Civil and military officials

It is common to speak in English of a “Sung civil service system.” But Sung officialdom differed in major ways from modern systems of professional civil service. First, the Sung system divided officials into two broad categories: civil (wen-kuan) and military (wu-kuan). The English term “Sung civil service” applies to all Sung officials, both civil and military, not just to the civil-side, or wen, officials. Second, Sung officials were not full-time employees in the modern sense. Most officials only spent about 50 percent of their careers in functional positions (ch’ai-ch’ien – often translated as “commission”). Half their time was spent actually working for the state, but a convoluted and time-consuming process of reassignment consumed the other half. This system generated long periods of downtime and also provided for sinecures between functional positions. The lengthy periods of voluntary and involuntary time off help explain the extensive and varied nonofficial activities of Sung bureaucrats. Some used this time to produce the copious amounts of literature and scholarship for which the period is renowned. Others devoted themselves to private business ventures that enriched themselves and their families. Many did both.

For the purpose of this chapter, “officials” will be defined as persons who held p’in (grade, rank, level) in the Sung personal ranking system for government employees. Such officials were called kuan-yüan in Sung parlance, a term usually rendered into English as official, functionary, mandarin, or bureaucrat. Sung society sharply distinguished between these graded officials and the lesser categories of ungraded government employees such as clerks and village officers, who in English are often referred to collectively as the “sub-bureaucracy.”⁷⁸ The Censorate, in conjunction with other relevant central government agencies, kept an “official register” (pan-pu) of officials on active service. The register was updated quarterly and included all graded officials who either held office or were qualified to hold office. It excluded those who were retired, in mourning, or “disenrolled” (ch’u-ming) for misdeeds. Officials received formal “patents of office” (kao-shen) for each functional position to which they were appointed. Variously colored official robes (kuan-fu) visibly marked their status as officials as well as their position in the hierarchy (purple the highest, scarlet the middle, green the lowest). More importantly, government census records designated families headed by an official as “official household” (kuan-hu). This status carried significant financial and legal advantages, including reduction or remission of certain tax obligations, the right to use legal proxies that buffered officials from normal court proceedings, and immunity from corporal punishment.⁷⁹

Formal court audience protocol emphasized the basic division of Sung officialdom into civil and military. As the emperor sat in formal audience and faced south, civil officials stood to his left on the east and military officials stood to his right on the west. Two separate systems of personal rank (kuan-p’in) – with different numbers of ranks and different names for the ranks in the two systems – also reinforced this division into civil and military. Furthermore, in the Ministry of Personnel, the appointment process for the two groups was divided into a “left selection” (tso-hsüan) for civil and a “right selection” (yu-hsüan) for military officials.

Each side, civil and military, was also divided vertically, by hierarchy, into two broad divisions. On the civil side, the lower division was called hsüan-jen (selection men), because the Ministry of Personnel determined their appointments through a “selection” process to be described below. Most scholars who write in English call this group “executory officials.” The upper division of civil officials was in Chinese ching-ch’ao kuan (literally, “capital and court officials”), and these are known in English as “administrative officials.” To be promoted out of the “executory” into the “administrative” division, an official had to undergo a process known as kai-kuan (change in official status), a lengthy bureaucratic ordeal that could take up to ten years to accomplish. Administrative officials were no longer subject to “selection” but “shot for vacancies” (she-ch’üeh) in a separate Ministry of Personnel process that was less demeaning and dilatory. Although kai-kuan was not as vital for military officials, their side too was divided into a lower division called hsiao shih-ch’en (servitors minor) and a higher division called ta shih-ch’en (servitors major).

There was a correlation between an official’s personal rank and the type of functional position for which he qualified and to which he was likely to be appointed. Administrative-class officials staffed the supervisory positions in the central administrative agencies in the capital, including all supervisory posts in the Three Departments. In the provinces, they served in top positions as circuit intendants, prefects (chih-chou), and controller-general (t’ung-p’an). Executory officials were mostly lower-level provincial officials such as county magistrates (chih-hsien) and county sheriffs (hsien-wei). In the capital, servitors major held posts related to court ceremony and security, and acted as guards, attendants, and ushers. Eunuchs were often appointed to servitors major positions, where they maintained the emperor’s storehouses and financial accounts and ran the secret service. Both at court and in the provinces servitors major also formed an elite officer corps for the army. In the provinces, they might also hold provincial positions as circuit intendants or military commissioners. Servitors minor were the bottom of the official ladder. They served as county police officials and inspectors of local militia, and staffed the local offices of the government monopoly bureaus that sold wine, tea, and salt.

These divisions of Sung officialdom arose from the dynasty’s origins among the militarized states of tenth-century China. These regimes were direct descendants of the independent provincial military governorships (chieh-tu-shih) of late T’ang. When these magnates made appointments, either to their own headquarters or to local monopoly shops, they used military titles under their own command and control structure. If one of these provincial magnates eventually claimed title as Son of Heaven (T’ien-tzu), he then laid a smattering of old T’ang civil titles for central government functionaries over his administration’s existing military structure. This practice continued in early Sung. But as T’ai-tsung’s policy of recruiting civil officials through the examinations began to bear fruit, and, as the civil structure of central administration began to cohere in the reign of Emperor Chen-tsung, the numbers of civil and military officials assumed a proportion they would retain for the remainder of the dynasty.

Military officials were not professional soldiers as such. Although they might serve tours of duty as army officers, their next position might not be with the army.⁸⁰ Rather they were officials whose p’in grades were in the military rather than the civil personal rank system. Wang Ying-lin insists that the dynasty maintained no prejudice toward either side and that officials moved back and forth between both systems.⁸¹ An equal number of civil and military officers, for example, stand as stone sculptures that line the approach to the tombs of the Sung emperors. However, a considerable disparity of perceived prestige did develop between the two systems. When literati, who were mostly civil officials, came to dominate the central administration early in the eleventh century, the functional positions for military officials tended to concentrate on security, finance, and low-level secretarial and accounting work. The two systems therefore developed different cultures. A majority of Sung officials in all four divisions came into office through yin privilege. However, an analysis of the entry methods of those who entered through avenues other than yin privilege reveals a stark difference between the civil and military sides. On the one hand, most civil-side officials who did not use yin privilege entered through the examination process. On the other hand, imperial clansmen, army transfers, and transfers from the clerical service occupied an equally strong minority position among non-yin military officials.

Over the course of the eleventh century, Sung officialdom developed a structure in which administrative-class civil officials occupied most positions of power, surrounded by support staff drawn from the ranks of military officials. There was an enormous social distance between the two groups. One often reads in the biographies of Sung officials that a high-ranking central-government official was transferred as punishment to a provincial office of the wine monopoly. In this case, the personal rank of the official might drop slightly, but his functional position would plummet from one normally held by senior civil administrative officials to one normally held by military servitors minor. One would be hard-pressed to find a parallel in American government, but a situation where an Assistant Secretary of State might be transferred to serve as a rural county sheriff would approximate the social distances between the top and bottom of these two divisions of Sung officialdom.

The question how many Sung officials there were at any given time is complex. The “official registers” do not survive. However, a variety of historical sources quote numbers from the “official registers” and provide scattered statistics for various time periods and different divisions of officials. Table 1 contains a representative sampling of these statistics, arranged according to the four classes of civil and military officials discussed above.⁸²

Table 1. Numbers of civil and military graded officials

^a Sources, in addition to standard abbreviations used in this chapter, are:

1. Li – Li Hsin-ch’uan, Chien-yen i-lai ch’ao-yeh tsa-chi.

2. Hung – Hung Mai, Jung-chai ssu-pi.

3. Han Piao – Han Piao, Chien-ch’üan jih-chi (TSCC ed.) 1, p. 3.

4. HCPSP – Huang I-chou et al., Hsü tzu-chih t’ung-chien ch’ang-pien shih-pu (Peking, 2004)

^b Figures in square brackets are not original and were derived from adding the separate totals for the four classes of official.

The numbers reveal several trends and patterns. First, from the late tenth through the middle of the eleventh century, the total number of officials doubled; then doubled again by the end of the Northern Sung. It decreased in the early Southern Sung, but rose to a new high in the late twelfth century, before declining slightly and remaining constant to the end of the dynasty. Since Southern Sung was physically smaller than Northern Sung and Sung administrative geography remained largely fixed (a population increase engendered neither a jurisdiction division nor an increase in staff), the Southern Sung figures represent a more drastic increase over late Northern Sung than the gross numbers suggest. Also, as will be explained below, the gap in data from 1127 through 1165 is probably related to political developments during this period. Second, the ratio of civil to military officials remained constant from the middle of the eleventh century to the end of the dynasty, with military officials slightly outnumbering civil officials. Third, within both civil and military categories, the ratio of lower- to higher-division officers varied greatly from period to period, with the rough average being about one higher-division official for every three in the lower division. Lastly, one must emphasize that these numbers represent only the number of individuals that were qualified to hold office. The number of actual, functional positions for them to occupy – the number of Sung officials on the job at any one time – probably never increased much above 16,000.

As noted above, based on the method of entry, the Sung civil service was more broadly based than that of T’ang. For the T’ang, the Hsin T’ang-shu (New T’ang history) preserves statistics on those who attended the yearly “selection examinations” (hsüan-chü), a process that qualified a candidate for appointment to a functional position. Forty-five percent qualified to attend because they had completed a course of study at a school, 40 percent by virtue of prior military service, less than 10 percent through yin privilege, and about 5 percent were clerical transfers.⁸³ For the Sung, Li Hsin-ch’uan preserves statistics on the total number of officials in 1213 and adds information that reveals by what entry method these officials qualified for their first position as a graded official. In the administrative class of civil officials, over 50 percent entered through yin privilege and 40 percent through regular examinations. In the executory class, almost 40 percent entered through yin privilege, 25 percent through regular examinations, and 30 percent through “facilitated degree examinations” (t’e-tsou ming).⁸⁴ Among the military classes, 44 percent of the servitors major entered through yin privilege and a third entered as transfers from the army. Of the servitors minor, half came through yin privilege, a quarter were imperial clansmen, 10 percent came as army transfers, and another 10 percent as clerical transfers.⁸⁵

Methods of entry

Li Hsin-ch’uan’s numbers confirm that 57 percent of graded officials on active duty in 1213 had entered the service through yin privilege. The graph yin means “to shelter, to cover, to protect.” A tradition dating to the Han period, “protection privilege” refers to the Chinese practice whereby the emperor granted to certain high officials the right to “protect” their families by directly appointing designated sons and grandsons to office. Although it was used sparingly in the T’ang, the early Sung rulers, as in so many other areas, took this moribund institution and transformed it into a major instrument of Sung policy. The Sung system of yin privilege, together with the growth of literati examination culture in the early eleventh century, transformed the character of Chinese officialdom.⁸⁶

T’ai-tsu, as the Sung founder, laid down basic rules that governed the use of yin privilege. T’ai-tsung, in order to staff the bureaucracy of his growing empire, so liberalized these policies that, by the reign of Chen-tsung, the system had already produced a drastic oversupply of officials.⁸⁷ As a result, the state attempted to impose tighter restrictions on who could designate and who could receive yin privilege. But, in one of the enduring contradictions of the Sung state, the few officials who warned against the abuses of the yin system were the very officials whose many colleagues benefited most from it. In 1043, a scale-back of yin privilege was the second of Fan Chung-yen’s ten proposals for government reform, but opposition to this item undermined support for the entire package.⁸⁸ Ninety percent of the sons of Northern Sung chief councilors entered officialdom through yin privilege, and only 10 percent through the examinations.⁸⁹

The Sung system of yin privilege granted to officials above a certain grade the right to designate not only sons and grandsons but also brothers, the sons of brothers and sisters, and, in some cases, even unrelated persons for “protection.” The recipients received an immediate personal-rank grade (and immediate “official-household” status), which then qualified them to compete for a functional position. The recipients thus bypassed the highly competitive and time-consuming examination process for civil officials or the active army service required for transfer to graded status for military officials.

The system, like virtually all aspects of the Sung civil service system, was based on a sliding scale. The higher the designating official, the more recipients he could appoint, the greater kinship distance from him the recipients could be, and the higher the rank grade his recipients would receive. For example, according to regulations in effect at the end of the Northern Sung, a chief councilor, at grade one of the personal-rank system, could appoint a son to grade twenty-eight, a grandson to grade twenty-nine, nephews and “clients” to grade thirty-six. But a lower official, perhaps a vice director (yüan-wai-lang) in one of the Six Ministries (Liu-pu), at grade twenty-two where the right to dispense yin began, could appoint sons, grandsons, and nephews only to grade forty.⁹⁰ These differences were considerable. Since administrative class began at grade thirty, the sons and grandsons of chief councilors began their careers as administrative-class officials, bypassing executory class and the kai-kuan hurdle altogether. However, their nephews, appointed to grade thirty-six among the lower echelons of executory class, were not so fortunate. At the bottom of the scale, recipients of grade forty held pre-executory rank and were not yet eligible for appointment to any functional position.

The policies allowed officials to name yin recipients on the occasion of the triennial suburban sacrifices (nan-chiao), on the emperor’s birthday, upon their own retirement from office, upon their own death through final testaments, and upon other special occasions such as the ascension of a new emperor or the naming of a new empress (huang-hou). Regulations concerning who could designate how many recipients on each of these occasions varied greatly during the dynasty. The surviving statute books from the Ch’ing-yüan era (1195–1201) indicate that a chief councilor could designate ten recipients at the suburban sacrifices, three upon his retirement, and five in his will. It appears therefore that a single long-lived, highly successful official could designate dozens, if not hundreds, of recipients during the course of his career. Furthermore, empresses, consorts, imperial princes, and clansmen (tsung-shih) were, as we shall see below, also high-ranked officials and major dispensers of yin privilege. Many literati placed blame for the excesses of the yin privilege on appointments to military rank for imperial relatives.⁹¹ Yet abuse of the system was widespread. Even if a dispenser had already designated all his eligible kin as recipients, social and financial pressures from distant kin and unrelated friends often pushed officials into questionable use of their remaining quotas. For example, a critic reports that each suburban sacrifice in the 1130s generated 4,000 yin nominations.⁹² Emperor Hsiao-tsung, who himself attempted an unsuccessful reform of yin privilege, well understood his official’s preference for family over state.

A full discussion of the Sung system of official recruitment through examinations is covered elsewhere in this volume and also in excellent, widely available secondary studies.⁹³The most important, socially and politically, of these examinations were the chin-shih (presented scholar) examinations. These were usually held biennially in the capital before 1070 and triennially after that date. Averaged on an annual basis over the course of the entire dynasty, these examinations produced about 200 graduates per year. The actual numbers ranged from less than a dozen early in the dynasty to 987 in 1226.⁹⁴ Those who passed the chin-shih examination entered officialdom “with formal qualification” (yu ch’u-shen). This designation opened career tracks unavailable to all those who attained graded status “without qualification” (wu ch’u-shen), largely those who entered through yin privilege. This distinction was fundamental to all aspects of the civil service. Those “with qualification” entered the system with higher personal-rank grades, received faster promotions, suffered less downtime between functional positions, and were given preference in the competition for better positions. As a result, the top decision-making echelons of government were attainable only to those who “had qualification.”⁹⁵

Yet even chin-shih graduates were not all created equal. Those who passed the examination were subsequently ranked and divided into five classes at the ensuing Palace Examination (tien-shih or yü-shih), conducted in the presence of the emperor. In the Northern Sung, the top three places received immediate administrative-class status, jump-starting their careers by at least ten years.⁹⁶ For example, during the reign of Jen-tsung, the chin-shih examinations were given thirteen times, and 4,570 candidates passed. From the total of thirty-nine men who took the top three places in these years, all but five eventually rose into the highest ranks of government.⁹⁷ On the other hand, as the oversupply of graded officials increased in the Southern Sung, steps were taken to lengthen entry into functional positions even for chin-shih graduates, and those who placed in the fifth class had to wait three years for their first positions.⁹⁸

The facilitated-degree examinations were also a significant entry method, especially in the Southern Sung. Facilitated degrees were awarded to those who had failed a minimum of five times to pass the regular chin-shih examinations and were over fifty years of age. They were given an easier version of the palace examination and also ranked into five classes. The higher classes obtained low-level executory positions, usually as provincial educators.⁹⁹ However, by the time they were appointed, these scholars were too old to accrue the time required for promotion to administrative class. The roster of 1213 pegs them at 30 percent of executory officials, but only 2 percent of administrative officials.

There were also other ways to become a Sung official. Transfers from the army into graded military positions were possible under a number of conditions, including as a reward for military accomplishment or after retirement from the regular army.¹⁰⁰ One could also purchase office by making a financial “contribution” (chin-na). Two percent of civil officials and 1 percent of the military officials on the 1213 roster attained office in this way. Such offices were always low-ranking and often ungraded, and prospects for promotion were not good. The government usually resorted to the sale of office only to raise cash for local or national emergencies. Surviving price lists from the early Southern Sung indicate that the highest prices purchased only the lowest executory or servitors minor rank.¹⁰¹ In 1180, Chu Hsi proposed the sale of similar offices to raise money for famine relief.¹⁰² Wealthy families took advantage of these opportunities to attain “official-household” status and to facilitate their interaction with local officials.¹⁰³

The roster of 1213 thus presents an image of the full Sung civil service. Contrary to what some secondary literature suggests, that image does not depict a homogeneous group of Confucian educated “literati.” Rather, the totality of Sung officialdom was a heterogeneous mixture of different cultural and social strata, each with its own criteria for admission to office, aspirations for accomplishment, and prospects for career advancement. The examinations played a role only for civil officials, and only top administrative-class officials were expected and required to possess a high standard of cultural literacy. A travel diary composed by Lu Yu (1125–1210) in 1170 provides a vehicle to confirm official statistics concerning the number of officials who had passed examinations. In his Ju-Shu chi (Diary of a journey to Szechwan), Lu Yu records the full title of each of the graded civil officials he met on a six-month journey from Shao-hsing (Yüeh-chou) on the east China coast to Kuei-chou (Ching-chiang fu) in Kuang-nan-hsi. Only eight of thirty administrative-class officials, or 27 percent, had passed an examination. For executory officials, the figure was 32 percent.¹⁰⁴ For military officials, promotion guidelines often required only that they be able to write out their family biography and read legal statute books. Two factors served to unite these disparate elements into an effective administrative force. First, a unified system of personnel management, especially after 1082, linked all officials together with a common set of procedures for personnel decisions. Second, intense competition for functional positions and for promotions, especially for civil officials, ensured that these procedures were widely understood and, in the main, rigidly enforced.

The personal-rank system

Personnel management in the Sung utilized a dual ranking system. There were two separate hierarchical structures, one that ranked the officials personally (chi-lu kuan, literally “stipendiary office,” sometimes also called “titular office”), and another that ranked the functional positions they sometimes occupied (tzu-hsü). The two systems interacted to form a sophisticated system of personnel management, a system whose leading modern student has called it “a crystallization of subtle thinking reminiscent of Sung ceramics.”¹⁰⁵ The personal-rank system, one for civil and one for military officials, provided each official with a grade, or after 1082 a “rank” (chieh), in a hierarchical structure that simultaneously ranked all officials. At any given point in his career, an official’s p’in grade fixed many outward manifestations of status and ranked him in relation to other officials. It determined, for example, where he stood at formal court ceremonies, the color of his official uniform, and the size of his funeral. More importantly, as we have seen, it determined the extent of his yin privilege. On the one hand, since an official always had personal rank – it is what made him an official – status and benefits from personal rank were permanent. On the other hand, benefits from a functional position ended upon termination of that particular tour of duty. The personal-rank system therefore provided for a certain standardization and unity of procedure among all Sung officials.

