Song officials inherited from earlier periods of Chinese thought a rich and varied discourse on the nature of human beings and how they should relate both to each other and to the state. A prominent feature of this discourse was its tendency to analyze complex phenomena into patterns of bipolar complementarity. “Bipolar” entails two contrasting poles or positions; “complementarity” presumes these opposites interact with each other over time in ways that produce recognizable, recurring patterns. In the best-known example, the permutations of yin and yang through the sixty-four hexagrams of the Book of Changes provided an analytical system that developed over the Zhou dynasty and had become a fixture in the intellectual and operational toolkit of Song thinkers and politicians.Footnote ¹

Most Song thinkers held with the Han philosopher Yang Xiong 揚雄 (53 BCE–18 CE) that human nature was a mixture of admirable and vile inclinations. Their cultivation would cause the admirable inclinations to predominate and result in a good person or junzi; cultivation of the evil would produce a deplorable person or xiaoren.Footnote ² This dichotomy – the former usually translated “gentleman,” with overtones of a noble or superior man; the latter as “petty men,” with overtones of inferior or ignoble men – became basic to Chinese, and especially Song, political thinking. The distinction first arose in the second half of the Spring and Autumn period. Originally, the entire hereditary aristocracy was considered junzi, “nobles” or “lords.” Eventually, however, declining standards of behavior forced its leaders to designate reprobate kinsmen as “petty men,” thus reserving the status of junzi for those presumed still able to demonstrate intellectual and moral leadership.Footnote ³ This transition was well under way when Confucius seized upon the terms as labels to distinguish those who had succeeded in his own program of moral education from those who had not. This junzi/xiaoren distinction pervades the Analects, Mencius, and even late Zhou non-Confucian discourse.

Despite long and prolific use over the centuries, these terms retained for Song officials at least two overtones of their early origins. First, the distinction between junzi and the xiaoren was based on moral, not on social or even intellectual, criteria. Second, both junzi and xiaoren belonged to the political or ruling class. As we shall see below, xiaoren had at least two senses in Song political discourse: first, fellow literati whom the writer opposed politically and attacked as xiaoren; and second, non-literati political actors such as eunuchs, female bureaucrats, military officials, or clerks. In both cases, however, the person’s perceived moral behavior triggered the xiaoren label. A Song official’s literati status did not automatically render him a junzi; nor did a commoner’s status necessarily render him a xiaoren. One indeed often encounters blanket condemnations of non-literati groups as xiaoren. Such uses arise not because all members of the group were considered xiaoren, but because literati perceived the majority of the group to behave badly. Thus, even though the latter groups were not literati, they were perceived as officers of the state, as members of a greater officialdom that comprised both junzi and xiaoren.

Modern scholars have focused on the junzi/xiaoren distinction largely in connection with studies of Song factionalism, giving far less attention to its implications for other areas.Footnote ⁴ The inadequate, almost comical character of any English rendering of these terms, plus the stark moral divide they reflect, have deterred both Western and Chinese historians from undertaking more comprehensive studies of how these terms actually functioned in Song. Yet the Renzong era Confucian revival brought these terms to the center of political discourse and generated a robust rhetorical system that aligned the junzi/xiaoren distinction with other bipolar Zhou era dichotomies such as moral duty versus profit (yi/li 義/利), virtue versus talent (de/cai 德/才), public versus private (gong/si 公/私), and forthright or straight versus skewed or biased (zheng/xie 正/邪). Aligning the axes and intersecting vectors of these pairs with those of the junzi/xiaoren dichotomy created a vigorous medium for Song Confucian discourse. This rhetoric provided, even for non-literati, a “common language” to discuss many aspects of governance for the remainder of Song, and influenced not only politics, but also historiography.

Few topics consume more space in Song political discourse than recruitment and personnel management. Two conversations between Emperor Taizong and his councilors reveal how the early Song maintained conventional, practical, and only minimally Confucian approaches to these issues. In 983/6, Taizong lauded the recent cohort of jinshi graduates and expressed his delight in recruiting so many junzi to his court. He went on to extol his efforts to recruit widely, to position his choices according to their individual talents, and to monitor their performance. Although he hoped all posts could be filled with “good talents,” he said he would prefer to find one truly excellent official rather than many mediocre ones. To this, Chief Councilor Zhao Pu 趙普 (911–981) agreed that the Great Peace would result when positions were staffed by officials best suited to carry out the duties of those positions. But he added, quoting Analects 4.7, that junzi and xiaoren can each be recognized by their mistakes. The then prevailing pre-Song commentary on this ambiguous passage held that xiaoren should not be faulted for lacking the abilities of junzi, thus implying that the wise ruler finds an appropriate position for each “talent.” In essence, Zhao Pu here reminds Taizong that junzi and xiaoren both have their uses and warns him not to neglect the talents of xiaoren in his quest to collect a court of junzi.Footnote ⁵ Zhao Pu’s advice contains no trace of the later Confucian disdain for the xiaoren and urges Taizong not to overlook their usefulness to the state.

