If there has been one fundamental issue that has continued to be a thorn in the side of Shi ji 史記 studies, it is that of authorship.Footnote 1 While Sima Tan 司馬談 (d. 110 b.c.e.) started the project, his untimely death meant that his son Sima Qian 司馬遷 (c. 145–c. 87 b.c.e.) inherited not only his official post but also his life's work. Beyond this basic knowledge, many unknowns and questions persist.
First, there is no clear demarcation between Sima Tan's efforts and those of his son. What was the degree of completion when the project was bequeathed to Sima Qian? What level of creative input did each have on its structure, scope, and content? Some have attempted to work through this issue by identifying the chapters that were composed by Sima Tan and those by Sima Qian using language analysis and the timing of the events recorded; however, this has largely been an effort in speculative scholarship, especially since both are identified by the appellation Taishigong 太史公 in the work, which effectively blends their voices into one.Footnote 2
Next, the number of people involved is not known. Were Sima Tan and Sima Qian the only two people who worked on the project? It is generally assumed that this was a father–son endeavor, and the text itself seems to give no indication otherwise. Still, with their court position, it is not entirely impossible that there was additional collaboration. William Nienhauser has posited as much: “It seems quite possible that such assistants with access to documents and working for the T'ai-shih ling could also have been involved in the composition of the Ssu-mas’ huge history.”Footnote 3 Yet, it is not easy to decisively demonstrate this.
Also, it is uncertain how many knew of the project. Shi ji has long been hailed as a private endeavor (sixiu 私修), unlike the officially commissioned histories of later dynasties.Footnote 4 With such a massive undertaking that took years to complete, did no one else know of this project? The fact that Sima Qian said he deposited the original in an official repository while keeping a copy at his home (Zhang Dake's 張大可 interpretation) casts doubt on this idea.Footnote 5 Furthermore, it is possible that Emperor Wu 武帝 (r. 141–87 b.c.e.) may have known about the project before its completion. According to a passage in Xijing zaji 西京雜記:
作《景帝本紀》，極言其短及武帝之過，帝怒而削去之。後坐舉李陵，陵降匈奴，下遷蠶室。有怨言，下獄死。
[Sima Qian] composed “The Annals of Emperor Jing,” speaking much of his shortcomings and of Emperor Wu's mistakes. The Emperor was incensed and cut those parts out. Later [Sima Qian] was punished because of his support for Li Ling after Ling surrendered to the Xiongnu. Qian was sent to the Silkworm Chamber and castrated. [Later] he made resentful remarks. As a result, he was imprisoned and died there.Footnote 6
From the above three aspects, it is clear that what we do know about the father–son effort to compose Shi ji is far less than what we need to know to fully understand the process. For the sake of expediency, most scholars simply cite Sima Qian as its author. Ban Biao 班彪 (c.e. 3–54) and his son Ban Gu's 班固 (c.e. 32–92) both did, with the latter's “Treatise on Literature” (yiwen zhi 藝文志) giving Sima Qian official authorship. This is perhaps the best that can be done with this authorial enigma, and it makes sense given the presumption that he had final editorial authority over the work.Footnote 7
If we take Sima Qian's textual lead, however, the aforementioned aspects of authorship are perhaps not the most important. Rather, it is Sima Qian's interaction with texts available to him and the resultant compilation. This can be seen from his arrangement of Shi ji 130, which is both a biography of the Simas and a summary of Shi ji. This entire chapter is text-centric—from his scribal ancestors in Zhou, to his father's post as Director of Court Scribes (Taishi ling 太史令) and discourse on the six schools, to his father's dying wish to remember and make up for the textual loss by completing Shi ji,Footnote 8 to Sima Qian's becoming Director of Court Scribes and organizing (chou 紬) the texts in the court repositories, to his discussion of the Annals with Hu Sui, to his reasons for composing each chapter of Shi ji, to his summary of the Shi ji project and its results. Moreover, in looking specifically at Sima Qian's autobiographic narration, his life seems tailored to this textual endeavor. Under his father's tutelage he mastered ancient writing forms at the age of ten—a full ten years earlier than those who entered scribal service—and at twenty he spent several years traveling about the empire, gathering documents and oral histories. When he became the Director of Court Scribes two years or so after his father's death,Footnote 9 the post provided access to all of the documents in the outer court's imperial archives. Then, his unfortunate castration resulting from the Li Ling affair made possible his appointment as the Director of Palace Writers (Zhongshu ling 中書令), thus affording him access to records of the inner court.
Deng Junjie 鄧駿捷 claims the repositories in the two courts are actually “two systems,” suggesting the possibility that the content of both might not be identical.Footnote 10 If this was the case, then a reinterpretation of the entire Li Ling affair may be in order, as it could be viewed as something of a blessing in disguise, making available material that was previously inaccessible (or even unknown) to him. At the very least, it afforded him an intimate look into the mind of Emperor Wu, as is reflected in Shi ji 28. Beyond that, it may have also provided him first-hand access to account books (jishu 計書) from across the empire. Han shu 漢書 commentator Ru Chun 如淳 records,
《漢儀》注：太史公，武帝置，位在丞相上。天下計書先上太史公，副上丞相，序事如古春秋。
An annotation in Hanyi states: [The office of] Grand Scribe was set up by Emperor Wu and was above the post of chancellor [in power]. All account books from throughout the empire were to be first sent to the Grand Scribe, with a copy sent to the chancellor. Events were ordered just as in the ancient histories.Footnote 11
Wu Changlian 吳昌廉 has argued that taishigong is clearly an error and concludes that the correct post to fit this statement is Director of Palace Writers, which was the highest in the inner court.Footnote 12 This access to court records, both inner and outer, was a critical element in the completion of Shi ji, since writings contained in the various repositories were the primary source material used in its compilation. Or as Sima Qian stated, with a degree of satisfaction, “Over [the last] one hundred years, all of the old writings and ancient happenings in the realm were collected [into the hands of] the Grand Scribe” (百年之閒，天下遺文古事靡不畢集太史公).Footnote 13
Given the text-centric nature of Sima Qian's autobiographical narrative and Shi ji 130 as a whole, perhaps the questions regarding authorship that should be asked are: What is the relationship between the editorial efforts of Sima Qian and the documents used as sources? How much of the materials available to him did he use? How did he use these materials? How much of Shi ji is his own writing? Bernhard Karlgren has looked at this last issue from the angle of archaic grammatical usages from Zhou-era texts. He analyzed chapters he denotes as “[Sima Qian's] independent chapters,” comparing content from Zuo zhuan 左傳, Guo yu 國語, Zhanguo ce 戰國策, Lüshi chunqiu 呂氏春秋, Xunzi 荀子, and the Analects cited by Sima Qian with chapters 7–12, 48–60, 88–116, 120–26, and 130. Karlgren found that Sima Qian wrote in the Han style in these later chapters, “deviat[ing] from the ancient patterns.”Footnote 14 He implies that we can see more of Sima Qian's own style in these chapters that deal with Qin–Han happenings. However, “independent” is somewhat misleading, as it can only be understood to signify independent of these ancient sources, but not all sources altogether. In chapters 7 and 48, and part of chapter 8, for example, Sima Qian drew on the work Chu-Han chunqiu 楚漢春秋 in writing about the fall of the Qin and the establishment of the Han. Stephen Durrant has observed that “Sima Qian felt quite free to rework his source either by reference to other written or oral sources or his own creative imagination” and that “many of the fragments from Seasons of Chu and Han have much less ‘literary polish’ than their parallels in Shi ji.”Footnote 15 Still, the delineation between “source material” and “own writing” is still not clear.
Furthermore, Taishigong yue 太史公曰 comments found at the beginning or end of most chapters indicate a strong authorial voice; but what about the main text of each chapter?Footnote 16 To put it another way, is the voice within each chapter a strong author in command of the source materials, or an overwhelmed one struggling to rein in the unwieldy amount of textual data?Footnote 17 And why is there tension between Sima Qian's clear scholarly voice pushing for accuracy and what reads like his desire to paint certain figures in a particular light?Footnote 18 From these questions and issues, it is clear that Shi ji sources are still very much unexplored. Perhaps this is due to the difficulty of large-scale textual analysis. Yet, given advances in information technology and archeology, this aspect has never been more open to exploration. What is needed first is a bibliography of Shi ji source texts from which to pursue these various avenues of inquiry.