The Sung personal-rank system was descended from the Five Dynasties practice of using titles of empty functional positions in the former T’ang central government to rank officials for the purposes of protocol and salary.¹⁰⁶ This practice became regularized in the early eleventh century into a hierarchical list of several hundred titles. These titles were eventually grouped and ordered into sixty-six steps that separated the lowest from the highest graded official. Yet promotion need not necessarily proceed one step at a time on this ladder. An examination graduate could theoretically advance to the top in thirty-six promotions.¹⁰⁷ Because of the number and redundancy of the titles (some steps contained a dozen essentially equal titles) and because, as Sung central government grew, many titles, especially in the Secretariat and Censorate, were no longer “empty,” there was continual pressure for reform. The Yüan-feng reforms of 1082 dispensed with the use of functional titles for personal rank and resulted eventually in a simple system of thirty-seven steps for civil and sixty steps for military officials. Each step was designated by a single “rank title” (chieh-kuan). As completed and regularized in the ensuing years, this revised personal-rank system remained in use until the end of the dynasty.¹⁰⁸

Table 2 lists the thirty-seven personal-rank titles for graded civil officials, as these became finalized in 1117. The Arabic numbers that equate to each grade (and which are used in this chapter), from grade thirty-seven (T’i-kung lang) at the bottom to grade one (K’ai-fu i-t’ung san-ssu) at the top, are an arbitrary device to help Western readers understand the system. Primary sources after 1082 always refer to these thirty-seven steps by their names, not by numbers. Numerical p’in rank was used to indicate the hierarchy of the functional pos-itions that most often corresponded to these steps on the personal-rank system (often rendered into English as 4a, 6b, etc.). As noted above, the most basic division was between executory class (ranks thirty-seven through thirty-one) and administrative class (ranks thirty through one). It is useful to think of the administrative class, as its creators did, in terms of a further subdivision into six groups of titles. These groupings represent a survival from the older, pre-1082, system, where each step of the promotion ladder was composed of a group of largely equivalent titles. The boundaries between these six administrative class subdivisions in Table 2 were real demarcators of power in Sung China. As we have seen, the power to dispense yin privilege began at grade twenty-two, the first of the vice director grades. With grade fourteen, the first senior director grade, came the power to place yin recipients into administrative class. To serve on the State Council required minister-in-attendance grade; that is, grade eleven or above.

Table 2. Personal-rank grades: civil-administrative and executory-class officials^a

^a This table combines data from Kung Yen‐ming, Sung‐tai kuan‐chih tz’u‐tien (Peking, 1997, 2008), pp. 568–74, 688, and 710–11, which rely, in turn, largely on SS, 169, pp. 4051–4, and 171, pp. 4110–11. Ranks in parentheses were added in 1109. Figures for monthly salary represent base salary in strings of 1,000 cash.

Table 2 also shows the numbers of officials appointed to administrative-class grades as recorded in the “official register” for autumn of 1085.¹⁰⁹ There may be gaps in these statistics, but they give a sense of the overall distribution of administrative-class officials across the various rank titles and groups. In addition to an obvious thinning of the numbers as one progresses into the upper ranks, the numbers also reveal important divisions between groups in the middle ranges of the administrative class. The distinct bulge in the court official group (grades twenty-five through twenty-three), with a total of 1,091 officials, the largest of any group, reveals the real promotion barrier between grades twenty-three and twenty-two. The wide disparity between directors (grades nineteen through fifteen) with a total of 160 officials and vice directors (grades twenty-two through twenty) with 619 officials also reveals the importance and the difficulty of promotion above grade twenty. These patterns formed because as early as 1066 the state effectively placed quotas on the number of officials who could be appointed to each group. Promotion across these major boundaries into the above group thus became more difficult.

Promotion

The Sung system of promotion in personal rank was called mo-k’an (promotion review), meaning literally to “to grind and examine.” The term derives from the rigorous scrutiny of documentation submitted by officials who requested promotion under the T’ang.¹¹⁰ Under the Sung, mo-k’an developed into a sophisticated, and perhaps the most important, mechanism for the personnel management of officials. In the early Sung, all graded officials received an automatic one-step upgrade in personal rank by act of imperial grace during the suburban sacrifices every third year. By 1007, this process had evolved into a separate performance review for each official, conducted at three-year intervals for civil and five-year intervals for military officials. Administrative-class officials submitted the required documentation to the Bureau of Personnel Evalu-ation (shen-kuan yüan). This documentation included patents of appointment for their present and past functional positions and promotions, a biography, annual evaluations, and, when required, recommendations. In 1066, after several failed attempts to slow down the promotion process, these intervals were increased to four and seven years respectively and “holds” were placed at vari-ous levels of the hierarchy by instituting quotas for certain grades. Officials who were qualified for promotion to these grades had to queue and wait until vacancies occurred to assume their promotion. The Yüan-feng reform revised these provisions slightly, abolished mo-k’an for officials above grade ten, and consolidated all reviews in the newly established Ministry of Personnel. As revised again in the 1130s, these regulations remained in force until the end of the dynasty.¹¹¹

Theoretically, the process of promotion review entailed two components: a review by the relevant central government authority of the yearly perform-ance evaluations written by the official’s immediate superior and a verification of the years in service necessary to qualify for the promotion. Every Sung official maintained his own dossier (yin-chih, literally “stamped papers”). Whenever he held a functional position, his immediate superior was required annually to annotate the dossier with a brief evaluation of his performance for the past year. One of these annual evaluations (k’ao) counted as one year of service credit towards the next mo-k’an review. Already by the middle of the eleventh century, however, these “evaluation scripts” (k’ao-tz’u) had become so standardized that they became useless for actual evaluation. The annual evalu-ations became a simple measure of years served. As Fan Chung-yen lamented, “now three evaluations means a promotion, and that’s what we call ‘performance review.’”¹¹² Hung Mai (1123–1202) has a good note on the decline of the annual evaluation system and quotes from evaluations written by the father of Huang T’ing-chien (1045–1105), who served as a provincial official in the mid-eleventh century. Hung praises the detail of these evaluations, but notes that Huang always gave his subordinates an “average” (chung) rating. Subsequently, the phrasing of the evaluations became so standardized that the annual annotation of dossiers was consigned to clerks.¹¹³ In 1091, Fan Tzu-yü (1041–98) complained that the annual evaluations had become useless: “The present Ministry of Personnel rankings of superior, average, and inferior are mere verbiage: there is neither reward to encourage nor punishment to dissuade anybody.”¹¹⁴

There were, however, several ways to receive service credits toward mo-k’an. The number of actual years served in a functional position, as measured by annual evaluations, was called “years of toil” (nien-lao). But actual service was only one way to accumulate service credit. Imperial acts of grace (ta-she), as in the beginning of the dynasty, routinely awarded service credits to all officials. The emperor also could specifically reward any meritorious official by conferring service credits toward his next scheduled mo-k’an review. And the emperor personally controlled promotions for his highest officials, since the mo-k’an process was considered too demeaning for ministers-in-attendance. Also, officials who could no longer themselves count service credits toward their own promotions because of the so-called “laws of halt” (chih-fa) were allowed to transfer those credits to their offspring or even to their deceased parents, since in Sung government even dead officials still maintained personal rank.¹¹⁵

Most importantly, mo-k’an credits were often either given as inducements and rewards or taken away as punishment. For example, in 1206 those willing to accept functional positions in outlying border areas were offered two years’ reduction in mo-k’an.¹¹⁶ Such reductions were highly attractive as inducements. Remarking to Wang An-shih on the decline in literati values over the course of his lifetime, Su Shih joked, “Today’s chün-tzu would kill for a half-year’s reduction in mo-k’an.”¹¹⁷ On the contrary, postponement or imposition of additional years toward the next promotion review was a common punishment for official transgression. The surviving Ministry of Personnel statutes from the end of the Sung contain a detailed table of correspondences between judicial sentences and mo-k’an postponements that range from three months to four years.¹¹⁸ Li Hsin-ch’uan cites an extraordinary example of the use of service credits to climb the personal-rank ladder. Mo Tzu-ch’un (1159–1215) took first place in the chin-shih examinations of 1196 and so received immediate administrative-class placement at grade twenty-eight. By 1203, he had been promoted sixteen times and attained grade twelve, a feat that required fifty-eight years of mo-k’an credits. But only seven of those credits came from actual “years of toil” in office. The remainder came from a combination of amnesties, merit rewards, credits transferred from his older relatives, and other inducements.¹¹⁹

Also considered a part of promotion review was the most important promotion of all, the “change in official status” (kai-kuan) from executory to administrative class.¹²⁰ The “change” was vital to any Sung civil official with aspirations to higher office, since virtually all officials who held court positions were administrative class. Even for examination graduates, failure to achieve kai-kuan meant years of drudgery in one provincial post after another. Also, given the large numbers of executory officials who entered via yin privilege or with facilitated degrees, kai-kuan served as a mechanism to separate the wheat from the chaff and to channel the truly capable into higher office. To be eli-gible for “change of status” an executory official normally required “three tours and six evaluations”; in other words, he must have completed tours of duty in three separate functional positions and accumulated a total of six years in those positions.

Unlike a usual mo-k’an review, however, kai-kuan required that the candidate collect five sponsorship endorsements from administrative-class officials on active duty in the provinces, one of whom had to be a circuit intendant.¹²¹ Strict limits were placed on the number of endorsements an official could make. A circuit intendant, allotted only five or six endorsements per year, might have several hundred executory officials working within his jurisdiction. Endorsements were not given lightly, and the competition was fierce. They were good for life and established a bond between the two men, whereby each was made legally responsible for the other’s actions (lien-tso). The requirement that an executory official collect five endorsements in different locales over an extended period of time was intended to discourage partisanship. However, the process of collecting endorsement was fraught with uncertainties and anxiety. If the endorsing official died or was convicted of a crime, the endorsement was no longer valid. In the late twelfth century, Chou Pi-ta (1126–1204) instituted a system whereby endorsements could be registered with the Ministry of Personnel upon issue and remained valid until used.¹²²

After a candidate for kai-kuan had accumulated his years in service, collected his endorsements, and survived review of his documents at the Ministry of Personnel, there was one remaining hurdle. The last step before actual “change of status” was an audience with the emperor. But already in 1065 there were 250 approved kai-kuan promotions with no positions open for them to occupy.¹²³ In 1086 a yearly quota for kai-kuan promotions was set at 100, but this limit was abandoned in the late Northern Sung. By 1117, annual kai-kuan promotions had ballooned to 370.¹²⁴ In the mid-twelfth century the quota fluctuated between fifty and ninety, and during the Hsiao-tsung period was eventually fixed at 100 per year.¹²⁵ Approved candidates, nevertheless, were still required to wait for a vacancy before they could be received in audience and appointed to their first administrative-class position.

Historically, soon after its inception, the mo-k’an performance review became a highly bureaucratic check on the accuracy of submitted documents. In 1098, Tseng Pu reminded the emperor that if the document review should determine that an official’s eligibility for mo-k’an “is just one day short, it won’t go forward.”¹²⁶ The scope and importance of the mo-k’an process encouraged a careful system of verification and control for the processing of personnel documents, and together with the memorial process, contributed both to the bureaucratic character of the Sung state and to the rich resources available to its modern historians. At least before the advent of the Tao-hsüeh critics in the latter twelfth century, many Sung officials viewed this bureaucratic and impersonal nature of performance review in positive terms. As Secretariat drafter, Su Ch’e (1039–1112) composed formal notices for mo-k’an promotions, and many of these texts survive in his collected works. A major theme is that the routine, bureaucratic nature of the process insures impartiality. Its rules apply to everyone, “eminent and humble alike.” Even promotions for imperial relatives and eunuchs are required to pass through the mo-k’an process. In drafting his notice for a military hero, Su implies that the man’s exploits would count for nothing unless they were properly credited to him through mo-k’an. The process protects both the official and the emperor from perceptions of favoritism.¹²⁷

Functional positions

The Sung system of personal rank was closely intertwined in practice with the separate system of functional rank. This latter system was based on the functional position which an official held.¹²⁸ The title of every such position had two parts: the first indicated its location and the second indicated the level of authority. For example, Su Shih’s first appointment was notary to the administrative assistant of Feng-hsiang prefecture (Feng-hsiang fu ch’ien-p’an), where Feng-hsiang designates the location of the duty, and notary to the administrative assistant designates the authority level. In titles of central-government positions, the locator is the name of the office where the duty occurs: Vice Director in the Ministry of Rites (Li-pu yüan-wai-lang). Appointment to a functional position was always for a fixed time period, which, depending on the position, might be from a minimum of two to a maximum of four years. Officials were prohibited from serving in jurisdictions where they lived, had relatives, or owned land. Also, the major portion of an official’s salary came from the function position he occupied and terminated when his commission expired.

The frequent rotation in functional position was a security and management tool that aimed to prevent officials from developing local ties or establishing too much influence over a given office. But this frequent rotation, and the impersonal nature of job assignment, frustrated the development of technical competence where it was really needed and decreased overall efficiency. For example, in 1145 the Diplomatic Office (Kuo-hsin so) of the Military Affairs Commission complained that its translators had all been reposted to commissions outside the capital. It requested that they be posted to shadow positions supervising the Lin-an city gates so they could once again be available in the capital to translate diplomatic correspondence.¹²⁹ Likewise, those few officials who took an interest in history complained frequently that the History Office was seldom staffed with qualified personnel.¹³⁰

Unlike the statistics for graded officials, statistics for the number of functional positions – the number of actual jobs available at any one time for these officials to assume – are extremely rare. The Chinese term ch’üeh (billet) denotes both a position and a vacancy in that position, or, in other words, a billet whether filled or unfilled. Documents for the year 1112 place the total number of billets at 14,000, against 43,000 graded officials – a three-to-one ratio of officials to available positions.¹³¹ A confirmation of this ratio comes from exact figures on the number of servitors minor and available billets for them in 1125: 7,086 billets and 31,082 servitors minor, or fewer than one in four.¹³² This ratio certainly represents a deterioration of the three-to-one ratio that was usual throughout the latter half of the eleventh century. Statistics from 1049 list a total of 17,300 officials and between 5,000 and 6,000 provincial billets.¹³³ A memorial from 1086 states that three officials sit on each billet: the one who presently occupies it, the one who has just been appointed to it, and the one whose appointment is now being processed.¹³⁴

The appointment process

In Sung there were essentially four authorities empowered to make appointments: the emperor, the office of the chief councilors, the Ministry of Personnel, and, under certain conditions, circuit intendants. The emperor personally appointed to functional office all officials with minister-in-attendance status (grade eleven and above). The chief councilors – more precisely the Secretariat for civil officials and the Military Affairs Commission for military officials – made appointments at the senior-director level (grades fourteen through twelve). Such appointments were called “hall appointments” (t’ang-ch’u). The Ministry of Personnel processed appointments for all officials of grade fifteen and below.¹³⁵

The actual appointment process, consolidated for most officials after 1082 in the Ministry of Personnel, was among the most complex aspects of the Sung civil service. This complexity arose from no central design but from many years of ad hoc solutions to administrative, personnel, and management problems. The most pressing problems were the drastic oversupply of officials and the need to create mechanisms that fairly rationed billets and regulated upward mobility. The appointment process matched officials with positions in a way that addressed these problems and at the same time gave officials some choice and flexibility in pursuing their careers. Four subsections (ssu) in the Ministry of Personnel processed appointments separately for officials in the administrative, executory, servitors major, and servitors minor classes. The ministry ranked all billets for each class into a tiered hierarchy and kept track of which billet was vacant or soon to be vacant.

There were five basic tiers for administrative class. Each tier was divided into a theoretical first and second tour of duty, thus creating a hierarchy of ten levels. The five basic tiers in ascending order were (1) county magistrate, (2) controller-general, (3) prefect, (4) judicial circuit commissioner (t’i-tien hsing-yü kung-shih), and (5) fiscal commissioner (chuan-yün shih). Hundreds of discrete, individual billets were grouped into each of these five basic tiers. Although the titles for each tier derived from provincial positions, central-government positions were eventually inserted into the hierarchy.¹³⁶ The result was an ordered structure for the progression of functional positions an official would hold over the course of his career. One tour of service at the lower level was required in order to advance from one level to the next. To advance from one tier to the next required, in addition to the service requirement, a number of sponsorship recommendations. For example, after two tours of duty in a county magistrate position, an official required two sponsors in order to advance to a first-tour controller-general position. The number of required sponsors varied with an official’s entry method. Clerical transfers could require an extraordinary seven sponsors for rank promotion. Promotion for administrative-class officials was a mechanical level-by-level progression; one could skip no levels. For executory-class officials, there were four tiers divided into seven levels. But one could initiate a “change of status” to administrative class from any executory level once the requirements for kai-kuan had been met. The position an official held on this tiered hierarchy was his functional rank.¹³⁷

All positions, however, even within the same tier, were not created equal. Geography played a major role. A position as “county magistrate” could vary considerably in power and influence depending on the size and location of the county. The state therefore subdivided both counties and prefectures into seven different grades using a combination of population and distance from the capital. Appointments were made on a sliding scale based on the official’s method of entry into service. Appointments into entry-level billets in the best locations were reserved for those “with qualification,” while the worst locations went to transfers from the clerical sub-bureaucracy.¹³⁸

Since there were, however, far more officials than billets, and far more billets in bad locations than in good, the Ministry of Personnel, in essence, presided over a sort of revolving auction. Officials used their functional rank and other qualifications such as entry method to bid for billets whose “price” varied according to the billet’s desirability. There were four categories of billet. Extraordinary positions (fei-tz’u) were those that filled within five days of posting; standard positions (ching-shih) filled from six days to three months after posting. After three months the position was downgraded to nonstandard (p’o-ko), and if not filled within ten days was then downgraded again to residual (ts’an-ling). Extraordinary positions were so attractive that officials whose functional rank qualified them for a higher position were nevertheless eager to take them. At the bottom of the scale, qualifications for residual pos-itions were lowered until they were filled.¹³⁹ The Ministry of Personnel also used this system to punish and reward officials by offering them preference for extraordinary billets or restricting them to nonstandard or residual. Those at the bottom of the pecking order, elderly yin recipients, clerical transfers, and those who had purchased office, were often restricted to residual positions.¹⁴⁰

The Ministry of Personnel held quarterly, and sometimes monthly, placement assemblies (chi-chu) for executory officials. Officials who had concluded their last tour submitted documentation to establish their functional rank and other relevant qualifications. Since there were many more officials than billets, officials between positions were required to queue up for the opportun-ity to attend the assembly. Already by 1058, this waiting time could extend up to two years, and by 1086 could stretch up to three years for executory officials.¹⁴¹ A coveted reward for meritorious service was a reduction in waiting time, which could advance one’s place in the queue by three months to a year. When an official’s turn finally came, he joined a group of thirty other office seekers at a ministry assembly. Those still waiting on queue were permitted to listen to the proceedings. The ministry official in charge read aloud the name of a position to which the official whose number was up had been matched. The official was free to reject this match, in which case he returned to the next assembly and was offered another option. He could decline three times, and, under certain conditions of demonstrated hardship, a fourth time, but was required to accept whatever position was offered after that point.

The procedure for administrative-class appointments, called “shooting for vacancies” (she-ch’üeh), was similar, except that the process was conducted in writing rather than verbally. Administrative officials were permitted to indicate in order of preference three circuits where they would like to serve their next position. The ministry took these preferences into consideration, matched their qualifications to available billets, and made an initial proffer. The official could reject two proffers, but was required to accept the third.¹⁴² These procedures allowed the ministry to match available positions with applicants, and still allowed the applicants a certain measure of choice. The ministry could attach a range of qualifications necessary for appointment to each class of billets, and these qualifications were reduced when the position was reduced from standard to nonstandard. For example, all open billets at a certain rank within the monopoly bureau system would be made available to administrative-class officials, who would take their pick of the best locations. The remaining positions would be downgraded to nonstandard and posted at the placing assemblies for executory officials. The system thus required officials to make wrenching decisions. Appointments made to an official who had exercised all his rejections were called “hard appointments” (ying-ch’ai) and were seldom preferable to the choices he had previously been offered.¹⁴³

There are random statistics on the numbers of officials on queue at the Ministry of Personnel and the number of open billets, and these statistics show wide variations. In the year 1111, there were over 400 administrative-class officials vying for seventy billets. In 1169, there were 500 executory officials and 340 open billets.¹⁴⁴ Yet many billets in remote locations still remained unfilled for long periods of time. Yang Wan-li (1127–1206) observed that openings for office managers in the military and judicial circuit intendancies in modern Kwangtung and Kwangsi remained open for as long as nine years. Those whose functional rank qualified them for these positions preferred to take nonstandard appointments as prefects or controllers-general in somewhat better, but still less than ideal, locations. As Yang summarized, “The high are unwilling to go; the low are unable to go.”¹⁴⁵

Criticism of the system

The statutes of the Ministry of Personnel appear to have provided flexibility both to the state and to its officials. But critics, especially in the Southern Sung, derided the process for its rigidity and its inability to match man and position in any except a mechanical way. The ever-increasing size of the statute manuals that regulated ministry procedures illustrates the haphazard growth of the bureaucracy. In the mid-1060s these manuals totaled about thirty chapters. Twenty years later, they had grown to 100 chapters. By 1149, they had ballooned to a staggering 425 chapters.¹⁴⁶ The surviving Li-pu t’iao-fa (Statutes of the Ministry of Personnel) and the Ch’ing-yüan t’iao-fa shih-lei (Classified statutes of the Ch’ing–Yüan period) are but tiny fractions of an enormous corpus of regulation and precedent that governed the lives and careers of Sung officials.