Over a decade later, in 995/1, Taizong and the Assisting Chief Councilor Kou Zhun 寇準 (961–1023) were discussing the recent execution of two minor officials for embezzlement of state funds. The emperor remarked that since junzi and xiaoren will always exist, the best the ruler can do is to attempt to identify the two: “If everyone was a junzi, what need would there be for punishments?” To which Kou Zhun observed that xiaoren had existed even under the sage-ruler Yao. And he continued, “Today, even among those who wear Confucian garb and hold high position there are those who attach themselves to xiaoren, plotting to secure their own comfort. Why worry about these vile clerks!”Footnote ⁶ This exchange reveals that Taizong had not changed his earlier opinions: (1) officialdom will always contain junzi and xiaoren, (2) the task of the ruler is to identify and punish the latter before they endanger effective governance.Footnote ⁷ In his response, however, Kou Zhun, a jinshi graduate of 980, suggests a greater danger and hints at the coming infusion of junzi and xiaoren as bipolar moral categories into Song discourse on factional politics. For him, xiaoren lurk even at the highest echelons of government, posing as (“wearing the garb of”) Confucians. Most probably, Kou had specific individuals in mind, but he sought more broadly to expand Taizong’s conception of xiaoren beyond the petty law breakers he had just executed.

This transition – from a conception that, because junzi and xiaoren will always exist, the wise ruler will find positions suitable for both, to a Confucian conception that xiaoren should be excluded from governance as morally inferior men – took decades to gain momentum and was never fully complete. The two conceptions, with many variations in between, continued to exist until the end of the dynasty. Nonetheless, in the Long Draft Li Tao located a seminal event in 1038/10, about which he commented “from this point the discourse on factions began.” The event in question was the issuance of an inner directive warning officials that the recent flood of memorials supporting Fan Zhongyan in his attacks on the Chief Councilor Lü Yijian 呂夷簡 (979–1044) constituted impermissible factional behavior. Assisting Chief Councilor Li Ruogu 李若谷 (970–1049) explained the warning: alluding to Zhao Pu’s advice that junzi and xiaoren both have their uses, he warned that factional attacks indiscriminately target junzi and xiaoren alike, and thus make it difficult for “upright officials to establish themselves.”Footnote ⁸ The edict and Li’s elaboration implied that Fan and his supporters lacked the standing to distinguish between junzi and xiaoren, a function that Taizong had clearly reserved for the emperor alone.

As the Confucian resurgence gained momentum through the late 1030s and early 1040s, Fan Zhongyan and Ouyang Xiu, in an effort to work around the edict of 1038, argued that junzi could form legitimate, beneficial political factions. Ouyang’s 1044 essay “Discourse on Factions” marked a seminal moment in the development of Song Confucian political thought. After the departure of Lü Yijian in 1043/3, the emperor turned to Fan Zhongyan, Fu Bi 富弼 (1004–1083), and Han Qi 韓琦 (1008–1075) to form a new administration. This group, with Ouyang’s support as remonstrator, had thwarted the appointment of Xia Song 夏竦 (985–1051), a Lü protégé, as Bureau of Military Affairs (hereafter BMA) commissioner in favor of one of their own, Du Yan 杜衍 (978–1057). In addition, Shi Jie 石介(1005–1045) composed a “Poem on the Sagacious Virtue of Qingli” that, while lauding the virtues of the new administration, lampooned Xia as a miscreant. Xia Song in turn charged that his antagonists had acted in collusion as a political faction. Ouyang Xiu wrote his “Discourse on Factions” and submitted it to the emperor in 1044/4 specifically to counter Xia’s accusation.Footnote ⁹

The core of Ouyang’s essay directly challenged Xia’s premise, as well as standard political theory, that because factions work against the interests of the sovereign, they are all composed of xiaoren and thus necessarily bad. In a few short lines, Ouyang made the following claims. First, junzi can also associate to form factions. Second, xiaoren factions are inherently false because their only motivation is short-term gain. They dissolve either fighting over the spoils or once the specific gain has been achieved. Third, junzi factions, on the contrary, adhere because their members are committed to common goals of individual moral self-cultivation and state service. Fourth, the stability of the state depends upon the sovereign’s ability to identify true junzi factions and advance them in office, while removing the xiaoren. The remainder of the short essay cited some basic historical precedents before its conclusion, insisting again that the emperor must “distinguish junzi from xiaoren.”Footnote ¹⁰