Bibliographic Studies on Shi ji
After Shi ji was completed and began circulating, a succession of scholars started composing additional biographies based on earlier histories, past happenings, and different accounts. The motivation was to make up for the gap of unrecorded history that had transpired since Sima Qian's demise and to correct errors perceived in Shi ji. One of the most prolific of these scholars was Ban Biao. Hou Han shu 後漢書 records,
彪既才高而好述作，遂專心史籍之閒。武帝時，司馬遷著《史記》，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。彪乃繼采前史遺事，傍貫異聞，作後傳數十篇，因斟酌前史而譏正得失。
Biao was both highly talented and fond of writing. After [leaving his post] he focused on historical works. During the reign of Emperor Wu, Sima Qian composed Shi ji; however, nothing was recorded from the taichu period onward. Later, officious persons expended some effort gathering together current happenings, yet most of these were amateurish [in writing] and quite inadequate as a continuation of Sima Qian's work. Because of this, Biao continued collecting previous histories and happenings of the past. To these he added other interesting tales and composed scores of biographies, thereby assessing the past histories and critiquing and correcting their merits and faults.Footnote 19
While this endeavor ostensibly led to him later composing Han shu, at this point his study and writings were focused on past histories, their merits, and their faults.Footnote 20 In a piece entitled “Lue lun” 略論, which could have been a preface to the biographies, he begins with the court scribes of Three Dynasties and the court histories of the vassal states of Zhou, and systematically weaves through the main historical records still extant from the different eras of ancient China, ending with Sima Qian's Shi ji. Of the latter he summarizes,
太史令司馬遷採《左氏》、《國語》，刪《世本》、《戰國策》，據楚、漢列國時事，上自黃帝，下訖獲麟 … … 遷之所記，從漢元至武則已絕，其功也。 … … 務欲以多聞廣載為功。
The Director of the Court Scribes Sima Qian selected from Zuoshi and Guo yu, pared down Shiben and Zhanguo ce, and relied on happenings in the various states in the Chu and Han period. His record goes back to the Yellow Emperor and down through the capture of the unicorn. … Of that which Qian recorded, his merit lies in the period from the beginning of Han up to [Emperor] Wu, where it breaks off. … He strove to achieve merit through its breadth of knowledge and extensive recordings.Footnote 21
Although Zuoshi chunqiu and the other works mentioned by Ban Biao were Sima Qian's principle sources for pre-Qin happenings, Shi ji contains citations from many more sources than these five books.Footnote 22 Furthermore, Ban Biao clearly states that Shi ji contains “extensive recordings,” yet for centuries scholars seem to have been content to leave the list of Sima Qian's citations largely as the few books of history Ban Biao enumerated. The only other sources commonly added are the Five Classics and perhaps one or two other works—despite Sima Qian's citing many other works by name. This thought has even persisted into Western Shi ji studies.Footnote 23 Nevertheless, it would be injudicious to assume that Sima Qian wrote the Qin–Han sections—or any section—without the aid of much textual material.
In the last century, however, seven scholars have pushed beyond these few works to compile more comprehensive bibliographies to ascertain the extent of Sima Qian's textual perusals and gleanings.Footnote 24 Table 1 shows the total works identified by each scholar.
Total number of works in each bibliography
While it is difficult to ascertain to what extent later bibliographic lists relied on earlier ones, it is clear that the lists grew in length with each iteration. What Table 1 does not show is that some works listed by Luo are not listed by Takigawa, some identified by Takigawa are not found on subsequent bibliographies, and so on. In fact, if a master list is generated, the total jumps to 142. A reason for this is the different criteria used to determine what works should be included, such as works that were quoted in Shi ji, works mentioned by Sima Qian but not quoted, or works mentioned in sources that he quoted. For example, in Li Si's memorial to Qin Shihuang advocating for banning certain books, he states that certain books should not be banned, namely, “books on medicine, divination, and horticulture” (醫藥卜筮種樹之書).Footnote 33 Jin, Lai, and Zheng listed these as a source for Shi ji; the others did not. Also, there is a large number of variant readings of the text, which stems from the fact that Sima Qian did not have a systematic way of citing the sources from which he culled. Some he merely mentioned by titleFootnote 34 or only listed the names of a few chapters or pieces,Footnote 35 others he mentioned the title and cited passages,Footnote 36 while yet others had passages copied (verbatim or in paraphrase) with no mention made of its source.Footnote 37 What's more, sometimes he even merged different sources, imbuing the passages with a different meaning.Footnote 38 All seemed to depend on the flow of the text. Because of this, many citations from other sources are not indicated clearly.Footnote 39
Furthermore, a general lack of fixed titles during Western Han, as well as many now-lost texts cited in the “Treatise on Literature” in Han shu, make it difficult at times to identify if a text is being cited, and if so what its title is—even in instances where the author or nature of a text may have been mentioned. This occasionally allows for more than one possible reading for a passage. As a result, all seven scholars have at least one unique listing. For example, Luo includes a work entitled Shijia yan 世家言 (Sayings of the Hereditary Houses); Takigawa lists Liehou gong ji 列侯功籍 (A Registry of Merits by Feudal Lords); Lu records Zhongshi wude zhi yun 終始五德之運 (Succession of the Five Elements); Jin Dejian lists Jian lun 劍論 (On Swords);Footnote 40 Lai Mingde argues that sections from the “Biography of Bo Yi” 伯夷列傳 come from a work call Bo Yi Shu Qi zhuan 伯夷叔齊傳 (The Biographies of Bo Yi and Shu Qi); Zheng Zhihong asserts that The Classic on Music 樂經 is still extant today (Zhang changes this to being lost); and Zhang lists the work Lizi 李子 (Master Li). The various, unique works listed in the bibliographies illustrate how contrasting their different readings of Shi ji are.
Differing scopes and variant readings are not uncommon issues in scholarship on ancient Chinese texts. However, within this range of “standard deviation” is room for increased accuracy. Although a comprehensive analysis of these biographical lists is beyond the scope of this present study, eleven source materials are listed below to highlight three principles that would lead to a more exhaustive and historically accurate bibliography of Shi ji sources. These principles are 1) listing actual “chapters” quoted and not book titles when dealing with sources found in books from the received corpus that were not compiled until after Sima Qian's time, 2) using unearthed texts to resolve contested interpretations and misreadings, and 3) including shorter, overlooked works which play an integral role in the composition of individual chapters. Two or three examples will be given for each principle. For each work discussed, a table is given showing which scholars included the work in their bibliographic lists and what title they gave it.
Works Formed after Sima Qian's Time
Texts were in a relatively high state of fluidity between the Qin and Eastern Han. This is due, in part, to the severe textual loss experienced during the Qin dynasty and its overthrow, and then the textual revival that occurred in Western Han. In addition, the widespread adoption of the clerical script over the official lesser seal script of the Qin meant a rewriting of old texts was necessary.Footnote 41 Due to these and other facets, many works had not yet been compiled into book form as they were known in later dynasties and even today.Footnote 42 In fact, it was not until late Western Han that bibliography, or the study and classification of books, became a formal pursuit of scholars. It was also around this time that paper saw its initial use.Footnote 43 These two factors contributed to books subsequently adopting a relatively stable and more condensed form during circulation.
For scholars from the Tang to the present, titles in the received corpus are most readily identifiable. However, since the purpose of these Shi ji source lists was to compile all known books/materials cited in Shi ji, listing titles that were not formed until later (while convenient and familiar) imbues them with anachronistic hues. One way to circumvent this is to list the single texts that Sima Qian cites, then annotate each piece's current location in the received corpus. The works Chunqiu fanlu 春秋繁露 (Luxuriant Gems of the Spring and Autumn) ascribed to Dong Zhongshu 董仲舒 (179–104 b.c.e.) and Da Dai liji 大戴禮記 (The Book of Rites by Dai the Elder) ascribed to Dai De 戴德 (fl. 43–33 b.c.e.) are cited here to illustrate this approach.