Southern Sung critics lamented that bureaucratic process had taken control of the man. Ch’en Fu-liang (1137–1203) argued, for example, that the regulations had become so complicated that the process reduced even the director and vice director of the ministry to mere signatures on documents prepared by ministry clerks. For their part, Ch’en related, the clerks merely processed papers in strict conformity with the regulations, fully ignorant of and unconcerned with the integrity, honesty, or intelligence of the official involved.¹⁴⁷ Yeh Shih (1150–1223) also decried the vast array of regulations that hamstrung top ministry officials and terrorized those who passed through the appointment process.¹⁴⁸ Stories of clerical corruption already begin to appear by the mid-eleventh century. Sources tell of clerks who withheld notice of vacancies, especially those created when the death of a parent would force an official into mourning and so create a sudden, unanticipated vacancy. The clerks would then sell the unexpected position “off the books.”¹⁴⁹

Other officials chose to satirize the clerks and the regulations. Hung Mai includes the following firsthand account of a Southern Sung personnel transaction in a note entitled “the laughable nature of clerical language.” The pacification commissioner (ching-chih shih) at Han-chou in Ch’eng-tu-fu circuit had conferred a temporary patent of nobility on a local deity and requested that the Ministry of Rites (Li-pu) make the patent permanent. The ministry denied the request on the ground that the deity had failed to appear in person within the one-year time limit for application of such patents. The commissioner was instructed to inform the deity of this rejection.¹⁵⁰

There was a subtle but important relationship between personal and functional rank. As long as an official continued to apply for and accept functional positions, he could expect to receive mo-k’an promotions in personal rank. There were many bottlenecks and roadblocks along the way, but progress proceeded at a known and relatively steady pace. Advancement in functional rank occurred with less regularity, only upon termination of a tour that could last from two to four years. By contrast, the quality of functional positions an official was able to obtain had significant impact on the rate of his promotions in personal rank. Good performance in highly visible positions generated rewards in the form of reductions in mo-k’an time. Rewards in the form of reductions in time on the ministry queue decreased downtime between appointments. As a result, officials with similar personal ranks could occupy functional positions at vastly different levels of real authority. Mo Tzu-ch’un, the superachiever who attained grade twelve in only seven years, finished his career as prefect of Wen-chou (Jui-an fu), a functional position that normally required only grade eighteen rank. One personal rank grade above Mo, at grade eleven, an official could serve as assistant chief councilor.

In 1086, Shang-kuan Chün (1038–1115) described the career of an average, upwardly mobile civil official. After passing the chin-shih examination, he entered the service at age thirty. Including downtime, he required seven years to complete one “tour” at the executory level and five years for each tour at the administrative level. And so, at age forty-five, he attained kai-kuan after two seven-year executory tours. With retirement at age seventy, he had time remaining for only five administrative-level appointments.¹⁵¹ Thus our theoretical official would have finished his career only halfway through the ten-level hierarchy of administrative-class functional positions, or as a first-term prefect, about where Mo Tzu-ch’un ended his career. Any official who rose above the level of prefect in Sung China was either extraordinarily gifted or extraordinarily well connected.

In Sung, the identification of either category of individual depended on the complex relationship between sponsored endorsements (chien-chü) and performance evaluations (k’ao-k’o). Various types of mutual responsibility, where one individual pledged surety for another (pao), were a basic aspect of Sung official life. Surety was used for minor matters, such as identity and document verification, and for major matters such as kai-kuan endorsements. Sponsored endorsements were vital to both kai-kuan and functional-rank promotions.¹⁵² The state maintained strict controls over the number of endorsements in both categories that ranking officials in any given government unit could dispense. The right to make endorsements was a basic commodity of power in the world of Sung officials. The Ministry of Personnel statutes contain detailed lists of quotas allotted to various units of government, and surviving records show provincial administrators negotiating with the ministry over quotas for endorsement rights.¹⁵³ The right to issue or to withhold endorsements from subordinates was a powerful personnel management tool for Sung officials, and a quota of kai-kuan endorsements was often included in the “budget” of new administrative units.¹⁵⁴

The role of performance evaluations is perhaps the most difficult single subject in the general area of Sung personnel management. Primary sources contain the usual wealth of data, but most scholars doubt how effectively any of the surviving schemes and systems of evaluation was ever implemented.¹⁵⁵ Although supervisors were required annually to annotate the personnel dossiers of their subordinates (k’ao), the content of these evaluations was perfunctory and not linked to the actual outcome of mo-k’an. For example, subsequent to the inauguration of revised standards for performance evalu-ation in 1061, the sources contain records of only two officials who were actually demoted for poor performance. In both cases, it would seem, the relevant agency had to memorialize the throne to fix the degree of demotion.¹⁵⁶ In short, although there were performance standards, the rigid, quantified structure of promotion review frustrated the development of any mechanism to align the given standards with actual personnel actions. Furthermore, the highly partisan nature of Sung political life also frustrated impartial performance evaluation. Li Hsin-ch’uan’s description of Emperor Hsiao-tsung’s attempt to impose real performance standards on the system highlights this problem. In 1181, Hsiao-tsung ordered circuit intendants to write annual evaluations of prefects within their jurisdictions, and he ordered the Censor-ate to monitor the fairness of the process. But both intendants and prefects so abused the system by using it to settle “private” scores that the Censorate was constantly investigating the veracity of their reports. Even verified negative reports still required consultation between the emperor and the State Council to determine an appropriate demotion.¹⁵⁷

Surviving records suggest that promotions or demotions based on actual performance were often related to ad hoc government attempts to raise cash. In other words, the state used reward and punishment as an incentive attached to specific government initiatives. For example, in 1135, performance standards offered inducements to officials who “improved tax registers, organized peoples’ militias, increased agriculture and sericulture, and encouraged filial piety (hsiao-hsing).”¹⁵⁸ The first three of these categories promoted activities that directly or indirectly increased state revenue. Often the threat for nonperformance was quite specific. In 1184, circuit intendants were ordered to include figures for shortfalls in salt quotas in the annual dossier reviews of tea and salt monopoly officials in Huai (Huai-nan) and Che (Liang-che) circuits. Those who fell more than 30 percent below quota were to be referred to the Ministry of Justice (Hsing-pu) for the assessment of fines before being allowed to attend their next placement assembly.¹⁵⁹

A final example illustrates the sudden, ad hoc nature of many Sung performance evaluations. In 1104 one decided to measure conformance to new regulations that governed the timely processing of documents through the various units of the Six Ministries. One Shih O (1058–1112), in six months and sixteen days as director of the Bureau of General Accounts (tu-chih lang-chung), correctly processed 51,015 documents and was awarded three years’ reduction in mo-k’an. Ho Ch’ang-yen (1067–1126), a vice director in the Ministry of Rites (Li-pu yüan-wai-lang), mishandled eleven documents from a total of 14,428. He was demoted one grade in personal rank.¹⁶⁰ Aside from marveling at standards of efficiency that would put any modern bureaucrat to shame, one may add several relevant facts to these figures. Ho Ch’ang-yen was the first-place chin-shih examination graduate of 1097, while Shih O was an older functionary with less renown but more experience.¹⁶¹ Also, the award-winning Shih O did not process the documents himself, but relied upon a staff of seventy clerks then attached to the Bureau of General Accounts (Tu-chih ssu).¹⁶² This works out to slightly over 100 documents per clerk per month.

Salaries

Official salaries (feng-lu) are another complex issue upon which scholars differ.¹⁶³ The bottom line appears to be that, especially in the lower and middle ranks, and especially in the years before the 1082 reform, Sung officials were well paid but not well off. Both Fan Chung-yen and Wang An-shih portray all but the top level of officials as so poor they are unable to marry off their children or bury their parents. The growing numbers of officials and the long waits between positions meant that “in seven years an official is paid only during three.” Both authors attribute pervasive under-the-table business activities and official corruption to inadequate salaries.¹⁶⁴ The Shen-tsung reforms addressed these problems and seem to have roughly doubled salaries for many officials. There was indeed a general trend toward higher compensation as the dynasty progressed. Writing in 1199, Hung Mai states that official salaries for entry-level provincial officials had risen by seven to eight times since the mid-eleventh century, a factor roughly twice the rate of inflation during the same period.¹⁶⁵ On the other hand, the eighteenth-century historian Chao I (1727–1814) describes Sung official salaries as among the highest in Chinese history and links official loyalty toward the dynasty to this generous compensation.¹⁶⁶

A major problem is that Sung officials received compensation in many forms, and the value of these compensations is often difficult to judge in modern terms. After the reforms of 1082, there were two broad forms of compensation – base salary (ch’ing-shou, literally “pure receipts”) and supplements to salary (t’ien-chih ch’ien). An official’s base salary had three components: (1) a monthly payment in strings of copper cash, with theoretically one thousand coppers to the string; (2) a monthly allotment of grain; and (3) twice-yearly allotments of various kinds of cloth. In all three categories of base salary, an official’s personal-rank grade determined how much he received. The salary figures in Table 2 represent only monthly cash payments. An official’s personal-rank grade determined his base salary, but he received that salary only when he held functional office.

Officials also received supplements to base salary. These were also paid only when an official held functional office, but the amount and form of compensation were more directly linked to the type of office he occupied. Officials whose functional positions were in the provinces received a portion of proceeds from “office lands” (chih-t’ien). These were agricultural landholdings that belonged to provincial jurisdictions; their proceeds were apportioned to the administrative-class officials assigned to that jurisdiction. The prefect of a major prefecture received proceeds from twenty ch’ing of land, almost 300 acres. The lowest-ranking officials received one-tenth of that amount.¹⁶⁷ In some cases, grain from all the office land in the jurisdiction was pooled and divided according to ratios that mirrored the acreage allotment of each official.¹⁶⁸

In place of office lands, officials at court received a supplement to base salary called “assignment office salary” (chih-ch’ien). This was essentially a cash supplement that was based on the functional office an official currently held. The amount was calculated on the differential between his personal rank grade and the rank of his functional office. For example, in 1087, Su Ch’e, then at personal rank grade twenty-two, was appointed vice director of the Ministry of Revenue (Hu-pu yüan-wai-lang). He received thirty strings of cash as his monthly base salary according to his personal rank and an additional forty-five strings for his vice directorship.¹⁶⁹ It was by this mechanism that official salaries doubled after the 1082 reform. Also counted as supplements to salary were allotments of foodstuffs and basic commodities such as tea, salt, wine, firewood, charcoal, and – that most basic commodity of the Sung official – paper. Horses were also provided, along with fodder, as well as servants. A chief councilor was allotted seventy servants and attendants. But even the lowest-ranking servitor minor had at least one servant. A system of vouchers provided transportation, board, and lodging for an official and his entourage during periods of official travel.

Another major supplement to salary was “public-use money” (kung-shih ch’ien). These were funds supplied to government units for entertainment and incidental expenses to be used at the discretion of the unit head. Such expenditures by central-government units were minor, but provincial officials, especially circuit intendants in areas with major concentrations of army units, received sizable amounts of public-use money. In the late eleventh century, such payments could exceed 50,000 strings per year.¹⁷⁰ From the modern perspective, public-use money looks suspiciously like slush funds or perhaps a suburban industrial development agency. Li Hsin-ch’uan has a long section detailing abuses of public-use money during the Southern Sung. In the provinces, a major purpose of the funds was for public-welfare projects. But officials used the funds to open pawnshops, pharmacies, and all manner of moneymaking enterprises, and skimmed off the profits for themselves. It also appears, since the funds were designated for entertainment, that neighboring officials often held banquets and then used public-use money to provide lavish gifts for each other.¹⁷¹

There was clearly an enormous difference between the highest- and the lowest-paid officials. The monthly salaries in Table 2 are deceiving in this regard. The Chia-yu era (1056–63) salary schedule ranges upward from 120 strings per month to a high of 400 strings for imperial princes on appointment as military commissioners, and to a low of 300 coppers (not strings but individual copper coins) for the lowest-ranking official.¹⁷² This differential accounts partially for the disparity in historical opinion on Sung salaries, since Chao I’s remarks focus on high-ranking officials. The appointment process, along with the vagaries of political fortune, also forced many Sung officials into cycles of economic boom and bust. Su Shih, for example, spent most of the 1070s in major prefectural posts. He earned income from 300 acres of “office land” and had at least twenty servants, in addition to his base salary. After his trial in 1079, he held no functional office, and thus received no salary, until 1085. In 1080, he budgeted his twenty-person household at 4,500 coppers, or 4.5 strings, per month.¹⁷³ To help gauge the extent of this poverty, six years earlier Wang An-shih had insisted to Emperor Shen-tsung that the lowest executory official required a minimum of 100 strings per month to support his household.¹⁷⁴

A final factor regarding Sung salaries was the high proportion of commodities to cash. Especially after 1082, an official usually received only one-third of the “cash” portion of his base salary in actual cash. The remaining two-thirds were paid in kind. Copper cash was the preferred method of payment, but because coinage was usually in short supply, salaries were often paid in whatever commodity the government had on hand. This practice required officials to sell the commodities to obtain their full cash salaries, and these sales were subject to market fluctuations in commodity prices. As a barom-eter of sensitivity on this issue, one of the many later complaints against Ts’ai Ching (1047–1126) was that he manipulated the document flow system and deceived the emperor to obtain his full salary in actual cash.¹⁷⁵

A group portrait

The above survey of civil service personnel, recruitment, and management paints a diverse portrait of Sung officialdom, more complex and variegated than the simple image of the “Confucian scholar-official” often presented in popular literature. If the goal of the Sung founders – T’ai-tsu, T’ai-tsung, and Chen-tsung – was to create a new, broad-based civil service that would support their monarchy, their success was quick and long-lasting. It may stretch the point to label this achievement as “representative” government, but the result, at least initially, was certainly inclusive rather than exclusive. The shih ta-fu encompassed a wide range of social classes, educational levels, professional training, and religious and intellectual orientation. Yes, the “literati” were officials, important and vital. But so also was the semiliterate accountant who kept the books in a provincial wine monopoly office, the merchant who purchased an office to evade taxes, the army officer who slaughtered Tibetans on the northwest frontier. Even the deities were officials, although their personnel management required special expertise and a certain latitude with regulations. Support for the monarchy and its policies was the common denomin-ator that determined initial inclusion and eventual career advancement. The greatest reward for good service was the ability to perpetuate that service through exercise of yin privilege. At any given moment, a majority of officials owed their own inclusion in the system to the prior good service of their elder relatives.

A common thread that connected the lives of Sung officials was their participation, both as exploiters and as victims, in the economic expansion during the Northern Sung and its contraction in the Southern Sung. The extensive downtime between appointments, the large core families to support, and the broad interface between government office and economic opportunity lured and forced many Sung officials and their families into business ventures. Once again, the gamut is startling. In the mid-eleventh century, Ssu-ma Kuang advocated stricter gender separation because poorer shih ta-fu families were forced to sell their daughters as concubines and maids.¹⁷⁶ Slightly later, Ho Chih-chung’s (1044–1117) daily income from his K’ai-feng rental property equaled his monthly salary as chief councilor.¹⁷⁷ Many officials led double lives as businessmen. Chu Hsi’s family had a printing business, and his letters reveal him using his students and associates to deliver orders and collect bills.¹⁷⁸ Despite the condemnation of commerce in Confucian texts, such activity was open and accepted.¹⁷⁹ The famous painting of Sung urban life known as Going up the river during the Spring Festival (Ch’ing-ming shang-ho t’u) shows retail shops with signs openly advertising the official status of their owners, for example the drugstore of “Imperial Defense Commissioner Chou.”¹⁸⁰

This diversity also extended to religious life. Without addressing the fractious issue of whether Confucianism was a religion or not, in their private and public lives Sung officials manifested a spectrum of religious belief and practice commensurate with their varied backgrounds. Study of the classic Confucian texts was widespread and almost universally honored as a worthy enterprise and directly related to government employment and practice. But many subsequent pillars of the “Confucian” establishment, such as Fan Chung-yen, Wang An-shih, Su Shih, and Chu Hsi, had extensive knowledge of Buddhism, and Buddhism influenced their thinking on social and governmental issues.¹⁸¹ Hung Mai’s I-chien chih (Record of the listener), a large twelfth-century collection of popular narratives, reveals an officialdom, especially in the lower ranks, whose religious beliefs and practices are identical to those of the general population whence they once had come.¹⁸²

The emperor

The Sung monarchy was unique in the long history of Chinese imperial rule. Although emperors ruled China from 221 bc until 1911, the people, the institutions, the rituals, and the protocols that comprised the totality of the Chinese monarchy varied enormously over the course of that history. The monarchy was not simply the emperor. It also included his immediate and extended family; the imperial clan; the eunuch (huan-kuan) and female palace bureaucracy; and those elements of the regular bureaucratic establishment, mainly the Institute of Academicians (Hsüeh-shih yüan), that were under the emperor’s direct control. In historical terms, none of these components of the Sung monarchy was new, but their size, their organization, their relationships with each other, and their interface with the outside world gave a distinctive function and feel to the Sung monarchy that distinguish it from any other in Chinese history. Once again, the genius of Sung was not to invent something new but to bring new order and structure to existing institutions.

The monarchy was among the most successful of Sung institutions and a major source of the dynasty’s continuity and stability. A few statistics reveal the source and depth of this stability. From 960 through 1276, there were fifteen Sung emperors. All ruled as adults. The average length of reign was twenty-two years. The longest was that of Jen-tsung, who ascended at age thirteen and reigned for forty-two years (1022–63), then Li-tsung at forty-one years (1224–64). The shortest reign was the unfortunate Ch’in-tsung (r. 1125–7), on the throne little more than a year before the Jurchen captured him in 1127. The average age of the Sung emperors at ascension was twenty-six; their average age at death was fifty-two. Che-tsung (r. 1085–1100) was the youngest when he ascended the throne at age ten; Kuang-tsung (r. 1189–94) the oldest at forty-three.¹⁸³ There were also nine regencies, periods when empresses ruled on behalf of, or in conjunction with, young emperors. Many of these regencies were for short periods of time. Two, however, the rules of Empress Liu (969–1033) from 1022 to 1033 and of Empress Kao (1028–93) from 1085 to 1093, lent stability to the monarchy during important periods of political growth and tension.