As is well known, this gambit to justify Confucian faction building was unsuccessful, and Ouyang himself seems to have backtracked to a more tempered position. But the real importance of the essay was to supercharge the Confucian character of the junzi/xiaoren dichotomy and to infuse the old terms with a highly contemporaneous, political meaning. Furthermore, the essay’s unstated implications challenged long-standing notions about the proper relations between the sovereign and his officials. First, Ouyang defined junzi and xiaoren by elaborating on the pithy pronouncement in Analects 4.16 that “gentlemen understand what is right; petty men understand what profits them.” Ouyang’s core definition of the junzi proceeds as a commentary on this maxim:

Ouyang does not write that only Confucians can be junzi, but his emphasis on “cultivating one’s person” (xiushen 修身) and the need to develop a sense of personal morality as a prerequisite for state service are basic Confucian doctrines: the qualities one needs to excel at state service must first be developed in private study and in communion with like-minded men intent upon duty, loyalty, and integrity. In other words, a junzi faction begins as a self-selecting and self-supporting group; and, by implication, its members will function within state service in the same way. Ouyang thus posed a stark challenge to the early Song idea of what a court filled with junzi would mean. For Taizong, the ideal court would contain officials he had personally selected and that were personally loyal to him. For Ouyang, the ideal court would be staffed by morally trained officials, unified in their commitment to the Way. They would be ultimately loyal to the sovereign but also to each other.

Because Ouyang has changed the definition of how junzi behave politically, the implication of his conclusion that the sovereign’s duty is to “discriminate between junzi and xiaoren” differs drastically from what Taizong would have understood such a phrase to mean. For Taizong, he himself as emperor would decide case by case who was a junzi and who not. For Ouyang Xiu, since the brotherhood of junzi was self-acknowledging, the duty of the emperor should be merely to sanction that acknowledgment. As we shall explore below in Chapters 6 and 8, the early Song emperors, as products of the military culture of the Five Dynasties, viewed themselves as the sole arbiters of individual talent. They chose their subordinates one by one to perform specific tasks based on criteria of their own choosing. But for Ouyang Xiu, “discrimination” means identifying and advancing groups of officials who have already self-identified as junzi. This idea of a pre-existing “faction,” legitimate because of its united commitment to the Confucian Way of governance, challenged in at least three ways the ultimate authority of the emperor over all personnel decisions. It implied that only the senior members of a “junzi faction” should be selected for leadership positions. It implied that these leaders, to ensure group unity, should be entrusted in turn with the appointment of their subordinates. And it implied finally that the sovereign’s only real duty was to determine the authenticity of the group as a legitimate junzi faction and appoint its leadership to senior offices.

Non-literati actors pushed back immediately and repeatedly against the premise that only junzi could be appointed to high office and against the corollary that, if xiaoren were to have any place in government, they should be subordinate to junzi. Among the vocal critics of Ouyang Xiu’s essay was the senior eunuch Lan Yuanzhen 藍元震 (d. 1077). In a secret memorial to the emperor, he pointed out that Cai Xiang 蔡襄 (1012–1067) in a 1036 poem had lauded Ouyang Xiu, Fan Zhongyan, Yin Zhu 尹洙 (1001–1047), and Yu Jing 余靖 (1000–1064) as “four worthies.” In return, when these officials came into power, they rewarded Cai with a top position. In Lan’s view, Cai’s appointment used state resources (Cai’s salary) to support a private coalition. He calculated that if each of the five or six members of this faction were to bring in ten associates, “within a few years they will fill all important positions in government … . No one will dare to speak against them. Wielding their hatred to pay back their enemies, what will they not do?”Footnote ¹¹ Although the eunuch’s response repeated conventional anti-faction rhetoric, his numbers provide an interesting perspective on mid-eleventh-century administration: namely, that a unified coalition of a half dozen senior administrators could, if given free rein to appoint their subordinates, dominate Song government. With this simple calculation, the eunuch effectively warned against and countered all three of Ouyang’s implied challenges to imperial authority.

