Literature Review

Discussions about the textual formation process of the Baihu texts mainly took place from the mid-Qing dynasty to late Republican eras, prior to 1949. The monumental Siku quanshu zongmu 四庫全書總目, as well as discrete works by the classical exegetes Zhuang Shuzu 莊述祖, Zhou Guangye 周廣業, Yao Zhenzong 姚振宗, Sun Yirang 孫詒讓, and Liu Shipei 劉師培 are prime examples (see below). When later scholars focused on the events connected with the White Tiger Hall conference, they rarely put forth any new ideas; instead, they selectively recycled ideas borrowed from these six works, making it possible in a brief literature review to summarize all the important findings relating to the Eastern Han conference.

A scholarly consensus today holds that the Baihu yizou is the first-stage result of the White Tiger Hall conference and also the basis for the Baihu tong compilation, for two reasons. First, the Baihu tong seems to be a different compilation from the Baihu yizou, given the number of bamboo bundles (juan 卷) in each. Cai Yong's 蔡邕 (132–192) “Bajun taishou xieban” 巴郡太守謝版 mentions three works, among them a Baihu yizou, which are 212 scrolls in toto.Footnote ⁹ Zhuang Shuzu's (1751–1816) “Baihu tongyi kao” 白虎通義考 then comments on Cai Yong's writing:

Although Zhuang clearly is wrong to conflate “bamboo bundles” (juan) with “chapters” (pian), as the two terms did not mean the same thing in the Han dynasty,Footnote ¹¹ we cannot fail to note the huge discrepancy between the six-bamboo-bundle Baihu tong recorded in the Sui shu 隋書 bibliographic treatise and the 100-bamboo-bundle Baihu yizou that Cai Yong mentioned.Footnote ¹²

Second, while the Baihu yizou has not survived, texts in the same genre afford enough evidence to probably distinguish it from the Baihu tong in terms of its function and structure. Liu Shipei's (1884–1919) “Baihu tongyi yuanliu kao” 白虎通義源流考 registers this important observation:

I believe this account by Liu Shipei to be more or less consistent with the facts of the matter, even if Liu never addressed further complications (principally, the pesky question of the Baihu tongde lun and what it was). Importantly, Liu distinguished phases in the textual production process, with the yizou (memorialized debating positions) preceding the production of the tongyi. If this is correct, the Baihu yizou in many bamboo bundles is a record transcribing, in part or in total, a number of administrative documents that took the form of memorials promoting one or more rival theories about the Classics, and this record was transcribed during, or immediately after the conference, and it subsequently received the emperor's written approval. By contrast, the extant Baihu tong is composed mainly of catechistic couplets in a question-and-answer format. As only one precise answer is given for each question in the vast majority of cases, naturally the Baihu tong content was considerably shorter than the Baihu yizou (more on this below). As an abridged version of the Baihu yizou, the Baihu tong summary would simply have conveyed the conference's main conclusions to its readers.

If discussions relating to the Baihu yizou seem complicated, the same is true of the Baihu tongde lun, only more so. The Siku quanshu zongmu claims, for example, that the Baihu yizou was named the Baihu tongde lun after it was compiled, and then Ban Gu edited it, making the Baihu tong.Footnote ¹⁵ Another view holds that the Baihu tongde lun is the Baihu tong. Yao Zhenzong's (1843–1906) “Hou Han Yiwen zhi” 後漢藝文志 says, for example,

Clearly, Yao believed the conclusion reached in the Chongwen Catalogue, that Baihu tongde lun and Baihu tongyi are alternative titles for the same compilation. Sun Yirang (1848–1908) shared a similar view with Yao, insofar as he, too, believed that the Baihu tongyi is the original name of the text compiled during Emperor Zhang's Jianchu reign period and Baihu tongde lun represents but a new name for the same text devised by someone unknown during the Six Dynasties (221–588).Footnote ¹⁷

Besides these two authoritative stances, another authority, Zhou Guangye (1730–1798), focused on the five graphs that make up the various titles (bai 白, hu 虎, tong 通, de 德, and lun 論). The Baojing tang 抱經堂 edition (1784) of the Baihu tong quotes Zhou to this effect:

Zhou Guangye was the first scholar to introduce the “Gong de lun” ascribed to Ban Gu into the scholarly debates, but he never really explains how and why these two texts came to be connected. To Zhou's speculation, Liu Shipei added:

All the foregoing theories seem reasonable enough, but all are equally speculative. The Siku quanshu zongmu editors nonetheless divided the short passage that appears in the Hou Han shu biography of Ban Gu in such a way as to confirm their sense of events. Believing that the Baihu tongde lun and Baihu yizou mentioned in the Hou Han shu “Annals of Zhangdi” are essentially the same text, they punctuate a single passage appearing in Ban Gu's biography in such a way as to establish the equivalence between the two texts, producing the following: “The Son of Heaven assembled the classicists to discuss the Five Classics, and compiled [or had compiled] the Baihu tongde lun.” They then couple the passage in Ban Gu's biography, which I would read “Ban Gu was instructed to compile it,” with a statement made in the “Rulin zhuan” (“Later the emperor instructed his archivists to compose the Baihu tongyi.”). But there is a problem: if the Siku quanshu zongmu editors are correct, it is hard to understand why the Baihu text mentioned in the Hou Han shu biography of Ban Gu was not compiled by Ban Gu, and was instead the Baihu yizou, whose compilation should have preceded Ban Gu's commission.

According to my alternative understanding of the relevant historical records, Yao Zhenzong and Sun Yirang were probably not correct when they asserted that the Baihu tongde lun represents no more than an alternative name for the Baihu tong. The Hou Han shu mentions the Baihu tongde lun as an independent work, but one may not conclude from this alone that this compilation was essentially the Baihu tong. Note that nearly all of the historical catalogues and literary works from the Six Dynasties to the Tang dynasty (618–907) quote a Baihu tong[yi]. Furthermore, no evidence exists for an unknown person during the Six Dynasties changing the name into the Baihu tongde lun, contra Sun Yirang. The Northern Song (960–1126) catalogue entitled Chongwen General Catalogue is the earliest extant text to identify the Baihu tong as the Baihu tongde lun, and since it cites no evidence for the equation, this may well be a Song-era view. Similarly, no supporting evidence supports the conflation of the “Gong de lun” with the Baihu tong, despite Zhou Guangye's proposed insertion of the graph (gong功) between tong and de. In the end, to follow all of Liu Shipei's deductions is simply too much of a stretch.