On average then, and in modern terms, the Sung monarchs were young or middle-aged adults who occupied their thrones for relatively long periods of time. Unlike in other dynasties, there were no baby emperors, no emperors poisoned by eunuchs, no recluse emperors, and none deposed or dominated by their affinal relatives. The Sung emperors were not figureheads but active monarchs fully engaged in the performance of their duties as heads of state. With the probable exceptions of Ying-tsung (r. 1063–7) and Kuang-tsung, they were generally in good health and worked long hours. Emperor Kao-tsung’s day, for example, began at dawn with the morning audience, after which he read the memorials that had been submitted at audience that day. In the afternoon, he read history, especially the Tso Chuan (Chronicles of Tso), which he perused continuously, completing the book every twenty-four days, then beginning anew. In late afternoon, he practiced calligraphy and archery. After dinner, he read memorials that reached him via non-audience channels. He retired during the second watch, between nine and eleven o’clock. Elsewhere, Kao-tsung relates that he often spent all day, even holidays, reading memorials.¹⁸⁴ How many is not certain, but already in 999, at the beginning of his reign, Chen-tsung complained that he was reading a hundred memorials per day and asked his ministers to screen them for him.¹⁸⁵

A Sung emperor carried a number of names and titles over the course of his life and death. After ascension, a committee of senior officials determined a formal honorific title. For example, in 978, two years after his ascension, T’ai-tsung was termed “In Accord with Fortune and in Unity with Heaven, the Sagacious and Enlightened, Civil and Martial August Sovereign” (Ying-yün t’ung-t’ien sheng-ming wen-wu huang-ti).¹⁸⁶ These titles – always carefully formulated to reflect contemporary political and cultural policy – ended with the phrase Huang-ti, the formal title of Chinese monarchs since 221 bc. Huang-ti and T’ien-tzu (Son of Heaven) were the most usual titles on the formal seals of Sung emperors.¹⁸⁷ During his reign, contemporary documents referred to a sitting emperor through euphemisms and circumlocutions such as the “ruler of men,” “the Sage,” “the Highest,” or “Heaven.”

After his death, another committee determined two further titles. The posthumous honorific was a long, formal title that also ended in Huang-ti. The temple title, for use in the imperial ancestral temple, was a two-graph phrase that ended with the graph tsung (ancestor). A first graph summarized his personality and the nature of the deceased ruler’s achievement. Sung emperors are known to subsequent history and to English readers by these temple titles, thus Chen-tsung (the Perfected Ancestor), Jen-tsung (the Benevolent Ancestor), Kao-tsung (the Lofty Ancestor).¹⁸⁸ The tomb of each emperor also received its own name, and subsequent writers sometimes referred to a deceased emperor by the name of his tomb. The tombs of seven Northern Sung emperors are located at Kung-hsien in modern Ho-nan. The tombs of the Southern Sung emperors are at Shao-hsing in modern Chekiang.¹⁸⁹ Twenty-seven Sung-era portraits of all the emperors and selected empresses survive in the collection of the National Palace Museum (Ku-kung po-wu-kuan) in Taipei. These are large formal portraits probably originally intended not for public display but for private ritual adoration within the palace. Similar in size and format, these Sung portraits show remarkable variation in realistic physical features such as body size and demeanor, facial features, and skin tone. In this way, they differ from the imperial portraits of later dynasties, which are usually highly stylized.¹⁹⁰

The emperor was first and foremost the primary religious officer of the Sung state. His function in this capacity was unique and irreplaceable. He was the principal officiant at a series of rituals that regulated time; offered sacrifices to deities; paid homage to stars, mountains, and rivers; and worshipped ancestors.¹⁹¹ The greatest of these was the triennial “suburban sacrifice” during which the emperor offered sacrifices to Heaven and other deities at a ritual altar complex south of the capital city. The ritual itself was a lavish display of pomp and power that reinforced the legitimacy of the dynasty, and served as occasion for the emperor to reward his officials with large supplements to salary and dispensations of yin privilege. Such grants, extending even to soldiers, clerks, and artisans, consumed 10 percent of the state budget in years when the ritual was performed.¹⁹² The Sung dynasty was also a high point in Chinese ideas concerning the “Hall of Light” (Ming-t’ang), a ritual building in which a version of the sacrifices to Heaven was combined with rituals to ensure the proper progression of the seasons. The performances were especially prevalent in Southern Sung and continued until the very end of the dynasty.¹⁹³ The emperor also performed an upscale version of the ubiquitous Chinese ancestral rites at which regular sacrifices were offered to statues and portraits of his imperial predecessors (tsu-tsung).¹⁹⁴

In addition to his religious and ritual duties, the emperor was also the political head of state. In theory, and judged from a modern Western perspective, his powers appear virtually unlimited. In practice, however, other elements within the monarchy, and especially the bureaucratic protocols that determined his relationship with the rest of Sung government, often limited his ability to exercise his theoretical powers. The political powers of the Sung monarch may be divided into four categories for convenience of presentation: legal, personnel administration, military, and financial.

The emperor was the ultimate legislative and judicial officer of the state. Major initiatives were undertaken and decisions issued in his name and often with his personal involvement. The “imperial will” was the basis for all subsequent lower-level administrative action. Expressions of the imperial will formed the basis for legal statutes, and compilations of such statutes were issued in his name. He was the final legal authority and often personally ruled on legal cases that involved important officials or on cases that had reached the apex of the appeals process.¹⁹⁵ He could intervene at any stage of a legal proceeding and resolve the matter at his discretion. He could also bypass lower-level judicial authority and order a so-called “mandated trial” (chao-yü), essentially an investigation and trial initiated on imperial authority.¹⁹⁶ In all such matters, the emperor worked in conjunction with the relevant legal and judicial organs of state. But he was the final authority.

The emperor also personally made all appointments and personnel decisions concerning top officials. Such officials, known collectively as ministers-in-attendance, had personal-rank grades of eleven or above. They numbered several dozen and included chief councilors and other members of the State Council, censors, and Han-lin academicians (Han-lin hsüeh-shih). This authority also extended to the appointment of empresses and heirs apparent, whose civil service grades were even higher than those for ministers-in-attendance. On the one hand, in 1012, Chen-tsung, against the advice of his councilors, appointed the daughter of a silversmith as empress, a woman subsequently known as Empress Liu. On the other hand, in 1033, Jen-tsung dismissed Empress Kuo (whom Empress Liu had forced the young emperor to marry during her regency) from her position as empress, once again, over the strong objections of senior advisers, such as Fan Chung-yen, whom he also dismissed for opposing his wishes on the matter.¹⁹⁷ Emperors jealously guarded their prerogatives in making such senior appointments. In 1061, Jen-tsung dismissed four censors for suggesting that several eunuchs had engineered the appointment of the Assistant Commissioner of Military Affairs – a State Council position – because these eunuchs had marriage ties to the appointee. Jen-tsung determined that the man was indeed related to the eunuchs, dismissed him, then also dismissed the four censors with the statement, “I make all appointments to the Council. How could I suffer to accept such advice from eunuchs!”¹⁹⁸ Theoretically, the emperor had authority to “review” (yin-tui, literally “invite for audience”) any appointment or promotion. In 1054, for example, Jen-tsung denied promotions for two officials because they had previously been convicted of minor bureaucratic violations. When Ou-yang Hsiu suggested that this rejection was unwarranted, he was dismissed from his post on the ground that his demurral contested the emperor’s ultimate right of personnel review.¹⁹⁹

The emperor was also the commander-in-chief. T’ai-tsu and T’ai-tsung were actual warriors who fought personally in combat. Although subsequent emperors rarely took the field against opponents, they created bureaucratic structures that severely restricted the independent authority of military commanders. The Sung began with a military coup, and subsequent emperors took pains to insure the same did not happen, in reverse, against them. Only the emperor had the authority to mobilize troops for war. Control of troop movements was centered in the Military Affairs Commission, but even this agency needed an imperial edict to move troops. Northern Sung emperors guarded this authority and were reluctant to share it, even with chief councilors. During the chaotic years of the early Southern Sung when communication with the crown was difficult, military commanders often took independent action. The famous struggle between Emperor Kao-tsung and the “martyred general” Yüeh Fei (1103–42) is best understood in this light. Kao-tsung bypassed normal administrative channels and sent letters directly to Yüeh Fei in the field. Filled with detailed instructions on strategy, logistics, and operations, the surviving letters, if genuine, show the emperor’s resolve to re-exert his predecessors’ tight control over the military.²⁰⁰ Kao-tsung executed Yüeh Fei because Yüeh was reluctant to acknowledge the emperor’s ultimate authority over Yüeh’s troops. A year later, a fellow general, Han Shih-chung (1089–1151), offered to donate three years’ revenue from his private landholdings to support army troops. Kao-tsung declined, and remarked to Chief Councilor Ch’in Kuei that the T’ang dynasty had lost control of the provinces by not exercising early, firm control over its military governors. “I have now brought military authority back under court control. If I wish to switch a general, he follows my order, no differently than do the civil officials.”²⁰¹

The emperor was also the chief financial officer of the state. The question of what revenues belonged personally to the emperor and what belonged to the state was complex and contentious. There was a constant struggle between the desire of the monarchy to control both its personal and state finances and the needs of the State Council for accurate financial information to support policy formulation and implementation. For example, representatives of the State Finance Commission (San-ssu), or its post-1082 successor, the Ministry of Revenue (Hu-pu), were never members of the State Council. On the one hand, early in Chen-tsung’s reign the Finance Commission was ordered to prepare “accounting registers” (k’uai-chi lu) that included nationwide data on population, revenues, and expenditures, and this practice continued until the end of the dynasty.²⁰² Information from the accounting registers was used to fix tax quotas for each provincial jurisdiction. On the other hand, the Palace East Gate Bureau (Nei tung-men ssu), the emperor’s eunuch-staffed financial-affairs office, kept its own books of the emperor’s personal finances. And an edict of 1003 threatened staff with decapitation if they revealed the numbers to the outside bureaucracy.²⁰³ In 1069, the emperor personally controlled 23 percent of total government income and expenditure.²⁰⁴ Although the crown lost some measure of control over state finance during the late Northern Sung, Emperor Kao-tsung was able to restore the system, such that by 1161 over half of state income entered the emperor’s Palace Storehouse (Nei-tsang k’u).²⁰⁵ The bulk of this money was spent on legitimate national expenses, such as military defense or disaster relief, but at the discretion of the emperor.

The female monarchy

A simple overview of the geography of its headquarters, the Great Inner (Ta-nei), or imperial city, provides a sense of the size and administrative scope of the Sung monarchy. This was a rectangular-shaped enclave, originally a modest 1.7 miles around, located in the northeast of the capital city, K’ai-feng.²⁰⁶ Two avenues bisected the area in each direction and separated the compound into four quadrants, each of which had its own inner walls and gates. The southwest quadrant was the administrative center of the court (ch’ao-t’ing), meaning not the emperor, but the Secretariat–Chancellery and the Military Affairs Commission, in essence the offices of the State Council. Behind these offices was the Hall of Civil Virtue (Wen-te tien), the main hall used for the outer audience. Ritual structures, especially the Hall of Light, occupied most of the southeast quadrant. The northwest quadrant contained a number of halls, pavilions, and gardens and was the actual residence of the emperor and of most members of the immediate royal family. The northeast quadrant was the administrative nerve center of the monarchy. It contained secretarial and financial services such as the Palace East Gate Bureau and the Institute of Academicians. The Palace Eunuch Service (Ju-nei nei-shih sheng) and the female-staffed Palace Domestic Service (Nei-sheng, Shang-shu nei-sheng), as well as the palace of the heir apparent (tung-kung) and weapons storage, were also located in this area.²⁰⁷

Women and eunuchs filled a majority of positions that administered the monarchy. Unlike other dynasties, however, the Sung monarchs devised mechanisms to control these groups and to prevent them from dominating either the sovereign personally or the monarchy as an institution. For both groups, the Sung created bureaucratic structures that, although distinct in terminology, paralleled those for male officials. Like other officials, palace women and eunuchs held graded bureaucratic positions, drew salaries, were promoted, and retired. The Sung monarchs placed ultimate control over the eunuchs in the hands of the State Council. The Military Affairs Commission processed appointments to the Palace Eunuch Service; chief councilors could also intervene to block eunuch appointments or to punish errant behavior.²⁰⁸ In addition, the elaborate safeguards on document processing and verification made it difficult for eunuchs to manipulate the imperial paper flow.²⁰⁹

The “female monarchy” was among the most remarkable and distinctive aspects of the Sung monarchy.²¹⁰ It comprised two distinct but related groups. First were women in the imperial harem (hou-kung) – the dowager empress (huang t’ai-hou), the reigning empress, imperial consorts, and minor wives – as well as their daughters, the imperial princesses. Second was a professional female bureaucracy that supervised its own affairs and oversaw many aspects of palace life. Both groups contained women from all social strata. Imperial consorts taken from outside the palace were always from socially prominent families, but those promoted into the imperial harem from the ranks of the female bureaucracy might just as well come from humble backgrounds. Like other major Sung institutions, the Palace Domestic Service dates from the reign of Chen-tsung. Organized in 1022 into six ministries with a plethora of subdivisions, it contained a total of 282 billeted positions. These six ministries were General Affairs, Ceremonies, Wardrobe, Food Service, Housekeeping, and Workshop Service.²¹¹ Although the ministries were directed primarily toward providing services for the empress and consorts, they also shared duties with the eunuchs in providing the same services to the emperor. These included the manufacture and organization of imperial clothing and regalia, the upkeep of the imperial apartments, the preparation and serving of food, and security and access control. By the mid-eleventh century, the Palace Domestic Service employed about 2,500 women. The Ministry of General Affairs co-ordinated the other ministries and maintained a centralized secretariat and record-keeping service.

In 1113, Emperor Hui-tsung, as a continuation of the Yüan-feng reforms of 1082, reorganized the Palace Domestic Service to mirror the organization of external government. The traditional six inner ministries of 1022 were reorganized into six divisions that corresponded to the external Six Ministries of the Department of State Affairs, and the top supervisory positions were recast as “inner councilors” (nei-tsai). As justification for this move, Hui-tsung’s edict declared that this new structure would facilitate “the disposition of matters submitted to the throne from the external Six Ministries.”²¹² This phrasing raises the interesting question whether, in addition to processing their own internal paperwork, the women also processed external documents coming to and from the emperor. There are tantalizing indications that this was probably the case. Already in 1041, the Records Office (Ssu-chi) and the Communications Office (Ssu-yen), the general-affairs units charged with secretarial functions, employed 120 women.²¹³ Li Hsin-ch’uan states directly that beginning under Emperor Hui-tsung palace women sometimes wrote imperial edicts, and the biography of Empress Cheng (1082–1133), who became Hui-tsung’s empress in the year 1111, states that the emperor admired her ability to compose official documents.²¹⁴ Writing in the mid-1230s, the literati official Wu Yung (b. 1181, chin-shih 1208) lamented that “nefarious concubines and duplicitous females” compose documents that purport to represent the intentions of the emperor.²¹⁵

The issue is not primarily whether female secretaries did or did not ghostwrite for the emperor (although it appears they often did) but rather the contribution of women to the political stability and the intensely literate, cultured character of the Sung monarchy. The Sung regency differed from the regencies of the Eastern Han period (25–220), during which powerful families of the emperor’s consorts usurped imperial authority. Essentially a new Chinese institution, the first Sung regency, that of Empress Liu from 1022 to 1033, set the pattern for those that followed.²¹⁶ Unlike the Han empresses, Empress Liu – and with her many other Sung empresses, consorts, and palace women – came from families without powerful political connections. This lack of access to an external power base forced Empress Liu to rely on female allies within the palace and to forge alliances with male court officials, especially with the chief councilors. The Sung regents did not draw external power into the vacuum that the lack of an emperor had created, but forged working coalitions of male court officials, other female power brokers within the palace, and eunuchs. Because each partner had a specific role to play, these coalitions perpetuated and even strengthened existing political structures. Alliances between Sung regents as well as empress dowagers and selected chief councilors became a feature of Sung court politics. But, at least in Northern Sung, these alliances were based on mutual political interests, not on family connections. Sung historians tended to praise these alliances when they agreed with the politics – Ssu-ma Kuang and Empress Kao – and to condemn them when they did not – Shih Mi-yüan (1164–1233) and Empress Yang (1162–1232). But the institutional base for such alliances remained similar.

Another factor contributing to the special nature of Sung regencies was the strong connection between female palace officials and the Sung military establishment. Four of the five women who served as regents during the Northern Sung came from military families, as did the principal consorts of all nine Northern Sung emperors.²¹⁷ This feature of the Sung monarchy – its extensive intermarriage with military families – goes back to T’ai-tsu himself, who offered his fellow generals wealth, security, and marriage ties to his family in exchange for their support of the fledgling dynasty. And this policy of intermarriage between the imperial clan and the military establishment continued at least into the Southern Sung.²¹⁸ On the one hand, the rigid separation of civil and military functions in the Northern Sung meant that regents, even those from powerful military families, had little access to political power, which was largely in the hands of literati officials. On the other hand, as high-ranking graded officials, the female bureaucrats of the Palace Domestic Service enjoyed the same yin privileges to appoint relatives to office as did male officials. In short, military families whose daughters did well in the palace service reaped many of the same benefits as did literati families whose sons did well in the chin-shih examinations. Literati writers constantly bemoan the excess of “military officials” (wu-kuan) created in this way, but this system of palace organization and the female bureaucrats’ role in it was a key factor in the stability of the Sung monarchy.

Senior palace women, sometimes in alliance with a chief councilor, also played key roles in deciding imperial successions. Seven emperors were not biological sons of their immediate predecessor (T’ai-tsung, Ying-tsung, Hui-tsung, Kao-tsung, Hsiao-tsung, Li-tsung, and Tu-tsung). Of the remainder, only two, Shen-tsung and Ch’in-tsung, were the eldest sons of their predecessors. The image, then, of a smooth imperial succession from father to eldest son is atypical for Sung. Ideally, the reigning emperor decided who would become heir apparent. Chen-tsung, Jen-tsung, Hsiao-tsung, Kuang-tsung, and Tu-tsung were duly appointed heirs apparent a number of years before their predecessors’ demise, and their ascensions to the throne were relatively smooth. For a variety of reasons, however, emperors were often reluctant to designate heirs, and many of these transitions appear to have been open questions. In two cases, there is strong evidence that dowager empresses, in consultation with inner and outer court officials, determined the succession. Tseng Pu’s diary records in detail the deliberations during which the views of Dowager Empress Hsiang (1046–1101) prevailed over those of Chief Councilor Chang Tun (1035–1105) in determining the succession of Hui-tsung in the year 1100.²¹⁹ And, according to one version of the story, the opinion of Dowager Empress Wu (1115–97) was crucial to the choice of Ning-tsung in 1195.²²⁰ In 1085, the Shen-tsung–Che-tsung transition resulted in the regency of Empress Kao, and later sources accused her of engineering the choice of the ten-year-old child as a vehicle to obtain a period of regency for herself.²²¹ Finally, there is no doubt Empress Yang collaborated with Shih Mi-yüan to arrange the ascension of Li-tsung in 1224.²²²

In all these cases, the political circumstances and personalities of the participants were different. Yet a common thread that runs through many of the surviving accounts of imperial transitions is the collaborative – almost “corpor-ate” – flavor of the decision process. Jen-tsung, for example, had no son, so his advisers suggested that he establish a school within the palace to train those clan boys who displayed imperial potential.²²³ The conversations in Tseng Pu’s diary reveal a calm and businesslike discussion whose participants weighed the pros and cons of several possible candidates in 1100. Apparently, virtually any male member of the imperial clan was eligible to be emperor. For example, during the chaos of 1127 an obscure clansman named Chao Shu-hsiang raised a small army to fight the Jurchen. Apparently harboring his own imperial ambitions, he was reluctant to support Kao-tsung as emperor and was immediately arrested and executed.²²⁴ In 1131, Kao-tsung, himself without male issue, commanded that ten boys from a branch of the imperial clan so distant that they barely qualified as relatives be brought into the palace and raised as his potential successor. The group was educated, observed, and tested until eventually, in 1153, Kao-tsung finally declared the future Hsiao-tsung heir apparent.²²⁵

The monarchy as a cultural institution

Historical sources present the Institute of Academicians as the emperor’s personal secretariat and think tank and, organizationally, a part of the monarchy.²²⁶ The academicians of the Institute composed the formal, important documents of imperial rule – notices of imperial appointments and promotions, amnesties, and foreign correspondence. Academicians might work from the emperor’s personal draft (tz’u-tou), which a eunuch delivered to their office in the palace, or they might be summoned into his presence and receive his text verbally. Since their final copy was read aloud at morning audience, they worked at night. Five or six academicians rotated this night duty, with at least one being on call at all times. The enormous pressure for literary elegance and speed, along with the possibility of frequent access to the emperor, meant that only the most talented literati were considered for these appointments, which the emperor made personally. Half of all Sung chief councilors had served earlier in their careers as academicians.²²⁷ Literati sources present glowing accounts of evening camaraderie between emperor and academician, exchanging poems and leisurely conversing about the politics and issues of the day.²²⁸

No doubt such events occurred. But their frequency and impact on the relationship between the emperor and his senior court officials is open to question.²²⁹ The imperial tutoring sessions known as the Classics Mat (Ching-yen) were more effective as a vehicle to convey informal literati sentiments to the emperor.²³⁰ Classics Mat sessions, which assumed their mature form in the mid-eleventh century, were tutoring sessions in which a group of senior officials read aloud, explicated, and answered the emperor’s questions on classic and history texts. Usually the emperor chose the texts. The form and scope of the sessions varied considerably over the course of the dynasty. The most common pattern was to hold sessions every other day during a spring and an autumn semester that each lasted about three and a half months. The number of lecturers was never fixed and ranged from two in the beginning of the dynasty to a high of twenty-four in 1058.²³¹ Lecturers were chosen personally by the emperor from among his senior officials, retained their primary appointments, and were given supplementary titles, designating them as Classics Mat lecturers (Ching-yen kuan). Occasionally, a renowned “private scholar” with no official position was invited to attend. The lecturers met with the emperor as a group, although they rotated the lead responsibility to read the text and answer questions. Disagreement and discussion were permitted. Minutes were taken, and versions of these records occur frequently in Li T’ao’s Long draft, especially during the Jen-tsung years.