Despite the immediate failure of Ouyang’s challenge, many influential Confucian literati held similar views. Over the next several decades, they restated each of his propositions explicitly as they developed the notion of “shared governance,” or as Wen Yanbo 文彥博 (1006–1097) put it to Emperor Shenzong 神宗 (1048–1085; r. 1067–1085) in 1071: “You rule together with us who are the officials.”Footnote ¹² Yet no Song emperor ever accepted all of the premises and implications of Ouyang’s essay. Rather, there developed between the monarchy and the Confucian literati a perpetual dialogue on the scope and execution of the sovereign’s authority over the recruitment and management of officials. Since the definition of junzi was now fixed, this conversation soon diverged along two lines, each following one of the two referents of the term xiaoren. On the one hand, in accordance with the Spring and Autumn period meaning of the term, the petty man could be a failed or apostate junzi, meaning in Song a technically qualified literatus whom other literati for some reason marked as morally unfit for junzi status. In this case, the distinction between junzi and xiaoren took place within a supposedly unified cultural group, all of whose members recognized each other as literati. Thus, Ouyang’s injunction to the emperor that he must “distinguish between junzi and xiaoren” was a charge that he must discern the true Confucian literatus from the false one. And there soon arose a plethora of petitions that the emperor do so, along with extensive advice on how to undertake this evaluation. On the other hand, since many Confucian literati held that most members of non-literati groups behaved as xiaoren, these groups could be classified en masse culturally and politically as xiaoren. The latter topic soon evolved into a debate about whether “virtue” or “talent” should be the preferred qualification for office, a topic that will be studied in Chapter 2.

The rhetoric of distinction that arose in the Qingli period permanently altered the framework of Song political discourse. Writing in the Southern Song, for example, Luo Dajing 羅大經 (jinshi 1229) praised the opinion of Lu Jiuyuan 陸九淵 (1139–1192) that the world was best treated as one large family. Thus, in antiquity, Shun and Confucius had settled conflicts between junzi and xiaoren as “family matters.” Luo held that this approach had persisted into the early Song, “when the terms junzi and xiaoren were not in use; it was as if we were one family, and junzi suffered no disasters. But after these terms first arose in Qingli, there was now a division between my family and his family, and junzi suffered from disasters, each more severe than the last.”Footnote ¹³

The rhetoric of junzi and xiaoren entered immediately into the vocabulary of personnel management. For example, in 1048/5, the censor He Tan 何郯 (1004–1072) framed his argument that Emperor Renzong should dismiss Xia Song, then BMA commissioner, as an example of the sovereign’s duty to distinguish junzi from xiaoren. He began his memorial by correlating junzi and xiaoren with other bipolar pairs such as “forthright/skewed” and “order/chaos.” Wise rulers learned long ago, he wrote, that junzi and xiaoren will always exist, but combining them together in the same administration has always resulted in failed governance; therefore, good rulers distinguished between them by advancing the former and demoting the latter. The memorial then launched into a description of Xia’s character and bureaucratic career that pegged him as a stereotypical xiaoren – shallow, avaricious, duplicitous, and conniving. Specifically, He Tan charged that Xia Song had conspired with his long-term eunuch ally, Yang Huaimin 楊懐敏, chief of the capitol police, to cover up responsibility for a recent breach in palace security. Although this specific accusation is the most concrete argument for Xia’s dismissal, He makes this instance of collusion representative of Xia’s entire career. The charges of moral depravity first made against Xia in Cai Xiang’s private poem of 1036 have now entered the official language of policy discourse. Lastly, He’s memorial exercises what will later become a major motif in the rhetoric of distinction. The intersection of the two definitions of xiaoren as self-serving profit seekers presumed that since personal benefit motivated both apostate literati (here Xia Song) and non-literati groups (here the eunuch Yang Huaimin), therefore the two groups will naturally make common cause with each other.Footnote ¹⁴

By the next decade, Confucian thinkers, especially the Qingli reformers who had been rehabilitated during the late Renzong reign, had coordinated these rhetorical motifs into the basis of a robust and comprehensive argument for how and why the sovereign should not appoint xiaoren to higher office. For example, in 1056/11, Ouyang Xiu, now a Hanlin academician, opposed the appointment of Jia Changchao 賈昌朝 (998–1065) as BMA commissioner. Although Jia’s father had passed the jinshi in 992, his son was granted jinshi status by imperial decree in 1017 as reward for submitting a poem that extolled new suburban sacrifice rituals.Footnote ¹⁵ Jia Changchao then rose to become chief councilor during 1040s, when he thwarted the Qingli reforms. Ouyang’s memorial began by noting that “literati public opinion” was in an uproar over the appointment. He attacked Jia as a mediocre scholar whose duplicitous ways and clever words had harmed good officials during his prior term as councilor. Many officials, Ouyang reported, feared Jia’s return. He faulted Renzong for failing to understand a vital element of the appointment process, namely the need to observe who supports and who opposes an appointment. He urged the emperor to advance appointments that junzi approve and xiaoren oppose, and to reject those that junzi oppose and xiaoren approve. How should one distinguish between the two groups? “The junzi stands firmly at court, speaks out forthrightly, does not flatter the ruler nor collude with powerful ministers; and so all praise his integrity and loyalty. These officials oppose Jia Changchao. Eunuchs, female officers, and close retainers are largely xiaoren. They all support him. From this Your Majesty may know what kind of man he is.”