Reviewing previous research on the formation of the Baihu tong compilations, it seems obvious that past scholarship relied on (1) the records preserved in the standard histories; (2) assumed certain models for textual genres; and (3) paid great attention to the bamboo bundle lengths for certain texts. Using the passage preserved in Cai Yong's “Bajun taishou xieban,” we can deduce that the original length of the Baihu yizou was much greater than that of today's Baihu tong. That seems as certain as any evidence from the distant past is likely to be. And since the three discrete Hou Han shu entries at present offer our only clues for this conundrum, given the paltry evidential record, we have no recourse except to return to the precise language of Hou Han shu, and carefully analyze it, for no reasonable conclusion can contradict the three Hou Han shu statements. With respect to genres, two main questions await resolution: (1) What is the Baihu tongde lun, if the textual record allows for some similarities and differences between it and two other works, the Baihu yizou and the Baihu tong? and (2) What does the graph lun 論 in its title indicate in Eastern Han times? While more evidence will be adduced below, for the moment I draw the reader's attention to Liu Shipei's attempt to compare the Baihu yizou with the Shiqu Li lun. Unfortunately, Liu ignored the fact that the Han shu bibliographic treatise records a Shiqu lun 石渠論, apparently an analysis of the Shiqu yizou, as I will endeavor to prove in the course of this essay. This bibliographic entry offers a critical clue, and Liu's failure to consider its implications prevented him from arriving at the correct solution, in my view. So even though historical records are the starting point and the end point of this research project, an analysis of textual genres is still a key factor underlying any analysis devoted to the formation of the Baihu texts. Finally, I would note that previous scholarship on the Baihu compilations have more or less ignored the Han institution of court conferences, and so created some of the difficulty, it seems.

Textual Formation and Circulation within the Institution of Conferences in the Han

To start our research, it is best to recapitulate all three records in the Hou Han shu regarding the White Tiger Hall conference and the literary works that followed it.

1. 1. In the Hou Han shu “Annals of Zhangdi”:

   On the renxu 壬戌 day of the eleventh month of the fourth year of the Jianchu reign period, … Emperor Zhang commanded the Superintendent of Ceremonial 太常, the Generals 將, the Counselors 大夫, the Academicians 博士, the Gentlemen Consultants 議郎, and the Palace Courtiers 郎官, together with the Academy disciples and the classicists, to convene a conference in the White Tiger Hall and have them expound upon the similarities and differences among [the pronouncements in] the Five Classics. He ordered Wei Ying 魏應, Leader of the Gentlemen of the Palace, All Purposes 五官中郎將, to receive the imperial edicts that posed questions, and Chunyu Gong, the Palace Attendant 侍中, to memorialize the exposition of proposals [to the emperor]. The emperor proclaimed the rulings, having attended in person, as in the preceding case of the Stone Canal Pavilion conference during the Ganlu 甘露 reign period of the Filial Emperor Xuan 孝宣帝 (r. 74–48). And Emperor Zhang it was who ordered the compilation of the Baihu yizou.Footnote ²⁰

2. 2. In the Hou Han shu biography of Ban Gu:

   The Son of Heaven convened the Classicists to expound and discuss the Five Classics, and put in writing the Baihu tongde lun, Ban Gu was instructed to compile it.Footnote ²¹

3. 3. In the Hou Han shu “Rulin zhuan”:

   In the Jianchu reign period, the emperor greatly assembled the Classicists in the White Tiger Hall, to examine minutely the similarities and differences among the Classics. This conference only ended after several months. Su Zong 肅宗 [i.e. Zhangdi] attended in person and proclaimed the rulings, as in the preceding Stone Canal Pavilion conference. Later, the emperor instructed his archivists to compose the Baihu tongyi.Footnote ²²

These passages are generally consistent in their description of the conference proceedings, but there are obvious discrepancies in their accounts concerning the formation of the literary work (or works) called Baihu. Of course, in historical writings historians sometimes present differing accounts of the same event, when citing different sources. Notably, however, these three records are consistent regarding the convening of the conference, suggesting that any discrepancies we find in passages discussing the post-conference stage may not simply be dismissed, let alone explained by contradictions in the sources available to Fan Ye. Moreover, these passages all provide different titles for the compilations, and, based on Cai Yong's account (summarized by Zhuang Shuzu) mentioned above, we imagine different bamboo bundle lengths for the two compilations entitled Baihu yizou and the Baihu tong. Thus, we may ask, what precisely is the reality reflected in these three different passages in Fan Ye's Hou Han shu? To answer this, we must trace the history of the court conferences as an administrative institution and consider as well the particular nature of manuscript compilation, textual genre formats, and literary circulation during the two Han dynasties.

During both the Western and Eastern Han, whenever the court confronted a difficult issue, it was wont to convene court officials to a disputation or debate regarding it, allowing the court to gather and then consider a range of policy suggestions. Such court conferences happened fairly often throughout the two Han dynasties, from 206 b.c.e. to 220 c.e., and there were two ways of holding conferences: one was predominately oral, and the second written (i.e. when policy positions were submitted in the form of written memorials). Written submissions certainly entailed prior circulation of administrative documents (many in edited form). Even in oral debates, with the emperor in attendance, the emperor convened a particular conference by sending out invitations to the conferees, in the form of imperial edicts summoning the participants. After a given conference was concluded, its main conclusions were then to be circulated in another form of literary production undertaken by the imperial administration. Thus, the conferences invariably entailed the preparation and circulation of many texts. As some excellent research has already explored the Han conference system,Footnote ²³ this essay will try to make use of that research in light of what we know about the formation process of the text(s) for the White Tiger Hall conference.