Chunqiu fanlu
Listings for Chunqiu fanlu in each bibliography
Chunqiu fanlu is a much-neglected work in Han studies, largely due to the belief that it was a forgery by, or at the very least the work of, Wang Su 王肅 (195–256) and his disciples. Despite this, Lai, Zheng, and Zhang listed Chunqiu fanlu as having been seen by Sima Qian. Lu listed instead an invented title based on Sima Qian's wording (see Table 2 above). Their various reasons for listing the work are surprisingly weak, however, which perhaps underscores the complexity of the text.
To understand their arguments, and to see how nuanced this issue is, a look at Sima Qian's discussion of Dong Zhongshu in Shi ji is in order. He does so in three main places. In Shi ji 14 he writes,
上大夫董仲舒推《春秋》義，頗著文焉。
Senior Counselor Dong Zhongshu expounded on the meaning of the Annals and in great measure wrote on it.Footnote 44
In Shi ji 121, Sima Qian writes in his biography on Dong Zhongshu,
今上即位，為江都相。以春秋災異之變推陰陽所以錯行 … … 行之一國，未嘗不得所欲。中廢為中大夫，居舍，著災異之記。 … … 終不治產業，以修學著書為事。故漢興至于五世之閒，唯董仲舒名為明於《春秋》，其傳公羊氏也。
After the present emperor assumed the throne, he appointed Dong Zhongshu to be the administrator to Jiangdu. By consulting ominous changes such as natural disasters and strange events recorded in the Annals, Zhongshu deduced the causes of disorderly interactions between yin and yang … When he carried out these techniques in this single state, he never failed to obtain the desired results. In the midst of his tenure as administrator to Jiangdu, he was dismissed from his post and appointed as a palace counselor. Residing at home, Zhongshu wrote records on disasters and anomalies … [When Zhongshu retired from his post and returned home to live,] he did not pay attention to [enhancing] his family's livelihood but instead occupied his time studying and writing books for the rest of his life. Thus from the time the Han arose to the reign of the fifth ruler, only Dong Zhongshu gained a reputation for elucidating the Annals; he taught [based on] the interpretations of Master Gongyang.Footnote 45
Then in Shi ji 130, Sima Qian records a rather lengthy discourse he had with Senior Counselor Hu Sui 壺遂, in which Sima Qian refers to teachings from a Dong-sheng 董生, or Master Dong. Given these teachings are on Confucius’ relationship to the Annals, the nature of the Annals, its position in the Six Classics, and its political application, it is generally accepted that Master Dong is indeed Dong Zhongshu.
These three passages combined provide a general framework for understanding Sima Qian's portrayal of Dong Zhongshu, namely that he was dedicated to in-depth research on the Annals and for a time explored yin-yang principles and natural disasters. More importantly, they show he wrote extensively on these topics, that these writings were available during Sima Qian's time, and that Sima Qian had read and was familiar with Dong's writings and teachings.
With the above passages as background, the rationale of each scholar's listing will be better understood. First, based on the section from Shi ji 121, Lai listed Chunqiu zaiyi ji. While this is likely not the title, listing it as thus is tenable, especially given Sima Qian's added anecdote of Dong being subject to Zhufu Yan's 主父偃 machinations and inadvertently denounced by his own disciple Lü Bushu 呂步舒.Footnote 46 Lai also cites two works by Dong listed in Ban Gu's “Treatise on Literature” as a basis for doing so and seemingly for listing Chunqiu fanlu:
〈漢志〉六藝略著錄有「公羊董仲舒治獄十六篇」，諸子略著錄有「董仲舒百二十三篇」，書名和《史記》所敘的「災異之記」不同，它是否即是今存的「春秋繁露」十七卷，尚無堅強的證據可以證明。
The Six Arts category in the Han shu treatise records “Gongyang Dong Zhongshu zhiyu in sixteen pian” and the Philosophical Masters category lists “Dong Zhongshu in 123 pian”; the names of these titles are different than Zaiyi zhi ji mentioned in Shi ji. There is still no firm proof to substantiate whether or not it is indeed the received Chunqiu fanlu in seventeen juan.Footnote 47
That is to say, Lai lists Chunqiu fanlu not because he thinks Sima Qian saw the work but because he holds there is no way to demonstrate that Chunqiu zaiyi ji and Chunqiu fanlu are not the same text.Footnote 48 Thus, he lists the work more out of uncertainty than based on proof from Shi ji.
Zheng cites the passage from Shi ji 14 as the basis for listing Chunqiu fanlu and then includes this gloss in Suoyin 索隱 as proof Sima Qian saw Chunqiu fanlu: “That is, he wrote Chunqiu fanlu” (作《春秋繁露》是[也]).Footnote 49 The problem is that Suoyin was composed in the Tang dynasty, much later than Wang Su's time. Therefore, Sima Zhen's 司馬貞 (679–732) claim holds little weight as supporting evidence.
Zhang bases his judgment on a line from Shi ji 130, wherein Sima Qian records Dong citing Confucius’ reason for compiling the Annals: “Confucius said, ‘My desiring to record it through empty theories would not be as profound and pronounced as viewing it through actual events.’” (子曰：我欲載之空言，不如見之行事深切著明也).Footnote 50 This is the first time one of the scholars linked the Shi ji text with an actual passage from the received Chunqiu fanlu. The “Yuxu” 俞序 chapter of Chunqiu fanlu cites Confucius as saying, “I rely on past events and apply my kingly mind to them because I consider that explaining things with abstract theories is not as good as the breadth and depth of past events for parsing and illuminating things.”Footnote 51 However, citing this passage is untenable for at least one, possibly two reasons. First, the parallel passage is a quotation of something that Confucius said, not Dong Zhongshu. Thus, there is no rationale for saying that this proves Sima Qian read the materials in our received Chunqiu fanlu. Second, according to Sarah Queen and John Major, “Yuxu” is actually a postface to a collection of Dong's writings put together by a “Mr. Yu” (俞先生),Footnote 52 meaning Mr. Yu would more than likely have lived after Sima Qian and would have had to cull this quotation from somewhere, which could then include Shi ji as a possibility.
Like Zheng, Lu cites Shi ji 14 and although he links Dong Zhongshu in 123 pian listed in Han shu with Chunqiu fanlu, he calls the work Dong Zhongshu Chunqiu yi, converting Sima Qian's wording into a title. This highlights the crux of the matter. In Shi ji 14, Sima Qian mentions Dong's writings on the Annals, but not by title. In Shi ji 121, he discusses Dong's writings on disasters and portents, but again not by title. Whereas in Ban Gu's “Treatise on Literature,” he lists a title Dong Zhongshu but gives no hint as to the content besides classifying the work under Confucian writers; he also lists Gongyang Dong Zhongshu zhiyu in the Chunqiu category but never lists a collection of Dong's writings on anomalies and disasters. Thus we have two genres of writing mentioned in Shi ji and two titles mentioned in the “Treatise on Literature” in Han shu, but they cannot be lined up. Thus, scholars are left either citing Chunqiu fanlu or inventing titles based on Sima Qian's wording in Shi ji.