Ouyang argues that Jia has colluded with eunuch allies inside the palace who have lobbied the emperor on his behalf. Ouyang maintained that Renzong, without any formal input from those who oppose the appointment, knew only what his retainers told him on the issue. Ouyang was careful not to suggest that the emperor had formally consulted with his inner retainers, but rather that they had utilized their constant proximity to the emperor to praise Jia and to lobby indirectly for him. “A word in the morning, a mention in the evening,” and Renzong had come to think that Jia was the right man for the position. As a result, the appointment was made via inner directive, by-passing the routine chancellery process that would include procedures for incorporating opinion from line officials (on this distinction see Chapter 4). Ouyang warned that the Censorate was already preparing to charge Jia and his eunuch allies with a variety of crimes, once he should take office. He concluded by urging Renzong to “reject the secret advice of your close retainers and accept the public opinion of the literati.”Footnote ¹⁶

Southern Song historians and anthologists rightly understood Ouyang’s memorial against Jia Changchao as an important milestone in the evolving conception of Confucian governance. If, on the one hand, Ouyang backed off from his earlier perilous defense of junzi factions, on the other hand, he had now further refined a contemporaneous definition of the junzi and how the junzi should interface with imperial decision making. The essay on factions, written in 1044, wrote simply that the junzi “served the state.” The 1056 memorial against Jia Changchao presented a sophisticated view of how the junzi works through institutional structures to achieve that goal. As an extension of his inner moral qualities, the junzi is now one who “stands firmly at court,” speaking out forthrightly and publicly on issues. His vehicle is “public opinion” (gonglun/gongyi 公論/公議). Unlike modern meanings of the term, Ouyang frames gonglun as the collective voice of the junzi expressed institutionally through government agencies and functions that he envisions as the domain of junzi and as vehicles for incorporating their forthright opinions into imperial decision making. Chapter 5 will further explore Ouyang Xiu’s functional definition of these “ministers of Confucian learning.”

These organs are the remonstrance agencies – the Bureau of Policy Criticism and the Censorate – as well as the court academic agencies. Also vital to Ouyang’s conception of public opinion is the integrity of the regular process for generating imperial edicts through the Secretariat and the Chancellery (Menxia sheng 門下省), where the Secretariat drafter (Zhongshu sheren 中書舍人) (in the former) and the supervising secretary (Jishi zhong 給事中) (in the latter) had authority to return the document for errors or omissions. These agencies and procedures stand at the core of the “public” communication channel, to be examined in detail in Chapter 4, and constitute the functional nexus of Confucian institutionalist governance. In Ouyang’s view, all these procedures and functions serve to assist the sovereign in his primary task of “distinguishing junzi from xiaoren.” Finally, in his view, efforts to undermine or by-pass these institutions and to influence imperial decision making without regard for the safeguards of public opinion strengthened the political hand of those non-literati actors, “mostly xiaoren,” who closely surrounded the sovereign and could thus exert informal influence over him.

By the end of the next decade, the political discourse that arose from the struggles of the Qingli period had evolved into a fully developed theory and practice of Confucian governance. The advent of a new sovereign, Emperor Shenzong, only nineteen at the time of his ascension in 1067/1, presented the opportunity and the need to restate the Confucian principles of institutionalist governance as they had evolved to this point in time. Two memorials written by the elder statesman and Qingli veteran Fu Bi set forth a systematic vision of governance that emphasized to the young emperor the centrality of his role as distinguisher of junzi from xiaoren. His memorial of 1067/8 treats this topic alone; a slightly later tract from 1069 places this imperial function within Fu Bi’s larger vision of governance. Both texts were written to explain general principles to a young ruler new to his station and thus offer particularly cogent exposés of their topics.

Writing only seven months after Shenzong had assumed office, in “On the Need to Distinguish between Junzi and Xiaoren when Gathering Advice,” Fu Bi predicts that if Shenzong cannot evaluate the quality of the advice he receives then his already proven virtue of gathering a wide spectrum of advice will not necessarily lead to better governance. He builds upon Ouyang Xiu’s definitions of junzi and xiaoren, pointing out that the latter always outnumber the former, that many xiaoren have a knack for appearing to be junzi, and that these factors make an official career difficult for the true junzi. To address these problems, Fu Bi maintains that the sovereign’s sole duty is to avoid appointing junzi and xiaoren together in top positions. In order to identify the true junzi, the emperor must listen carefully to the advice he receives. Since there is often financial gain for opinion that pleases the sovereign, the true junzi is likely to be the official who risks his career to offer advice that displeases him. Proffering such advice takes courage and loyalty; only two or three officials in a thousand will elect to do so. But these will be the true junzi. Shenzong should seek out such officials, nurture them carefully, and promote them into senior positions.Footnote ¹⁷