These records reveal fascinating glimpses of how the texts of the Classics and history were read against the backdrop of contemporary politics. Literati opinion differed about the primary function of the sessions. Some, usually chief councilors anxious to control what the emperor read and heard, urged that the lecturers “stick to the text” and refrain from comment on current issues. Others argued that since the ultimate purpose of the sessions was to educate the emperor to be an effective ruler, one should “go beyond the text” to relate its message to contemporary problems. For example, in 1045 Ting Tu (990–1053) used a passage from the Han-shu (Han history) to warn Jen-tsung against Fan Chung-yen and the dangers of factionalism.²³² Much depended on the age of the emperor and his relationship with individual lecturers. In 1086, Su Sung (1020–1101) used precedents drawn from the histories of the T’ang dynasty to illustrate “actions by rulers and ministers.” He related them to contemporary events and interjected his own opinions. Emperor Che-tsung was nine years old.²³³ Many of the texts read in the sessions were not the ancient Classics but compilations of Sung historical documents from earlier reign periods that were being combed for “precedents” to contemporary policy.

The Sung monarchy was also the country’s premier cultural institution. Beginning with T’ai-tsung, but especially under Chen-tsung, the Sung monarchs realized the power of visual and literary culture to attract and hold the allegiance of the emerging literati class. The early Sung monarchs supported a wide array of compilation and printing projects, including editions of the Confucian classics, the Buddhist and Taoist canons, and encyclopedias.²³⁴ Performance of imperial rituals required the constant production of cultural objects – buildings, implements, texts, music, paintings, and regalia. By the turn of the eleventh century the four divisions that would later make up the Sung Han-lin Academy (astronomy, calligraphy, painting, and medicine) had already been established.²³⁵ Chen-tsung initiated a dynastic policy of using art to record and celebrated the political achievements of the ruling house. He recorded his own 1008 performance of the feng and shan sacrifices in elaborate documentary paintings executed as public murals and on handscrolls. He also began the custom of constructing a hall on the palace grounds to house the writings, calligraphy, and other cultural artifacts of his imperial predecessor. One for each emperor, these halls served as libraries, museums, and centers to promote the cultural leadership of the royal house. A surviving eleventh-century painting shows Chen-tsung leading senior officials on a 1007 viewing tour of a new book collection and a display of the assembled manuscripts of Emperor T’ai-tsung. Records indicate twenty subsequent imperial visits to the same hall between 1007 and 1061 for viewings, literary gatherings, and banquets.²³⁶

Both imperial patronage of the arts and imperial use of art to promote the monarchy’s political goals reached their apogee in the late Northern Sung under Emperor Hui-tsung. A sovereign noted for his own distinctive callig-raphy and for a number of surviving paintings connected to him, Hui-tsung substantially increased the dynasty’s commitment to painting. Under his reign, painters attached to the Han-lin Academy were recruited through examinations and became graded officials with rank and status similar to members of the regular bureaucracy. The Han-lin Academy produced large number of paintings for imperial use as official gifts and as historical records of auspicious occasions. Hui-tsung also expanded the monarchy’s collection of past and contemporary painting and calligraphy. The catalogues of those collections, the Hsüan-ho hua-p’u (The Hsüan-ho period register of paintings) and the Hsüan-ho shu-p’u (The Hsüan-ho period register of calligraphy) both survive. Almost certainly compiled by eunuchs, these catalogues record the extensive involvement of imperial clan members in artistic pursuits.²³⁷

These Northern Sung traditions, which culminated with Hui-tsung, continued throughout the Southern Sung. Particularly important was the notion that a distinct imperial calligraphy, copied and propagated through stele inscriptions throughout the country, disseminated the emperor’s authority and leadership. Although Kao-tsung repudiated Hui-tsung’s distinct hand, he took equal pains to create his own style of imperial calligraphy and used it to reinforce the idea of a rebirth of Sung culture following the catastrophe of 1127.²³⁸ The cultural projects of the Sung monarchy combined the talents of all segments of the institution. As we have seen above, Sung biographies of imperial consorts, especially in the Southern Sung, stress their artistic ability. Not only Hui-tsung and Empress Cheng, but also Kao-tsung and Empress Wu, and Ning-tsung and Empress Yang, worked together on projects to enhance the dynasty’s cultural profile.²³⁹ The delicacy and balance so characteristic of Southern Sung Academy painting is an eloquent testament to their influence. Lastly, the eunuchs administered and provided support for the monarchy’s cultural programs. They acquired and organized supplies, compiled catalogues and kept records, and furnished logistical support to members of the Han-lin Academy.

One must also emphasize the intellectual and religious diversity of the monarchy’s support for cultural and religious institutions. The Sung emperors supported both Buddhist and Taoist organizations in return for church support of the monarchy and the state. The early Sung emperors, especially T’ai-tsung, reversed the anticlerical policies of the Later Chou dynasty and promoted the growth of Buddhist monasteries through imperial protection, land grants, and tax remissions. Buddhist monasteries were linked to the monarchy as places to pray for deceased soldiers, as sites locally to celebrate imperial birthdays and deathdays, and as repositories of imperial calligraphy and portraits. The state also used both Buddhist and Taoist institutions to funnel financial and spiritual aid to the population after natural disasters and warfare. Grants of money and land to monasteries were a significant government expense.²⁴⁰

It is difficult to learn much about the personal religious lives of the Sung monarchs – to distinguish public policy from private belief. But a fervent private commitment seems partially to underlie the intense public support for Taoism, especially under Emperors Chen-tsung and Hui-tsung. In his later years, Hui-tsung envisioned a state with dual and equal Confucian and Taoist underpinnings. In 1118, he established official government schools, teachers, and examinations in the Taoist classics. He also created an official interface between the civil service personal-rank system and the Taoist clergy in an effort to foster “Taoist officials” (tao-kuan).²⁴¹ There is strong evidence that such personal commitment to Taoism on the part of Sung emperors continued. In 1141, Emperor Kao-tsung was conferring with his Chief Councilor Ch’in Kuei on a proposal to discipline provincial clerks. The emperor demurred on the idea. In his personal life and as emperor he preferred to rule by being “limpid and still.” His allusion was to Lao-tzu, chapter 45, and its famous equation of Taoist perfection and political tranquility: “be limpid and still and the people will rectify themselves.”²⁴²

Historians often frame the Chinese monarchy during the Han and T’ang dynasties as a struggle between an “inner court” (nei-ch’ao – imperial kin, eunuchs, and informal advisers) and an “outer court” (wai-ch’ao – the graded bureaucracy). In this dichotomy, the triumph of the inner over the outer court led to dynastic collapse. The Sung emperors imposed controls on the major elements that would have comprised a menacing “inner court,” and so avoided this fate. They spent lavishly on subsidies for the imperial clan, and awarded its leaders high-sounding titles and salaries, but kept them far away from real power.²⁴³ They redefined the role of the empress and the female palace bureaucracy, but, at least for two hundred years, kept their affinal relatives largely at bay.²⁴⁴ In the ultimate choice between inner and outer, a continuation of their founders’ policies and the increasing vibrancy of literati culture pulled the Sung monarchs toward the outer court. And the relationship between the sovereign and his chief councilors became the axis of Sung political life. If this choice bought the Sung monarchs a measure of domestic tranquility, the price they paid was a slow devolution of their authority into the hands of the chief councilors.

The chief councilor

In 1071, the court was discussing the merits of Wang An-shih’s New Policies. The following exchange occurred between Emperor Shen-tsung and Commissioner of Military Affairs Wen Yen-po (1006–97):

In terms of modern corporate organization, if the emperor was the chairman of the board, the chief councilor was his chief executive officer. The analogy is hardly exact, since chief councilors, in theory and often in practice, exercised authority as members of a group, the State Council. But the office of chief councilor was the pinnacle of the Sung bureaucracy, the chief officer charged with formulating and executing policy. As symbol of his status as leader of “we who are the officials,” the chief councilors “led the formation” (ya-pan) of court officials at formal morning audience with the emperor. The emperor personally chose his chief councilors and determined their seniority vis-à-vis each other. Although he was theoretically free to choose whomever he wished, in practice he chose from a small group of elite officials whose career paths had groomed them for the top post and with whom he had direct experience. That path usually included a previous position as Finance Commissioner (san-ssu shih), Han-lin academician, prefect of K’ai-feng (chih K’ai-feng fu), or head of the Censorate (yü-shih chung-ch’eng).²⁴⁶

The State Council comprised the leadership of the “two headquarters” (erh-fu). These were the Military Affairs Commission and the Secretariat–Chancellery. After 1082, the Secretariat–Chancellery was divided and reorganized into the Three Departments, these being the Department of the Secretariat (Chung-shu sheng), the Department of the Chancellery (Men-hsia sheng) and the Department of State Affairs. Taken together the two headquarters made up the court or central administration of Sung government. The English term “chief councilor” translates the Chinese tsai-hsiang, an archaic term used in the Sung in a general and quasi-official way to refer to a chief executive. The actual titles of a Sung chief councilor reflected his position as a supervisory official in the Secretariat–Chancellery or, after 1082, in the Three Departments. These titles changed five times over the course of the dynasty and mirrored changes in the structure of these central organizations and in their relationship with the Military Affairs Commission. The details of these changes are highly technical and beyond the scope of this chapter. However, the chief councilors’ office is fundamental to Sung political history, and the following three paragraphs attempt to summarize these developments.²⁴⁷

The early Sung inherited the political structures of the militarized states of the Five Dynasties. Real power was concentrated in the Military Affairs Commission. The combined Secretariat–Chancellery of the T’ang survived as a document-processing center. The traditional T’ang Six Ministries within the Department of State Affairs were defunct. Initially, the Sung chief councilors were secondary in importance to the Military Affairs Commissioner. But, as soon as literati culture began to emerge in the early eleventh century, the first of the great chief councilors also emerged in the person of Lü I-chien. Sung chief councilors through 1082 were appointed Joint Manager of Affairs with the Secretariat–Chancellery (T’ung chung-shu men-hsia p’ing-ch’ang shih), a T’ang title that originally conferred authority to participate in deliberations in the Hall of Administration (Cheng-shih t’ang) or State Council. The number of chief councilors in this period was not fixed, but there were usually three at any one time. The Secretariat–Chancellery was subdivided into the Five Offices (Wu fang), one each for general administration, personnel, revenue, justice, and rites. These offices were staffed by clerks, and were known collectively as the Bureau of Edicts (Chih-ch’ih yüan), a designation that emphasized their secretarial and clerical rather than policy function. In 972, T’ai-tsu revived the old title, Participant in Determining Governmental Affairs (Ts’an-chih cheng-shih), often translated as assistant chief councilor. This title was conferred as a supplementary title on selected high-level officials and qualified them to participate in meetings of the State Council. There were usually two or three assistant chief councilors, and the position was often given to officials whom the emperor considered possible chief councilor material. Lastly, the State Council also included the supervisory officials of the Military Affairs Commission (shu-mi shih and shu-mi fu-shih), of whom there were two or three at any one time. Membership on the State Council thus fluctuated in the Northern Sung, but averaged about seven to eight members.

The Yüan-feng reform of 1082 attempted to restore the theoretical model of T’ang government outlined in the 738 Sixfold statutes of T’ang. This text described a central government with three more or less equal “departments,” each with a separate governmental function: the Secretariat would initiate policy, the Chancellery would review policy, and State Affairs would effect policy. To achieve this end, the reform divided the former Secretariat–Chancellery into two units, re-created the traditional Six Ministries under the Department of State Affairs, and transferred operations of the Five Offices into the Six Ministries. Additionally, the Six Ministries were divided into two groups. A “left” group included the Ministries of Personnel, Revenue, and Rites; a “right” group included the Ministries of War (Ping-pu), Justice, and Works (Kung-pu). There were two chief councilors. The Left Chief Councilor was Left Vice Director of the Department of State Affairs and concurrent Vice Director of the Chancellery (Shang-shu tso p’u-yeh chien men-hsia shih-lang). The Right Chief Councilor held a similar title but was concurrently Vice Director of the Secretariat (there were no directors). The position of Assistant Chief Councilor was abolished, replaced by Left and Right Assistant Directors of the Department of State Affairs (Shang-shu tso yu-ch’eng). This massive reorganization of government into left and right found concrete manifestation in a new administration building that housed the Department of State Affairs and the offices of the two chief councilors.²⁴⁸

For reasons to be described below, this system never worked and was revised a number of times in the late Northern Sung. In 1129, the new administration of Emperor Kao-tsung combined elements of the pre-1082 and post-1082 organizations. The rigid distinctions between the three departments were abandoned, and they became essentially one. “Left” and “right” chief councilors were retained, mainly to designate seniority (left over right); both were, however, again subtitled Joint Manager of Affairs with the Secretariat–Chancellery. The designation of assistant chief councilor was revived. Finally, in 1172, Emperor Hsiao-tsung changed the formal titles of chief councilors to simply Left, or Right, Councilor-in-Chief (Tso yu ch’eng-hsiang) and made them concurrent Military Affairs Commissioner. The early Sung emperors had insisted upon a strict separation of civil and military authority, and in the Northern Sung concurrent appointments – making the same man chief councilor and Military Affairs Commissioner – occurred only during wartime. In the Southern Sung, however, this strict separation began to erode. The reforms of 1172 marked the end of the independent Military Affairs Commissioner, although appointees to subordinate Commission positions continued to serve as members of the State Council.

The chief councilors’ working day was occupied – remarkably like a modern executive – with meetings and paperwork. We will discuss the paperwork below. Before 1082, the offices of the chief councilors were in the Hall of Administration, and afterward in the newly constructed building for the Six Ministries. Both buildings were located within the Great Inner, in other words within the imperial palace compound. In the mid-eleventh century, the support staff at the Secretariat–Chancellery numbered at least two to three hundred clerks, although this number ballooned after the Yüan-feng reforms. The councilorship came with an additional allotment of seventy personal attendants. Clerks, dressed in red robes, escorted the councilor on horseback from his residence outside the palace into the Great Inner. At the conclusion of the working day, often in the Northern Sung after a noon meal that was served to senior officials in the Hall, clerks escorted the councilor back to his residence.²⁴⁹ This custom is perhaps reflected in a number of Sung and Yüan paintings that depict red-robed grooms and horses.

That day began with the morning audience and the presentation of memorials. It continued at the Hall of Administration where the chief councilors met on a regular basis with subordinate staff. The councilors sat at desks facing south. Administrative-class officials were permitted to sit, facing north; executory-class officials stood to make their reports. To make such a visit to the office of the chief councilor was called “to pass the hall” (kuo-t’ang).²⁵⁰ The purpose of the meetings was to keep the chief councilors informed of issues and developments within the jurisdictions of the subordinate officers. Officially, the Sung state discouraged informal contacts between the chief councilors and other officials. Councilors were initially forbidden to receive guests in their own residences, so all contacts took place in their offices. Li T’ao writes that over a hundred people per day pressed to see the councilor, but that 80 to 90 percent involved “private” matters. Restrictions on non-office contacts were relaxed, but sanctions were imposed on officials who brought up private business as they “passed the hall.” In 1042, the emperor reinforced an old statute that fined officials who visited the Secretariat on nonofficial business one month’s salary.²⁵¹

The back offices of the chief councilor, as stated above, were not involved in policy but were simply centers for processing documents. They were staffed by professional clerks, who had their own hierarchy, but were not graded officials (shih ta-fu). After 1082, two or three graded officials in each of the Six Ministries and their twenty-four subunits (ssu) supervised a much larger number of clerks. Statistics on clerical staffing levels across the Six Ministries reveal a vast disparity in workload across these various units and hint at what the Sung central government actually did on a day-to-day basis. Li Hsin-ch’uan gives early Southern Sung clerical quotas for the Six Ministries as follows: Personnel 359, Revenue 288, Rites 56, War 135, Justice 63, and Works 19.²⁵² Clearly the Ministries of Personnel and Revenue were the main centers of action. Warfare and defense were also, of course, vital, and the Military Affairs Commission had its own quota of 327 clerks. But within the Six Ministries, the major concerns of the Sung central government were, first and foremost, itself, the processing of personnel actions; second, the raising and tracking of revenue. The actual number of surviving documents, as preserved in the Compendium of Sung documents, also supports this image. By far the largest section, at 1,900 pages, concerns government officials (chih-kuan); followed by trade and finance (shih-huo) at 1,659 pages; and rites (li) a close third at 1,461 pages.

Two patches of poetic doggerel from the mid-twelfth century confirm this statistical portrait of the workload in the Six Ministries. A popular ditty ranked each of the Six Ministries in terms of how hard its officials had to work:

But the clerks reworked the ditty to describe how much bribery they could extract and the lifestyle that an assignment in each ministry would support:

The remonstrance organs

The Yüan historians who compiled the Sung history (Sung-shih) wrote, “If we examine the foundations of the Sung state, then its heart and soul were the Censorate and Bureau of Policy Criticism.”²⁵⁴ The English word “Censorate” has unfortunately become the standard translation for the Chinese Yü-shih t’ai, literally, Terrace of Imperial Scribes. A basic unit of Chinese government since Han times, the “Censorate” seldom censored documents. Its closest modern parallel would be the internal-affairs division of a large urban police department. Its purpose was twofold. First, independently of other agencies, it gathered information and kept the emperor informed on conditions in the state. Second, it kept watch over the bureaucracy and enforced rules and standards for official conduct. In the metaphor of the state as a body, the censors were the “eyes and ears” of the sovereign. The Bureau of Policy Criticism was a unique eleventh-century institution. Its advocates pushed to turn the bureau into a formal agency that would critique the chief councilors and even the emperor himself. In the minds of these advocates, the bureau would watch the policy makers just as the Censorate watched the bureaucracy. Their failure to turn these aspirations into lasting political structures is among the most profound stories in Chinese politics and accounts for the intense interest in the Sung remonstrance organs among modern Chinese scholars, who often see the Sung experience as relevant to their own times.²⁵⁵

Like other units of central government, the Censorate had atrophied during the tenth century, and the early Sung emperors appointed censors only sporadically. The origins of the Sung Censorate as a freestanding agency are murky, but Li T’ao traced the beginnings to 1004.²⁵⁶ The origins of the Bureau of Policy Criticism, as well as an expanded Censorate, date from 1017, when Emperor Chen-tsung ordered that six billets be dedicated to censors and an additional six to policy critics. Before taking this action, he made two points in conversation with his advisers. He told them he had heard rumors that he was averse to remonstrance, and he intended to fill these positions with the best literary talents available.²⁵⁷ These conversations suggest that, at least in Chen-tsung’s mind, his 1017 action was largely public relations. He intended not to widen the channel of advice that flowed to him but to expand “public” awareness about how much advice he already accepted. In this vein, only two officials were actually appointed to the twelve new vacancies he created, and these appointments soon lapsed. It was not until 1023 under the Regent Liu (Dowager Empress Liu) that such appointments were again mentioned as a possible way to check the growing power of the chief councilors, and not until 1032 that a freestanding bureau was actually created.²⁵⁸ But Chen-tsung’s initial statements link the rise of the remonstrance function in Sung government directly to the rise of literati officials. Whether the literati pushed the emperor or the emperor used the literati, the greatest policy critics of the eleventh century were also the greatest literati, men such as Ou-yang Hsiu and Ssu-ma Kuang.

