual Criticism” and discussion below.

418
 In the examples I have considered so far, Sima Zhen was somewhat tentative

about ascribing to Chu authorship of unsigned passages. However, he was not so

agnostic about the next two examples. First, in Shiji ch.26, the “Treatise on the

Astronomical Offices”, he notes a change in dating format and comments, “After this,

from Taishi and Zhenghe down to the end of the chapter, the years and dates all take this

as a standard. It is all continued by Master Chu” [已後自太始、征和已下訖篇末，其

年次甲乙皆準此．並褚先生所續].109 Second, in glossing the description of the

“Treatise” genre in Shiji‟s last chapter, Sima Zhen had to reconcile the reference to

“military power” [兵權] with the fact that the Shiji as he knew it lacked a “Treatise on

Soldiers” [兵書].110 He wrote:

“Military power” referred to the Treatise on [Military Matters].111 After Qian‟s

death, this was lost. Chu Shaosun used the “Treatise on the Pitch-pipes” to fill it
in. The “Treatise on the Pitch-pipes” as we have it today does also have some
cursory remarks on military matters. [兵權，即（律）〔兵〕書也．遷沒之
後，亡，褚少孫以律書補之，今律書亦略言兵也．]112

Though Sima Zhen did unambiguously assign to Chu responsibility for the substitution,

he made no claims about the actual authorship or compilation of the “Pitch-pipes”

chapter.

109
SJ 26.1269.
110
See SJ 130.3319. Of course, at least one version of Zhang Yan‟s famous list of missing chapters does
include a “Treatise on Soldiers” [兵書] (see HS 62.2724; cf. SJ 130.3321 which has “Treatise on
Pitchpipes” [律書]).
111
In today‟s version of the Shiji suoyin, it reads 律書, but I have emended it to 兵書 because it makes
much better sense in context. Sima Zhen‟s picture of the textual situation with these chapters is clear from
the rest of the passage: Sima Qian‟s Shiji originally had no 律書, but did have a 兵書. The 兵書 was lost,
and Chu Shaosun used the 律書 to fill in for it. It is easy to see why an editor would have changed 兵書 to
律書, because of course the Shiji, as is so well-known today, contains no 兵書. The explanation for the
same variation in two versions of Zhang Yan‟s comment (HS 62.2724 and SJ 130.3321; see above note) is
no doubt similar.
112
SJ 130.3320.

419
 Zhang Shoujie, whose commentary seems to post-date Sima Zhen‟s, increased the

scope of Chu Shaosun‟s authorial role even further:

By the time of the Yuan and Cheng reigns, the Shiji had ten chapters which had
titles but no texts. Chu Shaosun filled in the Annals of Jing and Wu, the Table by
Years of the Generals and Ministers, the Treatise on Rites, the Treatise on Music,
the Treatise on the Pitch-pipes, the Hereditary Household of the Three Kings, the
Traditions of the Lord of Kuaicheng, and the Traditions of the Diviners by Days,
and of the Tortoise and Milfoil. The words and phrases of the Diviners by Days,
and of the Tortoise and Milfoil are exceptionally superficial and inferior, going
against the Honorable Senior Archivist‟s original intentions. [史記至元成閒十篇
有錄無書，而褚少孫補景﹑武紀，將相年表，禮書﹑樂書﹑律書，三王世
家，蒯成侯﹑日者﹑龜策列傳．日者﹑龜策言辭最鄙陋，非太史公之本意
也．]113

This passage is superficially very similar to Zhang Yan‟s list of the ten missing chapters

(the chapters are the same). However, it is abbreviated and subtly reordered, so that in

Zhang Shoujie‟s version, Chu Shaosun became responsible for replacing all ten chapters,

and not just the four that Zhang Yan attributed to him.114

This is not a mere textual accident. Of Shiji chapter 23, the “Treatise on Rites,”

Zhang Shoujie wrote, “For this treatise, Chu Shaosun made it by merely selecting from

Xun Qing‟s „Discussion of Rites‟ and joining it up” [此書是褚先生取荀卿禮論兼為

之].115 As for the compilation of the “Treatise on Music,” the case is a little more

complicated.

The Shiji “Treatise on Music” as we have it today seems to be made up of three

main parts. It begins with a long comment by the Honorable Senior Archivist, which

then blends almost seamlessly into an account of the origins and purpose of music. This

account is continued—albeit very sketchily and perhaps inaccurately—up to the reign of

113
SJ 128.3228.
114
Cf. SJ 130.3321.
115
SJ 23.1174.

420
the “present emperor.”116 Then there is an abrupt transition to a lengthy segment which

reproduces a reordered version of the Liji‟s “Record of Music.” Next, there is an

anecdote about music based on the Hanfeizi. A final remark by the Honorable Senior

Archivist closes the chapter.

Partway through the “Record of Music” section, Zhang Shoujie inserts a long

comment to explain the fact that the order of Shiji text differs from that of the Liji version:

This “Record of Music” is ordered and compiled by Gongsun Nizi.117 He made

the “Record of Music” to connect Heaven and Earth, string them together with
human dispositions, and debate about governance. Thus it is explained in detail.
In the past, the sequence of sections in Liu Xiang‟s “Separate Record” concurred
with the one in Zheng [Xuan‟s] “[List of] Titles,” but the ordering of sections in
the “Record of Music” does not match Zheng‟s “Titles.” Now in this [part of the]
text, the section are out of order because Chu Shaosun promoted some and
demoted others. Thus, it is now all in confusion. Now I have sought out the
former ordering so that the sections follow from one another, thus causing [what]
comes after to be somewhat understandable. I consider the latter text [i.e., the
ordering found in Shiji ch.24] to come from nothing more than Chu‟s own ideas.
[其樂記者，公孫尼子次撰也．為樂記通天地，貫人情，辯政治，故細解
之．以前劉向別錄篇次與鄭目錄同，而樂記篇次又不依鄭目．今此文篇次顛
倒者，以褚先生升降，故今亂也．今逐舊次第隨段記之，使後略知也．以後
文出褚意耳．]118

This passage shows three things. First, Zhang Shoujie considered that the original author

of the “Record of Music” had a profound intention, which was partly expressed through

the order of the sections. Second, the Shiji version gave the text in a different internal

order from other versions known to Zhang Shoujie. Third, Zhang held Chu Shaosun

responsible for these differences, which in Zhang Shoujie‟s view made the original

profound meaning of the text nearly incomprehensible.

116
For objections to the historicity of the account and a discussion of the authorship of this chapter
generally, see Martin Kern‟s “Note on the Authenticity and Ideology of Shih-chi 24.”
117
A Chunqiu period figure of semi-legendary status. The traditional attribution of the “Record of Music”
to Gongsun Nizi is of dubious historical value. See Kern, “Note,” 676-677.
118
SJ 24.1234.

421
 What did Zhang Shoujie make of the authorship of the first part of the “Treatise

on Music,” the part not taken from the “Record of Music”? That part does seem to have

been included in his version, for his glosses are found throughout. He certainly implied

(in his above quoted comment from chapter 130) that he thought Chu Shaosun wrote it.

However, he never stated this outright.

The tendency to expand Chu Shaosun‟s perceived contribution did not stop with

the Tang commentators.119 Rather than reviewing the debate, however, I want to

consider another factor in the Tang commentator‟s reaction to Chu Shaosun: even as they

expanded his alleged oeuvre, they denigrated both his writing ability and his moral

character. This was especially true of Sima Zhen.

In Chu‟s supplement to Shiji chapter 13, he discusses the problem of genealogy as

related to the semi-mythical figures of high antiquity. Because Chu flatteringly tied his

discussion to the ancestry of Huo Guang—a move whose political significance was very

blatant given the time in which Chu lived—Sima Zhen called him a “rotten ru” [腐儒]

and ranted, “What was he hoping to prove with this? [His] words do not follow, and are

like weeds or excrescences in the proper history. Commenting as an afterthought, „Is it

not mighty indeed!‟—oh, what falsity it is!” [竟欲證何事？而言之不經，蕪穢正史，

輒云「豈不偉哉」，一何誣也！]120

I have already cited Sima Zhen‟s complaint about Shiji ch.12, and what he saw as

Chu‟s lack of skill in compiling it. Regarding his work on the “Diviners by Days”, Sima

Zhen also hints at his dissatisfaction: “The „Traditions of the Diviners by Days‟ writes

that „that there is no way of knowing the customs of the various states.‟ And now Master

119
See Zhang Dake, Shiji yanjiu, 177-190.
120
SJ 13.507.

422
Chu has merely recorded some things to do with Sima Jizhu” [日者傳云「無以知諸國

之俗」，今褚先生唯記司馬季主之事也].121

Sima Zhen also lambasts Chu‟s work on the “Tortoise and Milfoil” (Shiji ch.128),

commenting, “His narration of events is prolix and disorderly, inferior and cursory; there

is nothing in it worth selecting from” [其敘事煩蕪陋略，無可取].122 His comments to

the “Self-Narration” give a more detailed critique.

[We know that] the kings of the three dynasties did not use the same tortoise-shell
[divination method], and that the barbarians of the four directions each had
different ways of divining. The writings having been lost, there is no
chronological record of their differences. Now Chu Shaosun has done no more
than select some miscellaneous sayings from The Great Diviner Prognosticates
on the Tortoise-shell. The wording is extremely prolix and vague, but he was
unable to edit it properly. [Instead he] erroneously gave it some forced
interpretation. This chapter is inept in the extreme! [三王不同龜，四夷各異
卜，其書既亡，無以紀其異．今褚少孫唯取太卜占龜之雜說，詞甚煩蕪，不
能裁剪，妄皆穿鑿，此篇不才之甚也．]123

Whether we agree with Sima Zhen as regards Chu Shaosun‟s ineptitude is beside the

point. I believe the important thing to notice about the changing view of Chu Shaosun‟s

authorship is that Chu was being used as a scapegoat, the hack-writer responsible for

parts of the Shiji that did not live up to Sima Qian‟s rising reputation as a historian and

stylist.

This was rather archly pointed out by Liu Zhiji, who quoted Zhang Yan‟s

comment about Chu Shaosun‟s writing being “superficial and inferior” [鄙陋], then

added his own note:

In the “Annals of the Five Emperors” and among the seventy arrayed traditions
that [Sima] Qian did write, he says that when Yu Shun encountered distress, he

121
SJ 130.3318.
122
SJ 128.3223.
123
SJ 130.3319.

423
 dug a hole and escaped thereby124; and after Xuanni [=Confucius] died, his
disciples took Youruo as their teacher.125 The inferiority of [Sima Qian‟s] words
sinks even lower than this,126 so how can one place all the blame on Master Chu
and whole-heartedly revere Master Ma [=Sima Qian]? [按遷所撰五帝本紀、七
十列傳，稱虞舜見阨，遂匿空而出；宣尼既殂，門人推奉有若。其言之鄙又
甚於兹。安得獨罪褚生而全宗馬氏也？]127

The major Tang commentators did perhaps want to sort out what was inferior in the Shiji

and assign it to someone other than Sima Qian. The project was more effective if that

other, scapegoat author had a name, so Chu Shaosun was slowly eased into the role. In

some sense, this is comparable to the process whereby Sima Tan later began to be

assigned (and/or blamed for) „Daoist‟ portions of the text. The process could, of course,

never be complete: as Liu Zhiji pointed out, there were parts of the Shiji that could not

reasonably be attributed to Chu Shaosun but nonetheless seemed to readers of the time

quite as objectionable as anything Chu wrote.

A DRASTICALLY DAMAGED SHIJI

As mentioned above, the Shiji was known to have suffered textual loss early in its

existence—namely the ten missing chapters. No chapters are completely missing from

the received Shiji, although the twelfth chapter is merely a copy of part of the “Treatise

on the Feng and Shan”, while various other chapters are sketchy or otherwise anomalous.

Yu Jiaxi, in his exhaustive study of the missing chapters issue, discussed whether Zhang

Yan‟s second century list of chapters is reliable, as well as tracing the “reappearance” of

124
See SJ 1.34.
125
The reason given in the Shiji that the disciples selected Youruo (whose name literally means „has
something similar‟) is because he resembled Kongzi in appearance. See SJ 67.2216.
126
I.e., the material supposedly written by Chu Shaosun.
127
STTS 26.502.

424
various chapters through citational and other evidence.128 One thing the “reappearance”

process reveals is the drive for a more complete and coherent Shiji, one which readers

could see as fully realizing the stated ambitions of its author.

Curiously, in the late Qing we also find the beginning of an opposite trend: a

tendency to attack the authorial coherence of the Shiji. I have already shown how Fang

Bao‟s analysis shifted considerable authorial responsibility onto Sima Tan. I turn now to

a far more extreme theory, this one regarding textual loss, which I call „the drastically

damaged Shiji‟ hypothesis. The basic idea is that the Shiji was drastically damaged at

some point in its history, and was “back-copied” from chapters in the Hanshu, or

replaced in other ways by unidentifiable forgers. I will argue two things about this

„drastic damage‟ theory: first, the evidence is not solid enough for sweeping

generalizations in either direction (nor, perhaps, even for confident judgments on

particular chapters); and second, it may be that part of the motivation for hypothesizing a

drastically damaged Shiji may be an attempt to avoid certain unfortunate complications of

the Sima Qian author function.

The Argument for Drastic Damage

Historically, the argument for a drastically damaged Shiji seems to have

originated with Kang Youwei 康有為 (1858-1927). In his Xinxue weijing kao 新學偽經

考 [A Study of the Forged Classics of the Xin Period] of 1891, he argued that Liu Xin 劉

歆 (46 BCE-23 CE) had tampered with the Classics. According to his argument, the Shiji

had also suffered such tampering, but could still provide evidence regarding what Liu Xin

128
I will not discuss the ten missing chapters in detail, as there is no need to repeat Yu Jiaxi‟s work.

425
did: “Although [Sima Qian‟s] text was very much tampered with by Liu Xin, its main

principles are clear and unadulterated” [ 雖其書多為劉歆所竄改，而大體明粹].129 His

main criteria for authenticity in the Shiji text is that any passages dealing with the “Six

Arts” 六藝 as a category was Sima Qian‟s original, while passages describing the “Old

Text” faction‟s rediscovery of variant texts were “disordering insertions made by Liu Xin

in order to mislead people” [劉歆竄亂以或人者].130

Kang‟s student, Liang Qichao 梁啟超 (1873-1929), disagreed strongly with

Kang‟s ideas about the Shiji: “As to Kang‟s arguments that Liu Hsin had secretly

introduced several tens of items into the Shih-chi... these were certainly completely

untenable even from a common sense viewpoint, and yet he maintained them stubbornly”

[乃至謂史記劉歆羼入者數十條。。。此實為事理之萬不可通者，而有為必力持

之].131 Though clearly Liang Qichao did not take up Kang‟s theory, he notes that another

scholar, Cui Shi 崔適 (1854-1924), did so with great enthusiasm: “Later, a certain Ts‟ui

Shih wrote Shih-chi t’an-yuan and Ch’un ch’iu fu-shih. These two books quoted and

developed K‟ang‟s ideas, making them more precise and succinct” [其後有崔適者，著

《史記探源》、《春秋復始》二書，皆引申有為之說，益加精密].132

129
Kang Youwei quanji, 586. Note that Liang Yusheng 梁玉繩 (fl. 18th c.), had compiled an extensive
record of passages from the Shiji that he considered problematic. His Shiji zhiyi 史記志疑 [Record of
doubts about the Shiji] points out problems of all sorts—especially inconsistencies—but tends to remain
neutral as to whether they should be attributed to Sima Qian‟s own carelessness or to later textual
corruption (see, for example, the Shiji zhiyi “Zixu” 自序, 2.)
130
Kang Youwei quanji, 600.
131
Liang Qichao, Qingdai xueshu gailun, 448; trans. Intellectual Trends of the Ch’ing Period, 93.
132
Ibid.

426
 Elaborating on Kang‟s ideas, Cui wrote: “All the Classics having suffered [Liu

Xin‟s] disordering alterations, the Shiji, being the gateway of the Five Classics, also had

to be disordered” [於是群經皆受其竄亂，而史記為五經門戶，則亦不得不竄亂

矣].133 Cui Shi claimed to have identified various signs of tampering by Liu Xin and his

confederates.134 Based on these criteria, he arrived at the conclusion that a great number

of chapters were not actually Sima Qian‟s work, including the “Basic Annals” of

Emperors Wen and Wu, six of the ten tables, all eight of the treaties, and 29 others.

This degree of textual damage far exceeds the ten missing chapters mentioned in

the Hanshu.135 But what could have happened to the rest of the Shiji? Li Kuiyao 李奎耀,

a later proponent of the „drastically damaged Shiji‟ theory, used a brief passage from the

Hanshu to introduce the idea that there was a great abridgement of the Shiji text.136 The

passage concerns a certain Yang Zhong 楊終 (1st c. CE), who “received an imperial edict

to reduce the Writings of the Honorable Senior Archivist into 100,000-plus words” [受詔

刪太史公書為十餘萬言].137 The Shiji describes itself as having 526,500 characters,138

while the current Shiji has some 50,000 more characters than that. An abridgement on

the scale of the one suggested by the Hou Hanshu passage corresponds to an

approximately seventy percent reduction in the text. If such a fate had indeed befallen

the Shiji, the implications for its authenticity on all levels would be shocking.

The question that immediately arises, however, is that even if Yang Zhong did

complete his abridgement, why would the original Shiji not continue to exist side by side

133
Shiji tanyuan, 2.
134
These are outlined in the introduction to Shiji tanyuan, 1-18.
135
HS 30.1714.
136
“Shiji jueyi” 史記決疑 [Resolving doubts about the Shiji].
137
HHS 48.1599.
138
SJ 130.3319.

427
with the abridgement? As modern scholar Chen Zhi 陳直 wrote, “Texts that were

produced through abridgement in ancient times all existed alongside the original. It is not

that, once the abridgement came out, the original suddenly ceased to exist or be

transmitted” [古代刪定的書，與原書皆是同時並存，不是刪本一出，原本湮沒不

傳].139

Li Kuiyao answered this challenge with the following argument:140 In the Western

Han, the Shiji was not well-known. The reason it got noticed in the Eastern Han did not

have to do with its own essential nature, but with the New Text/Old Text debates. For

example, in the first year of Jianwu (25 CE), Chen Yuan 陳元 and Han Xin 韓歆

petitioned for an Erudite position in Zuozhuan studies. New Text scholar Fan Sheng 范

升 and others violently rebutted the proposal, also implicating the Shiji in the process:

At the time, those who were rebutting [Fan Sheng] used the argument that the
Honorable Senior Archivist had many citations from the Master Zuo. Sheng
[therefore] also submitted that the Honorable Senior Archivist violated and
offended against the Five Classics and went against the words of Kongzi. [時難者
以太史公多引左氏，升又上太史公違戾五經，謬孔子言。]141

As Liu understood the process of the debate, because those who criticized Zuozhuan also

criticized Shiji, those who revered Zuozhuan also praised Shiji. According to Liu, this

unexpected attention turned out to have unfortunate consequences for the Shiji. Liu also

added a description of Ban Gu‟s public attack on the Shiji in the time of Emperor

139
“Han Jin ren dui Shiji de chuanbo ji qi pingjia” 漢晉人對史記的評價 [Han and Jin Dynasty people's
transmission and evaluation of Shiji], 221.
140
Paraphrased from “Shiji jueyi,” 1177.
141
HHS 36.1228.

428
Ming.142 Liu then made an argument for what I would call „punitive abridgement‟ of the

Shiji:

Emperor Zhang had personally inherited Emperor Ming‟s edict.143 After he

ascended to the position [of ruler], his tendency to be doubtful and suspicious of
[Sima] Qian‟s Shi[ji] would have been profound indeed. His ordering Yang
Zhong to abridge and cut down the Shiji cannot have been unconnected with
this.... Rather than saying that the Shiji was endangered by Yang Zhong, we
should say that the Shiji was endangered by Ban Gu, and it was just Yang
Zhong‟s hand that did the execution. [章帝親承明帝之詔，即位後，其疑忌遷
史之心，當為尤甚，下詔令楊終刪削，不為無因。。。。與其謂史記被厄於
楊終，不如謂其被厄於班固，特楊終親為劊子手耳。]144

At this point, it seems worthwhile to stop and consider what the actual physical

process of abridgement might have been. Was the original text “cut up” and the removed

parts “thrown away”? Or were the contents of the abridged text copied out from the

original, so that both continued to exist side by side? Liu Kuiyao implied that the

purported „punitive‟ abridgement was done according to the former method:

Probably the Shiji, having undergone the disaster of Yang Zhong[‟s abridgement],
almost completely lost its original appearance. Fortunately, at the same time, Ban
Gu stole [from the Shiji], seizing it for his own use. Thus, the authentic traces of
what was cut out of [Sima] Qian‟s Shi[ji] have instead been preserved by being
recorded in the Hanshu. [蓋史記經楊終之一厄，幾全失本來面目。幸同時有
班固竊之，攘為己有，俾遷史所被刪之真蹟，反得籍漢書以保存。]145

Thus, Liu‟s answer to the question of “where it all went” is that it went into the Hanshu.

Of course, as Chen Zhi and others have argued, the usual practice for abridgements

would have involved copying rather than culling.

The second question in the „drastically damaged Shiji‟ scenario is: What were the

sources for the reconstructed Shiji? One possible answer, mentioned above, is that the

142
Discussed in chapter 4 above.
143
I.e., Ban Gu‟s “Dian Yin” 典引, discussed in chapter 4 above. That text was sharply critical of Sima
Qian.
144
“Shiji jueyi,”1179
145
“Shiji jueyi,” 1188.

429
lost material was replaced based on the Hanshu. Various scholars, including Lu Zongli,

have pointed out that this is problematic.146 The Han portions of the Shiji, those that

overlap with the Hanshu, certainly do not amount to seventy percent of the Shiji text.

And besides, why would Yang Zhong only abridge Han chapters? Even leaving these

problems aside, the great abridgement would not seem especially consequential if it were

just a matter of the abridged chapters being temporarily “stored” in the Hanshu and then

brought back.

David Honey, the most recent Western proponent of the „drastically damaged

Shiji‟ hints at a more complicated picture, writing:

This drastically altered version [i.e., Yang Zhong‟s abridgement], together with
ten lost chapters from the Shih-chi which Pan Ku had mentioned, so reduced the
length and scope of the Shih-chi that later forgers had free rein to both supply the
missing chapters as well as augment the much-abbreviated text.147

Thus our list of purported sources for a drastically damaged—and later reconstituted—

Shiji would include back-copying from the Hanshu, Chu Shaosun‟s and Liu Xin‟s

interpolations, and further supplements or forgeries by unknown people.

There is little doubt that at least some of today‟s Shiji text differs from the original

text as written in the Western Han. The question raised by the „drastically damaged Shiji‟

theory is really one of scale. Just how much free rein did later forgers have? One way of

approaching this question is to try to pin down concrete details of the supposed

reconstruction, in particular when it was supposed to have taken place. Ban Gu and

Wang Chong seem to have had access to an unabridged Shiji text. A.F.P. Hulsewe, also a

believer in a drastically damaged Shiji extensively back-copied from the Hanshu,

146
See Chen Zhi, “Han Jin ren”; Lu Zongli, “Problems Concerning the Authenticity of Shih chi 123
Reconsidered,” 54.
147
“Textual Criticism,” 80.

430
hypothesized that the Shiji was unavailable and out of circulation between 100 and 400

CE.148 Lu Zongli has argued against this hypothesis using various lines of evidence,

including multiple citations of the Shiji from within the time period Hulsewe suggested.

Some of the more telling citations include those of Qiao Zhou 譙周 (201-270), who

demonstrated familiarity with large sections of the Shiji and quoted from or referred to

them by name.149 One particularly convincing example is a comment by Zhang Fu 張輔

(d. 305), who stated that the Shiji “narrates three thousand years worth of events in only

five hundred thousand words” [叙三千年事，唯五十萬言].150 The value of this

quotation, of course, is that it gives the size of the Shiji as Zhang Fu knew it, very close to

the size of the Shiji mentioned in the Sima Qian “Self-Narration.” In contrast, as Chen

Zhi noted, “Yang Zhong‟s abridged edition of the Shiji was never [mentioned as having

been] seen by any post-Han person” [案楊終所刪定的史記本，在漢以後的人，即未

見過].151

Western Debates over the Shiji Text

Western scholars, perhaps inspired by Liang Yusheng‟s record of the Shiji‟s

textual problems and by Cui Shi‟s skepticism, have brought Western text-critical

apparatus and methodology to bear on the issue of potential textual damage or loss within

the text of the Shiji. These articles tend to include both highly technical arguments and

148
also includes the quotation from Zhang Fu mentioned below; see “Problems Concerning the
Authenticity of Shih chi 123 Reconsidered.”
149
See Farmer, Talent, 95-119.
150
JS 60.1640.
151
Chen, “Han Jin ren,” 221.