Named chief councilor in 1069/2, but too ill to travel to Kaifeng, Fu Bi submitted to Shenzong a long memorial that combined an outline of his theory of governance with advice about how the emperor should proceed in his absence. The memorial, entitled “On Distinguishing the Straight from the Oblique,” begins with a summary of the classic analogy between the state and the human body, whereby components of the human body equate to components of Song officialdom. “The ruler is the head; the Council of State members are the arms, legs, heart, and spine; remonstrators, censors, and ministers-in-attendance who proffer counsel are the eyes and ears; and the multitude of officials in the capital and provinces are the muscles, joints, and veins.” Just as a robust organism demands that all body parts be healthy and work in coordination with each other, so does the body of the state demand harmony among its components. Fu Bi’s analogy affords outsize importance to the state councilors, equating them to arms, legs, heart, and spine. He emphasizes the need for harmony among these officials, about a dozen men who comprised the leadership of the Secretariat and BMA.Footnote ¹⁸ As Ouyang Xiu had first articulated, this harmony results not from collusion for short-term gain but as a natural result of common adherence to personal integrity and the public interests of the state. He stresses that unity among these officials sets a powerful example for the rest of officialdom and cites historical examples to show that personal conflicts among chief councilors weaken dynastic governance, just as diseases of the major organs lead to physical incapacitation.

Fu Bi invokes a series of quotations from the Classic of Changes to demonstrate the bipolar complementarity of junzi and xiaoren. Periods of misrule when xiaoren predominate have been the norm throughout history; periods of peace under junzi have been rare. Therefore, since even a single xiaoren among a leadership of junzi will spur disunity, these rare periods of good governance arise only when the ruler can eliminate all xiaoren from leadership. Once constrained in middle- and lower-level positions, the xiaoren have no option but to conform to the moral domination of junzi leadership. Fu Bi concludes that “the Son of Heaven has no official post, no set duties; his sole function is to distinguish between junzi and xiaoren and to employ or reject them accordingly.”

In response to the obvious question of how the ruler should make these decisions, Fu Bi cites the advice of Mencius that he should reject suggestions from close retainers and advisors and consider the appointment only of those who enjoy the widest possible support. But, taking nothing on faith, he must personally investigate and verify even these rare paragons before appointment.Footnote ¹⁹ Translating Mencius into Song political terms, Fu Bi explains that the proper method for obtaining informed input is to listen to “the public opinion of the entire empire.” This may include opinion from his close advisors, but the emperor cannot act on their opinions alone, nor rely only upon his own uninformed instincts. He must gather information from unbiased sources, evaluate independently, and make his own decisions. Fu Bi acknowledges that even Yao found this a difficult process, but he implies that Shenzong should avail himself of the full resources of his “eyes and ears” in making appointments to the “arms, legs, heart, and spine.”Footnote ²⁰

As Fu Bi’s two memorials make clear, by the late 1060s the rhetoric of distinction had evolved into a code language for the entire theory and structure of Song Confucian governance. It defined the principal and only duty of the sovereign (to appoint junzi to high office); it defined the ideal relationship between the sovereign and this leadership of junzi (harmony based on integrity and loyalty); and it identified the Secretariat, the remonstrance organs, and the academic agencies as the core components of such governance. So important had the rhetoric of distinction become that Fu Bi included the phrase in his final will and testament to Shenzong, an emphasis that found its way into Southern Song historiography and into Fu Bi’s official biography.Footnote ²¹ The premise that Song emperors preferenced junzi as civil leaders and thereby promoted “benevolent” governance became a narrative staple in the grand allegory of Song history.Footnote ²²

Fu Bi’s tracts, especially his memorial of 1069, were written in the shadow of the New Policies. Also in 1069, Cheng Hao 程顥 (1032–1085), using the same rhetoric of distinction, warned that the emperor could not employ both junzi and xiaoren together, for the latter would undermine the requisite commonality of purpose between ruler and servitor upon which successful administration depended. Citing the Book of Documents, he advised the young emperor to model himself on Yao and Shun and to “select servitors who share a single mind, united in virtue.”Footnote ²³ The reference to the sage-emperors Yao and Shun implies that before the emperor can exercise the discrimination necessary to effectively select officials, he himself must submit to Confucian moral training, in essence to prepare himself to recognize and then lead the self-identified junzi faction that Ouyang Xiu had earlier described.