The first stage in the formal proceedings was the initial edict (命議詔書) issued by the court that summoned various officials at different ranks (and sometimes technical experts) to attend the conference. These kinds of formal invitations were usually sent out in the emperor's name. For example, in the “Annals of Gaodi” (高帝) in the Han shu, we find an entry dated to the twelfth year of Gaozu's reign (195 b.c.e.), saying that Liu Bang 劉邦 wanted to reinstate King Wu (吳王, d. 154 b.c.e.), but sent an edict to ask his officials to “discuss and approve it.”Footnote ²⁴ Sometimes, the edict would take the form of a “ruling” (zhiyue 制曰) appended to a high-ranking official's memorial.Footnote ²⁵ The Han shu biography of Liu Chang 劉長 (198–174), King of Huainan (淮南王), says that the Chancellor (丞相) Zhang Cang 張蒼 (d. 152 b.c.e.) and others memorialized together that Liu Chang should be executed in accordance with the law for his improper behaviors, but when Emperor Wen (r. 180–157) refused their suggestion, he appended a ruling, which said, “We cannot tolerate extending the law to him, so you Nobles (列侯) and high officials commanding a 2000-bushel (二千石) salary are to discuss this issue further.”Footnote ²⁶ This is a good example of an emperor responding to an initial finding by his officials at a court conference.

In the second stage of the formal proceedings, participants at the conference were each to express their own positions in the debate. As Cai Yong says in his Du duan:

This account of the second stage of the process offers several surprises: that court conferences could be convened even when the court already held a fixed set of practices based on a specific premise, history of interactions, or principle; also, that in the end all those holding views that were contrary to the consensus view could compose rebuttals that would give their lines of reasoning. For example, the Han shu biography of Yu Dingguo 于定國 (d. 40 b.c.e.), who was the Chancellor at the start of Emperor Yuan's reign (48–33), says,

Here is the back story to that set of remarks: during Western Han, the Chancellor held supreme power over the court officials, power that in Eastern Han would devolve to the Secretariat. Yu Dingguo was adept at handling administrative affairs and often pushed his own proposals. But his nominal assistants, Chen Wannian and Gong Yu, acted very differently with respect to the proposals that Yu Dingguo expressed during court conferences. Gong often presented rebuttals to show his disagreement with Yu.

However, if the court was likely to be badly divided on an issue, a careful collection of reports and position papers needed to be compiled even before participants met at the conference, to help those participants sort through the range of options and positions presented by conference participants, so that eventually the assembly of participants could reach a decision, after due consideration of the entire range of debating positions. This required all the invited participants to formulate their individual views in writing, and it was common to refer to these written opinions as yi 議 (literally discussions,Footnote ²⁹ meaning debating positions or opinions). For example, Cai Yong's “Da zhai yi” 答齊議 says:

The passage above is a good example of part of a jointly-coauthored “discussion” (yi) put in writing by Minister Cai Yong and the Academician Ren Min, in reply to the question posed in an imperial edict.Footnote ³¹ One noteworthy aspect is Cai's use of the two formulae sizui dui 死罪對 and yugang sizui 愚戇死罪, phrases that typically accompany a public presentation or biao 表.Footnote ³² Apparently, after each of the invited respondents composed his own yi, they need not be directly submitted to the emperor for his approval;Footnote ³³ instead all the submissions were compiled together before being sent on to the emperor, a process called “compiling the dossier of [all the] texts of the yi [submitted for review]” (彙集議文).

During the Western Han, documents other than the yi dossiers were usually submitted to the Chancellor first and then sent on to the Imperial Counselor, who would present them with a memorial to the emperor.Footnote ³⁴ However, the extant historical records suggest that the process for submitting compilations of these debating positions or yi differed from that for other types of compilations, where the Chancellor almost always took the lead. The Shi ji biography of Yuan Ang 袁盎 (d. 148 b.c.e.) records a scene where Yuan wanted to discuss a certain matter in private with Chancellor Shentu Jia 申屠嘉 (155 b.c.e.). Shentu Jia replied, however, “If what you have to say concerns some public matter, you may go to the office and discuss it with the Senior Officer 長史, and in due time I will submit your proposal to the throne.”Footnote ³⁵ This remark shows that it was customary for chancellors to collect and then memorialize to the emperor what any lower-ranking officials had to say on the matter, as their duty.Footnote ³⁶ In Eastern Han, the Chancellor (now renamed the da situ 大司徒) was equally responsible for such memorializing, in theory, although over time we note a trend by which the Secretariat gradually took over this important duty. In the “Da zhai yi” cited above, the Secretariat received the edict and asked others for their positions, making it only natural that the Secretariat would then be responsible for compiling the participants’ responses in one or more memorials. A single piece of evidence suggests that other important figures could undertake this task as well: a Han shu entry in the “Annals of Xuandi” 宣帝 had Xiao Wangzhi 蕭望之 (47 b.c.e.), Senior Tutor for the Heir Apparent (太子太傅), discharging this responsibility during the Stone Canal Pavilion conference of 51 b.c.e.Footnote ³⁷

The administrative work of compiling the collective dossier of debating positions (yi) was not limited to simply collecting a formal statement in writing from every participant; there was further editing to be done to prepare the document for the imperial review. The editing work required the editors to offer fair summaries of the differing views offered by the participants, and calculate how many people were behind each position, so that the different positions could be presented in an orderly account. The Han shu biography of Wei Xuancheng 韋玄成 (d. 36 b.c.e.) records a case in which Wei along with others submitted the compilation summarizing their collective discussions (yi) one month after Emperor Yuan had ordered the removal of the ancestral temples in the commanderies and kingdoms in 40 b.c.e.:

How Chancellor Wei Xuancheng handled the ministers’ proposals can be inferred from the passage translated here. Wei first collected the views of every participant, and then organized all of their different views, counting the number of people who supported one position, and verifying the specific points where they agreed or dissented. On this basis he formulated the majority position in a memorial, and then offered what he considered the important dissenting views, attaching the names and the number of people who supported each of the dissenting views. Note that the significance of counting the number of people was clear: the view treated most extensively by the Chancellor usually had to be said to reflect a majority view.

The Han shu treatise on imperial sacrifices (Jiaosi zhi 郊祀志) records that in the beginning of Emperor Cheng's reign (33–7 b.c.e.), Chancellor Kuang Heng 匡衡 and Imperial Counselor Zhang Tan 張譚 memorialized that a conference should be held to determine the site on which to conduct the suburban sacrifices. At that conference, the participants’ vote split into two main positions, one of which was clearly in the majority. Wang Shang 王商 (d. 25 b.c.e.), General of the Right (右將軍), and forty-nine others thought that the sites for sacrifices to heaven and earth should be moved to the southern and northern suburbs of Chang'an respectively, basing their views on the Rites Record (Li ji 禮記). However, Xu Jia, Marshal of State and General of Chariots and Cavalry, with seven others thought that the earlier precedents should be followed. In response, Kuang Heng and others stated:

After noting this dissenting position, Kuang Heng concluded: “In such a case, the sites for sacrifices should be moved to the southern and northern suburbs of Chang'an, so as to be a solid foundation of all generations.” The Son of Heaven agreed to Kuang's suggestion.Footnote ⁴⁰ In this example, even though Kuang Heng was in favor of moving the sacrificial sites to Chang'an, he was still required to go through the lengthy process of convening a conference and gathering views first. As the overwhelming majority voted in favor of moving the sites, his policy proposal was finally deemed “suitable.”