431
larger methodological discussion of the principles by which one might decide the issue of

authenticity.

Perhaps the most extended series of articles concerns the “Traditions of Da Yuan”

大宛列傳 (Shiji ch.23). A.F.P. Hulsewé attacked the chapter in “The Problem of the

Authenticity of Shih-chi ch. 123, the Memoir on Ta Yüan.” As Hulsewé noted, the

debate over the authenticity of this chapter had already begun some time before (he cited

negative opinions by Cui Shi, Li Kuiyao, Paul Pelliot, and Gustav Haloun152). Pelliot for

example, had written,

In my opinion, legend must have quickly been substituted for history [in the story
of Zhang Qian, the “protagonist” of ch.123], perhaps in the form of an historical
romance which was accepted in China during the first century of our era. This
romance derived in part from the biography of Zhang Qian in the Qian Han shu,
and that which is found today in Shi ji (“Da yuan” chapter), far from being the
source of the Qian Han shu, was the same as that of the Han shu placed back into
the Shi ji by some forger after the first century of the Christian era. I am
announcing here somewhat revolutionary conclusions, but I do not see another
solution to the problem of the biography of Zhang Qian as it is given in the Shi ji,
and in the future I will return to this question in greater detail.153

Yet the authenticity of the chapter also had its defenders, including Edwin Pulleyblank,154

as well as Japanese scholars cited by Hulsewé who had argued for the primacy of the

Shiji account on geographical grounds.155

152
Hulsewé, “Problem of Authenticity,” 83, n.1-8.
153
“A mon avis, la légende avait dû ici se substituer très vite à l‟histoire, peut-être sous forme d‟un roman
historique qui était accepté en Chine au Ie siècle de notre ère; c‟est en partie de ce roman que dériverait la
biographie de Tchang K‟ien dans le Ts’ien Han chou, et celle qui se trouve aujourd‟hui dans le Che ki (ch.
du Ta-yuan), loin d‟être la source du Ts’ien Han chou, serait celle même du Ts’ien Han chou remise dans le
Che ki par un faussaire après le premier siècle de l‟ère chrétienne. J‟énonce ici des conclusions assez
révolutionnaires, mais je ne vois pas d‟autre solution au problème de la biographie de Tchang K‟ien telle
qu‟elle est donnée dans le Che ki, et je me réserve de revenir ailleurs sur la question en plus grand détail.”
Pelliot, “L‟Édition collective des oeuvres de Wang Kouo-wei,” 178 n.1. Unfortunately, Pelliot never did
return to the question.
154
See Edwin G. Pulleyblank, “Chinese and Indo-Europeans,” Journal of the Royal Asiatic Society 1966: 9-
39; and “The Wu-sun and Sakas and the Yueh-chih Migration,” Bulletin of the School of Oriental and
African Studies 1970: 154-160.
155
Hulsewé, “Problem of Authenticity,” 84 n.10.

432
 Hulsewé‟s own article may have been written partly in response to Pulleyblank,

or to the perceived need for a Western text-critical study of the long debated chapter.

The proximate cause of the article, however, was that Hulsewé was preparing a

translation for Han Social Structure156: since the Dayuan chapter was to be used as an

historical source, “the problem of the authenticity [of Shiji ch.123] had to be faced.”157

The context is important: clearly Hulsewé‟s aims differed from those of scholars like

Watson, Durrant, and Hardy. While the primary goal of their studies was to better

understand the nature of the Shiji, Hulsewé aimed first and foremost to better understand

the history of the period—a related but different project.

The bulk of Hulsewé‟s article consists of a comparison between Shiji and Hanshu

versions of the Dayuan material. Pointing to the existence of passages “where confusion

occurs in both texts in an identical manner,” Hulsewé questioned whether “a scholar of

Pan Ku‟s calibre and fame would have slavishly accepted the mistakes which it proves

possible to correct nearly two thousand years later.”158 It is worth noting here first that

Hulsewé‟s a high opinion of Ban Gu effectively becomes part of his evidence. Hulsewé

made two other observations: first, the material from Shiji chapter 123 is split up among

three or more different Hanshu chapters, or is not present in Hanshu at all; and second,

where the two are parallel, the Hanshu text is more concise. Hulsewé also brought up

two points of external evidence: first, the contents of Shiji ch.123 do not match the Shiji

ch.130 summary of that chapter;159 second, Shiji chapter 111 contains a group of small

156
By Tʻung-tsu Chʻü, edited by Jack Dull.
157
Hulsewé, “Problem of Authenticity,” 84.
158
Ibid., 85.
159
The “small preface” in Shiji ch.130 does seem to imply that the chapter would have more to do with
people from Dayuan who “wished to see the Central States” [欲觀中國] (SJ 130.3318). It is not clear that
we should automatically privilege the small prefaces themselves, however.

433
biographies, one of which is that of Zhang Qian. Since that group is prefaced by the

statement that these men‟s biographies do not appear elsewhere in the Shiji, Hulsewé

observed, it is odd to find Zhang Qian‟s particulars repeated in Shiji chapter 123.

The passages that Hulsewé was referring to occur on Shiji 111.2941 and 2943.

The first passages note that Li Guang has his own “Traditions” chapter (i.e., SJ ch.109),

but that “those who are without „Traditions‟ are…” [無傳者曰。。。]. Short

biographies of various figures then follow, with Zhang Qian‟s occurring on SJ 111.2943.

Although Hulsewé‟s argument is interesting, I might point out that Shiji ch.123 is not

devoted to Zhang Qian in the way that Shiji ch.109 is devoted to Li Guang. That is to say,

the apparent inconsistency may result from the fact that Hulsewé understood zhuan to

mean „biography,‟ whereas in fact it probably means biezhuan 別傳, an individual

chapter in the “Traditions” section of the Shiji. Wu zhuan 無傳 would therefore mean,

“without [separate] „Traditions‟ [chapters of their own].”160

One more aspect of Hulsewé‟s argument seems worth noting. Hulsewé discussed

the apparent lack of commentaries on the Shiji until Xu Guang (352-425). His oft-quoted

conclusion was that “this can only mean that during the years between 100 and 400 A.D.

the Shih-chi was not available.”161 As noted above, citation patterns weigh against this

statement. At least some scholars had access to the text during the centuries in question.

But the lack of commentaries is an interesting point162: Why did the Hanshu receive so

much more commentarial attention? As suggested by my discussion in chapter 1 above,

160
For a discussion of the problematic term zhuan 傳 within the formal structure of the Shiji, see chapter 3
above.
161
Hulsewé, “Problem of Authenticity,” 87.
162
As Lü Zongli has pointed out (see below), the lack of early Shiji commentaries was probably not as
complete as Hulsewé believed. Still, there is no denying that the Hanshu fared better in this regard.

434
the Hanshu appears to have been the more prestigious text, certainly for the first few

hundred years of its existence. Commentarial neglect of the Shiji is only surprising if we

assume that the text always enjoyed the high reputation it has today—an assumption that

is clearly problematic.

Hulsewé‟s attack on Shiji ch.123 has not been allowed to stand unchallenged.

The chapter was defended most thoroughly by Kazuo Enoki, in his article “On the

Relationship between the Shih-chi 史記, Bk. 123 and the Han-shu 漢書, Bks. 61 and 96,”

Memoirs of the Tôyô Bunko 41 (1983): 1-31. Like Hulsewé, Enoki too began with earlier

challenges to the chapter‟s authenticity, but by contrast Enoki pointed out the lack of

evidence—or other problematic issues—within each of these arguments. He then

proceeded to refute Hulsewé‟s arguments in a systematic, point-by-point manner. His

own arguments included the fact that the Shiji text contains older terminology; when

varying traditions are recorded, the Hanshu‟s version seemed to contain more and newer

information; that it is actually easier to split a chapter up (i.e., break Shiji ch.123 into

several Hanshu chapters) than to pull information together from several chapters,

especially given the pattern of cross-references in the Hanshu version; and that what

Hulsewé considers garbled text or “passaging errors” actually make better sense than

Hulsewé‟s proposed corrections. He ends with a summary of Pulleyblank‟s earlier

arguments.

Enoki‟s work was thorough and well-researched. However, while many of his

arguments to some extent undermined Hulsewé‟s evidence, they did not necessarily

prove the authenticity of the chapter. As with Hulsewé, much of Enoki‟s argument

operated on the level of individual words, words that an editor of any period might feel

435
perfectly justified in modifying. As for larger principles—for example, whether the

shorter reading should always be preferred, or whether it is more difficult to pull

information together into one chapter or split it apart into two—the debate between

Hulsewé and Enoki is strong evidence that these questions cannot be decided

categorically. Instead, the overall picture one derives from the entire issue is that the two

texts—i.e., the Shiji and Hanshu—preserve variant traditions and have had a complicated

relationship. Hulsewé recognized this when he wrote:

One can never be sure in how far even the most loyal copyists of our texts were
consciously or unconsciously influenced by the existing parallels, especially
because of their belief that one author had merely copied the other; nothing will
have been more natural than the tendency of attempting to fill a lacuna or to
correct obscurities in one text by consulting the other.163

It seems clear that, in the absence of other factors, textual evidence on a character-by-

character level could point either way. It is the weight one gives to these other factors

that in fact may play a deciding role.

Lü Zongli made many of these points in his article “Problems Concerning the

Authenticity of Shih chi 123 Reconsidered.” His stated goal was not to defend the

authenticity of Shiji 123, nor was it to “prove that the present Shih chi 123 could not be a

later reconstruction.” Rather, his goal was to reject Hulsewé‟s arguments on

methodological grounds.164 As mentioned above, one of his main points was to disprove

Hulsewé‟s claim that the Shiji did not circulate in the years 100-400 A.D. Among others,

163
“Problem of Authenticity,” 89.
164
This is not to say that he omitted to discuss his own conclusions: “Our own view is that the present
version of the Shih chi is an ancient text dating from the Han dynasty” (Problems,” 52). He noted that
“according to a traditional principle of Chinese textual criticism, an ancient text should be regarded as
authentic until there is sufficient evidence to warrant the claim that it is spurious” (ibid., 53), and seems
implicitly to have adopted the same principle.

436
he cited a reference by Sima Zhen (fl.745) to two early commentaries on Shiji, now long

lost:

Earlier, Yen Tu (d.167 A.D.) of the Later Han wrote the [Shih chi] yin I (The
Pronunciation and Meaning of the Words in the Shih chi) in one volume. There
was another work entitled Chang yin in five volumes with no author‟s name. In
recent times these two works are rare. [始後漢延篤乃有音義一卷，又別有章隱
五卷，不記作者何人，近代鮮有二家之本。]165

The second part of Lu‟s study rebuts Hulsewé‟s arguments from internal

linguistic grounds. Lu challenged many of Hulsewé‟s findings by referring to other parts

of the Shiji and Hanshu, taking Hulsewé to task for his normative judgments on the two

texts: “[Hulsewé‟s] comments on „bad grammar‟ and „poor understanding‟ are subjective

and without any reasonable supporting evidence.”166 He challenged the text-critical

principle of brevior lectio portior (the shorter reading is to be preferred) by pointing out

that abridgement was a standard practice in Chinese tradition: brevity, as also discussed

in chapter 1 above, was an important stylistic value. Finally, Lu made an important point

about the relationship between the two texts: “It is not simply the case that one is the

original and the other a duplicate,” he wrote of Shiji and Hanshu. “The two works were

written by two great authors, so their wording should never be exactly identical.”167

Again, the differences between Shiji and Hanshu do not come down only to

straightforward copying. Thus, as Lu concludes, text-critical principles (developed

mostly in relation to scribal copying in medieval Europe) may not be as useful as they

seem.

165
Qtd. and trans. Lu, “Problems,” 58. The original text can be found in Sima Zhen‟s Shiji suoyin
“Postface.”
166
Lu, “Problems,” 66.
167
Lu, “Problems,” 60.

437
 This should not be taken to mean that textual criticism is entirely useless for Shiji

studies, or that it should be ignored. But perhaps what the debate on Shiji ch.123 shows

is that we must have high standards for what constitutes evidence or proof—and that in

many cases, little or no evidence meeting those standards remains to us.

One chapter for which closer textual examination has proved more convincing,

however, is the “Traditions of Sima Xiangru,” Shiji ch.117. Just at the same time that

Hulsewé was working on the “Da yuan” chapter, Yves Hervouet was doing many of the

same things in his article, “La valeur relative des textes du Che ki et du Han chou.”

Hervouet‟s work was then extended and built upon in Martin Kern‟s, “The „Biography of

Sima Xiangru‟ and the Question of the Fu in Sima Qian‟s Shiji.” Adding to philological

evidence gathered by Hervouet and Derk Bodde, Kern also considered the anachronistic

content of the Sima Xiangru chapter and external evidence from other parts of the Shiji.

Though explicitly aware of the uncertainties inherent in any authenticity argument, 168 he

concluded, “There are several, and mutually independent, sets of data that in their

aggregate discredit the Sima Xiangru biography as a textual anomaly in both form and

contents.”169

It seems worthwhile at this point to go back to the question of motivations. Why

would one want to question—or defend—the authenticity of a Shiji chapter? Kern, for

example, was concerned with the role of the fu 賦 (rhapsody) in the Western Han, as well

as related issues involving literacy and oral performance. In considering the development

and function of that genre over time, it makes a significant difference whether accounts

168
Kern, “Sima Xiangru,” 315-316.
169
Ibid., 307.

438
of the fu in Shiji reflect actual conditions in Western Han, or merely later imaginations

about them.

William Nienhauser, on the other hand, has a very different agenda. Deeply

involved as he is in a large-scale effort to translate (and/or retranslate) the entire text of

Shiji, the defense of the Shiji‟s integrity is in some sense tied up with the justification of

his project.170 The introduction to volume II of the Grand Scribe’s Records also contains

a close comparison between the Shiji and Hanshu versions of a chapter, in this case, the

“Basic Annals of Gaozu.” It is a surprising choice, in that one rarely encounters doubts

as to the authenticity of this chapter. However, Nienhauser‟s concern is to actually to

argue against the more extreme versions of the “drastically damaged Shiji” hypothesis,

namely, “the general argument that most or many of the Han-dynasty chapters in the Shih

chi were in any fashion based on Han shu counterparts.”171 In that case, any and all of

the Han period chapters in the Shiji would be at issue. Based on his comparison,

Nienhauser concluded that “there seems little reason to accept” the drastically damaged

Shiji hypothesis. He goes on to maintain that “the burden of proof should always be on

those who want to change or exchange the Shih chi accounts.”172 This seems problematic,

insofar as the available evidence and the complex circumstances of textual transmission

do not admit of proof in either direction (see discussion below).

On the opposite extreme is David Honey‟s article on Shiji ch.110, “The Han-shu,

Manuscript Evidence, and the Textual Criticism of the Shih-chi: The Case of the „Hsiung-

nu lieh-zhuan‟,” mentioned above. Honey arrived at the opposite conclusion as

170
The translations of Nienhauser and his collaborators, as well as some of Nienhauser‟s own articles on
the Shiji are discussed in the Introduction, above.
171
Grand Scribe’s Records II.xlvii.
172
Ibid., II.xlviii.

439
Nienhauser, advocating that “unless philological evidence demonstrates

otherwise...primacy should be given to the Han-shu version of all parallel accounts,”

though he does include the caveat that “each parallel account must be examined

independently.” In some ways, Honey‟s position is not as extreme as it first appears. He

draws on the work of Vincent Dearing to make a useful distinction between text and

edition. Graphic evidence, he argues, tells us much more about the edition/record—the

physical objects that transmit the text—than about the text and its meaning. Thus, he

writes, “textual analysis is devoted to the physical, linguistic record of an edition, not an

original „text‟ as rhetorical or historical message.”173 This distinction may not be wholly

satisfying, but it is very useful in that it accords with reasonable intuitions about the

limits of philology in studying texts with two thousand years of transmission history.

What has happened to these texts over their long existence depends on many factors, not

least of which is their editors‟ own ideas about textual priority.

This is not to say that the philological evidence says nothing about the

text/message: another factor contributing to textual difference could still be differences in

the authors‟ ideologies and intents. Honey makes the point that many subsequent

mutations (Yan Shigu‟s deliberate archaicization of the Hanshu text, to cite one example)

should also be taken into account. That Honey prefers to limit his discussion to

edition/record rather than text/message shows that his project also differs fundamentally

from Kern‟s. He is not especially concerned with the issues of authorship and

authenticity, and wrote, “I personally see no reason that Pan Ku would want to reinvent

173
Honey, “Textual Criticism,” 75.

440
the historical wheel, so to speak, when a ready-made narrative on the Hsiung-nu was

there, or for the entire Han dynasty, for that matter.”174

Honey seems content to believe in Sima Qian‟s original authorship of the Han

dynasty chapters. Therefore, he does not need the “drastically damaged Shiji” theory to

explain the data he wanted to explain—that Hanshu versions of the Xiongnu chapter,

where there are parallels, seem to preserve an older version of the material than the Shiji

versions. Honey is little concerned about the effect of his argument on the field of Shiji

studies, but his acceptance of the Yang Zhong abridgement (as a replacement for the

original Shiji text) is potentially problematic not just for our text of the Shiji but for our

picture of Han dynasty textual practices.

Honey‟s conclusion, that “the original „countenance‟ of the Shih-chi is to be seen

not in the received version, nor in any extant printed version, nor in the scattered, isolated

MSS of the Shih-chi, but in the text of the Han-shu,”175 must certainly call into question

the “authenticity” of contrasting Sima Qian‟s style with Ban Gu‟s by placing parallel

passages side by side. If the parallel texts started off much closer to identical than they

are today, the “stylistic” differences apparent to earlier readers obviously came from

somewhere. Where, if not from the slow manipulations of later readers? Thus Honey‟s

conclusion that the Shiji account was loved to death, so to speak, by its readers:

The original Hsiung-nu account, transmitted in a secondary stage as chapter 94 of

the Han-shu, may have escaped more of the critical attention that was attracted to
the parent narrative in the Shih-chi by the fame of its author, attention that
introduced errors and interpolations over time.176

174
Ibid., 92.
175
Ibid., 90.
176
Ibid., 92.

441
On the other hand, given that Hanshu seemed to have received far more commentarial

attention in earlier periods, the conclusion is clearly a complicated one.

What, overall, does a discussion of this branch of Shiji studies teach us? In

questioning the state of the received text, it often casts doubt on the grand conclusions

and assumptions made by other scholars as to what “Sima Qian” was doing or intended to

do as he compiled the Shiji. But the text-critical approach also has problems of its own.

One is that “authenticity” is not wholly identical to “authorship.” Another is that while

philological evidence may be useful in argument, it tells us more about “edition/record”

than about “text/message.” As far as determining the actual origins of a given chapter,

the available evidence is often inadequate to form a conclusion—as seems to be the case

with the “Da Yuan” (ch.123). Rare cases like the “Sima Xiangru” (ch.117)—where it

seems we may be fortunate enough to have many sets of mutually independent

evidence—are the exceptions: by and large, the Shiji/Hanshu relationship is complex and

admits of no simple conclusions. Finally, rising above the cautious and optimistic

rhetoric employed by these scholars of philological problems in the Shiji/Hanshu

relationship, and considering their actual results, one must admit it: by and large

authenticity studies of the Shiji have reached a deadlock for lack of evidence.

The Appeal of a Badly Damaged Shiji

For late Qing and early twentieth-century scholars, embroiled in the at that time

much more momentous debate about the authenticity of the Classics, the textual

442
destabilization of the Shiji clearly gave them more freedom to use the parts of it that were

useful to them while disregarding what was inconvenient.177

The appeal of the badly damaged Shiji to Western scholars may also be amenable

to similar kinds of contextualization. First, there are methodological issues at stake. Can

we use Western text-critical methods to draw conclusions about Chinese texts? If so, as

Honey and Hulsewé were attempting to demonstrate, then it would provide a powerful

new tool with which to address issues of dating and textual priority. Relatedly, many

scholars who have been drawn to the „drastically damaged Shiji‟ hypothesis are also

concerned to explain the apparent primacy of the Hanshu versions of chapters where

there is overlap.

However, there are other possible explanations for at least some of the evidence.

The Shiji and Hanshu had a close and ongoing relationship dating back to long before our

earliest surviving editions. Editors frequently corrected one text based the other, and

meanwhile each version was also engaged in its own independent process of development.

Most traces of this are lost, but not all. For example, Hanshu commentator Yan Shigu 顏

師古 (581-645) wrote in his preface:

The old text of the Hanshu contained many archaic characters. In the course of
explanation and discussion, these later underwent repeated developments and
changes. Later people got used to the [updated] readings and made changes
according to the meaning [rather than the original wording of the text]. These
[sorts of editions] having been transmitted and copied in great numbers, and [the
text] has grown ever more shallow and popular. Now I have gone back around to
investigate an ancient edition, and returned [the Hanshu text] to its true
correctness. [漢書舊文多有古字，解說之後屢經遷易，後人習讀，以意刊
改，傳寫既多，彌更淺俗．今則曲覈古本，歸其真正．]178
177
The clearest case being Kang Youwei, who used the Shiji‟s references to the “Six Arts” while
dismissing as forgeries all references to texts found in walls. He offered little in the way of evidence, other
than that it is most convenient for his suppositions about the New Text/Old Text debate. These have been
called into question by Michael Nylan (see her article “The chin wen/ku wen Controversy in Han Times”).
178
Hanshu “Xu li,” 2.

443
With this kind of process potentially at work, the value of Western text critical

methodology on the level of individual characters—at least as far as Shiji/Hanshu

comparison is concerned—is open to question.

Asserting the primacy of the Hanshu text and suggesting that dubious later

material may have been inserted into the Shiji are ways of potentially questioning the

strength of Sima Qian‟s authorial role. It allows readers to avoid the entire messy issue

of autobiographical transference. In effect, it takes the Sima Qian out of the Shiji.

Potentially, one could proclaim the “death of the author.” Again, to quote David Honey,

if we “maintain the distinction between message and record, text and edition,” we can

also disregard traditional criteria such as “historiographical personality,” “intellectual

acumen,” and “power of expression.”179 For all the appeal of Sima Qian‟s apparent

strengths in these areas, the subtleties and uncertainties such subjective aesthetic criteria

introduce into the process of interpretation are terribly inconvenient, especially for

scholars more interested in the world „outside the text,‟ the actual events and

circumstances of the Western Han.

The foregoing review of three Shiji authorship controversies „beyond Sima Qian‟

is admittedly preliminary and far from complete. Nonetheless, it illustrates the processes

whereby, in the absence of new or particularly reliable evidence, readers have attempted

to reassign authorship of various parts of the Shiji. At least in part, these attempts were

variously motivated by the desire to elevate—or inversely, to side-line—the Sima Qian

author-function. Other motives were of course operative as well: ideological (in the case

179
Honey, “Textual Criticism,” 74.

444
of Sima Tan‟s authorship), stylistic (in the case of Chu Shaosun‟s), or methodological (in

the case of the drastically damaged Shiji). Nonetheless, the very strength of the impulse

to exclude parts of the Shiji from Sima Qian‟s oeuvre paradoxically reveals his centrality

as a towering symbol of authorship. The tradition had gradually made him into a larger-

than-life figure, and readers eventually felt the desire to escape from the too-long

interpretive shadow that he had begun to cast.

445
 Chapter 8

Sima Qian’s “Letter in Reply to Ren An” and the Idea of Authenticity

ON THE NOTION OF AUTHENTICITY

Sima Qian‟s “Letter in Reply to Ren An”1 has been celebrated both as a great

literary achievement and as an important interpretive key to Sima Qian‟s historical

magnum opus, the Shiji. When considered generally, the “Letter” seems straightforward.

However, in its details it is revealed as a surprisingly problematic text. Chinese scholars

have engaged in extensive debates over the “Letter”, especially as regards the linked

problems of its composition date and its underlying intention. Excellent scholars have

taken opposite positions in these debates. While there is abundant evidence, it never adds

up to a definite conclusion. Some Western sinologists have expressed doubts as to the

authenticity of the “Letter”: aspects of its textual history and its parallels with other texts

have struck them as problematic, although the issue has received little sustained attention.