The advent of the New Policies and their opposition intensified the political application of all aspects of the rhetoric of distinction. By 1071 Wang Anshi himself had begun to frame his opponents as xiaoren and urged Emperor Shenzong to purge the anti-reformers in order to form a cadre of officials united behind the New Policies.Footnote ²⁴ Likewise, the attacks of the anti-reformers against economic aspects of the New Policies heightened the association of xiaoren as officials intent on “profit” at the expense of “moral duty.”Footnote ²⁵ The rhetoric of distinction had indeed become, as Levine has so aptly described, a “common language” that came to divide Song officialdom. This language and this division would soon permeate not only political discourse but also historiography and endured for the remainder of the dynasty. The Confucian rhetoric of junzi and xiaoren soon became a tool of the inquisitions under Emperors Zhezong 哲宗 (1077–1100; r. 1085–1100) and Huizong, undergirded the autocracies under Cai Jing 蔡京 (1047–1126) and Qin Gui, and formed a staple of the language of daoxue politics and historiography.

A hint of these looming developments can be seen in a remarkable tract that the censor Fan Bailu 范百祿 (1030–1094) addressed to the young Emperor Zhezong midway through the Yuanyou administration in 1090/3 – “A List of Criteria for Distinguishing the Straight from the Oblique.” Fan advised the fourteen-year-old sovereign, then already demonstrating teenage impatience with his Confucian tutors, that the emperor could determine which of his officials were straight junzi and which were oblique xiaoren by observing their policy suggestions and their actions.

Fan then proceeded to map these two categories onto ten areas of official life and policy formation. The results are so mechanical they can be expressed in Table 1.1:

Although Fan Bailu’s rhetoric sometimes, no doubt intentionally, confounds the agency of the servitor and its imperial result, his message is plain enough: employing the straight produces policy outcomes different from employing the oblique. And mixing the two produces muddle. The influence of the political disputes over the New Policies is clear. The labels in Table 1.1’s leftmost column are not in Fan’s text but are my attempt to classify each of his ten distinctions. These categories combine traditional Confucian associations of the junzi as honest, virtuous, pious, and filial with Yuanyou characterizations of the New Policies as wasteful, pro-war, anti-agriculture, criminally harsh, and socially disruptive. Fan’s tract demonstrates how rapidly and thoroughly the disputes over the New Policies magnified the political impact of the rhetoric of distinction.Footnote ²⁷ According to Zhu Xi, Liu Zhi 劉摯 (1014–1081) supplied Empress Dowager Gao 高皇太后 (1032–1093) with ready-made lists of junzi and xiaoren, hoping no doubt to streamline the young emperor’s task, but Zhezong refused to consider his suggestions.Footnote ²⁸

What had begun fifty years earlier as an effort to interject Confucian morality into governance had morphed into a litmus test of political orthodoxy. When the Yuanyou period ended in 1094, the new reformist administration lost no time in applying “straight” and “oblique” classifications to purge the former Yuanyou administrators, including Fan Bailu, from office. In 1095 this new administration established the Agency for the Classification of Bureaucratic Documentation (Bianlei chenliao zhangshu ju 編類 臣僚章疏局), whose purpose was to review and classify Yuanyou period documents for potential use in judicial proceedings that removed Yuanyou officials from office.Footnote ²⁹ In 1102, this documentation was combined with another review of memorials that had been submitted in 1100 during a brief return to power of the anti-reformers, the so-called “Yuanfu period (1098–1100) submitters.” The result was a seven-tiered register divided into three “straight” and four “oblique” grades. Forty-one officials were determined to have submitted “straight” documents; 541 were deemed to be “nefarious and oblique.” Each level carried an appropriately calibrated reward or punishment. The names of the worst 120 oblique offenders were inscribed on a stone tablet erected at court.Footnote ³⁰ This stele was the precursor to the more famous “Stele Register of the Nefarious Yuanyou Faction” (Yuanyou jiandang bei 元祐姦黨碑) erected in 1106.Footnote ³¹

Both the language and the process of the decade of inquisitions from 1094 through 1104 can be found in Fan Bailu’s tract and grew from the Confucian rhetoric of distinction. Cai Jing’s preamble to the stele register of 1106 makes three points, all of which are implied in Fan’s text of fourteen years earlier. First, the faction list is a product of the imperial responsibility to demarcate clearly between good and evil. Cai Jing twice refers to this concept, each time using a quotation from the Book of Documents, where the sovereign “signalizes the good and separates the bad from them” and “distinguishes the good so as to cause ill for the evil.” Second, the list of 309 was obtained after a careful review of the official writings of a much larger group of officials. And third, since the oblique will be removed from office, the stele is erected as a warning to future generations of officials of the need to preserve the commonality of purpose between ruler and servitor.Footnote ³²