There are, of course, times when the compiler of various proposals (yi) would only list the various positions taken by the participants, without apparently highlighting the majority view. In the Shi ji biography of Gongsun Hong (公孫弘, 200–121), Noble of Pingjin 平津侯, it says, “Whenever court conferences were held, Gongsun Hong would simply state the pros and cons of the question and leave the emperor to make his own decision. He never ventured to contradict the emperor face to face or argue with the emperor in court.”Footnote ⁴¹ Put another way, Gongsun Hong only relayed the content of the various views expressed, as this allowed the emperor to decide for himself, even though such conduct abnegated his responsibility to act as Chancellor.Footnote ⁴²

As soon as the texts of the participants’ discussions (yi) have been compiled, edited, and submitted to the emperor, they became a new kind of textual record known as the yizou 議奏 [memorialized debating positions], which compilation was usually kept together with the initial edict convening the conferences.

Essentially, these two stages of the textual formation during written discussions—ministers writing out their debating positions individually and a high-ranking official (usually the Chancellor) compiling a dossier that represented all the debating positions—happened during so-called “oral arguments” at court conferences as well. During oral arguments, each minister expressed his own views, which were then transcribed by a court archivist or scribe, resulting in a formal record. Then, after oral arguments were concluded, a summary of the various arguments presented at the conference would then be reflected in a suitably worded memorial transmitted to the emperor, at which point it the entire memorial reflecting the range of important policy proposals became an yizou (memorialized summary of the debating positions). Because of the need, sooner or later, to summarize and edit arguments, whether written and oral, there was no major difference between oral and written debates in a conference devoted to one or more policy issues.

A final stage of the conference took place after the yizou had been sent to the emperor, for the emperor was still required to express his formal approval by appending a ruling signaled by phrase zhiyue 制曰 to the end of each document. Generally speaking, we know this from reports signifying four different levels of commitment by the emperor:

1. A unanimity of views meant that the emperor simply wrote “Approved” (ke). For example, the Han shu records an edict dated to the spring of the second year of Empress Gao's reign (186 b.c.e.) that says: “Now, order the court ranks of the nobles according to their various achievements, so that they will be conserved in the temple of Gaozu for all time, and their heirs may inherit the rank corresponding to such achievements. You shall discuss this and reach consensus with the nobles and then submit your conclusions.” Chancellor Chen Ping (陳平, ?–178 b.c.e.) memorialized: “The nobles have been fortunate to receive food and money from their fiefs, and your majesty now wants to provide them with even more benefits, by correlating their ranks with their merits. Your minister requests that this order be preserved in the temple of Gaozu.” Because this conference was reported after the agreement had been reached, the report of Chen Ping's views is simple and in line with the wishes of Empress Gao, so they were forthwith approved.Footnote ⁴³

2. Where there were a range of views, the emperor would write simply, “The imperial ruling is that So-and-So's views are approved,” selecting one view among many. The Han shu treatise on imperial sacrifices reports that a bronze tripod was found in Meiyang 美陽, after which it was offered to the court, which debated where the bronze tripod should be housed. A majority thought it should be placed in the ancestral temple, but Zhang Chang 張敞 (d. 47 b.c.e.), Governor of the Capital (京兆尹), submitted an alternative view, stating that, according to the engravings on the tripod, the vessel “was not suitable to be placed in the ancestral temple,” and the emperor's ruling was that he approved the view of the Governor of the Capital.Footnote ⁴⁴

3. The emperor occasionally not only announced his approval of a particular stance, but appended his own remarks to it, so that the content of the ruling would reflect the emperor's own position. For example, the Hou Han shu treatise on sacrifices (Jisi zhi 祭祀志) records a court conference on the establishment of the ancestral temple system during early Eastern Han, where Chancellor (大司徒) Dai She 戴涉 (d. 44 c.e.), a ritual expert, and others memorialized to Emperor Guangwu in the following yizou:

While there were dissents registered at the time on this tricky matter, the emperor finally approved of Dai She's view, saying: “As the sites of the ancestor temples have yet to be determined, from now on may the court sacrifices to all the ancestors be held together in the temple of Gaozu.”Footnote ⁴⁶ Guangwu agreed to let Emperors Cheng, Ai, and Ping, at least for the moment, receive sacrifices in the Temple of Gaozu in Chang'an, and he agreed to let Lord Nandun and Noble Jie of Chongling receive sacrifices in the mausoleum towns and temples annually and seasonally. Since these mausoleum towns and temples were located far away from the administrative headquarters in the commanderies and kingdoms, when it was inconvenient to hold sacrifices there, Guangwu gave his assent to having one or more local officials offer cult on behalf of the commandery Governor and kingdom Chancellor. Additionally, Guangwu reasoned, because the Filial Emperor Xuan had made great contributions to the dynasty, he should be given the ancestral title of Zhongzong 中宗 (Mid-dynastic Ancestor).Footnote ⁴⁷ Plainly, Emperor Guangwu drew upon the positions of Dai She and his allies.

4. If the emperor did not agree with any of the court participants’ ideas, he had the option of citing his own views after a formal ruling on the matter. In the example of the conference on Liu Chang's crimes (cited above), we saw that the members of the court had all urged the emperor to execute Liu Chang “in accordance with the law,” but the emperor, as paterfamilias, believed he had the right to not execute his family member; he decided he would merely depose him as king.Footnote ⁴⁸ In this case, Emperor Wen overturned the decision offered by his court for his approval.