In terms of personnel and organization, after 1082, the Censorate had three divisions. The Headquarters Bureau (T’ai-yüan) contained only two billets, the Vice Censor-in-Chief (Yü-shih chung-ch’eng), who was the chief supervisory official (there was no Censor-in-Chief; the post was always vacant), and his second-in-command, the Attendant Censor (Shih yü-shih). The Palace Bureau (Tien-yüan) had two billets for Palace Censors (Tien-chung shih yü-shih) respon-sible for the conduct of officials and palace ceremonies. The Investigation Bureau (Ch’a-yüan) had six billets for investigating censors (chien-ch’a yü-shih), one to oversee each of the Six Ministries. Within these three divisions there were eleven subdivisions with a total quota of forty-four clerks.²⁵⁹

The Bureau of Policy Criticism existed as a separate agency from 1032 through its dissolution in the Yüan-feng reform of 1082. During the 1020s and 1030s, nonfunctional Secretariat and Chancellery positions from the T’ang era, known collectively as “speaking officials” (yen-kuan), were rejuvenated and combined to make the new Bureau. Given its passion for T’ang organization charts, the 1082 reform, therefore, returned these officials to their original T’ang positions in the newly divided Secretariat and Chancellery. A “speaker” was a court official whose position duly authorized him to speak out and criticize policy either at audience or through the memorial process. All appointees to the Bureau were “speakers.” Other officials, often censors, could also be invested with temporary authority to function as “speakers.” Between 1032 and 1082, the Bureau had six billets: a “Left” (originally from the Chancellery) and a “Right” (originally from the Secretariat) Remonstrator (Ssu-chien), a Left and Right Exhorter (Cheng-yen), and two additional officials with other titles that could be appointed “acting policy critics” (chih chien-yüan). After 1082, these positions were among those gathered into the newly created Secretariat Rear Section (Chung-shu hou-sheng) and Chancellery Rear Section (Men-hsia hou-sheng), where the Yüan-feng reform concentrated officials with oversight functions of various types. This move diluted the independence of the “speakers” and marked an end to the Bureau’s existence as a freestanding agency. Shortly after restoration, in 1129, there was a brief attempt to restore the Bureau’s independent status. But the Bureau again fell victim to factional politics. By 1134, its billets were unoccupied, and its seal and records, then in the hands of the office clerks, were transferred back under Secretariat–Chancellery control.²⁶⁰

As the 1034 letter from Ou-yang Hsiu to Fan Chung-yen, cited above, points out, speakers could address the monarch on any policy issue. This was a remarkable power in a system that strictly forbade civil servants to write or speak on official matters outside the jurisdictions of their current positions. In the Northern Sung, censors and speakers enjoyed considerable autonomy to work free of pressure from the chief councilors, from other officials, and even from the emperor himself. Neither were individual censors required to inform their superiors or their colleagues about the nature of their investigations, and individual censors often weighed in on different sides of the same issue. To isolate them further from outside influence, they were forbidden during their tenure of office from writing sponsorship recommendations for other officials.²⁶¹ According to one source, Censorate or Bureau indictments forced twenty-three chief councilors from office during the Northern Sung.²⁶² However, by the late eleventh century, political factionalism and the growing power of the chief councilors had already eroded this autonomy, as Su Shih’s lament about the waning power of the policy critics reveals.²⁶³ Southern Sung literati sources, especially Tao-hsüeh works, are filled with purported Northern Sung anecdotes that stress the independence of censors and policy critics. But these anecdotes are often inventions to bolster the demand for greater literati autonomy within the contemporary Southern Sung political structure. They testify to what was absent rather than to what had actually existed.²⁶⁴

Key to understanding the failure of the Censorate and the Bureau to endure as lasting components of the Sung political power structure is to appreciate that no Sung official – not even councilors, censors, or policy critics – enjoyed any measure of statutory protection for what they wrote or said. The emperor had absolute and ultimate authority over all appointments. As we shall see below, the emperor had several options for how he could handle a censorial report. Tradition, however, maintained that if his report was not accepted, the censor had failed in his duty. He then submitted his resignation and remained at home “awaiting punishment” (tai-tsui). The emperor was then required, in writing, to accept or reject his resignation. If he rejected the resignation, the censor could resign once again, in writing, restating the findings of his original report. These protocols could lead to protracted standoffs. In an epic battle of wills in 1051, Jen-tsung refused to accept a policy report from Fan Chen (1008–89) concerning the designation of an heir apparent. Jen-tsung refused Fan’s resignation seven times; Fan wrote nineteen reports on the matter and grew gray “awaiting punishment” for almost a year. Jen-tsung acknowledged his loyalty, but refused to budge on the issue.²⁶⁵

Writing in 1061, almost thirty years after his letter to Fan Chung-yen, Ou-yang Hsiu was more sanguine about the role of the policy critic in the ever-shifting relationship among monarch, councilor, and critic. Much depended on the personality of the sovereign, Ou-yang observed. On the one hand, if he was hard-hearted and suspicious, he would welcome criticism of his councilors; and under such a sovereign, “criticizing the ruler is hard, but criticizing the councilor is easy.” On the other hand, if the sovereign, like Jen-tsung, is lenient and himself receptive to criticism, then more criticism is forthcoming and his councilors become more sensitive to that criticism. Under such rulers, “criticizing the ruler is easy, but criticizing the councilor is hard.” This latter formulation became a standard quotation in Southern Sung compendia and reflects the waning authority of the remonstrance function in the face of the growing power of the Southern Sung chief councilors.²⁶⁶

Although the Censorate and Bureau had extraordinary authority to investigate and gather information, they had no actual authority to punish. Their function was simply to present information before the emperor. If he wished to proceed with an indictment, the emperor referred the matter to the proper executive and judicial authorities for prosecution. For example, in Su Shih’s 1079 trial, which was conducted at the highest levels of Sung government, the emperor himself functioned as magistrate, the Censorate functioned as the “fact-finding” (t’ui-k’an) agency, but the Court of Judicial Review (Ta-li ssu) functioned as the “law-finding” (chien-fa) agency.²⁶⁷ Sung censors and policy critics relied upon generally applicable laws and regulations that were promulgated by authorized agencies of government. The Censorate also seems to have functioned as a collection point for the tens of thousands of regulations these agencies promulgated. In order to restore these collections after the loss of K’ai-feng, an order of 1133 required all agencies to forward copies of their regulations to the Censorate.²⁶⁸ The Censorate and Policy Bureau (Bureau of Policy Criticism) also had their own internal regulations and ordinances. These do not survive, but quotations suggest that they provided extensive and detailed guidelines for the conduct of investigations.²⁶⁹

These investigations relied upon three basic sources of information: rumor; copies of memorials routinely supplied to the Censorate; and, during certain periods, impounded documents. “Rumor” is an unhappy translation for the Chinese term feng-wen, literally “heard on the wind,” but conforms to the English notion of “unconfirmed information.” The authority to present an indictment based on unconfirmed evidence was among the most controversial aspects of the Sung Censorate. The conflict arose from the dual function of the Censorate as both an information-gathering agency and an agency that presented this information in the form of judicial indictments. As early as 1049 Jen-tsung, noting the increasingly litigious nature of his officials, forbade the Censorate to bring indictments based on rumor: there must be a signed, written accusation – with exceptions for cases involving “criticism of the court or the popular welfare.” But six months later, a remonstrator insisted that to fulfill Chen-tsung’s 1017 mandate establishing the Censorate and Bureau, these agencies needed the flexibility to indict on rumor. Since the 1049 edict, he maintained, few people had dared to offend the powerful by affixing their name to a public accusation, and submissions to the Censorate had declined in number. “Rumor” protected the source of information and so kept information flowing.²⁷⁰ However, this edict had broad exemptions and seems to have had little effect. The issue arose again in 1061, when Wang Ch’ou (d. 1065), then head of the Censorate, warned that unscrupulous officials were using the system to stir up trouble for their bureaucratic adversaries and to settle private scores. He defended the use of information “heard on the wind,” but maintained that the Censorate did not have the manpower to investigate every unsubstantiated accusation. He urged an edict imposing sanctions on officials who knowingly brought false charges before the Censorate, and the edict was produced.²⁷¹

In the Southern Sung, misuse of the Censorate for personal gain reached epidemic proportions and contributed to the institution’s decline. Regulations required that officials submit to the Censorate copies of some categories of official correspondence with other agencies (kuan-pao, literally “linked notification”; in modern terms, “carbon copies”).²⁷² When bureaucratic disputes arose, many took advantage of this provision to send “short scrolls” (tuan-chuan) to the Censorate. These were truncated versions of official documents, doctored to slant an issue in favor of the presenter and against his opponent. Some officials also submitted “short scrolls” anonymously in an effort to slander adversaries, and so the practice fell under the general category of information “heard on the wind.” Emperor Hsiao-tsung prohibited censors from accepting “short scrolls,” but the practice continued under a variety of guises until the end of the dynasty.²⁷³

The Censorate also routinely received copies of important memorials that arrived in the capital from the provinces. In 1043, Ou-yang Hsiu, then in the Bureau of Policy Criticism, requested that his agency also receive such copies. He objected to an executive order that had confined to the Military Affairs Commission circulation of a provincial report that described the rebellion of Wang Lun (1043). If their opinions were to be incorporated into decisions before they were promulgated as edicts, he argued, censors and policy critics must be informed of developments in a timely way.²⁷⁴ On the one hand, the 1082 reform removed the authority of the Censorate and Bureau to receive “copies,” and this was restored only to the Censorate in 1098.²⁷⁵ This provision accounts partially for the demise of the Bureau in relation to the Censorate after 1082. On the other hand, the reform granted authority to both bodies to impound documents from other court agencies, essentially the right to demand “copies” of specific documents.²⁷⁶

Officers of the Censorate and the Bureau presented their findings in two ways: at audience or through memorials. Both censors and policy critics each had one slot per month in the regular rotation for audience with the emperor. If the matter was urgent, they were required to request an audience through application to the Secretariat. This requirement caused delays, and opened the door for chief councilors to influence the process. In 1045, however, policy critics were allowed to take the next available audience slot upon simple notification at the Postern Office (ko-men ssu). In 1068, this provision was extended to censors.²⁷⁷ The normal time limits for audience presentation were waived for censors and critics, and other officials were required to cede time to them if the audience session was drawing to an end. They also enjoyed a certain measure of privacy during their audience with the emperor. Normally, officials were received at audience as part of a “group” (pan) that usually included other members of their agency. Censors and critics, however, for most of the Northern Sung, were received individually. Furthermore, at least after 1064, the emperor’s personal retainers (eunuchs and guards) were required to withdraw from earshot when a “speaker” had audience.²⁷⁸

A censorial report could also be submitted, without audience, as a memorial. In the early Sung, these were routed through the chief councilors’ office; but after 1017, they went directly to the emperor. Upon receipt, he had several options. He could “hold the report within” (liu-chung); that is, not refer the matter to the outside court. If the matter was held within, he could either do nothing, or take private action himself (if the matter concerned a chief councilor). If he chose to act publicly, he could annotate the report with his instructions for disposition and refer it to the Secretariat for implementation. Or he could refer the matter to the Secretariat to solicit consideration and advice. The State Council would then consider the matter, and the Secretariat would memorialize the emperor, suggesting a course of action and requesting permission to implement it. In any case, the censor who had submitted the report was informed of the outcome of his efforts.²⁷⁹

Censors and critics could also exercise their authority in less direct ways. A censor could route his report to the Secretariat, with or without a copy to the emperor. Although the Secretariat could not take action without memorializing the emperor, the goal of this tactic was to enlist support for the measure before it reached the emperor, and thus pressure him not to “hold it within.” Another technique that arose in the late eleventh century was to “copy” (fu-pen) the target of the indictment, as well as the Secretariat. The goal was to increase the indictment’s publicity and leverage its political effect in an effort either to pressure the emperor into accepting the report or to force the target to resign. A related technique that gained popularity in the Southern Sung was the practice of “open drafts” (lu-chang), whereby the target himself was sent a preliminary draft of the censor’s forthcoming indictment. In periods where the Censorate was under the control of strong chief councilors, receipt of an “open draft” was usually sufficient pressure to force the target to resign from office.²⁸⁰

The intense factionalism that arose in the latter half of the Northern Sung put severe pressures on the Censorate and the Bureau of Policy Criticism. Those who would control the court soon learned that they had first to control these sources of potential criticism. Between 1068 and 1071, Wang An-shih engineered the dismissal of nineteen censors and critics.²⁸¹ In 1086, Censor Sun Sheng (1038–99) warned that censors during Wang’s administration “were not the eyes and hears of Your Majesty but became rather the hawks and hounds of the chief councilor.”²⁸² The Bureau did not survive as a freestanding agency into the Southern Sung. The Censorate was so compromised that, by the mid-twelfth century, it had become, as Sun Sheng had warned, little more than an extension of the chief councilor’s office.²⁸³ The emperors themselves contributed to this decline. From the beginning, they were willing to use criticism when it suited their own ends. But the system allowed censors to comment on what the monarchs often considered internal matters, areas where criticism was less appreciated. In retaliation, an emperor sometimes reacted in ways that eroded the Censorate’s ability to function on any issue. For example, in 1051 censors railed against the cost of perquisites that Jen-tsung lavished on the family of his favorite consort. Jen-tsung responded by requiring them to petition the Secretariat for audience slots, thus effectively banning them, at least temporarily, from speaking at audience on any issue.²⁸⁴

Monarchs could also take other less direct measures. As long as he was willing and able to withstand the political pressure, the sovereign could “hold within” a report indefinitely. Empress Kao in 1088–9 held within and refused comment on twenty indictments submitted by Exhorter Liu An-shih (1048–1125) against Chief Councilor Hu Tsung-yü (1029–94).²⁸⁵ The empress was more forthcoming a year later when Exhorter Liu and Fan Tsu-yü memorialized that they had “heard on the wind” that Emperor Che-tsung, then aged twelve, was already having sex. When they urged the Empress to “protect the divine body,” she assured them and the chief councilor that the rumors were false.²⁸⁶ Custom forbade the monarch from inquiring about a censor’s sources, but he could return a report for further details and analysis, essentially accusing the censor of poor performance.²⁸⁷ By early Southern Sung, the system of documentary controls that insured the integrity of transmission from Censorate to monarch to Secretariat had broken down. In 1138, the head of the Censorate complained that Censorate reports were being edited and abridged beyond recognition: the text upon which the emperor took action no longer resembled the text which the censor had submitted.²⁸⁸ The thirteenth-century scholar Tu Fan (1182–1245) wrote that in his time Censorate reports that reached Emperor Li-tsung had been entirely rewritten and did not even possess the Censorate seal.²⁸⁹

Well before the end of the Northern Sung, the Censorate and the Bureau, despite a considerable body of supportive tradition and precedent, had failed to attain an institutional status that would allow them to compete at political parity with the monarch and the chief councilors. Both monarchs and councilors saw little to lose in the demise of their institutional monitors. Only the sovereign had power to preserve the strength of the remonstrance agencies, and he had his own motives for keeping them subservient. It was easier to allow the chief councilors to dominate these difficult agencies. The monarch, however, thereby lost a powerful check on the growing authority of the councilors. As early as 1055, Chao Pien (1008–84) warned Jen-tsung that the remonstrance agencies were the only effective check on the power of the chief councilors: if these agencies do not function, “you will not be informed, you will hear no sentiments from below, and ultimately your own position will be endangered.”²⁹⁰

But the literati and other officials also shared responsibility for this demise. When the chief councilor misused the Censorate for his own public, political ends, his techniques were simply a larger manifestation of ways the literati misused the institution for their own private ends. At the beginning of his reign, Emperor Hsiao-tsung was asked why he had been slow to fill vacant remonstrance positions. His excuse was perhaps self-serving but also caught a major piece of the problem: “Most literati (shih ta-fu) sold out their integrity long ago, so it’s hard to make these appointments.”²⁹¹ In the end, literati misuse of the remonstrance agencies destroyed their best chance to restrain the power either of their own leader, the chief councilor, or of their ultimate master, the monarch.

The audience

In theory, the Sung government made policy and implemented decisions in a simple way: one first “obtained the imperial will” (ch’ü sheng-chih) – that is, determined what the emperor wanted to do – then one issued an “edict” – that is, ordered the relevant authorities to implement these wishes. In practice, this simple process entailed a Byzantine mosaic of rituals, consultations, endless document shuffling and revising, all interlocked through ever-shifting compendia of bureaucratic precedent and regulation. The emperor could not simply tell someone what to do. To have legal force, the imperial will had to be expressed in writing and pass through proper oversight procedures. In the chaotic last days of the Northern Sung, an unruly street mob confronted Emperor Ch’in-tsung. A local tough offered to quell the mob in exchange for an appointment to office. On the spot, Ch’in-tsung took a sheet of paper and wrote, “make him Vice Minister of War (Ping-pu shih-lang).” Later, the chief councilor, even in the face of these extraordinary circumstances, refused to process the unorthodox appointment.²⁹² Even the emperor could not simply appoint someone to office without going through proper procedure. There were two ways that one could “obtain the imperial will” through a face-to-face encounter with the emperor at audience, or through the submission of a written “memorial” to him.

The English word “audience” translates the Chinese ch’ao, an old term that described the seasonal visits of feudal lords to the Chou dynasty sovereign. Both the Chinese and English terms cover a wide range of Sung pageants, cere-monies, and interviews at which a group of officials through ritual gestures paid obeisance to the sovereign. The time and location of the ceremony, the number and rank of the officials who participated, the length and depth of the interviews – all depended on the type of “audience” and varied continually over the course of the dynasty. At the upper end of the spectrum was the “grand audience assembly” (ta ch’ao hui), held three times yearly, including the first day of the year. Envoys from foreign countries, all court officials, and capital representatives of provincial governments offered simultaneous greetings and obeisance to the emperor and empress in an elaborate choreographed ceremony complete with musical accompaniment. A lavish banquet for the several thousand participants followed. At the other end of the “audience” scale, the emperor could issue an “inner invitation” (nei-yin) to an official for a private, open-ended discussion over tea, without guards.²⁹³

The daily audience between the emperor and his officials was the central decision-making vehicle of Sung government. However, the Sung inherited a hodgepodge of audience protocols from the T’ang and Five Dynasties and was slow to adapt these to its own needs. From a modern perspective, the basic problem was that the audience process served double duty as both a cere-monial ritual and a working session of government. These divergent demands often conflicted, and the two basic early Sung audience protocols reflect this conflict. The “regular audience” (ch’ang-ch’ao), or outer audience, was held each morning in the Hall of Cultured Virtue (Wen-te tien). The “regular obeisance” (ch’ang ch’i-chü), or inner audience, was held each morning in the Hall of Good Government (Ch’ui-kung tien). The former was descended from the T’ang period “front hall” audience (cheng-ya) and was revived in the early Sung. But by the middle of the eleventh century, this outer hall assembly of court officials retained only ritual significance as a manifestation of loyalty toward the emperor and as a leave-taking ceremony for officials departing to provincial assignment. Neither the emperor nor major officials with functional office, except censors, attended. Those who did attend – officials between positions or those whose positions carried no duties – assembled outside the Hall of Cultured Virtue in two groups, civil officials to the right, military officials to the left. After the two groups had bowed to each other, the censors led them into the hall, where guards and standards had already been displayed. There they waited until a chief councilor returned from the inner audience, at which time an usher announced that the emperor would “not sit.” The chief councilor then led the officials in a double kowtow to the vacant throne, a ritual known as “supervising the group” (ya-pan), and the outer audience was dismissed.²⁹⁴

There was also ritual and pageant but considerably more substance to the inner audience. At dawn, a party of imperial guards led by the Bearer of the Imperial Arms (Tai yū ch’i-chieh) met the emperor at the gate of the inner palace. They performed a double kowtow, and, after repeating the mantra “A myriad blessings upon the divine body,” escorted him to the Hall of Good Government. Attendance here was limited to civil officials with minister-in-attendance rank (about two dozen top officials) and a corresponding group of senior military officials. Officials attended in court regalia, which included their official robes, boots, and jade audience tablets (yü-tai). There were also officials from the Postern Office, who managed the audience, assorted ushers, and attendants. As the emperor ascended the dais, the guards took stations to the left and right of the throne, and the attendants repeated the mantra of greeting. At this point, officials who had received transfers and new appointments gave thanks to the emperor. After the palace guards, ushers, and eunuchs had made obeisance to the emperor, the presentation of memorials began. First came the chief councilors, followed by the supervisory officials of other major agencies. When the presentations were over, a eunuch announced, “No further public matters beyond the gate,” and the “morning audience” concluded.²⁹⁵

Depending on the period, the inner hall audience was held either every day or every other day. Afterwards, the emperor changed his robes, took breakfast, and often repaired to the Hall of Extending Harmony (Yen-ho tien), where he continued to hear reports from eunuchs and provincial officials, met with officials who had received an “inner invitation,” or viewed manuscripts and art objects newly acquired for the imperial collections. Such sessions were called “the second sitting in the back hall” (hou-tien tsai-tso). In addition to the outer, the inner, and the back hall audience, every fifth day all capital officials assembled in the inner hall to perform the “grand obeisance” (pai-kuan ta ch’i-chü), which consisted of a series of seven kowtows.