It seems very unlikely, however, that our received Chunqiu fanlu existed at the time of Sima Qian. Not only is it not mentioned in Shi ji or Han shu, but according to Queen and Major, the title did not appear in texts until the Southern Liang 南梁 dynasty (502–557), some six hundred years after Dong's death.Footnote 53 It might be possible to push that date earlier, as the title also appears in Xijing zaji, which is commonly attributed to Ge Hong 葛洪 (283–343), but this is still much later than Western Han.Footnote 54 Furthermore, Chunqiu fanlu is riddled with a plethora of lacunae, interpolations, taboo terms, and incongruences. Even the table of contents reads like a mélange of different texts and genres.Footnote 55 In fact, it was this textual turbidity that led Jin to not list it—one of the few works in any of the bibliographic lists to be discussed but determined unlistable.Footnote 56 All of this seems to point away from the possibility that Chunqiu fanlu is Dong Zhongshu, as Lu has suggested. Ban Gu appears to confirm this:
仲舒所著，皆明經術之意。及上疏條教，凡百二十三篇。而說《春秋》事得失，《聞舉》、《玉杯》、《蕃露》、《清明》、《竹林》之屬，復數十篇，十餘萬言，皆傳於後世。掇其切當世施朝廷者著於篇。
All of Zhongshu's compositions elucidated the meaning of the classical arts. As for his memorials submitted to the throne and items of instruction, they totaled [123] pian. His expositions of the success and failure of affairs in the Spring and Autumn, such writings as “Heard and Promoted,” “Jade Cup,” “Luxuriant Gems,” “Pure Brightness,” and “Bamboo Grove” came to an additional several tens of pian, amounting to more than a hundred thousand characters. All were transmitted to later generations. I selected portions which were pertinent at the time to his service in the court and thus wrote this chapter [of Han shu].Footnote 57
That is to say, memorials to the throne and other writings totaled 123 pian, which is the length of Dong Zhongshu; however, expositions on the Annals numbered “an additional several tens of pian.” It is curious that none of the scholars discussed this passage because Ban Gu lists here actual works penned by Dong Zhongshu elucidating the Annals. Of the pieces listed, “Luxuriant Gems,” “Jade Cup,” and “Bamboo Grove” are the first three chapters of the received Chunqiu fanlu.Footnote 58
Turning to Shi ji, surpassing the single tangential correlation that Zhang pointed out between Chunqiu fanlu and Shi ji 130, there are actually seven other passages in Shi ji 130, as well as another instance in Shi ji 14, that match up with sections from seven chapters in Chunqiu fanlu.Footnote 59 Of these, two correlate with content in “Jade Cup” and one with “King Zhuang of Chu” 楚莊王.Footnote 60 According to Su Yu 蘇輿 (1874–1914), “King Zhuang of Chu” was originally entitled “Luxuriant Gems”; however, since it was the first chapter in the collection, its title was appropriated as the title of the entire collection and the first three characters of the piece were used instead as its title.Footnote 61 Thus the parallel content between passages in Shi ji 130 and the first few chapters of Chunqiu fanlu, when connected to titles listed in Ban Gu's biography, gives credence to the claim that Sima Qian saw and used Dong Zhongshu's writings on the Annals. Listing the individual names of these chapters is preferable to Lu's title, as they are as close to reflecting what Sima Qian saw as present textual indications allow.
Da Dai liji
The case of Da Dai liji is similar to Chunqiu fanlu, only much more straightforward. In the summary remarks in Shi ji 1, Sima Qian writes, “I observed that the Annals and Guo yu explicate clearly ‘Wu di de’ and ‘Di xi xing.’” (予觀《春秋》、《國語》，其發明〈五帝德〉、〈帝繫姓〉章矣).Footnote 62 Then in the summary remarks in Shi ji 2, Sima Qian states, “Confucius followed the Xia calendar; men of learning, it seems, have [thus] widely circulated ‘Xiao xiao zheng’” (孔子正夏時，學者多傳〈夏小正〉云).Footnote 63 Sima Zhen glosses all three texts as being from Da Dai liji.
Listings for Da Dai liji in each bibliography
Zheng and Zhang were the only scholars who listed these several texts under this appellation, possibly taking their lead from Sima Zhen. Doing so, however, is clearly anachronistic, since “Da Dai” is a sobriquet of Dai De who lived during the reign of Emperor Xuan of Han 漢宣帝 (r. 74–48 b.c.e.), and Emperor Xuan's ascension to the throne was not until thirteen years after Sima Qian's assumed death. Ban Gu's “Treatise on Literature” states,
漢興，魯高堂生傳《士禮》十七篇。訖孝宣世，后倉最明。戴德、戴聖、慶普皆其弟子，三家立於學官。
When the Han arose, Gao Tangsheng from Lu transmitted Shili in seventeen pian. Up to the reign of Emperor Xuan the Filial, Hou Can was the most renowned. Dai De, Dai Sheng, and Qing Pu were all his students. The three of them each had their learning included in the imperial schools.Footnote 64
The “Treatise on Literature” does not record Da Dai liji, which has led many to the conclusion it was not compiled until later. Regardless of when it was compiled, we can safely conclude that when Sima Qian wrote of “Wudi de,” “Di xi xing,” and “Xia xiao zheng” these texts were not yet a part of its compilation.
What form these texts took before or during Sima Qian's time is not well known. Fang Xiangdong 方向東 believes that they were part of Guwen ji 古文記:
今之《禮記》和《大戴禮記》不是大、小戴所輯，而是他們用來教授生員的資料，源自《古文記》一百三十篇的內容，編輯成書約在東漢章帝時期。《大戴禮記》成書應在東漢鄭玄之前。
The present Book of Rites and Da Dai liji were not compiled by the two Dais. Rather they were materials the two Dais used in teaching their students. The material came from the contents of the 130 pian of Guwen ji. They were compiled into book form around the reign of Emperor Zhang in Eastern Han. Da Dai liji was likely compiled in the Eastern Han before Zheng Xuan's time.”Footnote 65
Not much is known about Guwen ji, at least not enough to list it as the main source for these three texts. What is known is that Sima Qian was quoting them as individual pian. And, as Jeffry Riegel observes,
It appears, in fact, that as late as the White Tiger Hall debates of 79 the critical editing which separated out the 49 p'ien of the present Li chi had not yet taken place, since sections from that text, from Ta tai li chi and others not occurring now in either of these, are simply quoted by the title of their p'ien and are not identified as comping from one or the other collection.Footnote 66
Given their uncertain relationship with Guwen ji and the misleading nature of citing Da Dai liji as the source, listing the three texts by their titles avoids all issues of textual pedigree and seems to reflect better Sima Qian's interaction with them.
On a related note, Lai Mingde cites an annotation in Suoyin that states,
〈五帝德〉、〈帝繫姓〉皆《大戴禮》及《孔子家語》篇名。以二者皆非正經，故漢時儒者以為非聖人之言，故多不傳學也。
“Wudi de” and “Di xi xing” are chapter names in Da Dai li and Kongzi jiayu. Since both of these works are not true classics, many classicists in Han times did not view them as containing words by the Sage. As such most did not transmit or teach them.Footnote 67
Lai then uses this quotation to assert that
由於「儒者或不傳」，史遷可能沒有閱讀到這兩篇文獻的原貌，但是他卻能夠從《春秋》、《國語》中領悟出其旨趣來。
Because “some classicists did not transmit them,” it is possible that Sima Qian did not read these two pieces in their original form (i.e., he did not see them); however, he could understand their gist from reading the Annals and Guo yu.Footnote 68
Dorothee Schaab-Hanke has outlined, however, that Sima Zhen likely was purposeful in his dismissal of these sources, given that he wrote “Sanhuang benji” 三皇本紀 to correct a deficiency and error that he saw in Sima Qian's work—that is, he took issue with the fact that Shi ji started with the Five Emperors and not with the Three Sovereigns.Footnote 69
Rather than reading this gloss as a possible indication that Sima Qian did not read these two texts (yet Lai lists them as sources anyway), Han Zhaoqi 韓兆琦 has outlined a more viable interpretation, namely that Zuo zhuan and Guo yu contain some parts that explain clearly points made in “Wudi de” and “Di xi xing.”Footnote 70 Had Sima Qian not read these two pieces, it would be hard to ascertain how the content of Zuo zhuan (Zuoshi chunqiu) and Guo yu actually explicated the arguments in “Wudi de” and “Di xi xing.”
Even more fundamental to the issue of Sima Qian's sources is the fact that a difference exists between textual transmission during scholarly instruction to one's disciples and the contents of archived records. That is to say, what texts were transmitted or not from teacher to pupil during Western Han does not have a direct bearing on Sima Qian's access to archival records, especially at a time when all the works under heaven were collected in the Grand Scribe's repositories. Sima Qian is indeed working on a fundamentally different plane than the average teacher-scholar in Western Han. It is for this reason that Shi ji contains occasional references to works being widely circulated or commonly transmitted, or the opposite. Sima Qian seems consciously aware of his unique access to a broader range of texts than others enjoyed.