Politicians of all persuasions in the Southern Song continued to utilize the rhetoric of distinction and appeal to its ensuing premises for political order. Even ministers subsequently judged anti-Confucian such as Qin Gui, Cai Jing’s successor in the “lineage of evil,” for example, when encountering in 1139/7 stiff resistance to his peace negotiations with the Jurchen, remarked to Emperor Gaozong on the need to “distinguish between junzi and xiaoren.” Once the process was complete, he went on to observe, “the way toward good rule can be fully attained.” Qin’s use of the cliché here signaled to Gaozong that they should begin the purges of those opposed to negotiation.Footnote ³³ Later in the century, Zhu Xi and other daoxue advocates embraced the distinction, especially the injunction against “mixing” of the two types. Emperor Xiaozong’s many failed attempts to coax contending factions into functional coalitions heightened daoxue sensitivity to the issue and especially the injunction against “mixing junzi and xiaoren.” Zhu Xi wrote to the new emperor Guangzong 光宗 (1147–1200; r.1189–1194) in 1189 that previous attempts to bring together junzi and xiaoren in the same administration had always failed, because the presence of one type always drives away the other. He argued that the Yuanyou administrators failed because they did not completely purge their opponents, who then later returned to power and acted against them.Footnote ³⁴

Zhu Xi repeated this advice again in 1191 in a series of letters to the then Chief Councilor Liu Zheng 留正 (1129–1206), whom Zhu accused of pursuing a centrist administration because Liu feared being labeled a factionalist. Zhu argued that the ultimate political consequences of not purging the current xiaoren from office would be much greater that the short-term cost of enduring the factionalist label. And he again criticized the Yuanyou officials for failing to fully purge their own ranks. They had correctly marked their opponents as xiaoren, but they had been too loath to acknowledge that many of their own associates were likewise xiaoren. Thus, they had failed to cleanse their own ranks. Zhu urged Liu to make a clean sweep of his own administration, whatever the immediate political cost.Footnote ³⁵ Zhu Xi’s letters to Liu Zheng reveal not only the vitality of the rhetoric of distinction but also its continued use to describe concrete issues and real personalities. Chapters 10 and 11 will describe the contemporary referents behind Zhu Xi’s language and show how the increasing polarization of Southern Song political culture enhanced for daoxue adherents the attractiveness of this rhetoric of distinction.

Traditional Chinese historians seldom draw a straight line from the junzi and xiaoren of Ouyang Xiu’s “Discourse on Factions” to Cai Jing’s inquisitions. But, as Levine has emphasized, by the end of the eleventh century the junzi and xiaoren of all political persuasions shared a common language, and that language was the rhetoric of distinction. This rhetorical bipolarity grounded three primary self-reinforcing and circular elements of eleventh-century Confucian political culture: (1) a definition of the Confucian literatus as a member of a self-identifying and self-sustaining group of morally committed officials; (2) a political language that defined this group in relation to those it excluded; and (3) an emphasis on the primary duty of the monarch being to govern by distinguishing between junzi and xiaoren among his top administrators.

The bureaucratic factionalism that resulted from implementation of the New Policies sharpened, expanded, and integrated these elements. A touchstone passage was Sima Guang’s rebuke in the Comprehensive Mirror of the Tang Emperor Wenzong 文宗 (809–840; r. 827–840) for his lament that he could easier manage rebellious governors than he could the feuding Niu and Li factions at his own court. For Sima, the remark betrayed Wenzong’s inability to distinguish junzi from xiaoren among his officials. Because the former were open and fair (gong 公), factual (shi 實), and virtuous (xian 賢), their conduct was straightforward (zhengzhi 正直). Because the latter were private and selfish (si 私), duplicitous (wu 誣), and venial (buxiao 不肖), they formed factions. But Wenzong’s inability to distinguish between the two permitted both to enter court service. And since their characters and actions were as fundamentally opposed as blocks of ice and bricks of hot charcoal, the result was bureaucratic feuding, diffusion of imperial authority, and dynastic weakness. Sima Guang’s Yuan dynasty commentator recognized immediately that the great historian was also commenting on the politics of his own time.Footnote ³⁶

It would be overly reductive simply to equate junzi with the Confucian institutionalism and xiaoren with the imperial technocracy of our models. Nonetheless, as we will explore in Chapter 8, the Confucian activism of the eleventh century brought to the center of Song political discourse a mindset that ruled out the possibility of a viable middle ground and that came to regard political compromise as moral compromise. The contrast between junzi and xiaoren fortified the Confucian penchant for bipolar thinking and dualistic oppositions. And, as the eleventh century advanced, the ancient Zhou terms came increasingly to reflect the functional polarization of the real world of Song politics. Accordingly, the political rhetoric of junzi and xiaoren became fundamental to the historiographical development of all three components of the grand allegory: the junzi of Song sought to bring about benevolence governance; for this reason, the Song founders chose junzi as political leaders; but a lineage of evil xiaoren thwarted these intentions