Both Western skepticism and Chinese debates seem to lead to impasse.

I approach the issue in a new way by reconsidering the whole notion of

authenticity in light of both Chinese and Western observations about the “Letter.” I

address two related questions: First, what exactly is at stake in preserving the “Letter” as

an „authentic‟ document? Second, is authenticity a concept to which the “Letter” can, or

should be expected to, answer? My study includes a careful examination of textual

variants in the “Letter,” its parallels with other texts, and a consideration of the many

types of evidence adduced by Chinese scholars.

1
The two main texts of this letter are to be found in HS 62.2725-2736 and in WX 41.1854-1866.

446
 It may be useful to begin by thinking about what kinds of things we might mean

by authenticity. There seem to be two major types of authenticity which matter to us: I

will call them accuracy and genuineness. By accuracy, I mean the quality of being true:

in accordance with facts, with reality. An extreme example of a document whose

accuracy is the important part of its authenticity is the phonebook. We care very little

who compiled the phonebook, but we care very much whether or not it gives us true

information about phone numbers and addresses. By genuineness, on the other hand, I

mean that something is what it claims to be, especially with regard to origin or authorship.

What we care about in considering a water-lily painting by Claude Monet is that it was

genuinely painted by Claude Monet, and not that it accurately portrays water-lilies.

Of course, these two aspects of authenticity are by no means independent from

one another in most cases. To understand how they interact, let us turn to the case which

concerns us here, Sima Qian‟s “Letter in Reply to Ren An.” We care about the (historical)

accuracy of this “Letter” in two main senses: as a third person account of historical

events (Li Ling‟s battle and defeat, Emperor Wu‟s reaction, etc.); and as a first person

account of its author‟s feelings, motivations, and intentions. We care about the

genuineness of the “Letter”—that it was written by Sima Qian in the early first century

BCE—for reasons that may be related to our concerns with its accuracy. First, the

genuineness of the “Letter” may have a bearing on the reliability of the third person

account: we consider Sima Qian to be a reliable witness for the events surrounding the Li

Ling affair because he was alive and at court during that time.2 Second, the genuineness

of the “Letter” is also essential to the legitimacy of the first person account: we naturally

2
That he might also be an unreliable witness for the same reasons is something we should also take into
account.

447
grant Sima Qian first person privilege regarding his own feelings, motivations, and

intentions. Naturally Sima Qian would know things about himself that a forger would

not know.

The third reason we care about the genuineness of the “Letter” includes—but goes

beyond—both these aspects however: we care very much whether the “Letter” was

written by the same person who wrote the Shiji. This is in part the same as caring that a

particular painting of water-lilies was done by Monet: because a famous name is a

generator and guarantor of value, and because the label of „forgery‟ carries with it so

great a taint that we avoid at all cost having it attached to a work of art we admire. In this

case, however, the issue is larger and more complex. For as long as the “Letter” has

circulated, the version of Sima Qian‟s life story it contains has been available for use as

an interpretive key to the Shiji‟s hidden meaning.3

The “Letter” is a major source for Sima Qian‟s life story, so crucial to our

understanding of it in fact that the Tang historiographical scholar Liu Zhiji 劉知幾 (661-

721) complained (about the Shiji‟s own account of Sima Qian‟s tragedy, found in the

final chapter):

How elliptical it is, him telling his story this way! From what he says—that “he
met with the catastrophe of the Li Ling affair and was hidden in darkness, bound
in black ropes”—it seems like he was captured along with Li Ling, and was
punished for that. And then again it seems like he was entangled with Ling, and
because of that was considered to have committed a crime against the state. In
this way, [Sima Qian] causes the reader to have a difficult time understanding the
details. We must rely on Ban Gu‟s having preserved the “Letter to Ren An,” as it
narrates the entire matter for which Sima Qian was punished. Supposing we did

3
Ban Gu 班固 (32-92), who first anthologized the “Letter” in the Hanshu 漢書 [History of the former Han],
also wrote in his “Dianyin” 典引 [Extension of Constant Models]: “Because [Sima Qian] himself suffered
punishment, he turned to subtle writing and piercing satire, denigrating and detracting from his own
generation” [以身陷刑之故，反微文刺譏，貶損當世] (WX 48.2158). This is the earliest example of a
type of reading that would be common, but not universal, throughout the history of Shiji interpretation. See
discussion in chapter 4 above.

448
 not have this record, would it have been at all possible to understand this matter?!
[自叙如此，何其略哉！夫云遭李陵之禍，幽於縲絏者，乍似同陵陷沒，以
置於刑，又似為陵所間，獲罪於國。遂令讀者難得而詳。頼班固載其與任安
書，書中具述被刑所以。儻無此録，何以克明其事者乎？]4

Of course, as discussed in chapter 6 above, not all scholars support this

interpretive approach to the Shiji. Taking as their slogan Yang Xiong‟s pronouncement

that the Shiji is “a true record” (實錄),5 scholars in this camp have argued for centuries

that Sima Qian was a good historian who did not allow his personal resentment to

contaminate or distort his portrayal of historical events.6 They have not always

succeeded in overcoming the more romantic views of Sima Qian‟s authorship, however.

The Shiji pinglin 史記評林 [Forest of comments on the Shiji], a compilation of

commentarial remarks about the Shiji edited by Ling Zhilong 凌稚隆, came into

circulation at the beginning of the Wanli 萬歷 period (mid-1570s), and marked the

almost total ascendance of autobiographical Shiji interpretation.

Eventually the predominance of autobiographical readings, together with other

factors, produced a reaction in Shiji studies, causing scholars to question whether parts of

the Shiji might not in fact have been written by someone other than Sima Qian.7 It is

important to notice that it was the Shiji‟s authenticity which was called into question by

4
STTS 16.460. See also discussion of this passage in chapter 4.
5
Found in FY 10.413.
6
See discussion in chapter 6.
7
Fang Bao argued that the authorial role of Sima Qian‟s father Sima Tan had been drastically under-
appreciated (Fang Bao ji 2.48, 2.60). Liang Yusheng and Cui Shi argued that the Shiji had suffered
extensive supplementing and interpolation (see Shiji zhiyi 史記志疑 and Shiji tanyuan 史記探源,
respectively). Yu Jiaxi produced an extensive study of the ten missing chapters, challenging the
presumption of the Shiji‟s integrity (see his “Taishi gong shu wangpian kao”). This pessimism about the
Shiji extended to the point that some Western scholars have seriously entertained the possibility that the
majority of the Shiji text was lost and replaced by later forgers (for example, Honey, “Textual Criticism”).
See discussion in chapter 7 above.

449
Chinese scholars, not the authenticity of the “Letter.”8 The “Letter” is a beloved

masterpiece, a ten million dollar Monet. No one, however much they objected to the

overuse of autobiographical interpretations of the Shiji, wanted to discredit the “Letter.”

Even Ban Gu, in some respects one of Sima Qian‟s harshest critics, could remain

unsympathetic toward the author of the “Letter” but could not help admiring his writing.9

Furthermore, in Chinese scholarly culture, texts are generally considered innocent

until proven guilty. What constitutes proof of guilt would in itself be the subject for an

entire monograph, but here two contrasting cases serve as excellent parallels. First, the

“Letter in Reply to Ren An” is not the only letter attributed to Sima Qian. The Gaoshi

zhuan 高士傳 [Traditions of lofty gentlemen], compiled by Huangfu Mi 皇甫謐 (215-

282), includes an exchange of letters, purportedly between Sima Qian and a recluse

named Zhi Jun 摯峻.10 Though the fragment of text here attributed to Sima Qian seems

to contain no glaring anachronisms or problems, I know of no scholar who takes it

seriously as coming from Sima Qian‟s hand. It is tacitly excluded from his corpus,

simply because the source text in which it appears is too dubious.

The second parallel I might briefly mention is with another letter, Li Ling‟s

“Letter in Reply to Su Wu.” Though the Wenxuan editors accepted it as authentic—it

precedes Sima Qian‟s letter in that anthology and stands there as the earliest example of

8
It should also be noted that the parts of the Shiji which fall under the most serious cloud of suspicion were
also rarely the most often-studied and well-loved chapters of the “core Shiji”.
9
Ban Gu‟s is the first recorded reaction to the “Letter”. He wrote:
“Alas! One with Qian‟s breadth of knowledge and experience still failed to understand how to
keep himself from harm. Having suffered the extreme penalty, he was sorrowful and poured forth
his resentment, and his writing rings true. The original traces of this self-inflicted suffering and
shame are found in the “Chief Eunuch” of the Lesser Odes (Mao #200). But what the Great Ode
(Mao #260) says, “Enlightened and wise, he keeps himself from harm” —this indeed is difficult.
烏呼！ 以遷之博物洽聞，而不能以知自全，既陷極刑，幽而發憤，書亦信矣．跡其所以自
傷悼，小雅巷伯之倫．夫唯大雅既明且哲，能保其身，難矣 ！ [HS 62.2738]
10
See discussion of this text in chapter 4.

450
the Chinese literary epistle—Liu Zhiji and Su Shi 蘇軾(1036-1101) both denounced it as

a fake on stylistic grounds. Perhaps the most damning evidence against it is that the

purported first-person description of Li Ling‟s battle with the Xiongnu seems to have

been derived from Sima Qian’s letter, which Li Ling could not have seen and of course

would not have used, in any case, to talk about something he himself had experienced!11

Sima Qian‟s “Letter in Reply to Ren An” suffers from none of these problems.

The Hanshu, which is where the “Letter” first appears, is as solid as any text in the early

Chinese tradition, and generally considered more solid than most. Nor does the “Letter”

seem to contain any glaring stylistic or historical problems (though as later discussion

will show, the devil may be in the details). It seems interesting, then, that Western

scholars from early on have been reluctant to accept the “Letter” as genuine. As I will

show in a moment, they put forth no very convincing arguments. It is not clear, then,

why they sought to question the “Letter‟s” authenticity. One possible motivation might

be the problem of the “Letter‟s” intended audience. Insofar as we believe that Sima Qian

once wrote a letter of this sort, we assume its intended audience was Ren An. Yet the

“Letter” we read today seems to address a far wider audience. The partial incoherence

generated by these different conjectures about the intended audience, at least to my mind,

does create a certain discomfort about the “Letter‟s” authenticity even as though it proves

nothing.

Some readers seem inclined to doubt the accuracy of the “Letter‟s” surface

meaning, and so suspect secret hidden meanings of various kinds. Others doubt the

“Letter‟s” genuineness, because the time and circumstances in which it was purportedly

11
For a thorough discussion of the arguments against the Li Ling letter, see K.P.K.Whitaker‟s “Some Notes
on the Authorship of the Lii Ling/Su Wuu Letters,” parts I-II, and my discussion in chapter 4 above.

451
produced do not seem to accord with the nature of the text. I will explore these problems

in greater detail throughout this section and in my conclusion. But first, I will turn to the

problems of textual overlap between the editions of the “Letter” and between the “Letter”

and various other texts.

THE PARALLEL PASSAGE PROBLEM

Paul Pelliot wrote a review, published in T’oung Pao (1932), which included the

following startling words: “I do not believe in the authenticity of the letter from Sima

Qian to Ren An.”12 The great French sinologist wrote this in response to a reference

from the introduction of Arthur Hummel‟s just-published translation of Gu Jiegang‟s

autobiography.13 Pelliot went on to write:

In addition, [Édouard] Chavannes‟ explication, which finds an allusion to a

passage from the Mu Tianzi zhuan in the phrase „hide it in a famous mountain‟
(藏之名山) runs up against the objection that that work was unknown in the
Han.14 In the absence of any other origin for the phrase, that connection too
would not stand in favor of the same phrase‟s authenticity in Sima Qian‟s
autobiography.15

12
“Je ne crois pas à l‟authenticité de la letter de Sseu-ma Ts‟ien à Jen Ngan” (“Review,” 132).
13
Hummel quoted the line, which appears in both the “Letter” and the “Self-narration,” in which Sima
Qian states his intention to store the completed Shiji in a “famous mountain (library)” (Autobiography of a
Chinese Historian, xxi).
14
For a detailed account of the circumstances surrounding the discovery of this text, see Rémi Mathieu‟s
“Mu t‟ien tzu chuan” in Loewe, Guide, 342-346. The general scholarly consensus is that at least a portion
of it is a genuine Warring States text, discovered in a tomb in the late 3rd century CE.
15
“Par ailleurs, l‟explication de Chavannes qui voit dans 藏之名山 une allusion à un passage du Mou t’ien
tseu tchouan, se heurte à l‟objection que cet ouvrage était inconnu sous les Han; si on ne trouve pas d‟autre
origine, le rapprochement même ne serait en faveur de l‟authenticité de cette phrase de l‟autobiographie de
Sseu-ma Ts‟ien.” (“Review,” 132). In an earlier review, Pelliot had also expressed other doubts about the
“Self-Narration”: “I am not certain that this autobiography, or rather, this postface to the Shiji, had the
absolute worth that one is accustomed to ascribe to it” (Je ne suis pas sûr que cette autobiographie, ou
plutot cette postface au Che-ki, ait la valeur absolue qu‟on est accoutumé de lui accorder), TP 1930: 78 n.4.
This remark, which appears in Pelliot‟s review of Richard Wilhelm‟s 1928 German translation of the Lüshi
chunqiu, seems to be based on nothing other than the liberties the “Letter” takes with chronology in the list
of authors whose misfortunes spurred their creative endeavors.

452
Pelliot never published any supporting evidence for his attack on the “Letter‟s”

authenticity. His use of “in addition” (par ailleurs) seems to imply that the problem with

the “famous mountain” allusion was not his only objection to the “Letter,” but it is

difficult to guess what the other(s) might have been.

As for Pelliot‟s complaint about the Mu Tianzi zhuan 穆天子傳 [Legend of King

Mu] and the “famous mountain,” it was probably occasioned by a footnote in the

introduction to Chavannes‟ translation of the Shiji: “This expression, „famous mountain‟,

designates the palace archives, through an allusion to the Mu Tianzi zhuan where it is said

that on the mountain Qunyu is found a place which the ancient kings called their

archives.”16 This way of understanding the “famous mountain” did not originate with

Chavannes, but goes back to a note by the Tang dynasty Shiji commentator Sima Zhen 司

馬貞 (fl. 745), who wrote, in an attempt to interpret the phrase 藏之名山:

This is saying that the primary copy is hidden in an archive, while the secondary
copy was left at the capital. The Mu Tianzi zhuan says, „The Son of Heaven
marched north. When he reached the Qunyu Mountains, the river was calm and
safe, and there was a place that could be entered in a straight line from any
direction, which the former kings called their archives.‟ Guo Pu said, „The
storehouse where emperors and kings of antiquity hid their texts.‟ Thus it is this
that [Sima Qian] means when he says „I will hide it in a famous mountain.‟ [言正
本藏之書府，副本留京師也．穆天子傳云「天子北征，至于群玉之山，河平
無險，四徹中繩，先王所謂策府」．郭璞云「古帝王藏策之府」．則此謂藏
之名山是也．] 17

It does not seem to me that the “famous mountain” phrase in the “Letter” and “Self-

Narration” has any necessary connection to wording of the Mu Tianzi zhuan. That is, it

16
“Cette expression « montagne célèbre » désigne les archives du palais par allusion à un passage du Mou
t’ien tse tshoan où il est dit que sur la montagne Kiun yu tien se trouvait un endroit que les anciens rois
appelaient leurs archives” (Mémoires historiques, I.cxcviii n.321).
17
SJ 130.3321.

453
surely need not be a direct allusion. As to whether Sima Qian was drawing on the same

tradition as the Mu Tianzi zhuan (and there are other ways of understanding 名山), and

how that tradition came to him, neither of these questions need to bear on the matter of

the “Letter‟s” authenticity. The best interpretation of Sima Zhen‟s comment is not, as

Chavannes understands it, as the identification of an allusion, but as an association in

Sima Zhen‟s mind, his best guess as to what the text might have been pointing to. Even

assuming that the “Letter” and the Mu Tianzi zhuan are both referring to the same sort of

“mountain,” both could have been referring independently to a phenomenon which

actually existed either geographically or—more likely—symbolically, as part of a living

tradition. The connection need not have been textual.

Édouard Chavannes himself also entertained doubts about the “Letter‟s”

authenticity, though he relegated them to a footnote in his introduction. He wrote:

Though Chinese critics have never had any reservations about the authenticity of
this piece [i.e., the “Letter”], it could give rise to certain suspicions in someone
with a rigorous mind. One finds there, in effect, a passage from the biography of
Li Ling and a passage from the autobiography of Sima Qian. The presence of the
first text proves little. This letter is cited in the History of the Former Han. There
is nothing which leads us to believe that the biography of Li Ling is anterior to it
or that the compiler of the biography did not just reproduce part of the letter. But
the presence of the second passage is more difficult to explain. One would have
to admit that Sima Qian was so inveterate a compiler that even in writing to his
friend he would copy from himself.18 Still, this is not absolutely impossible, and
this is why we will adopt the accepted opinion in China that the letter is
authentic.19

18
The generally accepted opinion among Chinese scholars today is that the “Letter” came first, and so the
parallel passage in the “Self-Narration” was copied from the “Letter” instead. See, for example, Zhang
Dake‟s Sima Qian pingzhuan 司馬遷評傳, pp.364-365.
19
“Quoique les critiques chinois n‟aient fait aucune réserve sur l‟authenticité de cette pièce, elle peut
exciter quelques soupçons chez un esprit exigeant ; on y retrouve en effet un passage de la biographie de Li
Ling et un passage de l‟autobiographie de Se-ma Ts’ien ; la présence du premier texte, ne prouverait à vrai
dire pas grand‟chose, car, cette lettre étant citée par l‟Histoire des premiers Han, rien ne peut faire croire
que la biographie de Li Ling lui soit antérieure, et que ce ne soit pas au contraire le rédacteur de cette
biographie qui a reproduit une partie de la lettre. Mais la présence du second passage est plus difficile à
expliquer ; il faudrait admettre que Se-ma Ts’ien était un compilateur si invétéré qu‟en écrivant à son ami il
s‟est copié lui -même. Cela n‟est pas d‟ailleurs absolument impossible et c‟est pourquoi nous adoptons

454
Chavannes‟ worries about the authenticity of “Letter” center on the problem of textual

overlap, an interesting consideration which I will discuss in detail below. The problem of

textual overlap in the Shiji was part of what led to Chavannes‟ disillusionment with Sima

Qian‟s authorial role—Chavannes had hoped to find an original creative genius and

instead found something he was less able to admire, an “inveterate compiler.”20

As the above discussion suggests, Western scholars‟ worries about the “Letter‟s”

authenticity tend to focus on its textual relationships and parallels. In fact, a careful

investigation of this issue shows more reason for concern than most Chinese scholars

might admit. There are two major recensions of the “Letter”—found in the Hanshu and

in the Wenxuan—and a third partial version, which appears in the Hanji. Beyond that,

there are also many passages in the “Letter” which have parallels in other texts. I will

begin by sketching out what these versions look like, and then try to suggest what kinds

of arguments might be made about the parallels and differences in these varying contexts.

As discussed above, the “Letter” makes its first appearance in Hanshu chapter 62.

There it is preceded by a slightly variant version of Shiji “Self-Narration” (ch.130) and by

an introduction, presumably written by Ban Gu, which reads:

After Sima Qian suffered punishment, he served as Director of Palace Secretaries,

a position of honor and favor. His old friend, the Regional Inspector of Yizhou,
Ren An, gave Sima Qian a letter, charging him with the duty [incumbent on]
worthy ministers of old. Qian answered him, saying: To the honorable

l‟opinion admise en Chine que la lettre est authentique.” (Mémoires historiques, I.xlii n.55). J.J.L.
Duyvendak, citing Chavannes‟ doubts, expresses similar reservations (see his “Review of Shryock, The
Origin and Development of the State Cult of Confucius,” 332).
20
See discussion of Chavannes‟ changing attitude toward the Shiji in the “Introduction” above.

455
 Shaoqing… [遷既被刑之後，為中書令，尊寵任職。故人益州刺史任安予遷
書，責以古賢臣之義．遷報之曰：少卿足下。。。]21

A vexing mystery in the textual history of the “Letter” is how it came into Ban

Gu‟s hands. How did it survive the tumultuous hundred and fifty years between its

ostensible composition date22 and Ban Gu‟s inclusion of it in the Hanshu? What might

have happened to the text during that time? And who might have read it? The responses

of certain early Shiji readers at least suggest that the “Letter‟s” version of Sima Qian‟s

story was not widely known or accepted during this period.23 But there is no direct hint

of provenance for the document that came into Ban Gu‟s hands, nor is there direct

evidence that anyone else had read it.24

Another version of the “Letter” was included in the much-later Wenxuan, an

anthology that was compiled in the court of Liang 梁 under the auspices of the Crown

Prince Xiao Tong 蕭統 (501-531). Though in many ways closely parallel, this second

version of the “Letter” has a few striking differences from the version found in the

Hanshu. These differences suggest that the Hanshu text was not the only version of the

“Letter” the Wenxuan compilers had to work with. This is not so unreasonable a

supposition if we keep in mind the large-scale extinction of predecessor anthologies

which eventually resulted from the Wenxuan‟s popularity. The Wenxuan text of the

21
HS 62.2725.
22
Generally accepted to be either 93 or 91 BCE. The controversy over the dating of the letter will be
addressed below.
23
See, for example, HHS 40.1327 and SJ 130.3321.
24
Only one circumstance strikes me as strange and potentially suggestive. Though various Hanshu
references prove that Ren An was an actual historical person, the vast majority of our information about
him comes from a short passage appended by Chu Shaosun 禇少孫, the infamous “third author” of the Shiji,
to the end of Shiji ch.104. Nearly every scholar of the “Letter” mentions this fact, but no one I know of has
really considered its possible implications. Why was it that Sima Qian did not write about Ren An? Why
was it that Chu Shaosun did? Is it not surprisingly convenient that Chu has given us so much information
about Sima Qian‟s correspondent? There is nothing conclusive about this circumstance, of course, but it
does strike me as a potential indication that Chu Shaosun saw the “Letter.”

456
“Letter” has a title, “Letter in Reply to Ren Shaoqing” 報任少卿書, and does not

reproduce Ban Gu‟s introduction. It does, however, include an opening phrase omitted

from the Hanshu, one which has proved most puzzling for scholars and commentators:

“Sima Qian, who [labors like] a horse or cow [in the cause of the] Honorable Senior

Archivist, repeatedly pays his respects, saying, to the honorable Shaoqing…” [太史公牛

馬走司馬遷再拜言，少卿足下…].25

It is not uncommon for editors of histories to make small or not so small changes

to the texts they anthologize.26 A reasonable explanation for this textual difference could

well be that Ban Gu removed the opening twelve characters because he found them

inappropriate or difficult to understand, or simply thought they did not fit the rhythm of

his compilation. Qian Mu 錢穆 (1895-1990) supports this possibility, writing, “Ban

[Gu‟s Han]shu preserves a record of this text, but he simply cuts out the opening phrase”

[班固書存錄此文，獨削去其首句].27 If this is correct, it is evidence that the Wenxuan

editors had a version of the “Letter” that was independent of the Hanshu text. If they

only had the Hanshu version of the text, how would they know what to supply?

Another striking difference between the two versions appears near the end of the

“Letter”. The context is a discussion of Sima Qian‟s purpose in compiling the Shiji. The

25
WX 41.1854. My translation relies partially on the interpretation of Ruan Zhisheng, “Sima Qian zhi xin,”
196-203, but the whole issue is quite controversial. Wu Renjie 吳仁傑 (ca.1204) first suggested that the 牛
be amended to 先. See discussion in Knechtges, “Key Words,” 5-6.
26
One notable example is that of the Hou Hanshu and Wenxuan versions of Ban Gu‟s “Dianyin.” The
Wenxuan version (WX 48.2158) includes Ban Gu‟s own preface (more than two hundred characters long),
while the Hou Hanshu omits it almost completely, except for a short sixteen character segments which Fan
Ye included apparently only for the purpose of making an ironic comment about it at Ban Gu‟s expense
(see HHS 40B.1375).
27
From “Taishigong kaoshi” 太史公考釋 [The Honorable Senior Archivist investigated and explained], 21.
Qian Mu goes on to add, “He did not know that these six characters actually contain Sima Qian‟s most
essential purpose in this piece and are not incidental or empty words” [顧不知此六字，乃遷此文最要用
意之所在,非偶浮文也].