Clarification from Unearthed Texts
In many summary remarks at the end of individual chapters, Sima Qian discusses scholarly issues and contemporary theories in an attempt to settle disputes and correct errors.Footnote 71 Despite this clear effort to imbue certain chapters with an authoritative scholarly tone, other passages in Shi ji are equivocal, contradictory, or otherwise unclear. This, combined with texts lost over the centuries, has led to misreadings and doubts by later scholars. Nevertheless, similar to the impact the discovery of oracle bones had in fleshing out Shang history and correcting related scholarly misinterpretations, many discovered texts over the last century have shed light on answers for issues in Shi ji scholarship.
After the clarification provided by unearthed texts, it is interesting to note that in certain issues, obfuscation stemmed more from scholarly doubt and misreadings rather than from the text itself. Shi ji often gives clues to a correct reading in these instances, especially when coupled with relevant passages from Ban Gu's “Treatise on Literature” and other later bibliographic compilations. This section will look at the instances of Sun Bin bingfa 孫臏兵法 (Sun Bin's Art of War) and “Kongshi shu” 孔氏書 (Mr. Kong's Writings).
Sun Bin bingfa
Sima Qian bundled into one chapter the biographies of Sun Wu 孫武, Sun Bin 孫臏, and Wu Qi 吳起 (Shi ji 65), all of whom exhibited exceptional abilities in military strategy. Therein he explicitly states that Sun Wu wrote Sunzi bingfa 孫子兵法 and Sun Bin wrote Sun Bin bingfa. Despite this, later scholarly doubts have swirled in a dizzying array of possible permutations, including whether Sun Wu was a real person, whether Sun Wu and Sun Bin might be the same person, and whether there was even an actual work called Sun Bin bingfa.Footnote 72 This last doubt is reflected in the fact that several bibliographies do not list Sun Bin bingfa as a source for Shi ji.
That the true nature of Sun Wu and Sun Bin, as well as the works attributed to them, has long been a vexed issue is due in part to Sun Bin's name. In a word, “Bin” does not seem like a real name. Having been deluded by Pang Juan 龐涓, he was punished by having his feet cut off but was identified by bin (a punishment where one's kneecaps were removed).Footnote 73 His using bin as a name opened the door for much interpretive leeway, leading some later scholars (such as Qian Mu 錢穆, 1895–1990) to assume that Sun Wu was, in fact, Sun Bin's name before his mutilation.
Listings for Sunzi bingfa and Sun Bin bingfa in each bibliography
Sima Qian's word choice added to the lack of clarity. In their respective biographies, uniform and distinct appellations were not used. Sun Wu's biography simply states, “Sunzi Wu was from Qi” (孫子武者，齊人也).Footnote 74 The rest of Sun Wu's biography identifies him as “Sunzi” a total of ten times. In Sun Bin's biography it states, “More than one hundred years after Sun Wu died there was Sun Bin” (孫武既死，後百餘歲有孫臏); then he is identified with the same “Sunzi” eight times. Moreover, the same appellation is used for both persons in the summary remarks,
世俗所稱師旅，皆道《孫子》十三篇、《吳起兵法》。世多有，故弗論，論其行事所施設者。語曰：「能行之者未必能言，能言之者未必能行。」孫子籌策龐涓明矣，然不能蚤救患於被刑。吳起說武侯以形勢不如德，然行之於楚，以刻暴少恩亡其軀。悲夫！
The common custom in talking of military matters is to speak of Sunzi in thirteen pian and Wu Qi bingfa. These are widely available in society, so I have not discussed these here. Instead, I have discussed how they went about doing things. It is said, “Those who can do something cannot necessarily explain it, and those who can explain something cannot necessarily do it.” Sunzi's schemes against Pang Juan were brilliant, and yet he could not save himself from the earlier disaster of being subject to corporal punishment. Wu Qi admonished Marquis Wu [of Wei] that difficult topography is not as important as virtue [in ruling a kingdom], and yet when he went to Chu, his cruelty and harshness cost him his life. What a pity!”Footnote 75
Then, in Shi ji 130 Sima Qian writes, “Sunzi had his feet cut off and then discussed military tactics” (孫子臏腳，而論兵法).Footnote 76 Sima Qian's wording led Jin to conclude that Sun Bin wrote Sunzi bingfa, arguing that Sima Qian lists two authors for Sunzi bingfa—Sun Wu in Shi ji 65 and Sun Bin in Shi ji 130.Footnote 77 The overuse of the honorary title Sunzi and a lack of clear titles for their works then seems to have given rise to these misunderstandings and misreadings.
Despite the fact that both Sun Wu and Sun Bin are referred to as Sunzi in the biography, we cannot conclude that Sima Qian made one of their biographies out of whole cloth; rather, he had a textual basis for including both persons in the chapter. Indications that point in this direction include the following: 1) Sun Wu's biography is brief while Sun Bin (of a later date) enjoys a more detailed treatment. This is in line with how Shi ji functions as a whole, with the increase of available historical records as events approach the Qin and Han dynasties. 2) The mention of Sun Bin and his bingfa in Shi ji 130 is part of seven texts Sima Qian lists as resulting from personal misfortune.Footnote 78 The other six works listed were clearly extant in the Western Han, leaving little room to doubt Sun Bin bingfa was as well. 3) Ban Gu records both works in his “Treatise on Literature,”Footnote 79 and given Yan Shigu's 顏師古 (581–645) annotations, it is likely that Sun Bin's work was still extant in the Tang dynasty. 4) Zhanguo ce also mentions Sun Bin, in conjunction with Wu Qi.Footnote 80
Of even stronger persuasion are unearthed documents that substantiate Sima Qian's biography. Specifically, a cache of bamboo texts was discovered in a Han tomb at Yinqueshan 銀雀山, Shandong 山東 in 1972 containing the texts Sun Bin bingfa and Sunzi bingfa. Sun Bin bingfa contained thirteen pian, totaling over ten thousand characters. While this number varies from what Shi ji and Han shu record, the editorial group believes it is because the copy is incomplete.Footnote 81 Also in the cache were copies of other military strategy texts, including Liudao 六韜 and Weiliaozi 尉繚子. That the tomb's occupant had a fondness for military strategy has finally put to rest the misunderstandings regarding Sun Wu's and Sun Bin's biographies. As both works are now known to be separate compositions, they should be listed in separate entries, as Lu, Zheng, and Zhang have done.
On a related note, Chen Zhi 陳直 holds that Shi ji 65 may contain some of the text that is now lost, citing the following passage from Sun Bin's biography: “Tian Ji desired to move troops to Zhao. Sunzi [Sun Bin] said, ‘To defuse an argument one does not employ fists and to stop a fight one does not bind or restrain’” (田忌欲引兵之趙，孫子曰：夫解雜亂紛糾者不控捲，救鬬者不搏撠).Footnote 82 He notes that
「捲」為「拳」字異文，「撠」為「戟」字繁文，兩字均從手者，是以字從義，蓋太史公據戰國古文舊簡直書者。
Juan 捲 is a variant of quan 拳 and ji 撠 is an elaborate form of ji 戟. Both characters have the hand radical in them. Thus the characters indicate the meaning. It is likely that the Grand Scribe directly copied ancient characters from old Warring States bamboo strips.Footnote 83
The passage does read like content in the first four chapters of Sun Bin bingfa, where Sun Bin answers questions posed by rulers. However, no corresponding passage exists in the unearthed text. Thus, it could very well represent a lost portion of the rediscovered text.
“Kongshi shu”
In Sima Qian's conclusion to Confucius’ biography in Shi ji 47 he states, “In reading Mr. Kong's writings, I have visualized what kind of person he was in my mind's eye” (余讀孔氏書，
想見其為人).Footnote 84 Despite this reference, none of the seven lists include “Kongshi shu,” with only Luo giving it the following annotative reference: “‘Kongshi shu’ is not the name of a book; whether it refers to the Analects or other books cannot be verified.” (孔氏書，非書之名，指《論語》抑他書，不可考).Footnote 85 Despite Luo's uncertainty, Shi ji 47 contains much-embedded information that aids in deciphering what Sima Qian is referring to with this generalized term.