Actually, there was probably no significant difference between the emperor's approval of an yizou and his approval of a regular memorial in terms of textual format. Just the same, after a set of position papers generated in connection with court conference were memorialized to the emperor and he had announced a formal ruling, that ruling by the emperor was transformed into an imperial edict. However, since the court conference was a mechanism designed to promote effective government, and not merely a tool for the execution of imperial power, an emperor could set a formal conference discussion into motion, without being entirely certain where it might lead.Footnote ⁴⁹ This suggests that, unlike approval for regular memorials, the emperor could exercise his decision-making powers in relatively limited ways when it came to approving or disapproving yizou. During both Western and Eastern Han, the emperor rarely vetoed the proposals put forward during court conferences, and the few cases where we see an emperor exercising a veto usually happened after legal cases involving members of the imperial family.Footnote ⁵⁰ As was shown in the four situations above, when the issue directly affected a member of imperial family, the emperor evidently had more freedom to wholly or partially reject a strong consensus position registered during a given court conference via memorialized discussions or yizou. Normally, the establishment of the imperial ancestral temple system was deemed to affect politics more than the deposal a king during Han. Therefore, Emperor Guangwu had only to approve Dai's proposal before he could add his own position into the ruling. By contrast, Emperor Wen made his decision independently because he was discussing his own close relative Liu Chang. In most cases, the emperor had the customary right to select among the proposals the court conference submitted; at the same time, custom mandated that he could not casually ignore the majority positions expressed, nor could he go well beyond the boundaries set by the participants at the court conference. In brief, the emperor could not exercise absolute free will, doing as he pleased. Rather, his rulings had to be aligned with the content of the court conferences and accord with the multiple forces at work there.

If an emperor or a regent ruling the court in his stead was deeply dissatisfied with the conference findings, or there was too much divergence among the participants’ positions to arrive at any clear conclusion after a single set of discussions, the conference could come to a halt before any clear conclusions were reached (寢議), and the debate participants would then be dismissed (罷議). Sometimes a conference would be reconvened (復議), even multiple times.Footnote ⁵¹ Under such circumstances, a set of position papers memorialized to the throne or yizou did not become an edict, needless to say, nor was the yizou further circulated. Thus, to reconvene a conference several times points to the seriousness of the issue undergoing debate, and sometimes the multi-stage process began several times, until some consensus was reached.

The quotation of a passage ascribed to Emperor GaoFootnote ⁵² (高皇帝所述書) in the Han shu biography of Wei Xiang 魏相 (d. 59 b.c.e.) tells us much about the institution of the Han dynasty conferences and the texts produced in connection with them:

The early Tang commentator Yan Shigu 顏師古 (581–645) quotes annotation by Ru Chun 如淳 (a third-century exegete commenting on the Han shu), which says: “Section Eight, rulings on the Clothing of the Son of Heaven, became the eighth edict to be promulgated.”Footnote ⁵⁴ This passage shows us that the final proposal of a court conference eventually became an edict, according to the regular conference process outlined above. First Liu Zhang received the order, in an initial edict, to convene a conference. Following this, the officials all expressed their individual debating positions (yi); then Chancellor Xiao had summaries of their debating positions compiled together, and this compilation was then transmitted to the Senior Imperial Messenger Liu Xiang and Liu Zhang, so that they might compose a memorial based on the conference proceedings (yizou). Finally, the Son of Heaven approved the yizou, giving it the status of a final edict (決議詔書).

In addition to the phases of this process when the debating positions are turned into the compilation of yizou,Footnote ⁵⁵ the final textual product issued in connection with the court conference took the form of a final edict. When we read a final edict, its text might consist of no more than a few hundred Chinese characters, but it was hardly composed on the spur of the moment nor was it singly authored, insofar as its content was shaped by multiple forces. And, as soon as one becomes aware of this, then the production of several texts from the White Tiger Hall conference is no longer so difficult to understand.

From the Baihu yizou, to the Baihu tongde lun, to the Baihu tong

According to the foregoing discussion, the debating positions that are gathered in a dossier and memorialized to the throne (yizou) were produced during a third phase of the court conference proceedings, after a summary of the participants’ different stances had been transcribed in response to the initial edict convening the conference. To return to the specifics of the White Tiger Hall conference, we have no indication that the production of a Baihu yizou was in any way irregular. Nonetheless, it has noteworthy features, as can be seen in Hou Han shu “Annals of Zhangdi.”

The “Annals” says:

According to this edict, the White Tiger Hall conference was not held at regular audience hall where the emperor routinely held court, but somewhere else in the palace,Footnote ⁵⁷ making the conference somewhat similar to judgments “sent down to the ministry for discussion” (xiayi 下議), which were often oral debates. Nor did the emperor feel the need to be personally present at through its month-long agenda.Footnote ⁵⁸ During the White Tiger Hall conference, two people, Wei Ying and Chunyu Gong, undertook the editorial work compiling the preliminary summaries of the court conference proceedings, as is very clear from the Hou Han shu: “Emperor Zhang ordered Wei Ying, the Leader of the Gentlemen of the Palace, All Purposes, to receive the imperial edicts that posed questions, and Chunyu Gong, the Palace Attendant, to memorialize the contents of those discussions [to the emperor].”Footnote ⁵⁹ It can be seen from the same history that “the Emperor personally made his rulings and pronounced the imperial verdicts.” This passage further emphasizes the emperor's role in the conference, as he examined and approved each of the memorials, possibly because occasionally, during the Han, it was the emperor's highest-ranking officials who drew up the formal administrative documents on “the emperor's behalf.” The White Tiger Hall conference went on for “several months before concluding,” and the participants touched upon many complex issues, in their wide-ranging discussions about the policy implications of the differences and similarities among the Five Classics. Plainly, the entire process extended over a long period of time, with various papers circulating, as men of various ranks and kinds and expertise assembled in the palace pavilion and engaged in prolonged discussions, after which summaries of the arguments and their conclusions reached by the participants were prepared, so that those might then be sent to the emperor for his approval. Evidently, the Baihu yizou referred to in the sources represents the compilation of the conference views by Chunyu Gong, eventually sent to the emperor for his approval.

So what, then, is the Baihu tong de lun 通德論, for that is the second title mentioned in connection with the White Tiger Hall conference in the Hou Han shu? Whereas the Baihu yizou was compiled by a process detailed in the Hou Han shu “Annals of Zhangdi,” the biography of Ban Gu mentions a compilation called the Baihu tongde lun, even if the passage is too laconic to be very clear. However, I suggest that the key to solving this conundrum is encoded in the genre dubbed lun 論.