457
following chart shows the parallels and differences between the two versions. Where the

phrases are identical, the two columns are merged into one.

Table 3: Comparison of HS 62.2735 and WX 41.1865 Variants

Hanshu 62.2735 Wenxuan 41.1865

Although I am not worthy, I have entrusted my meaning to useless phrases,
僕竊不遜，近自託於無能之辭，
gathering up and bringing together the old traditions of the world which were scattered
and lost,
網羅天下放失舊聞，
investigating conduct and events, generally investigating the conduct and
考之行事， events therein, putting together the ends
and the beginnings,
略考其行事，綜其終始，
and examining their principles of success and examining the guidelines of their
and failure, rise and decline. success and failure, rise and decline.
稽其成敗興壞之理， 稽其成敗興壞之紀，
Beginning from Xuanyuan and coming
down to the present, I am making ten
charts, twelve basic annals, eight
treatises, thirty hereditary households,
and seventy associated traditions,
上計軒轅，下至于茲，為十表，本紀十
二，書八章，世家三十，列傳七十，
in total, one hundred and thirty chapters.
凡百三十篇。

The main difference is that the Wenxuan version contains an outline of the Shiji‟s

structure, while the Hanshu version does not. Again, this is more easily explained as a

passage Ban Gu edited out, rather than as one that some subsequent editor added. The

reason is that Ban Gu‟s version of the “Letter” appears directly after a version of Sima

Qian‟s “Self-Narration”—and the latter text of course contains a detailed table of

contents and explanation of the Shiji‟s structure.

458
 Furthermore, a careful examination shows that the Wenxuan summary places the

sections in a slightly different order than they appear in today‟s Shiji: the tables here

appear first, before the basic annals. Ruan Zhisheng, Zhang Dake, and others have

suggested that this unusual ordering is evidence that a) the “Letter” was written before

the “Self-Narration”, and b) that the Shiji was roughly planned but not completed when

the “Letter” was written.28 If they are correct, and the ordering here does reflect an

earlier conception of the Shiji‟s structure, it provides even better motivation for Ban Gu

to have cut this passage: he would have been eliminating not mere redundancy but actual

contradiction. The editors of the Wenxuan, a literary anthology, had placed the “Letter”

outside a historiographic context. They may have cared more about the actual words of

the text than about whether it agreed in every detail with the Shiji‟s structure—

privileging genuineness over accuracy.29

I will say more about this section of the Wenxuan “Letter” when it comes to

discussing the “Letter‟s” parallels with other texts. At this point, however, it is worth

addressing the third, partial version of the “Letter” found in the Xun Yue‟s 荀悅 (148-

209) Qian Hanji 前漢紀 [Annals of the former Han]. The existence of this version is

rarely mentioned in Chinese scholarship,30 whether because it is not widely known or

28
See for example Ruan Zhisheng‟s “Sima Qian zhi xin,” 181-182 and Zhang Dake‟s Sima Qian pingzhuan,
365.
29
Of course another possibility is that there was some version of the Shiji in which listed the tables came
first. We do know that chapters of the Shiji tended to circulate separately, especially early in its history, so
a reordering is not inconceivable. However, the carefully enumerated Shiji table of contents, versions of
which existed in both the Shiji and Hanshu, tends to argue against any stable ordering or reordering in
which the tables appeared first. Furthermore, there is a convincing story to be told about Sima Qian‟s
original conception being to give a chronological overview followed by annalistic records of reigns—a
scheme which symbolically placed time and history above imperial power. This original conception could
well have been altered due to the very realities of that power—it might eventually have seemed wiser to put
the annals of rulers first in place and dignity, while relegating the chronological overview to second place.
30
Ruan Zhisheng is an exception, but he mentions it only in passing (“Sima Qian zhi xin,” 153).

459
because it is not considered important. However, Xun Yue‟s abridged version of the

“Letter” has a number of fascinating features.

Chronologically, the “Letter” appears in Tian Han 2 (99 BCE), directly after a

description of Li Ling‟s betrayal and its aftermath. Xun Yue thus fails to give us any

evidence about when the “Letter” was written. In addition, his version omits most of the

chronological clues by which scholars attempt to identify the date of the “Letter‟s”

composition, and most of text which seems like a direct reply to Ren An. Indeed, so

much is omitted that Xun Yue‟s version is not even one third the length of the Hanshu

and Wenxuan versions. The pattern of his omission and reshuffling is actually quite

complex, and worth considering closely.

Xun Yue‟s version of the “Letter” focuses initially on Sima Qian‟s official career.

It begins, “It has been more than thirty31 years since I inherited my father‟s office and

entered the service of the emperor” [僕賴先人緒業，得待罪輦轂下，三十餘年矣].32

There is the modest narrative about Sima Qian‟s failure to distinguish himself, leading up

to his mutilation. Xun Yue‟s version then skips back to a passage which appears earlier

in the Hanshu and Wenxuan versions, the one which cites historical examples of how

moral gentlemen refuse to have anything to do with eunuchs. Xun Yue then skips

forward again to Sima Qian‟s supposedly undistinguished early life and service, leading

directly into the description of the Li Ling affair.

The Li Ling narrative appears more or less in the conventional order, but is much

shortened. One effect of the abridgement is to confuse descriptions of the Xiongnu and

31
The Hanshu and Wenxuan versions have “twenty” instead of “thirty” here, making this an intriguing
further example of the ten year displacement problem which plagues Sima Qian‟s nianpu 年普
(chronological biography)—see discussion below.
32
HJ 14.146.

460
Han armies. In particularly, the “Letter” makes clear that the following lines were a

description of the Xiongnu response to Li Ling‟s campaign: “Those who tried to rescue

their dead and wounded could not even save themselves….All men who could hold a

bow were called out. The whole nation attacked at once and surrounded them” [虜救死扶
33
傷不給。。。舉引弓之民，一國共攻而圍之]. However, the recontextualization in the Hanji

version makes it possible and perhaps even likely that these lines be taken to refer to the

Han army:

The infantry that Li Ling commanded did not come up to five thousand. They
marched deep into barbarian territory….In fearless ranks they shouted a challenge
to the powerful barbarians, gazing up at their numberless hosts. Those who tried
to rescue their dead and wounded could not even save themselves. All men who
could hold a bow were called out. The whole nation attacked at once, fighting
their way along for thousands of miles until their arrows ran out and the road was
cut off. The relief forces did not come, and our dead and injured lay heaped up.
[且李陵提步卒不滿五千，深距戎馬之地，足歷王庭，垂餌虎口，橫挑彊胡，
挫億萬之師，虜救死扶傷不給。悉舉引弓之民，一國共攻之，轉 千里，
矢盡道窮，救兵不至，士卒死傷如積。]34

Whether this is an example of careless abridgement or intentional obfuscation, it shows

how seemingly innocuous editorial practices like elision can still have significant effects

on the portrayal of historical circumstances—and by extension, on the aspect of

authenticity that has to do with accuracy.

There are a number of textual variants between the Hanshu and Hanji versions of

the Li Ling/battle section of the narrative, but they are mostly of the omission or

reshuffling type. One interesting exception is the Hanji addition of an entire phrase:

“The losses Ling inflicted [on the Xiongnu] were such that his merit was still worthy to

33
HS 62.2729.
34
HJ 14.147. Compare with the Hanshu version of the same passage (differences in bold): 且李陵提步卒
不滿五千，深踐戎馬之地，足歷王庭，垂餌虎口，橫挑彊胡，卬億萬之師，與單于連戰十餘日，所
殺過當。虜救死扶傷不給，旃裘之君長咸震怖，乃悉徵左右賢王，舉引弓之民，一國共攻而圍之。
轉岗千里，矢盡道窮，救兵不至，士卒死傷如積。(HS 62.2729)

461
be proclaimed throughout the world. I believe that the reason Ling did not die was

because he intended to try to seek some future opportunity to repay his debt to the Han”

[身雖陷敗，其所摧破，亦足暴功於天下．僕以為陵之不死，直欲得當報漢

也].35 The Hanshu and Wenxuan versions have, in place of the emphasized phrase, “If

one looks broadly at his intention” [彼觀其意…].36 Clearly the Hanji‟s phrase is similar

in meaning and function. An interesting feature worth noting, however, is that the Hanji

has actually borrowed the added phrase from the parallel passage in Hanshu ch.54, “The

Arrayed Traditions of Li Ling” 李陵列傳. There, the Hanshu reuses passages from the

“Letter” to describe the Li Ling incident (discussed below). The HS ch.54 counterpart to

the italicized phrase is “the reason he did not die…” [彼之不死…].37 Xun Yue‟s change

accurately conveys the same meaning as the „original‟ phrase in the other texts, but does

not the genuineness of Xun Yue‟s text seem lessened?

Having finished describing the Li Ling affair, Xun Yue then brings in a passage

from earlier in the “Letter”, the one which describes the ruler‟s reaction to Li Ling‟s

defeat. Eliding the disorganized response of the ministers and Sima Qian‟s own

hesitation, Xun Yue skips directly to the point at which Sima Qian speaks his mind, also

including the ruler‟s suspicions. This leads directly to Sima Qian‟s punishment and his

decision not to commit suicide. Here again we find the passage that lists historical

figures who have created great literature because of (or in the course of) their misfortunes.

The list is truncated, however, and lacks any mention of Han Feizi, Lü Buwei, or the

35
HJ 14.147.
36
HS 62.2730; WX 41.1859.
37
HS 54.2456.

462
authors of the Shijing poems.38 Xun Yue‟s version of the “Letter” then ends with the

highlights of Sima Qian‟s work on the Shiji, concluding with the words, “When I have

truly completed this work, I shall deposit it in the Famous Mountain, so that it will be

handed down to later people. And though I should then suffer a thousand mutilations,

what regret would I have?” [僕誠以著此書，藏之名山，傳之後人。雖萬被戮，

豈有悔哉].39

It is easy to accuse Xun Yue of being an even more “inveterate compiler” than

Sima Qian. After all, his main claim to historiographical fame is that he chopped up and

reordered the Hanshu.40 Why should he not do the same with the “Letter in Reply to Ren

An”? Still, even being aware that Xun Yue was a heavy-handed editor, we might be

surprised by the extent to which he altered the text and still referred to it as Sima Qian‟s

letter. This might have something to do with our different expectations of different

genres. Historical accounts seem like fair game for heavy editorial interference. In both

Chinese and Western historiographical traditions, narrative reorderings and

manipulations are the tools which historians have always used to encode their own

judgements into their accounts, or—as in the case of the Hanji—to advocate the

advantages of a different structural model. The difference is that the Hanji was

considered a work in its own right, independent of the Hanshu, just as the Hanshu was

considered an independent work despite its heavy borrowing from the Shiji. Contrary to

modern Western intuitions, reordering was a type of authorial work in traditional China,

38
See discussion in chapter 1 above.
39
HJ 14.147.
40
For a study of Xun Yue‟s life and thought, see Chen Chi-yun‟s Hsün Yüeh (A.D. 148-209): the Life and
Reflections of an Early Medieval Confucian.

463
work for which the person doing it was credited and/or held responsible.41 We cannot

imagine Xun Yue presenting his reorganized text as “Ban Gu‟s Hanshu” but this is the

equivalent of what he seems to have done with Sima Qian‟s letter.

There are two ways to approach this issue. The first is to dismiss Xun Yue as an

anomalous case, someone who held an unusually irresponsible attitude toward the

integrity of any given text. The second is to admit that what Xun Yue did to the

“Letter”—while still presenting it as Sima Qian‟s work—was perfectly acceptable in his

time and perhaps earlier as well. The above comparison between the Hanshu and

Wenxuan versions certainly implies that Ban Gu made a few changes, though nothing on

the scale of the Xun Yue‟s. Do we take Ban Gu as typical and Xun Yue as extreme, on

the scale of invasive editing? Or, in light of other Han editorial projects, should we take

Xun Yue as typical and Ban Gu as fairly conservative? The latter option at least opens

the possibility that the “Letter” we have today is potentially quite dissimilar to whatever

text came from Sima Qian‟s brush near the beginning of the first century BCE.

I turn now to parallels between passages in the “Letter” and other texts. The

longest of these, noted by Chavannes in the passage quoted above, is between the

“Letter‟s” description of Li Ling‟s fateful battle and the Hanshu account of Sima Qian‟s

speech, included in the Li Ling chapter (HS ch.54). That account begins with the words,

“The emperor asked the Senior Director Archivist Sima Qian [about Li Ling] and Qian

said with great enthusiasm, „Ling is filial in serving his parents and trustworthy towards

his men…‟” [上以問太史令司馬遷，遷盛言：陵事親孝，與士信…].42 It then goes

41
See the discussion of Zheng Qiao in chapter 1 above.
42
HS 54.2456.

464
on with a somewhat shortened but clearly parallel version of the description in the

“Letter.”43 A curious feature of the way Ban Gu used (or reused) this material is that in

Hanshu ch.54 it is presented as direct speech.

I suspect that the most likely explanation for this parallel is that Ban Gu wanted to

dramatize this significant moment in history and, though he did not know exactly what

Sima Qian had said, he assumed it would have basically resembled what appears in the

“Letter.” Thucydides too engaged in a somewhat analogous form of imaginative

reconstruction,44 and certainly Sima Qian himself did so.45 David Knechtges, however,

has raised intriguing alternate possibilities:

We cannot be certain whether Ban Gu borrows directly from the Ren An letter, or
cites from a court account of Sima Qian‟s oral presentation to the emperor. It is
even conceivable that this section of the letter is derived from the proceedings of
the court council in which Sima Qian offered his defense of Li Ling.46

As Chavannes wrote, “there is nothing which leads us to believe that the biography of Li

Ling is anterior to [the „Letter‟], and that the compiler of the biography did not just

reproduce part of the letter.” Still, if Ban Gu had available to him the kind of court

account suggested by Knechtges, it could have been a source for the “Letter”. Of course

this is mere speculation. Nor do text-critical principles developed by Western

43
Many of lines contained in the “Letter” but lacking in the Hanshu 54 passage are also lacking in the
Hanji version of the letter, although the Hanshu 54 version is much more abridged even than the Hanji
account.
44
Thucydides was explicit about this, writing:
In this history, I have made use of set speeches some of which were delivered just before and
others during the war. I have found it difficult to remember the precise words used in the speeches
which I listened to myself and my various informants have experienced the same difficulty; so my
method has been, while keeping as closely as possible to the general sense of the words that were
actually used, to make the speakers say what, in my opinion, was called for by the situation.
(Peloponnesian War, I.22: 47)
45
One notorious example: the debate over the Qin succession at Shaqiu. See discussion in Bodde, China’s
First Unifier, 12-55.
46
Knechtges, “Key Words,” 8.

465
philologists offer much help in this instance— lectio brevior potior47 does not apply

reliably in a textual culture where editors were just as likely to abridge as to expand on

their source texts. While it does not seem particularly likely that a “Li Ling affair

transcript” was used to forge or expand the “Letter”, the possibility should not be entirely

discounted. This would not necessarily detract from the accuracy of the “Letter” (at least,

the Li Ling section), but it certainly has problematic implications for its genuineness.

Another extensive parallel is between a passage in the “Letter” and one in the

Shiji version of the “Self-Narration.” (In the Hanshu version of the “Self-Narration”, this

section has been eliminated, further evidence that Ban Gu used editorial license to excise

repetitions between the two texts.) This is one of the most famous passages attributed to

Sima Qian, regarding authors of the past who had encountered misfortune and produced

great works. The following table gives a comparison of the two versions.

47
I.e., “the shorter reading is the more probable reading.”

466
 Table 4: Comparison between “Self Narration” (SJ 130.3300) and “Letter” (WX
41.1864-5)

“Self-Narration” Passage (SJ 130.3300) “Letter” Passage (WX 41.1864-5)

西伯拘羑里演周易 蓋文王48拘而演周易
The Chief of the West was imprisoned at King Wen was imprisoned, and he expanded
Youli, and expanded the Changes. the Changes
孔子戹陳蔡作春秋 仲尼厄而作春秋
Confucius was in distress between Chen and Confucius was in distress, and he made the
Cai and he made the Chunqiu Chunqiu
屈原放逐著離騷 屈原放逐乃賦離騷
Qu Yuan was exiled, and composed the Li Sao Qu Yuan was exiled, and thereupon chanted
the Li Sao
左丘失明厥有國語
Zuo Qiu lost his sight, and produced the Guoyu
孫子臏腳而論兵法 孫子臏腳兵法脩49列
Sunzi had his feet amputated, and he set forth Sunzi had his feet amputated; the Bingfa was
the Bingfa compiled and arranged
不韋遷蜀世傳呂覽
Lü Buwei was banished to Shu; his Lü Lan has been handed down through the ages
韓非囚秦說難孤憤
Han Feizi was held prisoner in Qin; „The Difficulties of Disputation,‟ „The Sorrow of Standing
Alone‟
詩三百篇大厎聖賢發憤之所為作也
The three hundred odes of the poets were written when sages and worthies poured forth their
anger and dissatisfaction
此人皆意有鬱結
These men all had in their hearts a pent-up sorrow
不得通其道也 不得通其道
that they were unable to carry out their Ways and were unable to carry out their Ways
故述往事思來者
Therefore they transmitted events of the past, thinking of those to come.

Commentators‟ varying responses to this passage in its two different contexts

suggest that the passage fits more comfortably in the “Letter” than in the “Self-

48
Note that the Hanshu version and the Five Ministers edition of the Wenxuan both have 西伯, not 文王.
49
The Hanshu version has 修 instead of 脩.

467
Narration.”50 This may also be the reason why Ban Gu chose to cut the “Self-Narration”

version rather than that of the “Letter.” But what does this tell us about the relative

priority of the two passages? Again, by the principles of Western textual criticism, lectio

difficilior potior:51 the passage that seems more problematic (i.e., the “Self-Narration”)

would be preferred. Yet for parallels on this scale it is just as easy to imagine the more

organic structure of the “Letter” being first, and the more awkward fit of the “Self-

Narration” being an artifact of compilation. The uncertainties are numerous and daunting.

We could even speculate that Ban Gu‟s version of the “Self-Narration” is the more

faithful one in this respect, and some later aficionado took a passage from the “Letter”

and interpolated it into the Shiji “Self-Narration.” On the other hand, if we choose to

believe in the integrity of the “Self-Narration”, we could judge this passage in the “Letter”

as the one whose genuineness is most secure—these words at least came from Sima

Qian‟s own brush.

The other major parallel between the Wenxuan version of the “Letter” and the

Shiji version of the “Self-Narration” is far more complex. But though the dissimilarities

are great, phrase-level commonalities remain easily identifiable. This parallel occurs just

at one of the places where the Hanshu and Wenxuan versions of the “Letter” differ

significantly (see Table 1 above). The parallel in the “Self-Narration” occurs within the

framework of the “table of contents” section, where each chapter of the Shiji is poetically

summarized. The summary for the last chapter is much extended and here we find the

repeated material, beginning with the phrase, “I have gathered up and brought together

50
See, for example, objections to the historicity of the list by Takigawa (SKK 10.5209), Liang Yusheng
(Shiji zhiyi, 3.1470) to the “Self-Narration” version. In contrast, no Wenxuan commentator on the same
passage (WX 41.1864) raises any such objections.
51
I.e., “the more difficult reading is the more probable one.”

468
the old traditions of the world which were scattered and lost” (c.f. Table 1, second line).

The “Self-Narration” version then adds, “…the traces of kings and their deeds” [王跡所

興], which has no parallel in the “Letter”. The “Self-Narration” continues, “I have

sought out beginnings and investigated endings, looked upon flourishing and surveyed

decline, discussed and investigated conduct and affairs” [原始察終，見盛觀衰，論考之

行事]. This is loosely parallel to the Wenxuan version of the next few phrases,

“generally investigating the conduct and events therein, putting together the ends and the

beginnings,52 and examining their records of success and failure, rise and decline [略考其

行事，綜其終始，稽其成敗興壞之紀]. The ordering is different, however.53 The next

two phrases in the “Self-Narration” have no parallel in the “Letter”: “I have generally

extended [my history] to the Three Dynasties, and made a record of the Qin and Han” [略

推三代，錄秦漢]. The loose parallelism of this section suggests that the basic wording

of one text was reworked for the different needs and context of another, though I will not

venture an opinion as to which was prior.

The next two lines are almost directly parallel between the “Self-Narration” and

the Wenxuan “Letter”: “I begin my record54 with the imperial carriage and end by coming

down to the present” [上記軒轅，下至于茲]. This is a problematic line because the

“Self-Narration” contains two other different statements about the chronological scope of

the events recorded in the Shiji, causing endless occasion for speculation and

52
More comfortably translated into English as “cycles,” but I have chosen a literal rendering to preserve the
parallelism.
53
For differences in the Hanshu version, see Table 3.
54
C.f. the Wenxuan “Letter”: “I begin calculating from the imperial carriage…” [上計軒轅].

469
disagreement on the part of commentators.55 This is one of places in the “Self-Narration”

where its lack of integration shows through most clearly. However for the present

purposes, the best hypothesis is that this chronological statement belongs with the rest of

the passage around it, and was included in part because it was organically and historically

connected with its context. It is also the broadest chronological scope of the three.

What occurs next in the Wenxuan “Letter” is the structural outline highlighted in

Table 1 above: “I am making ten charts, twelve basic annals, eight treatises, thirty

hereditary households, and seventy associated traditions, in total, one hundred and thirty

chapters” [為十表，本紀十二，書八章，世家三十，列傳七十]. The “Self-Narration”

version is much longer, and reflects the order of today‟s Shiji. It contains remarks

justifying and discussing the various sections. In short, it is consonant with the

conventionally accepted view that the Wenxuan “Letter” reflects an earlier, less complete,

conception of the Shiji and the “Self-Narration” gives a completed version. Thus, this

particular overlap argues in favor of the “Letter‟s” genuineness (or the genuineness of

this portion of the “Letter”), at least insofar as we accept the genuineness of the “Self-

Narration.”56

In addition to the major parallels discussed above, there are numerous minor

parallels between the “Letter” and passages in the Shiji. For example, the Honorable

Senior Archivist‟s comment at the end of the “Tradition of Wu Zixu” (SJ ch.66) says, “If

earlier Wu Zixu had followed [his father Wu] She and died with him, how would it be

55
See SJ 130.3300, SJ 130.3320.
56
Why not entertain the possibility that the “Letter” is a shortened version of the “Self-Narration,” and that
the “Self-Narration” preceded the “Letter”? Certainly one could do so, but would then need to find some
explanation for the incorrect ordering of the sections. Furthermore, the “Letter” presents itself as preceding
the completion of the Shiji, while the “Self-Narration” purports to describe a completed text. If we want to
reverse the order of the two texts, we must accuse one or both of lying about themselves. This is not
impossible, of course, but one would not want to do so lightly.

470
different from an ant or a mole-cricket?” [向令伍子胥從奢俱死，何異螻蟻].57 This is

syntactically and lexically quite a close parallel to the phrase in the “Letter”: “If I had

submitted to the law and allowed myself to be executed, it would be like nine oxen losing

one hair; how would it be different from an ant or a mole-cricket?” [假令僕伏法受誅，

若九牛亡一毛，與螻蟻何以異？].58

Close but brief Shiji/“Letter” parallels like this (and they are numerous) provide

material for exuberant autobiographical readings. But this is because readers are

accustomed to accepting the integrity and genuineness of the “Letter.” The phrase from

the “Letter”, quoted above, is actually a rather awkward mixing of metaphors—is it like

both an ox-hair and an insect? It is easy to imagine an editorially-minded reader noticing

the similarity of sentiment and adding the last phrase to the “Letter” in reference to the

comment at the end of the “Wu Zixu” story. Again, I do not rule out the possibility that

that editorial hand could still have belonged to the “inveterate compiler” Sima Qian.

Though parallels like the one above are not particularly conclusive, they do tell us

something about the textual culture of early and medieval China, the fluidity with which

lines from one text could melt into lines of another. When this process reaches a certain

extreme of ubiquity, it actually becomes difficult to maintain a meaningful notion of

authenticity, at least in the sense of genuineness.