Given the amount of material in Confucius’ biography that does not correspond with the Analects, it is highly unlikely that “Kongshi shu” is solely referring to the Analects. The parts of Shi ji 47 that do not come from the Analects would have had to come from other sources. These would need to have content sufficient that, when coupled with content from the Analects, could induce the visualization Sima Qian experienced. Thus, they would need to be similar in nature, containing recordings of Confucius’ life, speech, actions, and so forth. That such records existed is demonstrable, as they were in ready abundance in the Warring States and Qin–Han periods.Footnote 86 For example, in the Mencius we see many citations of Confucius’ teachings and life experience that are not found in the Analects. The “Xici” 繫辭 and “Wenyan” 文言 sections of Yizhuan 易傳 contain frequent “The Master said” (zi yue 子曰) citations.Footnote 87 Moreover, during Western Han many other Confucius-themed works were in circulation, not the least of which are the documents reportedly found in Confucius’ old residenceFootnote 88 and newly discovered texts such as “Kongzi shi lun” 孔子詩論,Footnote 89 Rujiazhe yan 儒家者言,Footnote 90 and the “Zhidao” 智道 chapter of the Qi Analects.Footnote 91 Shi ji also records a register of Confucius’ disciples in circulation at the time.Footnote 92
Listings for “Kongshi shu” in each bibliography
The sources of these works can be clustered into three general categories. First, documents and writings that were passed down,Footnote 93 either by his disciples (as was the case with pre-Analects recordings) or by his descendants (as what ostensibly seems to be the case in the work Kongzi jiayu recorded by Ban GuFootnote 94). Sima Qian's discussion at the end of Shi ji 47 shows how Confucius’ tomb and residence gradually developed into a center of Confucian thought and teaching, which would also lend itself to serving as an archive for this purpose. Second, the unearthed documents in ancient script mentioned above. Third, writings that fall into the category of attribution (tuoming 託名) literature. The practice of attributing things to a higher historical authority was a widespread practice in the Warring States and Qin–Han periods. Confucius’ fame meant that his name was commonly utilized in this fashion.
Given his access to imperial repositories, Sima Qian would likely have had an abundance of material on Confucius available to him. In fact, Sima Qian's culling anecdotes and happenings regarding Confucius’ life from available writings is not unlike what Liu Xiang 劉向 (77–6 b.c.e.) did when commissioned to reorganize the imperial archives that had fallen into disarray. Liu's resultant Shuoyuan 說苑 also contains many anecdotes on Confucius’ life, further highlighting the amount of textual information that existed more than a half-century after Sima Qian died. Zuo Songchao 左松超 writes of Liu Xiang's work: “What is recorded in Shuoyuan came from other texts. Some of the recordings are from sources even earlier than Hanshi waizhuan or Lüshi chunqiu” (《說苑》所載，自有所本，其中部分資料來源較《韓詩外傳》、《呂氏春秋》等書更早).Footnote 95 Presumably, Sima Qian would have seen much of the same material.
A look at three examples of anecdotes from Shi ji 47 provides context on why “Kongshi shu” should be included in the bibliographic list. The first two anecdotes—Confucius discussing the two worthies from Jin and his comments about the attitude of a Gentleman—are best viewed through a study by Ning Zhenjiang 寧鎮疆 on the similarities of Rujiazhe yan and Kongzi jiayu, Shuoyuan, and Shi ji 47. Germane to the discussion here is his analysis of two phrases: “The Master said, ‘Duzhu and Zeming are worthy … from Jin’” (子曰：犢主澤鳴晉國之賢□) and “It was heard that a Gentleman is solemn about injury” (聞君子重傷□). Ning writes,
簡文「子曰：『犢主澤鳴，晉國之賢□ … … 』」，《說苑》為「孔子曰：『夫澤鳴犢犨，晉國之賢大夫也』」。「犨」與「主」音近可通，因此「犢主」實即「犢犨」，這樣看來，簡文與《說苑》基本相同。 … … 《史記‧孔子世家》作「竇鳴犢、舜華」，《家語》為「竇犨鳴犢、舜華」，二書明顯為同一系統。 … … 簡文「聞君子重傷□」，《說苑》作「丘聞之，君子重傷其類者也」。簡文前後有殘損，然就所存看，二者差別並不大。《家語》、《史記》對應部分作「何則，君子違傷其類者也」，《家語》、《史記》「何則」已不具備簡文、《說苑》「聽聞」（「聞」）的原始表述。《家語》「重」作「違」，《史記》作「諱」，「違」當是「諱」之訛，二者同出一源也很明顯的。
Text on the bamboo slips [from Rujiazhe yan records], “The Master said, ‘Duzhu and Zeming are worthy … from Jin,’” whereas Shuoyuan has, “Confucius said, ‘Zeming and Duchou are worthy officials from Jin.’ “Chou” is close in pronunciation to and can be used interchangeably with “zhu”; thus, “Duzhu” is actually “Duchou.” From this perspective, the bamboo text is basically the same as Shuoyuan … The “Hereditary House of Confucius” in Shi ji renders [their names] “Doumingdu and Shunhua.” Jiayu records, “Douchou Mingdu and Shunhua”; [so] the two works clearly belong to the same textual vein … In the bamboo text [it says], “It was heard that a Gentleman is solemn about injury,” whereas Shuoyuan renders it “[I] Qiu heard it said, a Gentleman is solemn about injury to those who are like him.” The bamboo slips are missing text before and after. However, from what is still extant, it is clear there is not much difference between the two. Corresponding parts in Jiayu and Shi ji are rendered “Why is this? A Gentleman avoids injury to those who are like him.” The “why is this” in Jiayu and Shi ji is markedly different the original “I've heard it said” expression in the bamboo text and Shuoyuan. In Jiayu, “solemn” (zhong) is rendered “to go against” (wei); in Shi ji it is “to avoid” (hui). Wei is likely a corruption of hui. That the two works [i.e., Shi ji and Jiayu] have the same textual source is also clear.Footnote 96
Scholars have determined that Han tomb no. 40 at Bajiaolang 八角廊 was sealed in Wufeng 五鳳 3 of Emperor Xuan (i.e., 55 b.c.e.), which means that it was only some thirty years after Sima Qian's death. Furthermore, the passages above are found in Shi ji 47 but not in the Analects. These anecdotes from Confucius’ life then can reasonably be assumed to have been recorded by Sima Qian via a different text.
A second example comes from a cache of texts discovered in 1977 in Han tombs at Shuanggudui 雙古堆, Fuyang 阜陽 county. Besides copies of Cangjie pian 倉頡篇, Shi jing 詩經, and Zhouyi 周易, there were also some damaged wooden and bamboo slips containing writings similar in style to the Analects.Footnote 97 The tomb belonged to Xiahou Ying's 夏侯嬰 son Xiahou Zao 夏侯灶. Xiahou Ying died in 172 b.c.e., meaning his son lived before Sima Qian's time. One damaged bamboo slip has on it “Confucius came to the river and sighed.” (孔子臨河而歎). This again correlates with passages in Shi ji 47, Jiayu, and Shuo yuan, but not the Analects.Footnote 98
Given Shuo yuan was produced after Sima Qian and scholars’ general belief that Jiayu was a product of Wang Su, these unearthed documents show that portions of Shi ji 47 that do not match the Analects do match other records and these records existed before or around Sima Qian's time. Furthermore, given the textual alignment of Shi ji and Jiayu, as well as Rujiazhe yan with Shuoyuan, they may show different textual traditions in Western Han regarding Confucius’ life.Footnote 99 This gives reason to believe that such records comprised or were included in Sima Qian's generalized term “Kongshi shu.” Since it can be demonstrated there existed actual records separate from the Analects, this source (or these sources) should be added to the bibliographic list of Shi ji sources using Sima Qian's term: “Kongshi shu.”
Overlooked Sources
Scholarly effort spanning nearly one hundred years has still not yielded a complete and comprehensive list of source texts for Shi ji. Granted, due to the number of texts that have been lost over the last two millennia, it is unlikely that a finalized and comprehensive list will ever be compiled. Despite this, there is still much more that can be mined from Shi ji based on the text itself.