In the Han shu bibliographic treatise, under the Documents classic (Shang shu 尚書) category, it records yizou, in 42 chapters; under the Annals classic (Chun qiu 春秋) category, it has a second yizou, this time in 39 chapters; in the Analects (Lun yu 論語) category, it lists a third yizou, in 18 chapters; and in the Classic of Filial Piety (Xiao jing 孝經) category, a Five Classics Assorted Discussions (Wujing zayi 五經雜議) in 18 chapters. All of these entries have notes identifying them with the Shiqu lun (“Concluding Discussions at Stone Canal”). Moreover, under the Rites category there are 38 chapters listed for an yizou whose annotation merely has the word “Shiqu.” Qian Dazhao's 錢大昭 (1744–1813) Hanshu bianyi 漢書辨疑 argues that the graph lun is missing.Footnote ⁶⁰ Qian is right, I believe, and therefore, I propose that the Shiqu lun be considered just another name for the Shiqu yizou.

May one then conclude that yizou and lun are always the same thing? Of course not. Textual evidence is required to come to such a conclusion, and too many early texts no longer survive. Happily some additional relevant information exists to help us elucidate the present problem. The Han shu bibliographic treatise records 155 chapters for the five yizou listed above, but sometime during the Six Dynasties, many were lost, so that only the Rites classic yizou survived. The Sui shu bibliographic treatise records a Shiqu Li lun, in 4 scrolls, compiled by Dai Sheng 戴聖,Footnote ⁶¹ which was later lost. Fortunately, some fragments drawn from this latter text have survived in the Tongdian collection, so that we can study them today:

Readers often have construed the sentence about wearing one-year mourning garments unless the male is his father's heir as the “personal view” of Senior Tutor Xiao. By that construal, the passage “Wei Xuancheng thought” begins another section where the two sides are discussing the difficulties of this question. However, comparative evidence from Han court conference proceedings suggests that the whole passage constitutes a single section and represents a portion of a memorialized set of debating positions or yizou. By the usual process of compiling an yizou, Xiao Wangzhi would first have assembled the proposals of the various officials who represented the majority position, to which he appended the dissenting view of Wei Xuancheng, and finally Emperor Xuan's reply to the debate is appended. So the whole foregoing passage in all likelihood reflects a procedure whose conclusion was disseminated after a conference. (The so-called lun may derive its meaning from the criminal trials which the Han called lunzui 論罪, “judging the crime,” in which case the title indicates the conclusions reached after due consideration at the conference.)

We find, too, that besides the Shiqu lun more examples exist where reports of court conference are called lun, during the Han and Six Dynasties periods. Two in particular prove interesting: Huan Kuan 桓寬 in late Western Han compiled the Yantie lun 鹽鐵論 in 60 chapters to “continue the discussions on salt and iron, to expand the items in the agenda, and to push the debate to its limits” as well as “to investigate the reasons of order and disorder, in order to form a model for his own thought.”Footnote ⁶³ (As readers before me have noted, the Yantie lun is not a precise transcript of the conference proceedings, but a summation produced after the debates had concluded, possibly many years after.) The formation of Wang Su's 王肅 (195–256) “Final Judgments on the Sage” (“Sheng zheng lun” 聖證論) is described in the Jiu Tang shu as follows. After Wang Su attacked Zheng Xuan 鄭玄 (127–200),

The textual production processes known from the Yantie lun and “Sheng zheng lun” cases need not be exactly identical with that which led to the Shiqu lun. What matters is that the texts of these two corpuses still originated in a court conference and still preserve the genre form of the yizou, and in the end there would have been final judgments recorded.Footnote ⁶⁵ As these three examples—the Shiqu lun, Yantie lun, and “Sheng zheng lun”—are all titled lun, it's entirely possible that the text of the approved yizou can also be called a lun.

The White Tiger Hall conference was convened more than a century after the Stone Canal Pavilion conference of 51 b.c.e., so we would expect the form of the Baihu tongde lun text to be modelled, at least loosely, on the Shiqu Li lun. If that was the case, the yizou reporting on the conference debating positions would have eventually been sent for approval to the emperor, so that he could produce a final edict whose text was then edited by the court archivists. At both the Stone Canal Pavilion and White Tiger Hall conferences, the yizou were memorialized to the emperor. By my account, the yizou represent as a single stage in the lengthy process of textual production, a phase before the final ruling took the form of an edict. Along with the conclusion of a conference, the entirety of the yizou would become the basis on which was composed final edicts which the emperor issued; at some point the emperor or the high-ranking members of his court would specifically order someone to compile the latest version of these into several volumes (ce 冊) to give them a relatively stable physical form, allowing them to be preserved for future reference. In this sense, yizou and the final form they took in edicts represent two stages in a text compilation process, with their content similar in some respects but changed in others. The Han shu bibliographic treatise, the “Yiwen zhi,” records some yizou compilations, but in this case the yizou, judging from the extant fragments, probably represent the final dossiers composed of imperial rulings relating to a particular conference. This has been understood by people such as Fan Ye, Liu Xie 劉勰 (465?–532?), Liu Zhao 劉昭, Wei Zheng 魏徵 (580–643) and Kong Yingda 孔穎達 (574–648), all scholars who lived after Ban Gu. Only Du You (735–812) seems to have been unaware of this convention and it was he, judging from the extant sources, who absurdly changed the title Shiqu Li lun to Shiqu Li yi in his Tongdian.

Looking back at Ban Gu's biography, we find that it says, “The Son of Heaven convened the classicists to expound and discuss the Five Classics, and put in writing the Baihu tongde lun, and he ordered Ban Gu to compile it.” This passage suggests that the texts of the final rulings on each question that came up for debate had already been drawn up before the conference ended and all that was left to do was to edit them into a compilation. If I am right, the statement in the Siku quanshu zongmu saying that “the White Tiger Hall conference yizou were [later] named the Baihu tongde lun” is probably correct, since the Baihu yizou was the title given to the collection of memorials during one or more production phases that culminated in the Baihu tongde lun, the details of which can be gleaned from the “Annals of Zhangdi.” This complex process was probably so well known to people of the time that Fan Ye only recorded some details relating to the post-conference period, after the Emperor had duly approved the final rulings based on the yizou.

The Hou Han shu “Rulin Zhuan” reports this: “Later, the Emperor instructed his archivists to compose the Baihu tongyi.” From this passage, we can see that the Baihu tong was a text composed some time after the conference, and its timing alone clearly distinguishes it from both the Baihu yizou and the Baihu tongde lun, by my reconstruction. Only the Baihu tong (in fragments) has been transmitted to present times, with the result that we can see its catechistic format. Awaiting explanation is how the archivists converted the text of the Baihu tongde lun into the text of the Baihu tong (see Table 1).