DEBATES ON THE DATING OF THE LETTER

The date of Sima Qian‟s “Letter in Reply to Ren An” has excited surprisingly

strong disagreements among Chinese scholars. The two main theories are Wang

57
SJ 66.2183.
58
WX 41.1860. C.f. HS 62.2732, where the final phrase is the slightly variant 與螻螘何異.

471
Guowei‟s 王國維 (1877-1927) argument that the “Letter” was written in 93 BCE (Taishi

太始 4),59 and Zhao Yi‟s 趙翼 (1727-1814) that it was written in 91 BCE (Zhenghe 征和

2) during the Wugu Affair.60 Excellent scholars are to be found on both sides of the

debate, and their positions do not follow in strict accordance with scholarly filiations:

Zhang Dake 張大可, for example, spends more than a page of his discussion lamenting

and apologizing for the fact that he is forced to disagree with his friend and teacher

Cheng Jinzao 程金造.61 The disagreement is all the more intense because most everyone

concerned considers it to be resolvable; they simply do not agree on the solution.

On the most basic factual level, there are three main pieces of evidence, all of

which occur very close to each other in the early part of the “Letter”:

1. I should have answered your letter, but it happened that I had just returned from
accompanying the emperor to the east (or: I had just come back to the east with
the emperor)… [書辭宜答，會東從上來]62

2. Now you Shaoqing stand accused of this incalculable63 crime. Weeks and
months have passed, and the end of winter will soon be upon us… [今少卿抱不
測之罪，涉旬月，迫季冬]64

59
See Wang, “Taishigong xingnian kao.”
60
Nian er shi zhaji 廿二史劄記 1.1. Chinese scholars almost invariably cite Zhao Yi as the source of this
argument, even though Wenxuan commentator Lü Xiang seems to have been the first to make a statement
on the matter. He wrote, commenting on Ren An‟s “incalculable crime”: “An was imprisoned in jail
because of the matter of Crown Prince Li” [安爲戾太子事囚於獄] (Liu chen WX 41.14a). Lü Xiang‟s
comment is mentioned in Burton Watson‟s excellent discussion of the “Letter‟s” dating (see Watson, Ssu-
ma Ch’ien, 194-198).
61
Shiji yanjiu, 91-92.
62
HS 62.2726; WX 41.1855.
63
Tang dynasty Hanshu commentator Yan Shigu 顏師古 (581-645) glosses bu ce zhi zui 不測之罪 as
“very serious” [不測謂深也] (HS 62.2727). Wenxuan commentator Lü Xiang, on the other hand, explains
it as meaning “whether he will live or die cannot be known” [不測謂生死不可知] (Liu chen WX 41.14a).
Though “incalculable” is not a perfect translation, it to some extent manages to capture both possibilities.
C.f. Chavannes, “crime insondable [roughly: „unfathomable crime‟]” (Mémoires Historiques, I.ccxxvii);
Watson, “terrible” (Ssu-ma Ch’ien, 58), Hightower, “an accusation whose outcome is uncertain” (Birch ed.,
Anthology, 95).
64
HS 62.2726; WX 41.1855.

472
 3. Furthermore, I am forced to go with the emperor to Yong, and I fear the sudden
occurrence of that which cannot be avoided… [僕又薄從上(上)雍，恐卒然不可
為諱]65

Ren An lost his life sometime during the Wugu Affair of 91 BCE.66 Based on the

apparently serious crime of which Ren An stands accused (in item 2), and the veiled

reference to his possible execution at the end of winter (item 3), Zhao Yi dated Sima

Qian‟s “Letter” to the eleventh month of Zhenghe 2 (91 BCE).67 He did not, however,

mention the other evidence related to Sima Qian‟s movements at the time (item 1). Zhou

Shouchang 周壽昌 (1814-1884), in support of Zhao Yi‟s theory, pointed out that the

emperor did journey to Yong at the beginning of 90 BCE (adequately fulfilling item 3).68

The main problem for the 91 BCE theory, then, was that the emperor made no trip to the

east in that year. Cheng Jinzao points out that this problem might be solved by adopting

the alternate interpretation of item 2, mentioned parenthetically above. Zhang Dake and

others have raised serious objections to this, however.69

The movements of Emperor Wu are of course well-documented in the Hanshu.

Based on these, Wang Guowei 王國維 (1877-1927)—following the first interpretation of

會東從上來 in item 1—suggested that Taishi 4 (93 BCE) best fit the evidence. Earlier in

that year, Emperor Wu had gone to Mount Tai and to Buqi, both places being located east

of Chang‟an (so fitting item 1). Then, in the twelfth month the emperor made a trip to

Yong (item 3). The latter part of item 2, which implies that Sima Qian was writing in the

eleventh month, sometime after returning from the east, but before the visit to Yong, also

65
Ibid. The Hanshu version of the “Letter” includes the extra 上 character, while the Wenxuan version
does not.
66
See HS 66.2881.
67
See Nian er shi zhaji 廿二史劄記 1.1.
68
Zhou, Hanshu zhu jiao bu 41.712.
69
See Sima Qian pingzhuan, 366.

473
presents no difficulty. The only piece of evidence that Wang Guowei might have

difficulty accounting for is the “incalculable crime” 不測之罪 that Ren An was said to

have committed. Wang solves this problem by citing part of Chu Shaosun‟s biography of

Ren An, in which the emperor complains that Ren An “had committed many crimes for

which he ought to have died, but I always let him live” [安有當死之罪甚红，吾常活

之].70 Though it seems a plausible suggestion, as Ruan Zhisheng has pointed out, it is a

bit speculative.71

A further piece of what I would call „factual‟ evidence is that Ban Gu‟s

introduction to the “Letter” refers to Ren An as “the Inspector of Yizhou” [益州刺史].

Proponents of the 93 BCE theory point to the fact that in 91 BCE Ren An was no longer

serving as the Inspector of Yizhou, but rather as “Commander of the Northern Garrison”

[監北軍使]. It was in that latter capacity that he committed the crime for which he

would eventually be executed. Defenders of the 91 BCE theory suggest that Ban Gu

merely referred to Ren An as the Inspector of Yizhou because it was in that capacity that

he was most well-known. Since in any case this piece of evidence falls outside the realm

of the “Letter‟s” actual text, however, it is difficult to take it as being conclusive either

way.

Given that the above factual evidence yields no definite answer, scholars have

turned to another type of evidence, namely, interpretation-based arguments concerning

which date and scenario best explain the curious nature of the “Letter” itself. The

problem is that Ren An‟s original letter has not survived. We have only two direct clues
70
SJ 104.2782-3. Wang Guowei emends the last part to “and I once before spared his life” [吾嘗活之], for
which he has been criticized by Ruan Zhisheng (“Sima Qian zhi xin,” 170). Wang‟s point may be stronger
with the emendation, but I believe it still stands even with the text as written.
71
“Sima Qian zhi xin,” 170.

474
as to what might have been in it. The first is that Ban Gu described its purpose as

“calling upon [Sima Qian] to fulfill the duties of worthy ministers of antiquity” [責以古

賢臣之義]. The other is in the “Letter” itself, where it says, “Formerly, I was honored to

receive your letter, instructing me to be cautious72 in my dealings and to take as my duty

the promotion of worthies and advancement of gentleman-scholars” [曩者辱賜書，教以

慎於接物，推賢進士為務].73 As Wenxuan commentator Li Shan pointed out, the last

phrase seems to be an allusion to the Liji “Ruxing” 儒行 chapter, which identifies one of

the duties of a ru as “promoting worthies and helping them advance successfully, without

hope of recompense” [推賢而進達之，不望其報].74 A second reference to the same

phrase occurs near the end of the “Letter”: “Now you Shaoqing on the contrary instruct

me to „promote worthies and advance gentleman-scholars,‟ but wouldn‟t this be false to

my private intention?” [今少卿乃教以推賢進士，無乃與僕之私指謬乎].75

The seemingly innocuous nature of Ren An‟s advice does not accord very well

with the impassioned tone of Sima Qian‟s “Letter.” This has led Shi Ding 施丁, for

example, to conclude that Ren An‟s letter was written much earlier, just after Sima Qian

got promoted to Director of Palace Secretaries, and was intended as generic career

advice.76 Others have suggested that it was a request for Sima Qian‟s help with Ren An‟s

own advancement, given Sima Qian‟s new and powerful position. However, as Ruan

Zhisheng has pointed out, it is hard to make sense of this latter scenario, since as far as

72
The Wenxuan text here has 順.
73
HS 62.2725; WX 41.1854.
74
LJ 59.978.
75
HS 62.2736; WX 41.1866.
76
See his “ „Bao shu‟ xie yu Taishi yuan nian dong” 《報書》寫於太始元年東 [The Reply Letter was
written in 96 BCE] in Sima Qian xingnian xinkao 司馬遷行年新考, 85-102.

475
history records, Ren An was far more powerful than Sima Qian until his fall during the

Wugu Witchcraft Affair.77 It is difficult to see why Ren An would need Sima Qian‟s

help with career advancement unless he was in trouble.

A theory which has considerable appeal to many scholars is that developed by

Bao Shichen 包世臣 (1775-1855). In a letter to his friend Shi Ganzhou 石贛州,

preserved in the Yizhou shuangji 藝舟雙楫 [Double Oars for the Artistic Boat], Bao

Shichen described the problem raised by Shi some time before, that the “Letter in Reply

to Ren An” seems to have nothing to do with promoting worthies and recommending

gentlemen. Bao describes his theory on the matter, in a passage which is now frequently

quoted in connection with this issue:

I venture to suggest that „promoting worthies and recommending scholars‟ was

not a phrase from Shaoqing‟s original letter. The Honorable Archivist is avoiding
mention of the fact that Shaoqing was pleading for rescue, and thus uses these
four characters to allude to the intention of the original letter. And he criticizes
Shaoqing by using the world‟s brave and talented in order to express his sense of
injustice. In the middle he relates the Li Ling affair, making it clear that he and Li
Ling were not close, and yet still he acted forcefully in order to rescue him. How
much more then should he be willing to face death in the case of Shaoqing! Yet
in truth it was connected with the fact that, after having been punished, the reason
he did not die was because the Shiji was not yet completed. The Honorable
Archivist‟s body was none other than the body of the Shiji. It was not that the
Honorable Archivist was acting selfishly. The Honorable Archivist could die for
Shaoqing, but the Shiji must not be thrown away for Shaoqing‟s sake. He
concludes by saying “Only after we are dead are right and wrong finally
determined,” and this is meant to refer both to the Honorable Archivist and to
Shaoqing: it is to console Shaoqing and to dissolve his own private regret. [竊謂
推賢薦士，非少卿來書中本語，史公諱少卿求援，故以四字約來書之意。而
斥少卿為天下豪雋以表其寃。中間述李陵事者，明與陵非素相善者，尚力為
引救。況少卿有許死之誼乎。實緣自被刑後，所 為不死者，以史記未成之
故。是史公之身，乃史記之身，非史公所得自私，史公可為少卿死，而史記
必不能為少卿廢也。結 “以死日是非乃定”，則史公與少卿所共者，以廣
少卿而釋其私憾。]78

77
See “Sima Qian zhi xin,” 160-161.
78
Yi zhou shuang ji 2.

476
Bao Shichen‟s hypothesis solves some problems but creates others. It solves the problem

of why so much of Sima Qian‟s letter is concerned with Li Ling and Sima Qian‟s

disastrous intervention on his behalf (because if Sima Qian tried to help Ren An the

results, Sima Qian intimates, would be similarly disastrous). It also goes some way

toward explaining the second reference to “promoting worthies and advancing

gentleman-scholars,” which Sima Qian claims would be so much in contradiction with

his private intention—because if Sima Qian risked his life for another worthy person, he

would not be able to complete the Shiji and all his humiliation would have been in vain.79

Where Bao Shichen‟s theory runs into difficulty, however, is explaining the

remarks at the beginning of the “Letter.” First, Sima Qian explains what prevented him

from answering Ren An‟s letter. Then he wrote, “Now you stand accused of this

incalculable crime” [emphasis added]. As Xu Shuofang points out, the “now” 今 does

seem to imply a contrast, especially with the “formerly” 囊者 with which Sima Qian

introduces his receipt of Ren An‟s letter.80 This suggests that Ren An did not stand

accused of the crime when he wrote the original letter, that the crime (or at least the

accusation) occurred between Ren An‟s letter and Sima Qian‟s. Furthermore, whichever

dating scheme one prefers, the idea that Ren An wrote his letter after committing the

crime would be chronologically a bit problematic. If Ren An‟s letter was written in 93

BCE, Sima Qian would have received it shortly before leaving for the east in the third

month (he returned to the capital in the fifth month), but a general amnesty was also

79
Note that Chavannes follows this suggestion, see Memoires historiques I.xliii.
80
Shi Han lungao 史漢論稿, 71. Luo Fangsong 羅芳松 (“Cong „nang‟ yu „jin‟ guan kui „Bao Ren An
shu‟”) and Shi Ding 施丁 (Sima Qian xingnian xinkao) both suggest that the contrast between the two
terms actually implies that Ren An‟s letter and Sima Qian‟s could not even have been written in the same
year.

477
proclaimed in the fifth month,81 so it is unlikely that Ren An could beg for rescue earlier

in the year and be still in danger in the eleventh month. If Ren An‟s letter was written in

91 BCE, Ren An did not commit his „crime‟ (i.e., involvement with the Wugu Witchcraft

Affair) until September. Chu Shaosun‟s account of events at least implies that the

emperor did not immediately consider Ren An‟s actions objectionable, so unless Ren An

was unusually prescient, he would not have identified the need to ask for help until later,

perhaps not even until Emperor Wu reversed the verdicts on the Prince Li rebellion.82

While not chronologically impossible, it seems strange for Sima Qian to be apologizing

for a delay in replying, if that delay were such a short time.83

Finally both David Knechtges and Zhang Dake have expressed doubts about

whether the phrase “promoting worthies and advancing gentlemen” makes any sense as a

coded message begging for help. Other scholars express a more agnostic view. If we can

assume that the 推賢進士 is indeed a reference to the Liji, a line later in the same passage

might support the pleading for rescue hypothesis: “There are those [gentlemen] who are

able to promote worthies and rescue people of ability in this manner” [其舉賢援能有如

此者].84 I have not seen any mention of this connection by scholars, but it seems at least

possible.

Debates among Chinese scholars over the dating of the “Letter” betray no concern

about the “Letter‟s” authenticity. Yet the intractability of the dating problem should give

81
HS 6.207.
82
Both Ruan Zhisheng (“Sima Qian zhi xin”) and Zhang Dake assume that Ren An did not actually lose his
life until early in Zhenghe 3 (90 BCE).
83
Ruan Zhisheng‟s argument (“Sima Qian zhi xin,” 171-172), that any delay is a long delay when a person
is afraid for his life, seems to me to be begging the question. Another possibility is that such an apology is
merely a conventional expression; still, it would be desirable to find a theory that did not resort to such a
dismissal.
84
LJ 59.978.

478
the reader pause. Is our position not like that of a detective, faced with a superficially

plausible account that in the end cannot quite stand up in court? Of course, the exact date

of the “Letter” is less important than its underlying meaning. Even that, however, is far

from clear.

DEBATES OVER THE LETTER‟S INTENDED PURPOSE

What was the “Letter” intended to say anyway? Chavannes believes that Sima

Qian was urging Ren An to commit suicide.85 Zhou Hong 周洪 has provocatively

suggested that Sima Qian was not friendly but rather hostile toward Ren An.86 Zhang

Dake, who does not believe that Ren An‟s original letter was a plea for help, nonetheless

conjectures that Sima Qian did successfully intervene on Ren An‟s behalf—in 93 BCE—

enabling him to survive long enough to commit his fatal mistake two years later. The

real problem is that no reasonable guess about the relationship between Sima Qian and

Ren An fully accords with the tone and content of “Letter in Reply to Ren An.”

Stephen Durrant and David Knechtges have suggested a different possible

motivation for the “Letter”. As Durrant wrote,

85
Chavannes, Mémoires Historiques, I.xliii-xliv.
86
Zhou puts forth the following arguments: 1) In addition to refusing Ren An‟s request, the “Letter”
strongly implies that Sima Qian has no one who truly understands him (知己 or 知音). Ren An would
therefore not fall into that category. 2) If Ren An were Sima Qian‟s close friend, why would not Sima Qian
write him a “Traditions” chapter—as he did, for example, for Tian Ren? We know something of Ren An‟s
story from an addition to the “Traditions of Tian Ren” by Chu Shaosun: in contrast to Tian Ren, Ren An‟s
character comes off as selfish and ignoble. 3) Sima Qian includes the details of his own misfortune not
because he thinks Ren An does not know them, but to clarify the difference between them, and to
emphasize that Ren An was one of the many near friends and relatives who stood by and did nothing to
help. 4) To send a letter refusing help to someone in Ren An‟s position would seem a cruel act if the
recipient were a friend—but an understandable one if the writer held a deep resentment toward the recipient.
5) The anger in the “Letter” is not principally directed against the emperor who misunderstood him but
against petty people who seek only to save their own skins. If the foregoing arguments about Ren An are
true, Sima Qian numbers him among the petty people.

479
 Perhaps the letter, which eventually found its way into the History of the Han,
was never intended simply as a private communication but was written as a final
testament to posterity in which Sima Qian allowed himself to be much more
direct and emotional than was appropriate in a more formal document like his
“Self-Narration.”87

Or in Knechtges‟ words, “Although the letter does have some features of a personal,

private communication, what is striking about it is its public character.”88 However

Knechtges notes that “the letter was not a recognized literary genre [in Sima Qian‟s time],

but was simply a form of personal and usually private communication.” Knechtges does

not seem troubled by the anachronism of Sima Qian‟s letter using a genre not previously

understood as public to “explain to posterity [his] reasons for accepting the most

degrading of all punishments instead of following the „honorable‟ course of committing

suicide.” Knechtges instead implies that Sima Qian‟s remarkable “Letter” may have

been single-handedly responsible for changing the nature of the genre. “With this work,

the letter as a literary form becomes as much a public expression as a private one.”89

Certainly this is how the “Letter” comes off in anthologies like the Wenxuan: as a

foundational literary work, a public self-justification masquerading as a private

communication. Yet if we really want to understand the “Letter‟s” origins, we have to

consider not just what it has become but what it originally was. How does a genre make

the transition from private to public? Is it because an author suddenly conceives of it as

public, or is it because a reader does? Surely the latter seems more likely. In that case,

we ought to also be able to make sense of the “Letter” in its private context? Is an

intention to create a public document enough to explain why Sima Qian wrote the “Letter”

to Ren An when and how he did? I would be inclined to argue that the story is more

87
Cloudy Mirror, 16-17.
88
Knechtges, “Key Words,” 14.
89
Ibid., 15.

480
complex than we will ever be able to find out for certain. At the very least, scholars

should think carefully about the relationship between the “Letter” and the Shiji. The

“Letter” may be a tantalizing interpretive key, but is it actually opening the right doors?

481
 Conclusion

Both author and work emerge

through the interpretation of a
text.
—Alexander Nehamas1

The authorship of the Shiji—the circumstances surrounding its creation—has

been the subject of many imaginative reconstructions, possibly even beginning with those

of (the historical person) Sima Qian himself. We have two documents purportedly from

his hand which discuss how the Shiji came into being: the “Letter in Reply to Ren An”2

and the “Honorable Senior Archivist‟s Self-Narration” 太史公自序. By and large, these

texts have been read very literally as a key to interpreting the Shiji. Scholars at times

manifest a certain uneasiness about the texts as sources, but, having nothing else to go on,

in the end wring from them as much evidence as possible.

An underlying argument of this study has been that we should not conflate the

historical writer (or writers) of these texts with the figures of the author which can be—

and have been—constructed from them. Alexander Nehamas, building on and critiquing

Michel Foucault‟s polemical essay, “What is an Author?”3 insists on maintaining the

distinction between writer and author which Foucault introduced but later (Nehamas

argued) seemed to abandoned.4 Nehamas wrote,

A writer is a historical person, firmly situated within a specific context, the

efficient cause of a text‟s production. Writers often misunderstand their own
texts… exist outside their texts and precede them in truth, not in appearance
only…. Precisely for this reason, writers are not in a position of interpretive
authority over their writings.5

1
“Writer, Text, Work, Author,” 288.
2
Anthologized in the Hanshu “Traditions of Sima Qian” (HS ch.62), and in the Wenxuan “Letters” (WX
ch.41). Hereafter, I will refer to this text as the “Letter.”
3
Language, Counter-Memory, Practice, 113-138.
4
See Nehamas‟ critique “Writer, Text, Work, Author,” 274-75 and 287-88.
5
Ibid., 272.

482
What then is an author, as distinct from a writer? Nehamas resisted the idea that the

author is a kind of fictional character produced by the text and no more real than any

other: “Though an author, too, is a character, it is a character manifested or exemplified

in a text and not depicted or described in it…. Unlike fictional characters, authors are not

simply parts of texts; unlike actual writers, they are not straightforwardly outside them.”6

In some sense, a version of this difficulty applies to all the „characters‟ depicted in

the Shiji and in historical narratives generally. Emperor Wu of the Han clearly has a dual

existence as an actual historical person and as a character in the Shiji. Cases like that of

Qu Yuan are murkier: certain stories or traditions about Qu Yuan surely preceded the

Shiji, but other than the Shiji‟s own portrayal, we have very little access to the historical

personage, if there even was one. As for figures like the Yellow Emperor, their historical

existence is highly dubious; in evaluating the portrayal of the Yellow Emperor in the Shiji,

scholars are limited to a discussion of the sources employed and the differences between

this and other versions of the myth.

Like Emperor Wu, Sima Qian has a dual existence. To put it another way, the

name “Sima Qian” potentially refers to two separable things: 1) the person who lived in

the Western Han, and 2) the author of the Shiji. How „truly‟ Emperor Wu was portrayed

is of course an important issue, one of deep concern to readers of the Shiji. Historical

matters aside, whether the Shiji was to be judged as a „true record‟ or a „defamatory text‟

often hung on the balance of this one question. Yet few scholars have shown a concern

for whether Sima Qian himself is „truly‟ portrayed, in part because the practical problem

is too great: there simply is no way to get at Sima Qian the historical person, outside of

6
Ibid., 273.

483
his existence in the texts he himself purportedly wrote. Nonetheless, the issue would

seem to be of central importance. Very few Shiji interpretations remained uninfluenced

by at least a background awareness of a Sima Qian author-figure. This was true whether

the readers in question subscribed to the „true record‟ theory (discussed in chapter 6), the

„defamatory text‟ theory (chapter 4), or the more positive versions of the „Sima Qian

romance‟ (chapter 5).

It has been my task here to question the straightforward identification of any Sima

Qian author figure with the writer of the Shiji.

Wayne Booth, writing about literary authors of fiction, developed the notion of

the „implied author‟ of a given text: “The implied author chooses, consciously or

unconsciously, what we read; we infer him as an ideal, literary, created version of the real

man; he is the sum of his own choices.”7 The implied author may be a useful concept for

rethinking the figure(s) of Sima Qian variously constructed from the „autobiographical‟

texts. Booth adds that there is an “intricate relationship of the so-called real author8 with

his various official versions of himself.”9 This is particularly relevant in considering the

dialogue with Hu Sui,10 a set-piece within the “Self-Narration” which can be read either

as an enthusiastic endorsement of the Han or as an ironic satire of profoundly critical

intent. In the case of Sima Qian, the matter is further complicated by the fact that even

the „official versions‟ found in the “Letter” and “Self-Narration”—while generally

accepted as belonging to the author Sima Qian—stand in uncertain relation to the

historical writer Sima Qian. Did the historical Sima Qian write the dialogue with Hu Sui?

7
Booth, Rhetoric of Fiction, 74-75.
8
By “real author,” Booth here seems to have meant what Foucault and Nehamas refer to as “the writer” (a
historical person).
9
Ibid., 71
10
SJ 130.3297-3300.

484
I see no way of answering this question; the most we can say is that, as it appears in the

“Self-Narration,” the dialogue is presented as being part of the implied author Sima

Qian‟s work.

A concern arises if we countenance the pulling apart of the now-inaccessible

person Sima Qian from the author figure of the same name. In Nehamas‟ formulation:

It may now appear that the figure of the author is seriously arbitrary…. If each
text can be interpreted, as it is often claimed, in different and even incompatible
ways, then the author appears to collapse into fragments. Each interpretation
generates its own author, and each text can give rise to many different and even
inconsistent authors.11

This is the problem I alluded to in my introduction (above) while reviewing Western

scholarship on the Shiji. If we are free to construct the author Sima Qian, we tend to

construct him in our own image.