The Yüan-feng reform brought a measure of order to this redundant system. The fossilized outer audience was abandoned and the inner audience in the Hall of Good Government became the formal “front hall” ceremony. The reform established a three-tiered hierarchy of attendance requirements based on an official’s rank and his importance in the decision-making process: (1) ministers-in-attendance and their military counterparts attended daily; (2) every fifth audience day, middle-ranking court officials were included, the proceedings were moved to a larger hall, and the “grand obeisance” was performed (since every fifth day amounted in theory to six times per month, such officials were called “six-attendance officials” (liu ts’an kuan)); (3) twice per month, on the first and the fifteenth, all capital officials joined the proceedings. Memorials were presented at all audiences.²⁹⁶ Complicated regulations governed conflicts caused by rain, official holidays (every tenth day), the suspension of audiences due to the death of prominent persons, or imperial illness.²⁹⁷

With slight modifications (six-attendance officials became “four-attendance” officials), these new audience procedures continued into Southern Sung. However, in the chaotic years immediately following 1127, the court was often on the move and seldom had access to locations that could accommodate the larger assemblies. Even after the peace of 1142 and the rebuilding of the imperial capital at Lin-an, for reasons we shall explore below, Emperor Kao-tsung and Ch’in Kuei showed little interest in reviving the larger Yüan-feng audience protocols. However, in 1162 Emperor Hsiao-tsung immediately ordered restoration of the “four-attendance” audience. By 1166 ritual guidelines were complete and officials were ordered to rehearse the new ceremony. Once the emperor had been escorted to the dais, the schedule for memorials to be presented that day was read aloud. Then various groupings of officials, beginning with the Postern Office, the imperial guards, the palace command, and the heir apparent, made a “regular obeisance” of two kowtows. There followed another group from the Military Affairs Commission that did likewise, then another group of imperial princes, army personnel, and military commissioners that did likewise. Then the palace censor entered the hall and announced the “grand obeisance” and took his station. At this point, the chief councilors and officers of the Three Departments led civil and military officials into the hall. They divided into two groups, civil and military, faced the throne and performed the “grand obeisance” of seven kowtows. The top officials of the Three Departments and the Military Affairs Commission then ascended the dais to memorialize, followed by those with new appointments, then other officials. When the presentation of memorials was finished, the emperor ended the audience and departed the hall.²⁹⁸

One should not downplay the ritual aspects of these audiences. The pageantry, military display, and ritual gesture emphasized the power of the sovereign and the loyalty of his retainers. The ceremonies also emphasized the diversity of these officers, their affiliation to the many divisions of officialdom, and their common solidarity in support of the sovereign. The audience also publicized an individual’s position in the hierarchy of officials and defined his relationship with his colleagues. An official’s personal-rank grade determined his exact placement within the two large audience divisions of civil and military officials. It is telling commentary that, among Compendium of Sung documents chapters on audience protocol, the longest subsection concerns “placement” (wei), where an official or group should stand in relation to others.²⁹⁹ Each attendee at audience had a precise position at which to stand, and that placement was a public manifestation of his “place” in Sung government.

Despite the pomp and ritual, the presentation of memorials remained the main order of business at audience. These presentations occurred under extremely formal conditions, in an enormous hall with hundreds of guards, soldiers, and assorted onlookers. Although most were too distant to hear anything actually said, the proceedings were nevertheless very public. Hsiao-tsung’s audience protocol called for a reading of the daily schedule of memorial presentations, even before the obeisance had taken place. There were essentially four ways to get on the presentation schedule. First, in early Sung, the supervisory officials of the Secretariat–Chancellery, the Military Affairs Commission, the Finance Commission (San-ssu), the K’ai-feng prefectural government (K’ai-feng fu), and the Judicial Control Office (Shen-hsing yüan) each led their “group” (pan) in the daily presentation of memorials. Later, daily presentation was limited to the Secretariat–Chancellery and the Military Affairs Commission, followed by three additional groups. Second, the emperor could directly order a specific official to come for audience (chao-tui), often because the official possessed information or expertise the emperor desired to access firsthand. Third, in the reverse procedure, an official could request an audience to submit his memorial to the emperor (ch’ing-tui). State Council members could submit with little advance notice. Other officials who ranked high enough to enjoy this privilege, however, were first required to petition the office of the chief councilor for an audience slot. This tradition arose early in the dynasty when Chief Councilor Chao P’u (922–92) required that his office first approve the content of all audience memorials before their presentation would be scheduled.³⁰⁰ Although these restrictions eased somewhat under Emperor Chen-tsung, some form of prior restraint seems to have remained in place through 1082, when Shen-tsung ordered that chief councilors not obstruct any memorial from an official who was qualified to submit.³⁰¹ Such open access to the throne, however, was an exception. Southern Sung scholars often lamented that technical restrictions placed on qualified presenters effectively excluded them from audience presentation.

Last was the “revolving audience” (chuan-tui; also “rotating audience,” lun-tui), an institution related in spirit to the “speaking” function of the remonstrance agencies: it was meant to expose the emperor to a wider range of information and afford him the opportunity to meet every court official in a personal face-to-face encounter.³⁰² In theory, at the “grand obeisance” on every fifth day, the memorial schedule was to include one or two officials from among the general population of court officials. These rotators were to present critical memorials on nonroutine issues, such as government failings, legal injustices, or hardship conditions among the people. In time, one would proceed through all court officials, and the rotation would begin again. A practice with T’ang antecedents, the revolving audience was a sporadic event in the Sung. Biographies often mention that an official attracted the emperor’s attention or presented a stunning memorial during his turn at the revolving audience. But the first attempt at serious implementation was not until the reign of Shen-tsung, and by the 1090s most upper-level officials were exempt from the rotation. Hui-tsung reduced the schedule to one rotator per month. Kao-tsung restored the plan, but Ch’in Kuei, chief councilor from 1138 through 1155, was so sensitive to criticism that most rotators, torn between their desire to impress the emperor and their fear of offending the powerful Councilor, took sick leave on the day of their rotation. By the early 1160s, the practice was nearly defunct.³⁰³ Under Hsiao-tsung, the rotation was restored, but the chief councilors, to avoid the possibility of criticism in open court, often transferred officials whom they did not trust just before their rotations were due. Many officials, therefore, never had their moment with the emperor, and the revolving audience became, in the words of Li Hsin-ch’uan, “a rotation of positions, not of men.”

During memorial presentations, each group mounted the dais, where the emperor was seated on a divan. Individuals read at least a portion of their memorials aloud. The emperor could stop the reading, question the official, or do nothing. The prolific records of discussions between the emperor and his senior advisers in the Long draft and the Chronological record doubtless derive from conversations that ensued following group submissions by the Secretariat–Chancellery and the Military Affairs Commission. A basic principle of Sung government was that no official (save remonstrance officials and rotators) could memorialize on matters outside the jurisdiction of the office he currently held. The Sung founders insisted upon a strict division of civilian and military responsibility between these two top agencies, and so they memorialized as two separate groups, neither knowing the other’s business with the emperor. During the Tangut wars in the 1040s, however, the emperor often ordered the agencies to memorialize jointly on common issues.

In Sung practice, a joint memorial implied that the parties who submitted the document agreed upon its contents. An edict of 1082, for example, ordered mid-ranking officials of the Six Ministries to memorialize together with their supervisors and forbade them from requesting private audiences.³⁰⁴ Surviving memorials confirm this understanding. They seldom provide a range of policy options, but rather forcefully argue for one option against the foolhardiness of others. Dissent was permitted, but the dissenter then had to submit his own memorial. As the edict of 1082 implies, an official’s rank sometimes precluded this option. When the Secretariat and the Commission were ordered to jointly memorialize, if they could not agree, then they submitted multiple memorials. This practice eroded the boundary between the civilian and the military authority of the two agencies, and opened the door for each to memorialize on the other’s business.³⁰⁵ Hsiao-tsung’s reading of the daily memorial schedule, therefore, was more than ritual pomp. The very composition of the list outlined the major political divisions of the moment. Tseng Pu’s diary, for example, shows him engaging in a wide array of submission vehicles, including joint submissions, joint memorials that resulted from deliberations ordered by the emperor, and individual submissions.

Irregular individual submitters were required to deliver two copies of their text to the Postern Office the day prior to their audience, along with a capsule biography and summary of their official career (in modern terms, a curriculum vitae).³⁰⁶ This raises the question whether the emperor had foreknowledge of the memorials that were presented to him. Kao-tsung’s statement that he read audience memorials after the session strongly suggests that in many cases he had no such knowledge and that the public reading of the document was itself a formality. There were obvious time constraints on the number of memor-ials that could be presented. Primary sources speak not in terms of the number of individual memorials submitted but rather of the number of “groups” (pan) that could be accommodated in one session. This number appears to have varied considerably with each emperor, his age, and the political situ-ation. Audiences generally began at dawn, somewhere between five and seven in the morning. In 998, Chen-tsung had finished hearing memorials by nine, then ate, and repaired to the back hall for the “second sitting” till noon, during which he did not hear memorials. In 1006, he complained that audiences were taking too long and limited submissions to five pan, one each from the Secretariat–Chancellery and the Military Affairs Commission, and three additional pan whose composition the Postern Office scheduled for each session. In 1068, Shen-tsung reduced the number of additional pan from three to one.³⁰⁷ Allowing an hour for initial pageantry, Chen-tsung’s schedule would leave about two hours (from seven to nine) for five pan or about twenty-five minutes per pan. In 1029, however, Empress Liu and Jen-tsung heard nineteen pan in one session; nine before the break at nine o’clock and an additional ten before adjourning in the early afternoon.³⁰⁸ By the same calculations, this would allow only about fifteen minutes per pan. In the early Southern Sung, audience submissions were often limited to only two pan. Such constraints could have left little time for anything but a token reading of each memorial.

Critics of the audience submission system – and these included emperors themselves – lamented its rigidity and bureaucratic nature. In 1066, Emperor Ying-tsung complained that he met every day with his ministers “yet we never have time for a leisurely discussion of the principles of government but rather are constantly harassed by this plethora of documents.” He ordered the Secretariat to handle more matters internally, and established a triage system: (1) on major issues, the Secretariat was to continue to “obtain the imperial will” through audience memorials; (2) on matters of detail, the Secretariat should submit a “routine petition” (shu-chuang) through non-audience channels for imperial approval; (3) on matters already covered by statute, the Secretariat should issue its own directives to subordinate agencies.³⁰⁹

In the late twelfth century, Chu Hsi also criticized the audience process. He confirmed that presenters were able to read only a few sentences before their time was up. There was little discussion and consequently nothing was decided; even small matters were referred down for additional consultation and recommendation, which could take months or years to complete. He made two suggestions for improvement. First, the emperor should sit behind a long table on which the presented memorial could be spread out for his immediate perusal. Second, subordinates of the presenting agency should be available in the hall for immediate consultation. In this way, if the emperor had a question, he could confer with knowledgeable parties and, on the spot, approve or deny the memorial. Chu cites a Six Dynasties precedent for this procedure and claims himself to have witnessed its effectiveness when he was provincial administrator at Chang-chou. There, no one would volunteer to speak first at staff meetings, he reported. But, if a written text was presented for comment, active discussion ensued and led to a speedy decision. In other words, the written text should be a base for oral conversation; it should not become an end in itself.³¹⁰

The general inefficiency of the audience memorial system led to several mechanisms to preserve the precedent but work around the difficulties. Most prominent was the “stay-behind” (liu-shen), essentially a private session accorded to a senior official, at the request either of the emperor or of the official himself. Normally, all ministers-in-attendance could request a stay-behind, but the scope of the privilege varied with the emperor. In 1118, Hui-tsung forbade stay-behind requests from anyone except Ts’ai Ching. In 1133, Kao-tsung ordered the privilege limited to State Council members.³¹¹ Eligible officials were required to request a “stay-behind” in open court; therefore, although the audience was private, the fact of its occurrence was not. Some sources suggest that these stay-behind sessions occurred at the conclusion of group presentations, in other words, between the pan; others suggest that they occurred mainly during the “second sitting.”³¹² Tseng Pu’s diary confirms that memorials submitted during stay-behind sessions were secret. On one occasion, during his own stay-behind, Tseng saw memorials left with the emperor during a stay-behind by Chief Councilor Chang Tun, and Tseng suspected that the emperor’s questions to him concerned the contents of Chang’s memorials.³¹³

Memorial and edict

Audience presentation was part of a larger system of governmental communication via memorial (in modern terms, a memo). This system funneled written documents upward through proper channels to the monarch and the chief councilor. Memorials submitted from “inside,” from within the court, often served as the base text for the edicts, decrees, and orders that emanated from the offices of the monarch and the chief councilor and that counted as “law” in the Sung state.³¹⁴ Memorials submitted from “outside,” from provincial authorities, provided regular reports on local conditions. In 984, all graded officials in the provinces were granted permission to memorialize through channels (that is, their memorials had to be submitted first to their super-iors for eventual transmission to the court). Provincial officials above the level of prefect were permitted to address the emperor directly; those below were required to address the Secretariat–Chancellery or the Military Affairs Commission.³¹⁵ Top provincial officials were required to report regularly upon the termination of their tours or, in Southern Sung, six months after their arrival on location.³¹⁶ In 995, the secretariat drafters were ordered to screen the flow of provincial documents for those that warranted action.³¹⁷ In 1070, between 400 and 500 memorials from the provinces arrived in K’ai-feng every day.³¹⁸

There were many restrictions on how a memorial could be written and what it could contain. They had to be signed and could not be submitted anonymously. They could neither address matters outside the jurisdiction of the writer, nor request that the emperor “keep them within” or thwart court opposition to a proposed course of action. They could not request that the writer come to K’ai-feng to make his case in person.³¹⁹ Most memorials therefore concerned routine, noncontroversial matters. The edict of 984 stated that “those whose suggestions are accepted will be rewarded; those whose are not will incur no punishment.”³²⁰ This specific dispensation suggests that writers often did incur formal or informal sanctions for unwelcome suggestions.

As we have seen above, the legal base for the purges of the late eleventh century rested largely upon written submissions to the court. In a system where negative reports could generate Censorate investigations and endanger the career advancement of all officials in the chain of command, it took a very brave or a very foolhardy soul to submit a critical memorial or even one that suggested a new way of doing things. So rare, in fact, were such memorials that the emperor sometimes designated when they were permitted. Natural disasters, negative astrological phenomena, or military crises often precipitated these calls for criticism. Edicts “seeking speakers” (ch’iu-yen) or efforts to “widen the speakers’ way” (kuang yen-lu) often stipulated that the writers would suffer no sanctions for their criticism. But there was no “law” to protect the “speaker” and no guarantee that a future administration would not use the text against its writer.³²¹

There were two forms of memorial. Ministers-in-attendance and officials with “senior director” status (about eighty officials) were entitled to use the “administrative memorial” (cha-tzu); all others used the “submitted petition” (tsou-chuang).³²² The former was confined to senior court administrators, former chief councilors, and top provincial officials on urgent matters. Differences in the color of paper, the number of characters per line, signing protocols, and routing distinguished the two forms. Either form of memorial could be submitted either sealed (shih-feng) or unsealed (t’ung-feng). The former was used when the official himself reported or was ordered to report on secret or urgent matters, natural disasters or anomalies, and legal issues. Sealed memorials went directly to the emperor. An unsealed memorial was used for all other matters and contained a summary of its contents written on the outside of the envelope. The penalty for unwarranted use of the sealed memorial was 100 blows of the bamboo rod.³²³

The transmission, copying, and distribution of memorials and edicts were a major enterprise at the capital and involved several agencies. The Capital Memorials Office (Tu chin-tsou yüan) was located just outside the Great Inner and served as a central post office and copy center for the inward receipt of memorials from the provinces and the subsequent outward distribution of edicts. Descended from a T’ang agency that housed capital liaisons between the court and the provincial governors, at its founding in 982 the office already employed 200 clerks and scribes under the general supervision of the Secretariat–Chancellery. The same office also distributed a daily administrative gazette (ti-pao) that reprinted appointment and promotion notices, texts of important edicts and memorials, and general court news of interest to officials.³²⁴ After receipt and cataloguing, the memorials were transferred to the Transmission Office (Yin-t’ai ssu) within the palace. This office logged the memorials once again, sorted them, and routed them to their ultimate destinations: emperor, Secretariat–Chancellery, Military Affairs, or Finance. If the memorial was unsealed and addressed to the emperor, copies were made and distributed to relevant agencies.³²⁵ Provincial memorials destined for the emperor were finally delivered to the Memorial Presentation Office (T’ung-chin ssu), where they were combined with those from court officials who deposited their memorials in person at the Postern Office. A eunuch-staffed agency, the Memorial Presentation Office was between 993 and 1082 under joint supervision with the Transmission Office and co-ordinated the flow of paperwork to and from the emperor.³²⁶

The above network was for memorials submitted through regular channels. There is evidence that both provincial and court officials sought to avoid oversight as well as the delays and information leaks that often resulted from submission through the Capital Memorials Office. An edict of 1126, for example, forbade provincial officials from submitting memorials through the Palace Eunuch Service. This action suggests that such officials hoped, through eunuch intermediaries, to bypass the preliminary offices and slip their texts directly into the document flow to the emperor through the Memorial Presentation Office.³²⁷ Another text of the same period hints at a similar strategy. The various “drum offices” were transmittal agencies for petitions for redress of injustices from officials and the general populace – topics that precluded use of regular channels – and from lower-ranking officials with no other avenue for submission.³²⁸ However, Ch’ao Yüeh-chih (1059–1129) relates in his memoirs that there was a fee of 2,000 cash payable to every official on the drum office staff who handled the petition and that the total cost per submission averaged between 10,000 and 20,000 cash. A submission to the drum office also required a guarantor, although this requirement seems to have been abolished in 1133.³²⁹ Since this sum represented the average monthly salary of a mid-level court official, many sought alternative channels. Most popular was to use connections among the emperor’s female palace staff to bypass even the Memorial Presentation Office.³³⁰