Table 1.

A Comparison of the Three Texts for the White Tiger Hall Conference

As for the organization of the Baihu tong, each chapter revolves around a theme, and there are smaller sections within each chapter that take a question-and-answer format. It is not hard to see that this textual format results from the work of condensing the Baihu tongde lun, that is the collection of the final rulings of the conference. The “questions” in the Baihu tong without a doubt correspond to the questions discussed in the final ruling in edict form, and the answers’ portion probably most often reflects the imperial response (though it may but need not reflect the imperial personal opinions, which will be further argued below) to the question, thus preserving the conference conclusion even if much of the argument supporting that conclusion was cut. Sometimes, after the questions and answers, there is appended a phrase “another saying” (yishuo) or “someone says” (huoyue). Presumably, these represent important objections to the approved positions at the conference, as recorded in the Baihu yizou. What is interesting to consider is this: why were these necessary to include in the final report? Judging from the histories, they were retained for future reference, as sometimes they were referred to in later court proceedings.

As we have seen, the Baihu yizou contained the memorialized statements in circulation, which could be added to or subtracted from, or otherwise revised, as the court conference proceeded. The positions gathered by Chunyu Gong (either orally or in writing) may have been continuously memorialized and continually revised, so all we can say at this remove is that these yizou were in all likelihood organized chronologically (i.e. by the date of submission of the memorial to the throne). Only after the conference ended and the yizou were finally given a final form, on the basis of which final imperial rulings were produced in the form of edicts disseminated beyond the closed circle of the court conference participants, would these rulings become law and therefore be subject to later critique by members of the court and by experts outside.

By way of comparison, we know that the Shiqu lun was organized according to the Classics and Zhang Rong hosted the discussion on the “Sheng zheng lun” by sequentially referring to individual passages in each of the classics which were topics of discussions. From this, it can be seen that the conclusions of conferences on the Classics were often initially organized by the texts of the Classics. This undertaking was relatively easy and did not necessarily require a great number of people to undertake it. In all likelihood, then, in the case of the White Tiger Hall reports and summaries, Ban Gu was able to complete his task of summarizing the main points within a short period of time. At the same time, turning the Baihu tongde lun into the Baihu tong would not only have required the reworking of the small question-and-answer sections, but also a complete reorganization of the order of the whole text so that the arrangement would not reflect the Classics per se, but the various issues raised in connection with the Classics. The so-called “arrangement by the Classics” more often than not focused on the citation text, rather than considering the issue within the larger context of classical learning. The conversion of the Baihu tong to catechistic format arranged by questions and answers on a theme allowed for the import of the discussions to be more readily accessed by court officials trying to apply the court conclusions to their administrative work. Here, the arrangement of the thematic chapters of the text is even more nuanced and its intellectual content more systematic than the arrangement by the Classics.

In any case, during the conversion process that eventually created the Baihu tong compilation, a great deal of historical information was apparently stripped away, such as when the discussion was held, who was involved, what their titles were, and how they participated in the conference. What remained in the text is just the content that the throne, which commissioned the Baihu tong production, wished to have conveyed. There were many voices at the White Tiger Hall conference and their debates were varied, wide-ranging, and complex,Footnote ⁶⁶ but after much of the background information about the discussions was stripped away, what remains in the Baihu tong is the question-and-answer sections conveying the ideals that the Eastern Han court wished to promote among its subjects. After the editorial process that the Baihu memorials went through, the finished product that we see today in the Baihu tong was a polished piece of political theory.

Because of the widespread assumption that Ban Gu compiled the Baihu tong, many believe that the “archivist” mentioned in the Hou Han shu “Rulin zhuan” refers to Ban Gu, perhaps as head archivist. However, this commonplace assumption is probably inaccurate. Let us examine the evidence at hand. In his own final chapter of the Han shu (Postface, Xuzhuan 敘傳), Ban Gu himself says: “In the Yongping 永平 reign period, I was appointed the Gentleman in charge of collating the palace writings.”Footnote ⁶⁷ Then, the Hou Han shu biography of Ban Gu claims that, after Emperor Zhang took the throne and before the White Tiger Hall conference, Ban Gu had already been promoted to Xuanwu Major (玄武司馬).Footnote ⁶⁸ The biography of Yang Zhong 楊終 (d. 100 c.e.) also says:

This shows that the White Tiger Hall conference happened after Ban Gu had been promoted to the Xuanwu Major, and while he still held the position of Gentlemen Collating Texts. The Taiping yulan 太平御覽, juan 233, quotes the Song shu 宋書 “Treatise on the Bureaucracy” (“Baiguan zhi” 百官志) as saying,

The Gentleman Collating Texts is a specific kind of archivist responsible for the work of compiling and editing court documents. This explains why Ban Gu received the order to compile the Baihu tongde lun. However, many people were hired to be the Gentlemen Collating Texts, not just Ban Gu, and we know, without a doubt, that more than just Ban Gu participated in the White Tiger Hall conference. In addition to Jia Kui and Yang Zhong, there was also a distinguished official named Ding Hong, who, because of his oratory displayed during the White Tiger Hall conference, received praise from the Emperor several times, and eventually became a Gentleman Collating Texts afterwards.Footnote ⁷¹ The saying “later, the Emperor instructed his archivists to compose the Baihu tongyi” probably points to a time not long after the conference ended when Emperor Zhang ordered several of his archival experts, including the Gentlemen Collating Texts, who had participated in the conference to complete the work of compiling the conference records, the Baihu tongde lun, into the intellectual work known as the Baihu tong.

Because Emperor Zhang particularly favored Ban Gu,Footnote ⁷² and because he finished the Baihu tongde lun ahead of schedule, it would hardly be surprising if Ban Gu was a member of these archivists and copyists. However, the extant materials do not suffice to allow moderns to determine whether Ban Gu played a significant role in this work. All we can say at this remove is this: when we compare the two Hou Han shu passages about the activities of Chunyu Gong and Ban Gu, we see that the Hou Han shu “Rulin zhuan” passages which specifies that the “archivists” composed the Baihu tong is clearly different, which suggests that the formation process of the Baihu tong was not an effort of a single person. It is still worth pointing out that the Baihu tong preserves the final rulings of the White Tiger Hall conference. Therefore, no whether it was compiled individually or collectively, the fact that its content originated from discussions would not change.