In the foregoing study, I proposed to think through this difficulty with reference to

traditional readings of the Shiji, and in particular, traditional constructions of its author.

In chapters 1 and 2, I contextualized and to some extent categorized the better-known

readers of the Shiji and the interpretations they produced. The different contexts in which

the Shiji was read can seen as roughly corresponding to what Stanley Fish has called

“interpretive communities.”12 In the Chinese contexts I have described, Alexander

Nehamas‟ caveat, that “each person belongs to a number of different interpretive

communities,”13 is highly a propos. Differences in interpretive community led to

differences, not only in how Sima Qian as author was evaluated, but even in how he was

constructed. As the cases of Zhang Lei and Su Zhe showed, the same reader could even

11
“Writer, Text,” 284.
12
Is There a Text in this Class, 147-174.
13
“Writer, Text,” 276. See also ibid., 290 n.12: “It is precisely this fact that makes it possible to criticize
the conventions accepted by each community and to provide rational alternatives to them.”

485
come to opposing evaluations of Sima Qian depending on the context in which they were

operating.

In chapter 3, I discussed the form of the Shiji and in particular how it was read as

an aspect of Sima Qian‟s authorial work. It could be argued that this aspect (actually a

part of what I called “the new historical tradition,” discussed in chapter 1) was part of

what made Sima Qian more than just an author. In relation to the development of

historical writing in China, Sima Qian comes closer to being what Foucault refers to as an

“initiator of discursive practice,” defined as someone who “produced not only their own

work, but the possibility and the rules of formation of other texts.”14

Foucault was careful to contrast the “initiator of discursive practice” with the

founder of a genre. Discussing Ann Radcliffe and the formation of the genre of Gothic

romance, he wrote that while “her function as an author exceeds the limits of her work,”

that really only means “there are certain elements common to her works and to the

nineteenth-century Gothic romance.”15 By contrast, Foucault‟s paradigm “initiatiors”—

Marx and Freud—“cleared a space for the introduction of elements other than their own,

which, nevertheless, remain within the field of discourse that they initiated.”16

Sima Qian, as the „Father of Chinese History‟ (though at some times and in some

contexts, he shared this distinction with Ban Gu) can be seen as playing a role similar to

that of Marx or Freud in this respect. For history-writing in the traditional Chinese

context, as for Marxist or psychoanalytic discourse, “it is inevitable that

practitioners…must „return to the origin‟”17: in this case rethinking and reevaluating the

14
“What is an Author?” 131.
15
Ibid., 132.
16
Ibid.
17
Ibid., 134.

486
Shiji, in formal and other terms. It was possible to disagree with Sima Qian‟s choices,

and those who took the Shiji as their inspiration did so in a variety of ways. Nonetheless,

the Shiji came to be seen as an origin to which these later thinkers always did return.

I have attempted to show that these “returns” resulted in a process of influence

that was mutual rather than unilateral: it is not just that the Shiji influenced later works;

issues that arose in the production of later works also brought out or created new

interpretations of the Shiji itself. What (and how) the Shiji means today differs greatly

from the meaning its very first readers found in it. This process of change has been

brought about by the Shiji‟s many readers, each of whom returned to the text from a

slightly different starting point, and each of whom left behind a slightly different

understanding. As Foucault describes it: “This return, which is part of the discursive

mechanism, constantly introduces modifications… the return to a text is not a historical

supplement… [but] rather, it is an effective and necessary means of transforming

discursive practice.”18 The continuous and influential presence of the Shiji transformed

the discursive practices of those who wrote about history and played an influential role in

literary contexts as well. But at the same time that the Shiji was changing the world, the

world was also changing the Shiji.

Small but significant changes could have occurred at the level of the text itself,

brought about by editors and commentators who „corrected errors‟ and „eliminated

ambiguities‟ to bring the text in line with their own understanding. Traces of such

activity are extremely difficult to locate, however, and even more difficult to identify

with certainty. Yet the meaning of a text can change in other ways as well: Foucault

wrote, “There are no „false‟ statements in the work of the initiators; those statements
18
Ibid., 135.

487
considered inessential or „prehistoric,‟ in that they are associated with another discourse,

are simply neglected in favor of the more pertinent aspects of the work.”19 In a massive

text like the Shiji, the issue of selection assumes crucial importance. Because no reading

can comprehend the whole of the text, every discussion must take some sections as

„representative,‟ while the vast majority of the text is silently ignored. While there are

parts of the Shiji that nearly all readers tend to privilege, the actual composition of the

„core‟ or „representative‟ Shiji has inevitably varied. Though I have not been able to do

justice to it in the current study, the differing notions of a „core Shiji‟—taken in

conjunction with selection principles (explicit or implicit) and the reader‟s purpose in

making the selection—would certainly make an interesting subject for future study.

In Part II, I addressed what has historically been a fundamental division in Shiji

interpretation: the split between readings that emphasize tragic incidents from Sima

Qian‟s biography as being key to Shiji interpretation (chapters 4-5), and readings which

construct Sima Qian as a “good historian” who was seeking only to produce a “true

record” (chapter 6). Though for the purposes of analysis I have treated these two types of

readings as separate and distinct, I should emphasize that they are not mutually exclusive.

For most readers of the Shiji, both constructions of Sima Qian as author are

present, either simultaneously or successively. We read, for example, the evaluation at

the end of Shiji‟s first chapter:

The Documents are often incomplete and have gaps, but what has been lost from
them can sometimes be found in other records. If one is not fond of study or deep
reflection, if one‟s minds cannot recognize the true significance of things, it is
indeed difficult to explain such things to men of shallow views and scant
knowledge. I have brought together these records and put them in order, selecting
those parts which seemed most in accord with classical standards, composing this

19
Ibid., 133-134.

488
 as the head of the texts of the Basic Annals. [書缺有閒矣，其軼乃時時見於他
說。非好學深思，心知其意，固難為淺見寡聞道也。余并論次，擇其言尤雅
者，故著為本紀書首。]20

Even if we must admit that the author‟s criteria might not match our own, how can we

help but recognize in this statement a deep commitment to truth? This Sima Qian, in his

very first evaluation, was the initiator of a discourse about how an individual author can

produce a true historical account.21 Later historians would expand the discourse greatly,

but Sima Qian remained the paradigmatic „true historian.‟

If we turn to other passages in the Shiji, however, the tragic romance of the Sima

Qian biography is a subtext whose insistent presence is impossible to ignore. In the

evaluation at the end of the “Traditions of Ji Bu and Luan Bu” (SJ ch.100), the author

comments:

Ji Bu made a name for bravery in Chu… yet he suffered punishment and disgrace
and became a slave, and did not commit suicide. Why did he stoop to this?
Because he chose to rely upon his abilities. Therefore he suffered disgrace
without shame, for there were things he hoped to accomplish and he was not yet
satisfied…. Truly the wise man regards death as a grave thing. When slaves and
scullion maids and lowly people in their despair commit suicide, it is not because
they are brave; it is because they know that their plans and hopes will never again
have a chance of coming true. Luan Bu wept for Peng Yue and faced the boiling
water as though it were his true destination. This is indeed what it means to know
the right place to die, not counting death as important in itself. [季布以勇顯於
楚。。。然至被刑戮，為人奴而不死，何其下也！彼必自負其材，故受辱而
不羞，欲有所用其未足也。。。賢者誠重其死。夫婢妾賤人感慨而自殺者，
非能勇也，其計畫無復之耳。欒布哭彭越，趣湯如歸者，彼誠知所處，不自
重其死。]22

20
SJ 1.46; translation adapted from Watson, Ssu-ma Ch’ien, 184.
21
Many other passages in the Shiji also support this picture. The evaluations at the end of the “Traditions
of the Assassins” (SJ 86.2538) and the “Traditions of Dayuan” (SJ 123.3179) are notable examples.
22
SJ 100.2735; translation adapted from Watson, Ssu-ma Ch’ien, 192-3.

489
Though these comments ostensibly refer to the specific cases of Ji Bu and Luan Bu, no

well-informed reader can avoid hearing in them strong echoes of the “Letter in Reply to

Ren An”:

A man has only one death. That death may be as weighty as Mount Tai, or it may
be as light as a goose feather. It all depends upon the way he uses it…. How
could I bring myself to sink into the shame of ropes and bonds? If even the
lowest slave and scullion maid can bear to commit suicide, why should not one
like myself be able to do what has to be done? But the reason I have not refused
to bear these ills and have continued to live, dwelling in vileness and disgrace
without taking my leave, is that I grieve that I have things in my heart which I
have not been able to express fully. [人固有一死，死有重於泰山，或輕於鴻毛，
用之所趨異也。。。且夫臧獲婢妾猶能引決，況若僕之不得已乎！所以隱忍
苟活，函糞土之中而不辭者，恨私心有所不盡，鄙沒世而文采不表於後
也。]23

The reader seeking to construct Sima Qian as an author who deliberately marked his text

with personal, autobiographically-motivated concerns finds ample justification for doing

so in correspondences of this type. The above-cited passages suggest only one among

many such correspondences.24

It is a matter for individual readers to decide whether these two very different

constructions can be envisioned as coexisting within a single persona. Historically, they

were sometimes treated as incommensurable and sometimes seen as complementary. Of

course these two modes of reading, whose development I explored in Part II, are by no

means the only ways of interpreting the Shiji or constructing its author—they were

merely two of the most prominent in pre-modern times.

A gap in the current study, which I hope to fill at a future time, is a discussion of

how Sima Qian came to be seen as a fore-runner (if not fully an initiator) of quite another

23
HS 62.2732, 2733; translation Watson, Ssu-ma Ch’ien, 63, 65.
24
For a list of passages that Takigawa, for example, considers to be autobiographically motivated, see SKK
10.5271-5272.

490
mode of discursive practice: that of fiction criticism. Though in some ways a

development of the „literary‟ uses of Sima Qian I discussed in chapter 2 and the

„romantic‟ readings I outlined in chapter 5, this late-Ming and early-Qing phenomenon in

truth placed Sima Qian in an entirely new context and interpretive community. To give

just one striking example of how much this new context changed the way Sima Qian‟s

authorial role was constructed, consider Li Zhi‟s 李贄 (1527-1602) impassioned defense

of Sima Qian in the face of Ban Gu‟s long-ago “three criticisms”:

If Qian had not been mutilated and humbled, if he had not been careless and
negligent, if he had not been quick to believe, if he had not contradicted the
judgements of the Sage—how would he have even deserved to be Qian? ….If I
went along with the judgements of the Sage, then what I said would be nothing
more than what the Sage said. It would not be the words of my heart. If the
words do not come forth from my heart—if the phrases are not brought about by
that which cannot be suppressed—then they are utterly lacking in flavor. [使遷而
不殘陋，不疏略，不輕信，不是非謬於聖人，何足以為遷乎？。。。夫按聖
人以為是非，則其所言者，乃聖人之言也，非吾心之言也。言不出於吾心，
詞非由於不可遏，則無味也。]25

The appeal of such a portrayal is undeniable, but is it a Sima Qian that the Western Han

figure of the same name would even be able to comprehend, let alone identify with? And,

how much does this question matter?

Perhaps it is helpful to once again appeal to Nehamas‟ discussion and critique of

Foucault (as well as other theorists who attempt to “do away with the author”). Nehamas

wrote, “In general, the author is to be construed as a plausible historical variant of the

writer, as a character the writer could have been.”26 Though admitting that this way of

constructing the author is in some sense arbitrary, Nehamas added that it is “a well-

established practice” and that alternative practices (such as deliberately anachronistic

25
Cangshu 藏書 ch.40, cited on SJYJJC 6.18.
26
“Writer, Text,” 285.

491
readings), may be interesting but in reality tend to “confine themselves to partial

interpretations of parts of texts.”

This applies to the problem of Shiji interpretation in two main ways. First, I

would argue that traditional Chinese readers, in their construction of Sima Qian as an

author, did follow a version of the practice Nehamas described. Yet the Shiji is a truly

ancient text, produced in an age far distant from our own. The boundaries of plausibility

are by no means well-defined. To make matters worse, a great deal of the information we

do have about the Western Han comes from within the Shiji itself, the very text we are

trying to interpret.

Perhaps the circularity concealed beneath the innocuous word “plausible” is less

worrying than it might be, however. The second connection I would make between

Nehamas‟ discussion and Shiji interpretation is that, as Nehamas suggests, an acceptable

test of plausibility might well be how extensively a given authorial construction can be

used in interpreting the work. That is, how much of the Shiji text can a given

construction of „Sima Qian‟ explain? Or alternatively, simultaneously, how much of the

text went into that „Sima Qian‟s‟ construction?

Readers less inclined toward such theoretical complexities may at this point be

tempted to turn away from the entire enterprise. In chapter 7 above I included a

preliminary discussion of one way that readers of the Shiji have used textually

problematic aspects of the Shiji to go beyond or away from the idea that Sima Qian was

its (only) author. The authorship controversies I reviewed there were of course not solely

motivated by such a desire. Nonetheless, to claim that “the historical Sima Qian was not

the writer of some portion of the Shiji” is to be absolved of responsibility for explaining

492
the connection between that portion of the Shiji and Sima Qian the constructed author.

This was sometimes done for the benefit of the Sima Qian author-figure, as when Tang

commentators attributed shoddier chapters or passages to Chu Shaosun instead.

Sometimes it was an attempt to achieve greater ideological clarity: when the „Daoist‟ and

„Confucian‟ strands of the Shiji began to seem irreconcilably contradictory, Sima Tan

appeared as a second author to whom readers assigned different philosophical

commitments. Sometimes, finally, it is a concern about the plausibility criterion I

discussed above: when it seemed that the author of some portion of the Shiji could not

plausibly be a person of the Western Han, then the historical writer Sima Qian and the

author of (some larger or smaller portion of) the Shiji must begin to come apart.

In my final chapter I discussed a text that is unavoidably central to all

autobiographical readings of the Shiji, the “Letter in Reply to Ren An.” Foucault wrote

that “a private letter may have a signatory, but it does not have an author.”27 But is this

text a private letter? As I argued above, the Ren An letter inhabits a strange space

between public and private. Any author-figure we might try to construct from it is

superficially plausible but actually full of contradictions. Nehamas maintained that the

project of interpretation involves “formulat[ing] a series of hypotheses about the actions

we must attribute to the author… in order to account for the features of this work.”28 On

this model, interpreting the famous “Letter” is problematic at best. Not least among the

problems is the fact that the text itself suffers from a certain instability, especially in its

degree of overlap with the Shiji “Self-Narration” and the Hanshu “Traditions of Li Ling.”

27
“What is an Author?” 124.
28
“Writer, Text,” 287.

493
 In short, it is difficult to construct, from the text of “Letter,” a believable Sima

Qian who could have authored it. Yet as a key for autobiographical interpretations of the

Shiji, the “Letter” seems made to order. In some sense, this conclusion is a troubling one.

On the other hand, whatever its relation to the historical Sima Qian, the “Letter” has

become an inseparable part of the author Sima Qian and of how we understand the

creation of the Shiji.

Where does all this leave us, then? Where do we go from here, both as regards

the future of Shiji studies and our understanding of authorship? Nehamas observed that

It is often said (and more often believed) that interpretation is required when a
particular text conceals an implicit and, ideally, profound meaning differing from
the meaning that text appears to have…. Such views hold that interpretation is
needed when the meaning of a text is somehow “beyond” or “behind” it.29

This view of interpretation would not have seemed alien to traditional Chinese readers of

the Shiji. As discussed in chapter 1, the commentator Sima Zhen wrote of the Shiji that it

“was subtle and had the ancient [virtue of] substantiality. Thus the famous worthies of

the Han and Jin did not yet know to value it” [微為古質，故漢晉名賢未知見重].30

Sima Zhen‟s own commentary, tellingly entitled “Explication of the Hidden in the Shiji”

[史記索隱], took as its explicit purpose just such a construal of the interpretation process.

Its overall focus was on “metaphors of depth and concealment.”31 Similarly, Cheng Yi

wrote, “The subtle nature and marvelous meaning of Zichang‟s [=Sima Qian‟s] writings

are lodged outside the path of the written characters” [子長著作，微情妙旨寄之文字蹊

徑之外].32 In both cases, meaning is being sought “beyond” or “behind” the actual

29
Ibid., 276.
30
“Preface to the Explication of the Hidden in the Shiji,” 7.
31
“Writer, Text,” 276.
32
Cited in Jiao Hong‟s Jiaoshi bi cheng 2.37.

494
words. Qin Guan went even farther and made an effort to specify, characterize, and

explain these implicit and profound meanings (see discussion in chapter 5).

Whether we are satisfied or not with this model of interpretation, we are unlikely

to be able to engage in it as competently as these earlier readers. They were steeped in

the tradition to a degree of which we can hardly conceive. Instead, perhaps what we

might offer to Shiji studies is a different way of understanding the entire project of

interpretation, especially as it relates to the figure of the author. Perhaps we should, as

Nehamas suggested, “avoid the view that to understand a text is to re-create or replicate a

state of mind which someone else has already undergone… Such states belong to writers

but not to authors.”33 Nehamas proposed as an alternative to “metaphors of depth and

concealment” (and to “naively psychological conceptions” of the author-figure). Instead,

he proposed that interpretation be seen instead “in terms of breadth and expansion.”34

This allows us to avoid the metaphysical problem of separating the „apparent‟ and the

„real.‟ Instead

When we say that an action or a text means something other than what it appears
to mean, we do not have two meanings, one real and one apparent. All we have…
is a series of progressively more complicated, detailed, and sophisticated
hypotheses aiming at construing a text as an action, at trying to find the meaning
it does have in its relationship to its agent and to that agent‟s other actions, or
texts.35

The authorial figure or figures of Sima Qian constructed in the course of reading are

merely tools to aid in interpretation. They should not be conflated with the historical

figure they are modeled on. They are useful insofar as they lead to better, more

interesting, and more broadly applicable readings of the Shiji.

33
“Writer, Text,” 285-86.
34
Ibid., 277.
35
Ibid.

495
 If a more self-conscious awareness of such issues is what my research has to offer

to Shiji studies, what does the development of Shiji studies as outlined in this dissertation

have to offer to the discourse of authorship? First, as I indicated in the introduction, the

richness of extant reader responses to the Shiji provides an unusually full case study. We

have a nearly continuous record detailing readers‟ changing conceptions of the author

Sima Qian and his relation to his text. While I have not been able to include much

comparative work within the scope of this dissertation, such work seems a potentially

fruitful direction for further study. What do reader responses to Sima Qian have in

common with responses to comparable Western figures like Herodotus or Thucydides?

Second, the changing uses to which Sima Qian has been put are themselves interesting in

that they reveal much about the interpretive communities that originated them. Since

traditional Chinese readers often defined themselves in relation to history, it is not unjust

for us to „read‟ them through the medium of this relation.

In this study, I have suggested much but offered little in the way of certainty. In

some sense, this is appropriate to the work of interpretation. As Nehamas wrote,

“interpretation ends when interest wanes, not when certainty is reached.”36 There is a

great deal about the Shiji that can never be known for certain. And yet interest in the

Shiji has never wholly waned, and shows no signs of doing so. The work of interpreting

the text—and its author—is ongoing. It will continue to shed light on both the Shiji and

on the very process of interpretation.

36
“Writer, Text,” 278.

496
 Abbreviations

Abbreviations: Journals

AcO(B) Acta Orientalia (Budapest)

AM Asia Major
ArOr Archiv orientální
AS Asiatische Studien
BMFEA Bulletin of the Museum of Far Eastern Antiquities
BSOAS Bulletin of the School of Oriental and African Studies
CEA Cahiers d’Extrême-Asie
CLEAR Chinese Literature: Essays, Articles, Reviews
CRI China Review International
EMC Early Medieval China
EtC Études chinoises
HJAS Harvard Journal of Asiatic Studies
JAOS Journal of the American Oriental Society
JAS Journal of Asian Studies
JRAS Journal of the Royal Asiatic Society
MS Monumenta Serica
OE Oriens Extremus
TP T’oung Pao

Abbreviations: Primary Sources

CCBZ Chu ci buzhu 楚辭補注 [Songs of the South, annotated and supplemented]
CQZZ zhu Chunqiu zuozhuan zhu 春秋左傳注 [Commentary on the Spring and
Autumn with the Zuozhuan]
FY Fayan yishu 法言義疏 [Exemplary sayings with glossees and sub-
commentary]
GSZ Gao shi zhuan 高士傳 [Traditions of Eminent Men]
Han Changli Han Changli wenji zhushi 韓昌黎文集注釋 [The literary collection of
Han Changli, annotated and explained]
HHS Hou Hanshu 後漢書[History of the latter Han]
HS Hanshu 漢書 [History of the former Han]
JS Jinshu 晉書 [History of the Jin]
JTS Jiu Tangshu 舊唐書 [Old history of the Tang]
LH Lunheng 論衡 [Discourses weighed in the balance]
LSCQ Lüshi chunqiu jiaoshi 呂氏春秋校釋 [The Spring and Autumn of Master
Lü collated and translated]
LZYJ Liu Zongyuan ji 柳宗元集 [Collected writings of Liu Zongyuan]
NQS Nan Qi shu 南齊書 [History of the Southern Qi dynasty]
NS Nanshi 南史 [History of the Southern dynasties]
QTW Quan Tang wen 全唐文 [Complete Tang prose]
QSW Quan Song wen 全宋文 [Complete Song prose]

497
SBCK Sibu congkan 四部叢刊 [Four divisions collected edition]
SGZ San Guo Zhi 三國志 [Records of the three kingdoms]
SJ Shiji 史記 [Archivist‘s records]
SJYJJC Shiji yanjiu jicheng 史記研究集成 [Collected research results on the Shiji]
SKK Shiki kaichū kōshō 史記會注考證 [Collected commentaries on the Shiji,
researched and investigated]
SKQS Siku quanshu 四庫全書 [Four treasuries complete books]
SSXY Shishuo xinyu 世說新語 [New Accounts of the World‘s Tales]
STTS Shitong tongshi 史通通釋 [Understanding history, with comprehensive
commentary]
SuiS Suishu 隋書 [History of the Sui]
WSTY Wenshi tongyi jiaozhu 文史通義校注 [Comprehensive principles of
prose and history, collated and annotated]
WX Wenxuan 文選 [Literary selections]
WXDL Wenxin diaolong zhu 文心雕龍注 [Literary mind and the Carving of
Dragons, annotated]
WXTK Wenxian tongkao 文獻通考 [Comprehensive Investigation of Documents]
XTS Xin Tangshu 新唐書 [New history of the Tang]
Zhuangzi Zhuangzi jinzhu jinyi 莊子今注今譯 [The Zhuangzi with modern
commentary and translation]
ZZYL Zhuzi yulei 朱子語類 [Categorized sayings of Master Zhu]
ZZTJ gangmu Zizhi tongjian gangmu 資治通鑑綱目 [A general outline of the
Comprehensive mirror as an aid to government]

Bibliography

Allen, Joseph Roe, III. ―An Introductory Study of Narrative Structure in Shiji.‖ CLEAR 3
(1981): 31-66.
An Pingqiu 安平秋, Zhang Dake 張大可, and Yu Zhanghua 俞樟華. Shiji jiaocheng 史
記教程 [A course of study on the Shiji]. Beijing: Huawen chubanshe, 2002.
Ban Gu 班固. Hanshu 漢書 [History of the Former Han] (HS). Beijing: Zhonghua shuju,
1962.
Bauer, Wolfgang. Das Antlitz Chinas: Die Autobiographie Selbstdarstellung in der
Chinesischen Literatur von ihren Anfängen bis Heute [The face of China:
autobiographical self-expression in Chinese literature from its beginnings to
today]. Munich: Carl Hanser, 1990.
—. ―Time and Timelessness in Premodern Chinese Autobiography.‖ In Bieg, Lutz, et al.,
eds. Ad Seres et Tungusos: Festschrift für Martin Gimm. Opera Sinologica 11.
Wiesbaden: Harrassowitz, 2000. 19-31.
Beishi 北史 [History of the Northern Dynasties]. Compiled by Li Yanshou李延壽 et al.
Beijing: Zhonghua shuju, 1975.
Berkowitz, Alan. Patterns of Disengagement: The Practice and Portrayal of Reclusion
in Early Medieval China. Stanford: Stanford University Press, 2000.