Historically there has existed a strong bias against non-canonical and “apocryphal” books; in these lists there is also a bias against materials not in book form or of book length. This aspect is manifest in Jin, Zheng, and Zhang specifically focusing on “books seen by Sima Qian”; and while Luo, Takigawa, and Lu all used cai 材 (materials) as their scope of inclusion, their lists only contain titles of books too. Lai alone lists “books and other documents” (shuji he wenxian 書籍和文獻). Yet, texts falling under the category “other documents” are numerous in Shi ji. These include the lyrics of ballads, poems, transcriptions of stele inscriptions (especially in Shi ji 6), court records, imperial edicts, memorials to the throne, and epistolary writings. These play a unique role in the construction of individual chapters. Examples of overlooked works include “Tiying shang Han Wendi shu” 緹縈上漢文帝書 (Tiying's memorial to Emperor Wen of Han), Sima Rangju bingfa 司馬穰苴兵法 (Sima Rangju's Art of War), and “Zigong shui Tian Chang” 子貢說田常 (Zigong Persuades Tian Chang).
“Tiying shang Han Wendi shu”
Lai Mingde is the only one to include “Tiying shang Han Wendi shu” in his bibliographic list. He gives no explanation for his reasoning; rather he simply quotes a passage from the “Biographies of Bian Que and Cang-gong” 扁鵲倉公列傳:
文帝四年中，人上書言（淳于）意，以刑罪當傳西之長安。意有五女，隨而泣。意怒，罵曰：「生子不生男，緩急無可使者！」於是少女緹縈傷父之言，乃隨父西。上書曰：「妾父為吏，齊中稱其廉平，今坐法當刑。妾切痛死者不可復生，而刑者不可復續，雖欲改過自新，其道莫由，終不可得。妾願入身為官婢，以贖父刑罪，使得改行自新也。」書聞，上悲其意，此歲中亦除肉刑法。
In the fourth year of Emperor Wen, a memorial was submitted denouncing (Chunyu) Yi [i.e., Cang-gong]. He was to be transferred to Chang'an in the west for punishment. Yi had five daughters, who wept [at the news]. Enraged, Yi shouted, “I have fathered no sons and thus have no one to send in times of crisis!” His youngest daughter, Tiying, was distraught over his words, and thus followed her father west. She submitted a memorial, saying, “Your maidservant's father was a low-level official who was known throughout Qi for his integrity. Now he is sentenced to corporal punishment. Your maidservant is pained most poignantly that the dead come not back to life, and that those mutilated cannot live as before. Despite their desires to start anew, no path lies before them and they are forever damned. Your maidservant desires to offer herself as a slave of the court to atone for her father's crimes, thus enabling him to the opportunity to change his ways.” When the memorial had been read, the Emperor was moved to compassion by its meaning. That year, he abolished corporal punishment.Footnote 100
Listings for “Tiying shang Han Wendi shu” in each bibliography
Such a contribution to Western Han society is exemplary in many facets and should not be overlooked nor left unlisted. The lexical markers shangshu yue 上書曰 and shu wen 書聞 seem to indicate that Sima Qian recorded verbatim the memorial of this girl speaking truth to power, underscoring the importance this act had in effecting the abolishment of mutilating punishments in the capital and perhaps throughout the empire.Footnote 101
It is worth noting that a comparable passage in Han shu shows another important aspect of this affair. The “Treatise on Punishment and Law” 刑法志 records,
書奏天子，天子憐悲其意，遂下令曰：「制詔御史： … … 其除肉刑，有以易之；及令罪人各以輕重，不亡逃，有年而免。具為令。」
The memorial was submitted to the Son of Heaven, who was moved to compassion by its meaning. He then issued an edict to the imperial censor: … corporal punishment should be abolished and something instituted in its stead. [The emperor] also commands that in finding people guilty, it should be done according to the gravity of their crimes. If they do not flee the punishment, when their sentence is up they will be stripped of their office and be made a commoner. These are to be drafted formally into law.Footnote 102
Mutilating punishments were abolished and lashings selected as a replacement. The result, as Ban Gu records, was, “Outside the court, the law was known for its light punishments. Inside the courts, however, criminals still died. Those who would have had their right toes cut off were executed instead. Those who would have had their left toes removed were now given five hundred lashings. Those who would have had their nose removed were given three hundred. Most died.” (外有輕刑之名，內實殺人。斬右止者又當死。斬左止者笞五百，當劓者笞三百，率多死).Footnote 103
Thus, while the memorial did bring about reform, the imperial censor and other officials who held the power to decide on punishments were reluctant to relinquish control that comes from such measures. The court merely switched mutilating punishments for what mostly ended up being capital punishment. Regardless, Lai Mingde is right in including it in the list of source texts. Not only does it have great historical value, it is also an important element in Sima Qian's crafting a benevolent and virtuous image of Emperor Wen.
Sima Rangju bingfa
Luo and Takigawa believed that Sima Qian read Sima fa 司馬法, while Jin, Lai, Zheng, and Zhang listed the work as Sima bingfa 司馬兵法. Lu Nanqiao instead lists Sima Rangju bingfa. These discrepancies stem from the mention of these two books in “The Biography of Sima Rangju” 司馬穰苴列傳 (Shi ji 64), and the fact that in Shi ji 130 Sima Qian used the title Sima fa instead of Sima bingfa.
Listings for Sima bingfa and Sima Rangju bingfa in each bibliography
That Sima Qian both saw and read the work Sima bingfa is clear from his summary remarks in Shi ji 64:
余讀《司馬兵法》，閎廓深遠，雖三代征伐，未能竟其義，如其文也，亦少褒矣。若夫穰苴，區區為小國行師，何暇及《司馬兵法》之揖讓乎？世既多《司馬兵法》，以故不論，著穰苴之列傳焉。
I read Sima bingfa. It was vast and profound. Even all of the punitive expeditions from the Three Dynasties together could not exhaust its implications. Its writing style, however, is slightly excessive. As for Rangju, he merely led the army of a small state. When would he ever have an opportunity [to use] the courtesies in Sima bingfa? Since Sima bingfa is in wide circulation, I have not discussed it here. Instead, I have written the biography of Rangju.”Footnote 104
The title of the work, however, may actually have been Sima fa. In addition to its mention in Shi ji 130, a passage from it is directly quoted in a memorial by Zhufu Yan 主父偃 in Shi ji 112. The quotation is prefaced with “Sima fa states” 司馬法曰.Footnote 105 Fayan 法言 also cites the work as Sima fa.Footnote 106 Given this textual evidence, Sima (bing)fa should be listed as a source for Shi ji.
What is less clear is whether Sima Rangju bingfa should be listed. The key to understanding the nature of this work lies in Sima Qian's view of Sima Rangju, a general from Qi 齊 known as Tian Rangju 田穰苴. In the quotation above, Sima Qian's query “When he would ever have an opportunity [to use] the courtesies in Sima bingfa?” seems to invite a pejorative reading. However, its meaning is just the opposite. The term hexia 何暇 is used four times in Shi ji. Twice it is used in its standard meaning of “unable to give attention to something due to other pressing matters.”Footnote 107 There also is an extended meaning indicating “an opportunity or occasion to do something (not necessarily urgent) that is outside one's scope of attention or duty.” An example of this is found in Shi ji 56, which records that Liu Bang 劉邦, still King of Han 漢, is upset that Wei Wuzhi 魏無知 recommended Chen Ping 陳平 to him, given rumors that Chen Ping was less moral than would be expected. Wei's response was, “Behaving like Wei Sheng or Xiao Ji [historical figures known for their moral behavior] will not help in winning battles. [Given that we are at war] when would my Lord have occasion to make use of such?” (今有尾生、孝己之行而無益處於勝負之數，陛下何暇用之乎？)Footnote 108 Sima Qian's comment that Tian Rangju would not be able to make use of the courtesies in the work due to the small scope of his environs parallels this. Tian Rangju has the ability to use them, but not the circumstances. Furthermore, Sima Qian records Yan Ying's praise that Tian Rangju is “a man whose refined manner [could] win the loyalty of his hosts and whose martial spirit [could] awe his adversaries.” (其人文能附眾，武能威敵).Footnote 109 Thus, his portrayal is of one who is refined, martial, awe-inspiring, and talented in elucidating the ancient compendium of military strategy. In other words, Sima Qian found the embodiment of military strategies and techniques in him.