By my analysis, the Hou Han shu recorded different phases in the production process of several texts generated on the basis of the White Tiger Hall conference, and the chronological sequence of the production phases should be clear. These chapters corroborate each other, and once the entire lengthy production of the several Baihu texts is clarified, the collective nature of the Baihu tong during its textual formation process can be made certain and we can realize that the view ascribing all the phases of textual production to Ban Gu is not likely to be accurate. With the same evidence in mind, we can also, with a fair degree of certainty, assign Baihu tong to the last phase of textual production produced by the conference, at which point all previous textual productions relating to the conference were superseded and could be discarded. That an earlier phase of textual production, the Baihu tongde lun, was not transmitted and quoted in later times, so far as we know, is understandable, since the Baihu tong, in some sense, would have superseded earlier versions.

Did Emperor Zhang Really Approve of his Rulings for the White Tiger Hall Conference?

Above, we have reviewed the formation process leading to the production of several texts produced in connection with the White Tiger Hall conference. This is crucial for scholars to understand, if they wish to analyze the contents of this book and situate those within the context of the White Tiger Hall conference discussions. Besides, considering Emperor Zhang's position on the White Tiger Hall conference in addition to other historical records would be useful for better understanding this event and its text.

Because the Hou Han shu account says that that Emperor Zhang chengzhi linjue 稱制臨決 (“proclaimed the rulings, having attended in person”) during the White Tiger Hall conference, past research uniformly holds that Emperor Zhang dominated this discussion and then established a unified ideology for the court.Footnote ⁷³ Obviously enough, these assumptions go well beyond what the four-character phrase chengzhi linjue says. Moreover, such assumptions ignore the implications of the standard histories regarding the court controversies and debates over the best readings (plural) of the Classics and masterworks. Parts of the Baihu tong show a strong inclination to restrict the powers of the emperor,Footnote ⁷⁴ so it is hard to imagine that the conference conclusions faithfully reflect the imperial will;Footnote ⁷⁵ besides, the Hou Han shu shows us Emperor Zhang's dissatisfaction with at least some of the conference proceedings.

In 85, six years after the Conference was held, Emperor Zhang commissioned Academician Cao Bao 曹褒 d. 102 c.e.) to make a set of new rites for his court, showing that he was not exactly thrilled with the results of the White Tiger Hall conference. Superintendent of Ceremonial Chao Kan 巢堪 opposed this plan, and he argued, significantly, that Cao Bao, a single person, could not determine the dynastic rites on his own, implying that this could only be done through court conferences. When Ban Gu suggested holding another conference to have classicists discuss the rites again, Emperor Zhang refused to consider that option: “We could convene the ritual masters and call it a ‘court debate,’ but that would just raise more doubts and divergent readings without coming to any firm conclusions. Previously, when Yao 堯 wanted the Dazhang 大章 composed, all he needed was a single person, Kui 夔 to do it!”Footnote ⁷⁶ From this we may deduce the following: the White Tiger Hall conference findings displeased Emperor Zhang, so he decided to choose Cao Bao to devise new rites more in line with his own beliefs.

Clearly, the emperor, who was classically trained himself, disputed some of the conference findings, but in the end, as was not uncommon with such court conferences, the officials’ ideas prevailed over those of the emperor and he was persuaded (or forced?) to have them written into the conference record. A citation of the earlier Shiqu Li lun in Tongdian, juan 77, records a discussion that took place, asking why the musical accompaniment is mentioned during the rite of district archery (鄉射), but it is not mentioned during the rite of imperial archery (大射) in the Rites classic; in reply, we are told, Wenren Tonghan 聞人通漢, Dai Sheng, and Wei Xuancheng championed their own readings. This citation does not include a passage about a ruling (zhiyue); it only states that the officials finally agreed with the arguments of the ritual master Wei Xuancheng.Footnote ⁷⁷ Although this one case is not necessarily a precise counterpart to the Baihu tong, it still suggests that the support by a majority of participants was critical to determine what would be approved and written into the final account of the court conference proceedings.Footnote ⁷⁸ This means that even if the emperor could weigh in on the court discussions (chengzhi linjue), he found it hard to ignore the consensus and simply mandate that the conference participants agree with him.

Tjan Tjoe Som points out in his translation of the Baihu fragments that, to all appearances, the White Tiger Hall conference had achieved nothing, since it failed to unify disputes regarding classical learning that Tjan ascribed to the opposing schools of Modern Script (今文) and the Archaic Script (古文). To Tjan's way of thinking, Emperor Zhangdi was powerless against the band of his own Modern Script Academicians.Footnote ⁷⁹ Tjan's view does not reflect the Eastern Han view of things, it seems. I would emend Tjan's assessment to say that the White Tiger Hall conference could not possibly have reached unified conclusions about the policy implications of the Five Classics, given how important those policy implications were to every action performed and contemplated by Emperor Zhang and all his officials. In this regard, the Baihu tong cannot be considered a synthesis of the court's findings, establishing a single court ideology. Rather, it is best to see the text we have now as evidence for vigorous debates among the conference participants, who included the Emperor himself and a range of other officials. Thus we should read the extant Baihu tong as the work of the assembled experts, rather than the expressed will of a single person, no matter how eminent the status of that person.

As we can see, the White Tiger Hall conference was a significant event not only in politics but also in the study of the Classics in the Eastern Han, even if the emperor was not satisfied with all of the findings to which he gave nominal approval. Because it represents one attempt to provide greater political legitimacy for the Han ruling house through the cultural capital invested in classical learning, it is a valuable object for further investigation. Sadly, at present the Baihu tong is the only major piece of writing available to us. The textual formation process of this book was gradual, and the solutions I have derived for the problems posed in the text may strike uninformed readers as abstract or unduly simplistic. Nevertheless, underlying the current text are the realities of the early Eastern courts’ policymaking by a diverse group of political actors. Ergo, my attempt to uncover the facts about the court conference via the Baihu tong, reversing the imagined process of textual formation, to glean information about the probable historical basis for the disputes recorded in the text. While we are still quite far from being able to demonstrate the full complexity of the textual formation processes, the foregoing account lays the groundwork for an improved understanding of an influential Eastern Han text, as well as the complex sociopolitical realities that underlay the seemingly bland final report that we hold in our hands today.