498
Bielenstein, Hans. The Bureaucracy of Han Times. Cambridge: Cambridge University
Press, 1980.
—. ―An Interpretation of the Portents in the Ts‘ien-Han-Shu.‖ BMFEA 22 (1950): 127-
143.
Birch, Cyril, ed. Anthology of Chinese Literature: Volume I: From Early Times to the
Fourteenth Century. New York: Grove Press, 1994.
Bodde, Derk. China’s First Unifier; A Study of the Ch’in Dynasty as Seen in the Life of
Li Ssu, 280?-208 B.C. Hongkong: Hong Kong University Press, 1967.
—. ―Review of The Grand Scribe’s Records Vol.I.‖ CLEAR 17 (1995): 137-142.
Bol, Peter. ―Culture and the Way in Eleventh Century China.‖ PhD dissertation,
Princeton University, 1982.
Booth, Wayne. The Rhetoric of Fiction. Chicago: University of Chicago Press, 1961.
Borges, Jorge Luis. Labyrinths. New York: New Directions Publishing, 1962.
Cang Xiuliang 倉修良. Hanshu cidian 漢書辭典 [Dictionary of the Hanshu]. Jinan:
Shandong jiaoyu chubanshe, 1996.
Chao Gongwu 晁公武. Junzhai dushu zhi 郡齋讀書志 [Junzhai‘s bibliographic record].
Shanghai: Shanghai guji chubanshe, 1990.
Chavannes, Édouard, trans. Les Mémoires historiques de Se-ma Ts’ien. 5 vols. Paris:
Ernest Leroux, 1895-1905; rpt. (with 6th vol.), Paris: Adrien Maisonneuve, 1969.
Chen, Chi-yun. Hsün Yüeh (A.D. 148-209): the Life and Reflections of an Early
Medieval Confucian. Cambridge: Cambridge University Press, 1975.
Chen Fuliang 陳傅良. Zhizhai xiansheng wenji 止齋先生文集 [Collected Writings of
Master Zhizhai]. SBCK. Shanghai: Shangwu yinshuguan, 1929.
Chen Guoqing 陳國慶. Hanshu Yiwen zhi zhu shi huibian 漢書藝文志注釋彙編
[Compilation of Commentary and Exegesis on the Hanshu Treatise on the Arts
and Writings]. Beijing: Zhonghua shuju, 1983.
Chen Shidao 陳師道. Houshan shihua 後山詩話 [Back Mountain poetry talk]. In Biji
xiaoshuo daguan 筆記小說大觀 , Volume 6. Taibei: Xinxing shuju, 1975.
Chen, Yu-shih. Images and Ideas in Chinese Classical Prose: Studies of Four Masters.
Stanford: Stanford University Press, 1988.
Chen Zhi 陳直. ―Han Jin ren dui Shiji de chuanbo ji qi pingjia‖ 漢晉人對史記的評價
[Han and Jin Dynasty people‘s transmission and evaluation of Shiji]. Sima Qian
yu Shiji lunji 司馬遷與史記論集 [Collected essays on Sima Qian and Shiji].
Xi‘an: Shansi renmin, 1982. 215-242.
Cheng Jinzao 程金造. Shiji guankui 史記管窺 [Peering at the Shiji]. Xi‘an: Shanxi
renmin chubanshe, 1985.
Chu ci buzhu 楚辭補注 [Songs of the South, annotated and supplemented]. Qu Yuan 屈
原 et al., commentary by Wang Yi 王逸, Hong Xingzu 洪興祖, ed. Taibei:
Tiangong shuju, 1989.
Chʻü, Tʻung-tsu. Han Social Structure. Dull, Jack, ed. Han Dynasty, China, volume I.
Seattle: University of Washington Press, 1972.
Chung, Eva. ―A Study of the Shu (Letters) of the Han Dynasty.‖ Ph.D. dissertation,
University of Washington, 1982.
Clark, Elizabeth. History, Theory, Text: Historians and the Linguistic Turn. Cambridge:
Harvard University Press, 2004.

499
Cordier, Henri. ―Obituary for Édouard Chavannes.‖ TP 18 (1917): 115.
Csikszentmihalyi, Mark and Michael Nylan. ―Constructing Lineages and Inventing
Traditions through Exemplary Figures in Early China.‖ TP 89.1-3 (2003): 59-99.
Curry, Anne. Agincourt: A New History. Stroud, Gloucestershire: Tempus, 2005.
DiCosmo, Nicola. Ancient China and Its Enemies: The Rise of Nomadic Power in East
Asian History. Cambridge: Cambridge University Press, 2002.
Durrant, Stephen W. The Cloudy Mirror: Tension and Conflict in the Writings of Sima
Qian. SUNY Series in Chinese Philosophy and Culture. Albany: State University
of New York Press, 1995.
—. ―Creating Tradition: Sima Qian Agonistes?‖ In Steven Shankman and Stephen
Durrant, ed. Early China/Ancient Greece. SUNY Series in Chinese Philosophy
and Culture. Albany: State University of New York Press, 2002. 283-298.
—. ―Self as the Intersection of Traditions: The Autobiographical Writings of Sima
Qian.‖ JAOS 106 (1986) no.1: 33-40.
—. ―Ssu-ma Ch‘ien‘s Conception of Tso chuan.‖ JAOS 112 (1992) no.2: 295-301.
—. ―Ssu-ma Ch‘ien‘s portrayal of the first Ch‘in emperor.‖ In Brandauer, Frederick P.;
Huang, Chun-chieh, eds. Imperial Rulership and Cultural Change in Traditional
China. Seattle: University of Washington Press, 1994. 28-50.
Duyvendak, J.J.L. ―Review of Shryock, The Origin and Development of the State Cult of
Confucius.‖ JAOS 55.3 (1935): 330-338.
Egan, Ronald C., trans. Limited Views: Essays on Ideas and Letters [Selections from
Guan zhui bian by Qian Zhongshu]. Cambridge, MA: Harvard University Asia
Center: Distributed by Harvard University Press, 1998.
—. The Literary Works of Ou-yang Hsiu (1007-72). Cambridge U.K.: New York:
Cambridge University Press, 1984.
—. ―Narratives in Tso Chuan.‖ HJAS 37.2 (1977): 323-352.
—. Word, Image, and Deed in the Life of Su Shi. Harvard-Yenching Institute
Monograph Series, 39. Cambridge: Harvard University Press, 1994.
Elman, Benjamin. ―History in Policy Questions from Southern Provincial Civil
Examinations During the Late Ming and Early Qing.‖ Proceedings of the
Conference on the Achievements of Historical Figures in South China during the
Late Ming and Early Qing [明末清初華南地區歷史人物功業研討會論文集].
Hong Kong: Chinese University of Hong Kong, History Department, 1993. 179-
206.
Enoki, Kazuo. ―On the Relationship between the Shih-chi 史記, Bk. 123 and the Han-
shu 漢書, Bks. 61 and 96.‖ Memoirs of the Tōyō Bunko 41 (1983): 1-31.
Fan Wenlan 范文瀾. Zhengshi kaolue 正史考略 [Outline of an investigation of the
standard histories]. Shijiazhuang: Hebei jiaoyu chubanshe, 2002.
Fan Ye 范曄. Hou Hanshu 後漢書 [History of the Later Han] (HHS). Beijing:
Zhonghua shuju, 1963.
Fang Bao 方苞. Fang Bao ji 方苞集 [Fang Bao‘s writings]. Liu Jigao 劉季高, ed.
Shanghai: Shanghai guji chubanshe, 1983.
Farmer, J. Michael. The Talent of Shu: Qiao Zhou and the Intellectual World of Early
Medieval Sichuan. Albany: State University of New York Press, 2007.
Forke, Alfred, trans. Lun heng: Wang Chung’s Essays, 2 Vols. New York: Paragon
Book Gallery, 1962. (Reprint of 1907 edition.)

500
Foucault, Michel. ―What is an Author?‖ Trans. Donald F. Bouchard and Sherry Simon. In
Donald F. Bouchard, ed., Language, Counter-Memory, Practice. Ithaca, New
York: Cornell University Press, 1977.
Fuehrer, Bernhard. ―The court scribe‘s Eikon psyches: a note on Sima Qian and his letter
to Ren An [literary rhetoric].‖ Asian and African Studies (Bratislava) 6, no.2
(1997) 170-183.
Ge Hong 葛洪 (284-364). Xijing zaji 西京雜記 [Miscellaneous records of the Western
capital] (XJZJ). In the series Gujin yishi jing bian 古今逸史精編. Chongqing:
Chongqing chubanshe, 2000.
Geuss, Raymond. Morality, Culture, and History: Essays on German Philosophy.
Cambridge, U.K.; New York: Cambridge University Press, 1999.
Giles, Herbert, trans. Gems of Chinese Literature. London, B. Quaritch, 1926.
Graham, A.C. ―The Nung-chia ‗School of the Tillers‘ and the Origins of Peasant
Utopianism in China.‖ BSOAS 42 no.1 (1979): 66-100.
Gu Yanwu 顧炎武. Ri zhi lu 日知錄[Record of Knowledge Daily Gained]. Taibei: Wen
Shi Zhe chubanshe, 1984.
Han Yu 韓愈. Han Changli wenji zhushi 韓昌黎文集注釋 [The literary collection of
Han Changli, annotated and explained] (Han Changli). Yan Qi 閻琦, ed. Xi‘an:
San Qin chubanshe, 2004.
Han Zhaoqi 韓兆琦. Shiji jianzheng 史記箋證. Jiangxi: Xinhua shudian jingxiao, 2005.
Hanshu 漢書 [History of the Former Han] (HS). Compiled by Ban Gu 班固 et al.
Beijing: Zhonghua shuju, 1962.
Hanshu cidian 漢書辭典 [Dictionary for the Hanshu]. Cang Xiuliang 倉修良 et al.
Jinan: Shandong jiaoyu chubanshe, 1996.
Hardy, Grant. ―Can an Ancient Chinese Historian Contribute to Modern Western Theory?
The Multiple Narratives of Ssu-ma Ch‘ien.‖ History and Theory 33, no.1 (1994):
20-38.
—. ―Form and Narrative in Ssu-ma Ch‘ien‘s Shih chi.‖ CLEAR 14 (1992): 1-23.
—. ―His Honor the Grand Scribe Says…‖ CLEAR 18 (1996): 145-151.
—. ―The Interpretive Function of Shih chi 14, ‗The Table by Years of the Twelve Feudal
Lords‘.‖ JAOS 113, no.1 (1993): 14-24.
—. Worlds of Bronze and Bamboo. New York: Columbia University Press, 1999.
Hartman, Charles. Han Yü and the T’ang Search for Unity. Princeton: Princeton
University Press, 1986.
—. ―Poetry and Politics in 1079: The Crow Terrace Poetry Case of Su Shih.‖ CLEAR 12
(1990): 15-44.
Hawkes, David. The Songs of the South. New York: Penguin Books, 1985.
He Bingsong 何炳松. Zhedong xuepai suyuan 浙東學派溯源 [The origins of the
Zhedong school]. In He Bingsong wenji 何炳松文集. Beijing: Shangwu
yinshuguan, 1996.
Henderson, John B. Scripture, Canon, and Commentary: A Comparison of Confucian
and Western Exegesis. Princeton: Princeton University Press, 1991.
Hervouet, Yves. ―La valeur relative des textes du Che ki et du Han chou.‖ In Mélanges
de sinologie offerts à Monsieur Paul Demiéville, vol. 2. Paris: Presses
Universitaires de France, 1974. 55-76.

501
Hightower, James. ―The Fu of Tao Ch‘ien.‖ HJAS 17.1, June 1954: 169-230.
—, trans. ―Letter to Jen An (Shao-ch‘ing).‖ In Cyril Birth ed., Anthology of Chinese
Literature: From Early Times to the Fourteenth Century. New York: Grove Press,
1965.
Holoch, Donald. ―Melancholy Phoenix: Self Ascending from the Ashes of History.‖ In
Wolfgang Kubin ed., Symbols of Anguish: In Search of Melancholy in China;
Helmut Martin in memoriam (1940-1999). Schweizer asiatische Studien.
Monographien, Bd. 3. Bern; New York: Peter Lang, 2001. 169-212.
Honey, David. ―The Han-shu, Manuscript Evidence, and the Textual Criticism of the
Shih-chi.‖ CLEAR 21 (1999): 67-97.
Hsu, Yeong-huei. ―Song Gaozong (r.1127-1162) and his Chief Councilors: A Study of
the Formative Stage of the Southern Song Dynasty (1127-1279).‖ Ph.D.
dissertation, University of Arizona, 2000.
Huang Zhen 黃震. Huangshi richao 黃氏日抄 [Daily copying of Master Huang]. SKQS.
Huang Zongxi 黃宗羲. Song-Yuan xue’an 宋元學案 [Record of Song and Yuan scholars]
(SYXA). Beijing: Zhonghua shuju, 1986.
Huangfu Mi 皇甫謐. Gao shi zhuan 高士傳 [Traditions of Eminent Men] (GSZ). Zhou
Guangpei 周光培 ed., Han Wei bi ji xiao shuo 漢魏筆記小說, vol.1.
Shijiazhuang: Hebei jiaoyu chubanshe, 1994.
Huangfu Shi 皇甫湜. ―Biannian jizhuan lun‖ 編年紀傳論 [Discussion of chronicles and
annals-traditions]. Chizheng wenji 持正文集 [Chizheng‘s collected writings]. In
Congshu jicheng xubian 叢書集成續編 [Grand compendium series, continued],
vol.183. Taibei: Xin wen feng chubanshe, 1989.
Hucker, Charles O. A Dictionary of Official Titles in Imperial China. Stanford: Stanford
University Press, 1985.
Hulsewé, A.F.P. ―Founding fathers and yet forgotten men: a closer look at the tables of
the nobility in the Shih chi and the Han shu.‖ TP 75, nos.1-3 (1989): 43-126.
— . ―The Problem of Authenticity of Shih chi ch.123.‖ TP 61 (1975), 83-147.
Hummel, Arthur, trans. The Autobiography of a Chinese Historian: Being the Preface to
a Symposium on Ancient Chinese History (Ku shih pien). By Gu Jiegang 顧頡剛.
Leyden: Brill, 1931.
Itano Chōhachi. ―The t’u-ch’en Prophetic Books and the Establishment of
Confucianism.‖ Memoirs of the Toyo Bunko 34 (1976): 47-111.
Iser, Wolfgang. Act of Reading: A Theory of Aesthetic Response. Baltimore: Johns
Hopkins University Press, 1978.
Ji Xusheng 季旭昇, Shuowen xinzheng 說文新證 [New evidence regarding the Shuowen].
Taibei: Yiwen yinshuguan, 2003.
Jin Dejian 金德建. Sima Qian suo jian shu kao 司馬遷所見書考[An investigation of
books seen by Sima Qian]. Shanghai: Renmin chubanshe, 1963.
Jinshu 晉書 [History of the Jin] (JS). Compiled by Fang Xuanling 房玄齡 et al. Beijing:
Zhonghua shu ju, 1975.
Jiu Tangshu 舊唐書 [Old history of the Tang] (JTS). Compiled by Liu Xu 劉昫 et al.
Beijing: Zhonghua shu ju, 1975.

502
Kang Youwei 康有為. Xinxue weijing kao 新學偽經考 [A Study of the Forged Classics
of the Xin Period]. In Kang Youwei quanji 康有為全集 [Complete Works of
Kang Youwei]. Shanghai: Guji chubanshu, 1987.
Kern, Martin. ―The ‗Biography of Sima Xiangru‘ and the Question of the Fu in Sima
Qian‘s Shiji.‖ JAOS 123.2 (2003): 303-316.
— . ―A Note on the Authenticity and Ideology of Shih-chi 24.‖ JAOS 119.4 (1999): 673-
677.
— . ―Religious Anxiety and Political Interest in Western Han Omen Interpretation.‖
Studies in Chinese History 10 (2000): 1-31.
Knechtges, David. ― ‗Key Words,‘ Authorial Intent, and Interpretation: Sima Qian‘s
Letter to Ren An.‖ Conference on ―The Power of Words: The Interpretation of
Pre-modern Chinese Literature.‖ Harvard University (Harvard-Yenching
Institute), Cambridge MA. 27-28 Oct. 2006.
—, trans. and intro. Wenxuan, or Selections or Refined Literature vols.I-III. Princeton:
Princeton University Press, 1982.
Knoblock, John and Jeffrey Riegel, trans. The Annals of Lü Buwei: a Complete
Translation and Study. Stanford, CA: Stanford University Press, 2000.
Kroll, Jurij L. ―Ssu-ma Ch‘ien‘s Literary Theory and Literary Practice.‖ Altorientalische
Forschungen 4 (1976): 313-25.
Kundera, Milan. Immortality. Trans. Peter Kussi. New York: HarperCollins, 1992.
Lau, D.C., trans. Confucius: The Analects. New York: Penguin Books, 1979.
—, trans. Mencius. New York: Penguin Books, 1970.
Legge, James, trans. The Chinese Classics, 5 Vols. London: Trübner & Co., 1861.
Leslie, Donald et al., eds. Essays on the Sources for Chinese History. Canberra:
Australian National University Press, 1973.
Lewis, Mark Edward. Writing and Authority in Early China. Albany: State University
of New York Press, 1999.
Li Changzhi 李長之. Sima Qian zhi ren’ge yu fengge [司馬遷之人格與風格]. Shanghai:
Kaiming shudian faxing, 1949.
Li Fuyan 李福燕. ―Jianlun Sima Tan dui Sima Qian de yingxiang‖ 簡論司馬談對司馬
遷的影響 [Brief discussion of Sima Tan‘s influence on Sima Qian]. Nanping
shizhuan xuebao 23.1 (2004): 73-77.
Li Ji 禮記 [Record of Rites] (LJ). In Shisan jing zhushu 十三經注疏 [Commentaries and
sub-commentaries on the thirteen classics]. Taibei: Yiwen yinshuguan, 1955.
Li, Wai-yee. ―The Idea of Authority in the Shih chi.‖ HJAS 54 (1994), no.2: 345-405.
—. The Readability of the Past in Early Chinese Historiography. Cambridge: Harvard
University Asia Center, distributed by Harvard University Press, 2007.
Liang Qichao 梁啟超. Intellectual Trends in the Ch’ing Period. Trans. Immanuel C.Y.
Hsü. Cambridge: Harvard University Press, 1959.
—. Qingdai xueshu gailun 清代學術概論 [Intellectual Trends of the Qing Period]. In
Zhongguo lishi yanjiu fai 中國歷史研究法 [Research Models for Chinese
History]. Shijiazhuang: Hebei jiaoyu chubanshe, 2000.
Liang Yusheng 梁玉繩. Shiji zhiyi 史記志疑 [A record of doubts about Shiji]. Beijing:
Zhonghua shuju, 1981.

503
Liu chen zhu wenxuan 六臣註文選 [Literary Selections with Six Ministers Commentary]
(Liu chen WX). 2 volumes. Taibei: Guangwen shuju, 1964.
Liu, James J.Y. Chinese Theories of Literature. Chicago: University of Chicago Press,
1975.
Liu, James T.C. China Turning Inward: Intellectual-Political Changes in the Early
Twelfth Century. Cambridge: Council on East Asian Studies, Harvard University,
distributed by Harvard University Press, 1988.
Liu, Shih-shun. Chinese Classical Prose: The Eight Masters of the Tang-Song Period.
Hong Kong: Chinese University Press, 1979.
Liu Xie 劉勰. Wenxin diaolong zhu 文心雕龍注 [Literary mind and the Carving of
Dragons, annotated]. Fan Wenlan 范文瀾, ed. Beijing : Renmin wenxue
chubanshe, 1958.
—. Wenxin diaolong zhuding 文心雕龍註訂. Ed. Zhang Lizhai 張立齋. Taibei:
Zhengzhong shuju, 1971.
Liu Yiqing 劉義慶 (403-444). Shishuo xinyu 世說新語 [New Accounts of the World‘s
Tales] (SSXY). Shanghai: Shanghai gu ji chu ban she, 1982.
Liu Zhiji 劉知幾. Shitong tongshi 史通通釋 [Understanding history, with
comprehensive commentary] (STTS). Commentary by Pu Qilong 浦起龍.
Shanghai: Shanghai guji chubanshe, 1978.
Liu Zongyuan 柳宗元. Liu Zongyuan ji 柳宗元集[Collected writings of Liu Zongyuan].
Beijing: Zhonghua shuju, 2000.
Loewe, Michael. A Biographical Dictionary of the Qin, Former Han and Xin periods
(221 BC - AD 24). Leiden: Brill, 2000.
— , and Edward Shaughnessy, eds. The Cambridge History of Ancient China.
Cambridge: Cambridge University Press, 1999.
—. Crisis and conflict in Han China, 104 BC to AD 9. London : Allen & Unwin, 1974.
—, ed. Early Chinese Texts: A Bibliographical Guide. Berkeley: University of
California Press, 1993.
—. ―Review of Grant Hardy, Worlds of Bronze and Bamboo.‖ TP 87 (2001): 221-230.
Lü Buwei 呂不韋 et al. Lüshi chunqiu jiaoshi 呂氏春秋校釋 [The Spring and Autumn
of Master Lü collated and translated] (LSCQ). Commentary by Gao You 高誘,
collation and modern Chinese translation by Chen Qiyou 陳奇猷. Taibei:
Zhongzheng shuju, 1988.
Lü Xisheng 呂錫生. ―Sima Qian gongxing xiyi‖ 司馬遷宮刑析疑 [An analysis of
doubts about Sima Qian and the ‗palace punishment‘]. Zhongguo shi yanjiu 中國
史研究 4 (1983): 68.
Lu Zongli. ―Problems Concerning the Authenticity of Shih chi 123 Reconsidered.‖
CLEAR 17 (1995): 51-68.
Lü Zuqian 呂祖謙. Dashiji jieti 大事記解題 [Chronicle of major events with an
explanation of headings]. In Wang Yi 王禕 cont., Dashiji xubian 大事記續編
[Chronicle of major events continued]. Siku quanshu zhenben fifth series, vols.
62-71. Taibei: Taiwan shangwu yinshuguan, 1974.

504
Luo Fangsong 羅芳松. ―Cong ‗nang‘ yu ‗jin‘ guan kui ‗Bao Ren An shu‘‖ 從囊與今觀
窺報任安書 [Peering at the ‗Letter in Reply to Ren An‘ from the use of
‗formerly‘ and ‗now‘]. Chengdu daxue xuebao 1985 no.3: 37-41.
Ma Duanlin 馬端臨. Wenxian tongkao 文獻通考 [Comprehensive Investigation of
Documents] (WXTK). Beijing: Zhonghua shuju, 1986.
Ma Zong 馬總. Yilin 意林 [Forest of intentions]. SBCK.
Mather, Richard B., trans. Shih-shuo Hsin-yu: A New Account of Tales of the World,
Second Edition. Ann Arbor: Center for Chinese Studies, University of Michigan,
2002.
McMullen, David. State and Scholarship in T’ang China. Cambridge: Cambridge
University Press, 1988.
Min Zeping 閔澤平. ―Handai shuxinti sanwen yishu gailun‖ 漢代書信體散文藝術概論
[A discussion of the art of the Han dynasty epistolary prose genre]. Tianzhong
xuekan 天中學刊, vol.16 no.1 (Feb.2001): 81-83.
Nan Qi shu 南齊書 [History of the Southern Qi dynasty]. Compiled by Xiao Zixian 蕭子
顯 et al. Beijing: Zhonghua shuju, 1975.
Nanshi 南史 [History of the Southern dynasties]. Compiled by Li Yanshou 李延壽 et al.
Beijing: Zhonghua shuju, 1975.
Needham, Joseph. Science and Civilisation in China (Volume 3): Mathematics and the
Sciences of the Heavens and the Earth. Cambridge: Cambridge University Press,
1959.
Nehamas, Alexander. ―Writer, Text, Work, Author.‖ In A. J. Cascardi, ed., Literature
and the Question of Philosophy. Baltimore: Johns Hopkins University Press,
1987. 265–91.
Ng, On-cho and Q. Edward Wang. Mirroring the Past: The Writing and Use of History
in Imperial China. Honolulu: University of Hawai‘i Press, 2005.
Nienhauser, William H. ―A Century (1895-1995) of Shih chi Studies in the West.‖
Asian Culture Quarterly 24.1 (1996): 1-51.
—. ―Historians of China.‖ CLEAR 17 (Dec 1995): 207-216.
—. ―An Interpretation of the Literary and Historical Aspects of the Hsi-ching tsa-chi
(Miscellanies of the Western Capital).‖ PhD dissertation, Indiana University,
1972.
—. ―A Note on a Textual Problem in the Shih Chi and Some Speculations Concerning
the Compilation of the Hereditary Houses.‖ TP 89, nos.1-3 (2003): 39-58.
—. ―On reading the ‗Basic Annals of Gaozu‘: Liu Bang and biography in Sima Qian‘s
eyes‖ Neder, Christina; Roetz, Heiner; Schilling, Ines-Susanne, eds. China and
her biographical dimensions: commemorative essays for Helmut Martin.
—. ―A Reexamination of ‗The Biographies of the Reasonable Officials‘ in the Records
of the Grand Historian.‖ Early China 16 (1991): 209-234
—. ―Worlds of Bronze and Bamboo, Sima Qian‘s Conquest of History.‖ CLEAR 22
(2000): 155-168.
—. ―The Study of the Shih-chi (The Grand Scribe‘s Records) in the People‘s Republic of
China.‖ In Helwig Schmidt-Glintzer, ed., Das andere China. Wiesbaden:
Harrassowitz Verlag in Kommission, 1995. pp.381-403.