According to Sima Qian, all the expeditions of the Three Dynasties could not exhaust the content of the ancient compendium of military strategies, tactics, and protocol known as Sima (bing)fa.Footnote 110 Since Tian Rangju could master (but not apply) them all in his own limited sphere, we can only conclude that his skill was transcendent. Due to Rangju's military acumen, “King Wei of Qi commissioned officials to research and expound on the ancient Sima bingfa and to add to it Rangju[’s tactics]. Because of this, it was called Sima Rangju bingfa.” (齊威王使大夫追論古者《司馬兵法》而附穰苴於其中，因號曰《司馬穰苴兵法》).Footnote 111
Of the seven scholars, Lu was the only one who listed Sima Rangju bingfa; however, he did not list Sima bingfa. His annotations show that he combined the sentence “I read Sima bingfa” with King Wei of Qi's request, and thus listed Sima Rangju bingfa. However, it is not clear if he viewed them as the same work or as two versions of the same text. Based on Sima Qian's description, it would be difficult to argue they are the same work. There could, however, be an argument made that Sima Rangju bingfa is an “updated and expanded” edition of Sima bingfa, with the primary premise being text from Sima bingfa was used in Sima Rangju bingfa. Nevertheless, the phrase zhuilun 追論, rendered here as “to research and expound on,” is rather non-descriptive in terms of understanding the extent of textual reworking that occurred. Since we do not have both works at our disposal to compare, it is impossible to determine how much the latter resembles the former. Given the change in title and in content, this study views them as two separate works and argues they should be listed as such.
“Zigong Persuades Tian Chang”
This final example is one that is not mentioned in any of the source lists.Footnote 112 In Shi ji 67, Sima Qian lists seventy-seven disciples of Confucius. Biographical details, however, are only available for the first half of the disciples and are quite limited in length. The most noticeable exception to this is Zigong, whose lengthy biography is comprised primarily of an account of his persuading Tian Chang in Qi to not attack Lu 魯 but rather Wu 吳 instead.
This vignette is found (with some varying details) in several other works, including Kongzi jiayu, Yuejue shu 越絕書, and Wu Yue Chunqiu 吳越春秋. As John Lagerwey has noted, Wu Yue chunqiu from Eastern Han is largely a reproduction of content from Zuo zhuan, Guo yu, and Shi ji.Footnote 113 As such it cannot be considered a source text. Also, as mentioned earlier, Kongzi jiayu is not widely accepted as an authentic source of historical material (although unearthed documents are substantiating an increasing portion of its contents) or even a Western Han production. However, research on Yuejue shu, despite its being compiled in Eastern Han, seems to point to a proper place in the genre of pre-Qin and Han accretion literature.Footnote 114
The passage in Shi ji 67 correlates with the “Chen Cheng Heng” 陳成恆 chapter of Yuejue shu. The ostensible compilers Yuan Kang 袁康 and Wu Ping 吳平 divided up content in Yuejue shu into inner and outer chapters and also into canonical texts (jing) and traditions (zhuan) based on the source's reliability. “Chen Cheng Heng” falls in the inner/traditions quadrant. As Olivia Milburn points out, this means it is not only “one of the core texts of the Yuejue shu” but also “seems to be the product of just one hand … almost certainly dating to the Qin dynasty.”Footnote 115 She states that philological evidence points to the chapter being a product of the Qin dynasty, as it uses the term qianshou 黔首 (the common people) and characters that were taboo in Western Han.
A relatively high degree of similarity exists between the accounts in Yuejue shu and Shi ji. This has led Axel Schuessler and Michael Lowe to state “‘Ch'en Ch'eng heng’ in chüan no. 7, is identical, almost verbatim, with text that appears in Shih chi 67” and to further speculate that “in view of its identity with text in the Shih chi, it is likely that ‘Ch'en Ch'eng heng’ did not derive from an original Yüeh chüeh shu.”Footnote 116 A close reading reveals that although some recorded speech is the same in the two passages, the passages as a whole are far from “verbatim.” The Shi ji version exhibits the telltale signs of editing common in other of Sima Qian's rewritings—the passage is shorter, the flow of the narrative more to the point, and language more economical. In short, it is highly more likely that “Chen Cheng Heng” (or an earlier version) was the source text used for the Shi ji passage rather than the opposite.
To further substantiate the possibility that this was a pre-Han account, Zigong's traveling to Qi to prevent an attack on Lu is mentioned in Han Feizi 韓非子 and in Mozi 墨子.Footnote 117 The Hanfei zi account is brief and somewhat dubious.Footnote 118 However, the Mozi account, while also brief, parallels with Sima Qian's version and contains details also found in Shi ji that are not found in the Yuejue shu or Kongzi jiayu versions.Footnote 119
Given the various evidence outlined above, it is likely that Sima Qian primarily used the “Chen Cheng Heng” account (or its prototype) in crafting the main portion of Zigong's biography. As such, this chapter should be listed as a source text.
Conclusion
Griet Vankeerberghen has asserted, “Authorship, in such cumulative texts, is located in discriminating selection and superior organization.”Footnote 120 And Stephen Durrant has stated that Sima Qian “is responsible for the narrative choices that have led to the completed text, and those choices, even when they involve little more than quotation from older sources, can be taken to reflect the intentions of the Han historian.”Footnote 121 Yet, if scholarship continues to take (and misquote) Ban Biao's statement on Zuoshi chunqiu, Guo yu, etc. as Sima Qian's source list, it will be impossible to know the length of his “discriminating selection” or fully understand “the intentions of the Han historian.” Thus, working toward a more complete knowledge of the texts that Sima Qian referenced and assimilated into his Shi ji is critical. Simply put, the degree to which the sources of Shi ji as a whole, and of individual chapters in particular, are identified is the degree to which we can assess his authorship and understand his message.
Over the last century, seven scholars have worked to compile bibliographic lists of works mentioned or used in Shi ji. Over time these lists have grown in length. While an overall expansion has occurred, it has come via varying additions and deletions among the different lists. If all bibliographies were combined, Zhang's—the most populous to date—would only contain seventy-five percent of that total. If Sima Qian's use of source texts is to be effectively explored, the next step would be to compile a master list of the seven bibliographies and then fine tune it using the principles outlined in this study. Additional texts cited or mentioned in Shi ji need to be added to the list. Those added will go beyond the limited scope of “books” to include memorials to the throne, poems, letters, documents, and other shorter texts—as much as they can be identified. The importance of these individual pieces of writing in historical studies is especially manifest when one considers the effect of the many pieces Ban Gu added to Han shu when copying from Shi ji. Footnote 122 The clarity and angles these shorter works provide to the historical figures recorded therein make them critical to assessing Sima Qian's view and use of source materials.
Such a list will provide a foundation upon which the actual work of understanding this aspect of authorship—further exploring the way Sima Qian incorporated his sources into the Shi ji narrative—can be fully realized. In this regard, we can turn again to Ban Biao's summary comments on Shi ji. As far as can be ascertained, Ban Biao was the first to give us a preliminary and brief analysis of Sima Qian's adaptations of source materials:
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，甚多疏略，不如其本。
Of that which Qian recorded, his merit lies in the period from the beginning of Han up to [Emperor] Wu, where it breaks off. As for his collage of the Classics and their commentaries, with the doings of the hundred schools scattered throughout, it is sketchy and full of holes. It is nothing like what is in the original records.Footnote 123
This assessment is but an extremely brief synopsis of Ban Biao's reading of the text. Much more exploration and explanation remain to be conducted. That is not to say that this is wholly unexplored territory. Several articles have been written on this aspect of Shi ji. Karlgren's exploration of grammatical appropriation and Nienhauser's article on the “Hereditary House of Jin” are two such examples. Yet, much more of the Shi ji text awaits to be dissected and analyzed by mapping each chapter and their use of source texts. It is within the appropriation, paraphrasing, and combining of source materials—along with the passing over of these materials—where Sima Qian's creative genius remains to be fully uncovered.