505
—. ―Tales of the Chancellor(s): the Grand Scribe‘s Unfinished Business.‖ CLEAR 25
(Dec 2003) 99-117.
—. ―Travels with Édouard—V.M. Alekseev‘s Account of the Chavannes Mission of
1907 as a Biographical Source.‖ Asian Culture Quarterly 22 (1994): 81-95.
— et al., trans. The Grand Scribe’s Records (four volumes: I, II, V pt.1, VII, VIII).
Bloomington: Indiana University Press, 1994-.
Nylan, Michael. ―The chin wen/ku wen Controversy in Han Times.‖ TP 80 (1994): 83-
145.
—. The Five “Confucian” Classics. New Haven: Yale University Press, 2001.
—. ―Sima Qian, A True Historian?‖ Early China 23-24 (1998-1999): 203-246.
Ouyang Xiu 歐陽修. Ouyang Xiu wenji 歐陽修文集. Beijing: Zhonghua shuju, 2001.
—. Shi ben yi 詩本義 [Fundamental Principles of the Odes]. SBCK.
— et al. Xin Tangshu 新唐書 [New history of the Tang] (XTS). Beijing: Zhonghua shuju,
1975.
Pelliot, Paul. ―L‘Édition collective des oeuvres de Wang Kouo-wei.‖ TP 26 (1929):
113-182.
—. ―Review of Arthur Hummel, trans., Autobiography of a Chinese Historian.‖ TP 29
(1932): 130-135.
—. ―Review of Richard Wilhelm, trans., Frühling und Herbst des Lü Bu We.‖ TP 27
(1930): 68-91.
Peterson, Willard J. ―Ssu-ma Ch‘ien as Cultural Historian.‖ In Peterson et al., The Power
of Culture: Studies in Chinese Cultural History. Hong Kong: Chinese University
Press, 1994. 70-79.
Pines, Yuri. ―Chinese History Writing Between the Sacred and the Secular.‖ In John
Lagerwey and Marc Kalinowski eds., HdO Early Chinese Religion: Shang
through Han (1250 BC - 220 AD). Leiden: Brill, 2008.
— . Foundations of Confucian Thought: Intellectual Life in the Chunqiu Period, 722-
453 B.C.E. Honolulu: University of Hawaii Press, 2002.
Pokora, Timoteus. ―The Chronological Tables in the Shih-chi and Their Interpolations,‖
Orientalistische Literaturzeitung 82.3 (1987): 229-44.
— . ―Ch‘u Shao-sun—The Narrator of Stories in the Shih chi.‖ Annali dell’Istituto
Orientale di Napoli 41 (1981): 403-430.
— . ―The First Interpolation in the Shih chi,‖ ArOr 29.2 (1961): 311-15.
— . Hsin-lun (New Treatise) and Other Writings by Huan T’an. Ann Arbor: Center for
Chinese Studies, University of Michigan, 1975.
— . ―Ironical Critics at Ancient Chinese Courts (Shih chi 126),‖ OE 20 (1973): 49-64.
— . ―The Less Eminent Followers of Kao-tsu: Shih-chi 98,‖ Rocznik Orientalistyczny
44.1 (1985): 21-40.
— . ―Review of Burton Watson‘s Ssu-ma Ch’ien, Grand Historian of China.‖ TP 50
(1963): 294-322.
— . ―Shih chi 127, the Symbiosis of Two Historians.‖ Charles Le Blanc and Susan
Blader, eds. Chinese Ideas about Nature and Society:Studies in Honour of Derk
Bodde. Hong Kong: Hong Kong University Press, 1987. 215-34.
Puett, Michael. The Ambivalence of Creation: Debates Concerning Innovation and
Artifice in Early China. Stanford: Stanford University Press, 2001.
— . ―Review of The Cloudy Mirror.‖ HJAS 57.1 (1997): 290-301.

506
Pulleyblank, E.G. ―Liu K‘o, A Forgotten Rival of Han Yü.‖ AM 7 (1959): 145-160.
Qian Mu 錢穆. ―Taishigong kao shi‖ 太史公考釋 [The Honorable Senior Archivist
investigated and explained.] Zhongguo xueshu sixiang shi luncong 中國學術思
想史論叢 [Collected essays on the Chinese intellectual and academic history].
Taibei: Dongda tushuguan gongsi, 1977.
Qian Zhongshu 錢鍾書. Guan Zhui Bian 管錐編. Beijing Shi: Shenghuo, dushu, xin zhi
sanlian shudian, 2001.
Qin Guan 秦觀. Huaihai ji jianzhu 淮海集箋注 [The annotated Huaihai collection].
Shanghai: Shanghai guji chubanshe, 1994.
Qin Jiamo 秦嘉謨. Shiben jibu 世本輯補 [Basic Generations compiled and
supplemented]. In Siku quanshu weishou shu jikan 四庫全書未收書輯刊 [Texts
not collected in the Siku quanshu, compiled and reproduced]. China: Linlang
xianguan, Jiaqing 23 [1818].
Qinding Siku quanshu zongmu 欽定四庫全書總目 (SKQS ―Zongmu‖). Beijing:
Zhonghua shuju, 1997.
Quan Tang wen 全唐文 [Complete Tang prose] (QTW). Dong Gao 董誥 et. al. ed.
Beijing: Zhonghua shuju yingyin, 1983.
Quan Song Shi 全宋詩 [Complete Song poems]. Beijing: Beijing daxue chubanshe, 1991.
Quan Song Wen 全宋文[Complete Song prose]. Shanghai: Cishu chubanshe, 2006.
Quirin, Michael. Liu Zhiji und das Chun Qiu. Frankfurt am Main: P. Lang, 1987.
Rolston, David. Traditional Chinese Fiction and Fiction Commentary: Reading and
Writing Between the Lines. Stanford: Stanford University Press, 1997.
Ruan Zhisheng 阮芝生. ―Sima Qian zhi xin—‗Bao Ren Shaoqing shu‘ xilun‖ 司馬遷之
心《報任少卿書》析論 [Sima Qian‘s heart—analysis and discussion of the
‗Letter in Reply to Ren Shaoqing‘]. Taida lishi xuebao 臺大歷史學報 26 (2000):
151-205.
Sargent, Stuart H., trans. ―Understanding History: the Narration of Events, by Liu Chih-
chi.‖ Renditions 15 (1981): 27-33.
Schaab-Hanke, Dorothee. ―Did Chu Shaosun Contribute to a Tradition of the Scribe?‖
OE 44 (2003), no.4: 11-26.
Schaberg, David. A Patterned Past: Form and Thought in Early Chinese Historiography.
Cambridge: Harvard University Asia Center, distributed by Harvard University
Press, 2001.
Scholes, Robert and Robert Kellogg. The Nature of Narrative. New York: Oxford
University Press, 1975.
Shao Yiping 卲毅平. ―Han Mingdi zhaoshu yu Ban Gu‖ 漢明帝詔書與班固 [Ban Gu
and the edicts of Han Emperor Ming]. Fudan xuebao 1985, no.6: 65-68.
Shen Fen 湛芬. Zhang Lei xueshu wenhua sixiang yu chuangzuo 張耒學術文化思想與
創作 [Zhang Lei‘s thought on intellectual culture and his creation]. Bashu shushe,
2004.
Shi Ding 施丁. ―Sima Qian xie ‗Bao Ren An shu‘ niandai kao‖ 司馬遷寫《報任安書》
年代考 [An investigation of the year in which Sima Qian wrote the ―Letter in
Reply to Ren An‖]. Xinan shifan xueyuan xuebao 1985 no.4: 41-47.

507
—. Sima Qian xingnian xinkao 司馬遷行年新考 [A new investigation into the life and
activities of Sima Qian]. Xi‘an: Shanxi renmin jiaoyu chubanshe, 1995.
Shih, Vincent Yu-chung, trans. The Literary Mind and the Carving of Dragons. New
York: Columbia University Press, 1959.
Sima Qian 司馬遷. Shiji 史記 [Archivist‘s Records] (SJ). Beijing: Zhonghua shuju,
1982.
Shiji yanjiu jicheng 史記研究集成 [Collected achievements in Shiji research] (SJYJJC).
Eds. Zhang Dake 張大可, An Pingqiu 安平秋, Yu Zhanghua 俞樟華. Beijing:
Huawen chubanshe, 2005. 14 volumes.
Shisan jing zhushu 十三經注疏 [Commentaries and sub-commentaries on the thirteen
classics]. Taibei: Yiwen yinshuguan, 1955.
Smith, Kidder. ―Sima Tan and the Invention of Daoism, ‗Legalism,‘ et cetera.‖ JAS 62.1
(2003): 129-56.
Solomon, Bernard S. The Veritable Record of the T’ang Emperor Shun-tsung.
Cambridge, MA: Harvard University Press, 1955.
Songshi 宋史 [History of the Song]. Tuo Tuo 脫脫 et. al., ed. Beijing: Zhonghua shuju,
1975-1981.
Su Shi 蘇軾. Dongpo zhilin 東坡志林. Shanghai: Huadong Shifan Daxue, 1983.
—. Su Shi wenji 蘇軾文集. Beijing: Zhonghua shuju, 2004.
Su Xun 蘇洵. Jiayou ji jian shu 嘉祐集箋注 [Annotated collection of the Jiayou period].
Shanghai: Shanghai guji chubanshe, 1993.
Su Zhe 蘇轍. Su Zhe ji 蘇轍集. Beijing: Zhonghua shuju, 1990.
Suishu 隋書 [History of the Sui] (SuiS). Compiled by Wei Zheng 魏徵. Beijing:
Zhonghua shuju, 1975-1981.
Sukhu, Gopal. ―Yao, Shun, and Prefiguration: the Origins and Ideology of the Han
Imperial Genealogy.‖ Early China 30 (2005-2006): 91-153.
Sun Deqian 孫徳謙. Taishigong shu yifa 太史公書義法. Taibei: Taiwan Zhonghua shu
ju, 1969.
Takigawa Kametaro 瀧川龜太郎. Shiki kaichū kōshō 史記會注考證 [Collected
commentaries on the Shiji, researched and investigated] (1934) (SKK). Reprinted
as: Shiji huizhu kaozheng fu 史記會注考證. Taiyuan Shi: Bei yue wen yi chu ban
she, 1999.
Tao Qian 陶潛. Tao Yuanming jijian zhu 陶淵明集箋注. Commentary by Yuan Xingpei
袁行霈. Beijing: Zhonghua shuju, 2005.
Thucydides. History of the Peloponnesian War. Rex Warner, trans. New York: Penguin
Books, 1954, revised 1972.
Tillman, Hoyt. Utilitarian Confucianism: Ch’en Liang’s Challenge to Chu Hsi.
Cambridge, MA: Council on East Asian Studies, Harvard University: Distributed
by Harvard University Press, 1982.
Tuo Tuo 脫脫 et al. Jinshi 金史 [History of the Jin]. Beijing: Zhonghua shuju, 1981.
Twitchett, Denis. ―Chinese Biographical Writing.‖ In W.G. Beasley and E.G.
Pulleyblank, eds. Historians of China and Japan. London: Oxford University
Press, 1961. 95-114.

508
—. The Writing of Official History under the T’ang. Cambridge: Cambridge University
Press, 1992.
Van Ess, Hans. ―Recent Studies on Sima Qian.‖ MS 49 (2001): 517-528.
Vankeerberghen, Griet. ―The Tables (biao) in Sima Qian‘s Shiji: Rhetoric and
Remembrance.‖ In Bray et al., ed., Graphics and Text in the Production of
Technical Knowledge in China. Leiden: Brill, 2007.
Vervoorn, Aat Emile. Men of the Cliffs and Caves: The Development of the Chinese
Eremitic Tradition to the End of the Han Dynasty. Hong Kong: Chinese
University Press, 1990.
Wang Anshi 王安石. Wang Jing Gong shi zhu 王荆公詩注 [Commentary on the poems
of Wang Jing Gong]. Commentary by Li Bi 李壁. SKQS.
Wang Chong 王充. Lunheng 論衡 [Discourses weighed in the balance] (LH). Beijing:
Zhonghua shuju, 1995.
Wang Guanguo 王觀國. Xuelin 學林 [A forest of learning]. Beijing: Zhonghua shuju,
1988.
Wang Guowei 王國維 ―Taishigong xingnian kao‖ 太史公行年考 [An investigation into
the life and activities of the Honorable Senior Archivist]. In Guantang jilin 觀堂
集林. Taibei : Yiwen yinshu guan, 1956.
Wang Ruoxu 王若虛. Shiji bianhuo 史記辨惑 [Refuting errors regarding the Shiji].
Hunan Yilao ji 滹南遺老集. Shanghai: Shangwu yinshuguan, 1937.
Wang Tong 王通 (584-618). Wen Zhongzi zhong shuo yi zhu 文中子中說譯注
(Zhongshuo). Zheng Chunying 鄭春穎 trans., commentary. Heilongjiang renmin
chubanshe, 2003.
Wang Yun 王惲 (1227-1304). Qiujian xiansheng daquan wenji 秋澗先生大全文集
[The great complete writings of Master Qiujian]. SBCK. Taibei: Taiwan
shangwu yinshuguan, 1967.
Wang Zhengde 王正徳 (fl.12th c.). Yushi lu 餘師錄 [Records from an abundance of
teachers]. SKQS.
Watson, Burton. Ssu-ma Ch’ien, Grand Historian of China. New York: Columbia
University Press, 1958.
—, trans. Records of the Historian: Chapters from the Shih chi of Ssu-ma Ch’ien. New
York: Columbia University Press, 1958.
—, trans. Records of the Grand Historian: Han Dynasty (Han I, Han II). New York:
Renditions-Columbia University Press, 1993. 2 volumes.
—, trans. Records of the Grand Historian: Qin Dynasty (Qin). New York: Columbia
University Press, 1993.
Whitaker, K.P.K. ―Some Notes on the Authorship of the Lii Ling/Su Wuu Letters,‖ parts
I-II. BSOAS 15 (1953): 113-137, 566-587.
—. ―Some Notes on the Background and Date of Tsaur Jyr‘s Poem on the Three Good
Courtiers.‖ BSOAS 18 (1956): 303-311.
White, Hayden. ―The Value of Narrativity.‖ In W.J.T. Mitchell ed., On Narrative.
Chicago: University of Chicago Press, 1980. 1-24.
Wöfflin, Heinrich. Principles of Art History: The Problem of the Development of Style in
Later Art. M.D. Hottinger, trans. New York: Dover, 1932.

509
Wu Jiansi 吳見思. Shiji lunwen 史記論文 [Essays on the Shiji]. Shanghai: Guji
chubanshe, 2008.
Wu, Pei-yi. The Confucian’s Progress: Autobiographical Writings in Traditional China.
Princeton: Princeton University Press, 1990.
Xiao Li 肖黎. Sima Qian pingzhuan 司馬遷評傳 [Critical biography of Sima Qian].
Jilin: Wenshi chubanshe, 1986.
Xiao Tong 蕭統 et al., comp. Wenxuan 文選 [Literary selections] (WX). Commentary
by Li Shan 李善. Beijing: Zhonghua shuju, 1977.
Xu Qianxue 徐乾學. Zizhi tongjian houbian 資治通鑑後編 [Later compilation of the
Comprehensive mirror as an aid to government]. SKQS.
Xu Shen 許慎. Shuowen jiezi 說文解字 [Explanation of Writing and Analysis of
Characters]. Beijing: Zhonghua shuju, 1978.
Xu Shuofang 徐朔方. Shi Han lungao 史漢論稿 [Preliminary discussion of the Shiji and
Hanshu]. Nanjing: Jiangsu guji chubanshe, 1994.
Xunzi jiaoshi 荀子校釋 [The Xunzi collated and explained.] Taibei: Xuesheng shuju,
1981.
Yiwen leiju 藝文類聚 [Categorized compendium of arts and literature]. Ouyang Xun 歐
陽詢, ed Taibei: Wenguang chubanshe, 1974.
Yan Kejun. Quan Jin Wen 全晉文 [Complete Prose of the Jin Dynasty]. Beijing:
Shangwu yinshuguan, 1999.
Yang Bojun 楊伯峻. Chunqiu zuozhuan zhu 春秋左傳注 [Commentary on the Spring
and Autumn with the Zuozhuan] (CQZZ zhu). Beijing: Zhonghua shuju, 2000.
—. Lunyu yizhu 論語譯注 [The Analects with translation and commentary]. Beijing:
Zhonghua shuju 1989.
Yang Haizheng 楊海崢. Han Tang Shiji yanjiu lungao 漢唐《史記》研究論稿
[Preliminary discussion of Shiji research from Han to Tang]. Jinan: Qi Lu shushe,
2003.
—. ―Wang Ruoxu de Shiji bianhuo‖ 王若虛的《史記辨惑》[Wang Ruoxu‘s Criticizing
Errors in Shiji]. Beijing University Journal of the Institute of Chinese Ancient
Texts. Beijing: Beijing Yanshan Chubanshe, vol.1 (1999): 299-331.
Yang Xiong 揚雄. Fayan yishu 法言義疏 [Exemplary sayings with glossees and sub-
commentary]. Beijing: Zhonghua shuju, 1987.
—. Yangzi fayan 揚子法言 [Exemplary sayings of Master Yang]. Commentary by Liu
Zongyuan 柳宗元 and Sima Guang 司馬光. SKQS.
Yang Yanqi 楊燕起 et al. Shiji jiping 史記集評 [Collected comments on the Shiji]. Shiji
yanjiu jicheng 史記研究集成 [Collected research results on the Shiji], vol.6.
Beijing: Huawen chubanshe, 2005.
Ye Shi 葉適. Shuixin xiansheng wenji 水心文集 [Master Heart of Water‘s Collected
Writings]. SBCK. Shanghai: Shangwu yinshuguan, 1929.
Yu Jiaxi 余嘉錫. Taishi gong shu wangpian kao 太史公亡篇考 [Investigation into the
lost chapters of the Honorable Senior Archivist] (Wangpian). In Yu Jiaxi lunxue ji
余嘉錫論學集 I.1-108. Taipei: Wenhai, 1979.

510
Yu Zhanghua 俞樟華. Shiji xintan 史記新探 [Shiji newly investigated]. Beijing: Minzu
chubanshe, 1994.
—. ―Song ren dui Shiji de yanjiu‖ 宋人對《史記》的研究 [The research done by Song
dynasty people on the Shiji]. 浙江師大學報, 1995.4.
—. ―Tang Song badajia yu Shiji‖ 唐宋八大家與《史記》 [The eight masters of the
Tang/Song and the Shiji]. Jiangsu shehui kexue 江蘇社會科學 (1996): 134-137.
—. Tang Song Shiji jieshou shi 唐宋史記接受史 [Shiji reception history in the Tang and
Song]. Jilin renmin chubanshe, 2004.
Yuan Chuanzhang 袁傳璋. ―Wei Wei Hong zhi Sima Qian ‗xia yu si shuo‘ bian wu bu
zheng‖ 為衛宏之司馬遷下獄死說辨誣補證 [Disputing slander and
supplementing evidence regarding Wei Hong‘s ―theory that Sima Qian died in
jail‖]. Anhui shixue 安徽史學 1984 no.3: 20-29.
Yuanshi 元史 [History of the Yuan]. Song Lian 宋濂 et.al., ed. Beijing: Zhonghua shuju,
1975-1981.
Zeng Gong 曾鞏. Zeng Gong ji 曾鞏 集. Beijing: Zhonghua shuju, 2004.
Zhanguo ce 戰國策 [Persuasions of the warring states]. Compiled by Liu Xiang 劉向.
Shanghai: Guji chubanshe, 1978.
Zhang Dake 張大可. Shiji yanjiu 史記研究 [Research on the Shiji]. Lanzhou: Gansu
People‘s Press, 1985.
—. Shiji wenxian yanjiu 史記文獻研究 [Textual research on the Shiji]. Beijing: Minzu
chubanshu, 1999.
—. Sima Qian pingzhuan 司馬遷評傳 [A critical biography of Sima Qian]. Nanjing:
Nanjing daxue chubanshe, 1994.
Zhang Hao 張淏 (fl.1216). Yungu zaji 雲谷雜紀 [Miscellaneous records from Cloud
Valley]. Beijing: Zhonghua shuju, 1991.
Zhang Lei 張耒. Zhang Lei ji 張耒集 [Collected works of Zhang Lei]. Beijing:
Zhonghua shuju, 1990. 2 volumes.
Zhang Peiheng 章培恒 and Liu Jun 劉駿. ―Guanyu Li Ling ‗Yu Su Wu shi‘ ji ‗Da Su
Wu shu‘ de zhenwei wenti‖ 關於李陵《與蘇武詩》及《答蘇武書》的真偽問
題 [On the problem of the authenticity of Li Ling‘s ―Poems to Su Wu‖ and
―Letter of Reply to Su Wu‖]. Fudan xuebao 復旦學報. 1998 no.2: 71-77.
Zhang Tingwang 張廷王 et.al. Mingshi 明史 [History of the Ming]. Beijing: Zhonghua
shuju, 1975-1981.
Zhang Weifang 張偉芳. ―‗Bao Ren An shu‘ dui ‗Bao Sun Huizong shu‘ de yingxiang‖
《報任安書》對《報孫會宗書》的影響 [The influence of the ‗Letter in
Response to Ren An‘ on the ‗Letter in Response to Sun Huizong‘]. Changzhou
shifan zhuanke xuexiao xuebao, vol. 22 no. 1 (2004): 8-10.
Zhang Xuecheng 章學誠. Wenshi tongyi jiaozhu 文史通義校注 [Comprehensive
principles of prose and history, collated and annotated] (WSTY). Beijing:
Zhonghua shuju, 2005.
Zhao Yi 趙翼. Gaiyu congkao 陔餘叢考 [Collected investigations of Gaiyu]. Hebei
renmin chubanshe, 2007.

511
—. Nian er shi zhaji 廿二史劄記 [Notes on the twenty-two histories]. Beijing:
Zhonghua shuju, 1963.
Zhou Hong 周洪. ―Shi lun Sima Qian yu Ren An de guanxi‖ 試論司馬遷與任安的關系
[An attempt to discuss the relationship between Sima Qian and Ren An]. Lishi
jiaoxue 歷史教學 [The teaching and study of history], 2000 no.4: 51-54.
Zhou Shouchang 周壽昌. Hanshu zhu jiao bu 漢書注校補 [The Hanshu with collation,
commentary, and supplements]. Shanghai: Shangwu yinshuguan, 1936.
Zhou Zizhi 周紫芝. Taicang timi ji 太倉稊米集 [Collected rice shoots in the great
granary]. SKQS. Taibei: Taiwan shangwu yinshuguan, 1971. Vol.267.
Zhu Xi 朱熹. Zhuzi yulei 朱子語類 [Categorized sayings of Master Zhu] (ZZYL).
Beijing: Zhonghua shuju, 1986.
—. Zizhi tongjian gangmu 資治通鑑綱目 [A general outline of the Comprehensive
mirror as an aid to government]. SKQS.
Zhuangzi jinzhu jinyi 莊子今注今譯 [The Zhuangzi with modern commentary and
translation] (Zhuangzi). Commentary by Chen Guying 陳鼓應. Hong Kong:
Zhonghua zhuju, 1990.

512