Once the imperial will had been determined, an edict was formulated and promulgated for implementation. The Sung history lists seven categories of imperial edict, classified in a hierarchy according to the importance of the topic and the status of the person or group to whom it was directed.³³¹ The category determined where the edict was composed. The emperor’s academicians (hsüeh-shih) composed the highest categories. The chief councilors and other State Council members or their staffs composed the lesser categories. In the early Sung, the emperor simply marked his approval on the submitted memorial and returned it to the relevant agency. In 990, however, all approved memorials were routed through the Secretariat–Chancellery, the Military Affairs Commission, or the Finance Commission, as appropriate, where an edict text was composed. Tseng Pu’s diary, for example, records that after audiences he “composed the imperial will” (tso sheng-chih) both individually and with other members of the State Council.³³² Once these drafts were complete, these agencies then “re-memorialized” (fu-tsou) for permission to promulgate the final text.³³³

“Directed edicts” (nei-chiang, also chung-chih, nei-p’i), as mentioned above, were edicts generated within the monarchy itself and issued directly to the relevant agency. They often concerned appointments for relatives of palace women and eunuch requisitions for palace supplies. Depending on the period and the emperor, directed edicts may or may not have been routed through the Secretariat.³³⁴ In order to forestall abuse, Emperor Jen-tsung in 1059 author-ized an agency that received a directed edict to delay its implementation for one day and to submit to him a subsequent “retained memorial” (chih-tsou, also tsou-ping) to verify that the directed edict indeed reflected his intention.³³⁵ Directed edicts were not uncommon. In 1082, Ts’ai Ching completed a collection that dated from 1067 through 1079 and included 1,346 such edicts, an average of more than one every three days.³³⁶ The “retain memorial” was an important tool in the literati struggle against imperially dispensed patronage. In 1193, for example, Chief Councilor Liu Cheng (1129–1206) received a directed edict to appoint one of Emperor Kuang-tsung’s affinal relatives to office. Liu objected with a “retain memorial.” When the emperor refused to accept the memorial, Liu left both documents on the emperor’s divan and left the audience.³³⁷

The Yüan-feng reforms of 1082 – especially separation of the Secretariat and the Chancellery and reconstitution of the Department of State Affairs – introduced major changes that affected how edicts were composed and promulgated. Emperor Shen-tsung, as described above, attempted to create three equal “departments” (Secretariat, Chancellery, and State Affairs) headed by two chief councilors, one “left” who was also concurrent head of the Chancellery – the “Chancellery Councilor” – and one “right” who was also concurrent head of the Secretariat – the “Secretariat Councilor.” With each of the three departments performing a distinct and separate function, the new organization was to work in the following way. (1) On important matters, the Secretariat, after consultation with the emperor in audience, “composed the imperial will” on yellow stationery; these drafts were called “approved yellows” (hua-huang). For lesser matters, the Secretariat submitted non-audience “routine petitions,” which, after imperial approval (hua), were transcribed as “recorded yellows” (lu-huang). The Military Affairs Commission followed a similar procedure, but used white stationery. (2) These drafts were then forwarded to the Chancellery, where the “yellows” were kept as base copies. After checking for errors, these drafts were again recopied and submitted to the emperor for final approval (hua-wen), a process known as “re-memorializing” (fu-tsou). The “approved yellows” were then again recopied and submitted to the Supervising Secretary for final Chancellery clearance. (3) At this point, the document became an “edict” (ch’ih), bore the signatures of both chief councilors and other Secretariat and Chancellery personnel, and was transmitted to State Affairs for recording, distribution, and implementation.³³⁸

In addition to creating the three separate departments, the Yüan-feng re-organization also enhanced the authority of the Secretariat drafters to “return” imperial scripts, a practice that, as seen above, was first employed in the late 1030s. Also known as “return for correction” (feng-po), Shen-tsung extended this authority to the Supervising Secretaries in the Chancellery. Both the Secretariat and the Chancellery now had key officials empowered to detain documents. Shen-tsung also physically separated these officials with “return authority” in the Secretariat Rear Section and Chancellery Rear Section, where they became supervisory officials. These measures gave the Secretariat drafters, for whom there were six billets in the Secretariat Rear Section, and the Supervising Secretaries, for whom there were four billets in the Chancellery Rear Section, a certain measure of independence from the chief councilors. They were empowered to return to its source any document they found wanting either for technical reasons or on grounds of content, including personnel appointments, policy deliberations, and legal matters.

The division of the Secretariat and Chancellery also entailed a new audience protocol. Before 1082, all chief councilors, as supervisory officials of the combined Secretariat–Chancellery, memorialized together as one “group.” Shen-tsung, however, insisted that, as heads of the new Chancellery and Secretariat, the left and right chief councilors memorialize during audience in two separate groups. His intention was to balance the two councilors against each other, an intention also apparent in the “checks and balances” built into the relationship between the three departments.

There was, however, a major flaw in Shen-tsung’s design. The left “Chancellery Councilor,” with control over the important ministries of Personnel and Justice, was intended as the senior of the two councilors, yet only the right “Secretariat Councilor” had authority to initiate policy by “obtaining the imperial will.” The Chancellery had a powerful oversight function, yet it could not memorialize on new initiatives. This organization forced the two chief councilors to co-operate. Yet, if they did not, it produced instant gridlock. In the early Southern Sung, Yeh Meng-te (1077–1148) penned a trenchant critique of this original Yüan-feng design. When the councilors agreed on an issue, they requested permission to memorialize together. This made the oversight function of the Chancellery superfluous, since the initial consensus already signaled the concurrence of the Chancellery Councilor. When they did not agree, the Chancellery was required to “return for correction” through the Secretariat. The Secretariat Councilor would then often detain the matter rather than negotiate a compromise. Rather than effecting a balance between the two councilors, the Yüan-feng reform set them at institutional loggerheads, and the Secretariat Councilor, the junior partner, had the bureaucratic advantage over his senior colleague. Thus the Secretariat soon became the most powerful of the three departments, and Shen-tsung’s design was thwarted almost from inception.³³⁹

Already in late 1082, Shen-tsung exempted from Chancellery review documents that involved military activity on the borders.³⁴⁰ As soon as Shen-tsung died in 1085, Ssu-ma Kuang and Lü Kung-tso (1018–89) requested from Empress Kao, and were granted, permission for the chief councilors to jointly memorialize. As Lü maintained, “we are all in the same boat crossing the stream,” implying that the oversight function of the Chancellery was superfluous in a unified administration.³⁴¹ Ssu-ma Kuang went further and advocated a recombination of the Secretariat and Chancellery. His son submitted his memorial posthumously in 1089. It was not acted upon at the time but is among the most revealing political documents of the age and served as the foundation for the eventual recombination of the two agencies in 1129.³⁴²

Ssu-ma objected to the excessive delay and paperwork that the round-robin nature of continual deliberation and consultation produced:

Ssu-ma argued that the authority to “return for correction” in the Chancellery was redundant and created tension among the top officials. “It serves no purpose except to double the number of clerks and multiply paperwork.”³⁴³

The autocratic councilor

The autocratic councilor, or in Chinese terms the “weighted minister” (ch’üan-ch’en), featured prominently in Sung political life. Certainly, dynasties both before and after Sung had their share of autocratic politicians. But the Sung was unique in Chinese history for the long periods when sole councilors dominated the political scene and for the dynasty’s repeated return to this pattern of political organization. “Sole councilor” (tu-hsiang) refers to a political circumstance in which, instead of the statutory two or three, there was only one chief councilor. During the 316 years of Sung rule, almost half, or precisely 145 years, were periods when only one chief councilor held the office. In the Northern Sung, sixteen of the seventy-two men who served in the office functioned for a time as sole councilor. These periods constituted 37 percent of the time between 960 and 1126. Twenty-two of the sixty-one councilors in Southern Sung served as sole councilor, equivalent to 62 percent of the years between 1127 and 1276.³⁴⁴ Writing in the early Yüan, Wang Ying-lin, the last of the great Sung scholars, directly attributed the fall of his dynasty to the increasing periods of autocratic rule by sole councilors.³⁴⁵

Traditional – and most modern – historians treat the issue of autocratic rule in moral terms. The Sung history of 1345 gathered the biographies of many of these councilors in a section entitled “nefarious ministers” (chien-ch’en). Ou-yang Hsiu devised this historical category in the mid-eleventh century for his New T’ang history (Hsin T’ang-shu), a creation not unrelated to the increasingly partisan politics of his own day. The Tao-hsüeh movement in late Southern Sung turned this concept into a guiding principle for the telling of Sung history and determined which councilors were to be included in the category.³⁴⁶ Political figures consigned to this “nefarious” category were “petty men” (hsiao-jen) who acted from base and selfish motives at the expense of the sovereign and the state.³⁴⁷ The most prominent “nefarious ministers” in the Sung history are Chang Tun, Ts’ai Ching, Ch’in Kuei, Han T’o-chou (1152–1207), and Chia Ssu-tao (1213–75).

However, if one leaves aside the Tao-hsüeh compulsion to explain the past as an eternal morality play, and examines the Sung propensity for autocratic councilors from an institutional standpoint, a rather different picture emerges. There is a strong relation between the Yüan-feng reforms and the rise of the sole councilor. The first period of long-term sole councilorship, that of Chang Tun, alone in office from 1094 through 1100, began little more than a decade after the Yüan-feng reforms. Some evidence suggests that Emperor Shen-tsung, perhaps with his past experience with Wang An-shih in mind, forced the separation of the Secretariat–Chancellery, with its counterpoised left and right councilors, in order to check the growing power of the chief councilors and to restrain the political factionalism that arose in the wake of the New Policies. Li Ch’ing-ch’en (1032–1102), a major contemporary politician, once remarked, “in his late years Shen-tsung set up the three departments so he could divide the authority of the councilors and have them watch and check on each other. These were farsighted plans.”³⁴⁸

Most modern historians, following Chu Hsi, note that Shen-tsung pushed through the Yüan-feng reforms against the advice of senior advisers, including the retired Wang An-shih.³⁴⁹ But, as early as 1058, Jen-tsung had appointed a commission to study proposals for institutional reform. A comparison of the commission’s findings and Shen-tsung’s 1082 reforms is revealing. Shen-tsung adopted many of the commission’s suggestions, including the creation of the three departments, but with two striking exceptions. The earlier commission had advocated abolishing the Military Affairs Commission and moving its functions into the Secretariat and the Six Ministries, a move Shen-tsung refused to consider. It had also recommended a looser, “corporate” definition of the chief councilorship, rather similar to the existing mid-century State Council than to the rigidly defined, interlocking roles that Shen-tsung eventually imposed.³⁵⁰ Both these aspects of Shen-tsung’s 1082 reform sought to exert the power of the monarch over that of the chief councilors. When Li Ch’ing-ch’en commented that Shen-tsung’s plans were “farsighted,” the immediate political backdrop was Chang Tun’s sole councilorship. In other words, Shen-tsung’s reforms precipitated precisely what they had been intended to prevent: the concentration of too much power in the hands of one official.

As the above description of the document flow through the three departments illustrates, the reforms created institutional gridlock if the chief councilors did not agree. Despite Empress Kao’s reluctance to undo her son’s reforms, she acquiesced almost immediately – perhaps from simple bureaucratic necessity – in ways to circumvent the problems they created. In 1086, Wen Yen-po was named “manager of vital military and national issues” (p’ing-chang chün-kuo chung shih). This was an old T’ang designation that conferred upon Wen an honorary, elder-statesman status but theoretically ranked him above the chief councilors Ssu-ma Kuang and Lü Kung-tso. Two years later, however, Empress Kao named Lü Kung-tso, then Right Chief Councilor, “manager of military and national issues” (p’ing-chang chün-kuo shih). In the opinion of the greatest of Southern Sung historians, this simple omission of the graph “vital” from his title afforded Lü total control over all matters in the three departments and the Military Affairs Commission.³⁵¹ He became in effect a super-councilor with authority to transcend the limitations that the Yüan-feng reform had imposed on regular councilors. In the view of his adversaries, the man who three years earlier had said “we’re all in the same boat” had managed to commandeer the entire fleet. Although Lü died the following year, subsequent “nefarious ministers,” such as Han T’o-chou, cited the precedent that Lü had created to justify their own accumulation of extraordinary authority. After the dismissal of Ts’ai Ching in 1125, a long edict restored the original 1082 configuration of the three departments and specifically stated that Lü Kung-tso had begun the deformation of those institutions in 1088.³⁵²

Most historians dismiss Hui-tsung’s rule from 1100 through 1125 as a time of stagnation and moral decline – in short, as prelude to the debacle of 1126. Accordingly, the period is little studied. But this quarter-century was pivotal to the development of Sung political institutions. It marked both a culmination of trends set in motion since the mid-eleventh century and laid groundwork for the political structures of the ensuing Southern Sung. Hui-tsung’s reign witnessed the struggle between two competing notions of government. On the one hand, as outlined above, Ou-yang Hsiu, Ssu-ma Kuang, and colleagues promoted a government of closely interlinked agencies, each with a carefully described function, each headed by shih ta-fu, their interaction regulated through an elaborate balance of function, with strong oversight powers accorded to the remonstrance agencies. Ou-yang Hsiu’s chapters on institutions in the New T’ang history illustrate aspects of this concept. Emperor Shen-tsung basically accepted this vision, but, in an effort to defend the monarchy against the growing power of the literati, he pushed the concept to the extreme, and created gridlock and disjunction.

On the other hand, in 1069 Shen-tsung allowed Wang An-shih to form the Bureau for the Implementation of Fiscal Regulations (Finance Planning Commission) as an ad hoc agency to co-ordinate and streamline the New Policies. In defense of the new bureau, Wang An-shih noted in exasperation, “At present, if the Secretariat needs to pay for anything that costs more than 100 cash, it must proceed through Finance Commission clerks, who then memorialize before the funds are disbursed.”³⁵³ His opponents, however, saw more sinister motives. Han Ch’i (1008–75) objected that there was no precedent in Sung history for an agency that decided issues outside the Secretariat and without memorializing. “One has here a ‘Secretariat’ that is beyond the jurisdiction of the Secretariat.”³⁵⁴ Although short-lived, the bureau created a powerful precedent for an alternative vision of how government could work. The emperor, in close collaboration with one trusted adviser, rules through temporary and fluid bureaucratic substructures that largely bypass the elaborate organization chart of established government. Essentially, as Han Ch’i foresaw, the emperor allows one adviser to create a government within the government. Clerks, not shih ta-fu, staff these governmental substructures. Emperor Hui-tsung and Ts’ai Ching developed this model during the first quarter of the twelfth century because it offered a powerful promise of better efficiency and filled the bureaucratic void left by gridlock of the regular system.

Ts’ai Ching proposed the Advisory Office (Chiang-i ssu) as early as 1094 and specifically cited its 1069 precedent.³⁵⁵ Formed in 1102 to re-establish the New Policies, the office was divided into seven units, each charged with devising and implementing solutions in a specific problem area: the imperial clan, state finance, commerce, salt production, tax collection, the excessive number of officials, and livestock management. Ts’ai Ching headed the office with the adjunct title of Supervisor of the Advisory Office (Ti-chü chiang-i ssu) appended to his title as chief councilor. He established a sub-office in the Military Affairs Commission, headed by his brother, Ts’ai Pien (1058–1117), to co-ordinate with that agency. He staffed the central headquarters with seven trusted lieutenants, appointed three officials on temporary assignment from other agencies to head each subunit, and provided the whole with a support staff of several hundred clerks.³⁵⁶ The Advisory Office could also dispatch personnel to the provinces to overrule local officials and oversee implementation of its policies. Viewed against the backdrop of both pre- and post-1082 organization charts, the Advisory Office was unorthodox. Staffed by carefully chosen officials assembled to solve specific problems, the office sliced through traditional bureaucratic lines of authority with a power that Ts’ai Ching derived from his unique association with Hui-tsung. In 1104, Ts’ai Ching and Ts’ai Pien closed the office, claiming that its goals had been accomplished. This fact suggests the office was not, as is commonly portrayed in secondary literature, a vehicle for building Ts’ai’s personal clique, but rather a temporary mechanism to expedite a complex policy agenda.

The popular image of Hui-tsung’s reign as one long period of political decline and autocracy is misleading. Ts’ai Ching served four times as chief councilor: he was appointed four times, but also dismissed four times. The new conception of streamlined government faced constant opposition from many quarters, mainly from those officials who were cut from the loop.³⁵⁷ Hui-tsung’s vacillation toward Ts’ai Ching mirrors this opposition. In 1122 Wang Fu (1079–1126) established the ad hoc Frontier Defense Office (Ching-fu fang) to circumvent the Military Affairs Commission and co-ordinate the alliance with the Jurchen against the Liao. His long proposal to establish this office reveals that many of the 1082 checks and balances built into the relationship among the three departments were still in force.³⁵⁸ As late as spring 1126, after the retreat of the Jurchen army from the first siege of K’ai-feng, certain officials, including the subsequent Tao-hsüeh paragon Li Kang (1083–1140), revived the Advisory Office as a last-ditch effort to co-ordinate policy against the Jurchen. But the Remonstrator Ch’en Kung-fu (1077–1142) objected that the organization was unorthodox and violated mid-eleventh-century models. Other officials called the idea “a joke” (k’o-hsiao): “We submit that each of the Six Ministries of the Department of State Affairs should attend to its own affairs, then forward their documents to the Secretariat to obtain the imperial will.”³⁵⁹ Less than a year later, the Jurchen invaded the capital for a second time, and the Northern Sung came to an end.

As we have seen above, when Emperor Kao-tsung revamped the central administration in 1129, he adopted Ssu-ma Kuang’s plea to recombine the Secretariat and Chancellery and brought back the post of assistant chief councilor. But he retained the Six Ministries under the Department of State Affairs, in essence combining elements from both pre- and post-1082 organizations. However, the first councilor appointed to the new post, Lü I-hao (1071–1139), also held adjunct appointment as imperial defense commissioner (yü-ying shih). The Imperial Defense Command had been created as an emergency, unified military command following the collapse of the Northern Sung army in 1127.³⁶⁰ In 1130, however, censors attacked Lü for using the agency as a personal power base, and he was dismissed from office.³⁶¹ The Imperial Defense Command was reborn as the Office for Emergencies (Chi-su fang). A central clearing house for military correspondence with frontline forces, this office was designed to circumvent the delays and leaks that ensued if these documents were processed through the Military Affairs Commission. The Office for Emergencies became a standard component of sole-councilor rule: Han T’o-chou established one in his own residence in 1205, and Chia Ssu-tao established another within the Secretariat in 1273.³⁶²

It is also misleading to view the sole ministers – even the “nefarious ministers” – as usurpers of imperial authority. The Tao-hsüeh historians created the chimera of usurpation to absolve the imperial house of responsibility for failed initiatives, and eventually for the fall of Southern Sung, under the sole councilors. But in every case, these councilors, as established in the example of Shen-tsung and Wang An-shih, worked in close collaboration with the monarch. Often a sovereign turned to the sole councilor, and to his fluid administrative arrangements, as the only alternative to the ever-feuding shih ta-fu and regular bureaucratic paralysis. The relationship between Emperor Kao-tsung and Ch’in Kuei, the first great sole councilor in the Southern Sung mold, presents the clearest example. Early in his reign, Kao-tsung struck a balance between the competing forms of government structure, revived the Censorate and policy critics, and laid ground rules for proper bureaucratic documentation. But the endless bureaucratic feuding of the 1130s drove him in desperation to Ch’in Kuei, and once that decision was made, Kao-tsung stuck with his choice, as Ch’in Kuei negotiated the 1142 peace with the Jurchen, built the new capital at Lin-an (Hang-chou), and laid the economic foundation for the restored Sung state.³⁶³

These accomplishments were achieved through a wholehearted embrace of the Hui-tsung/Ts’ai Ching model of government, where fluid bureaucratic substructures circumvented routine government organization. In this context, no document in Sung political history reveals more than Hung Mai’s note on Southern Sung staffing levels in the Six Ministries. He relates that in the mid-1150s, during the last years of Ch’in Kuei’s administration, there was exactly one regularly appointed, graded official in the entire twenty-four subunits of the Six Ministries – a director (lang-chung) in the Ministry of Justice named Sun Min-hsiu. The entire Ministry of Personnel was run by a clerk named Chang Yün.³⁶⁴

Hung Mai links these stunning statistics to Ch’in Kuei’s disdain for the regular bureaucracy: “the longer he was in office, the more he disliked having graded officials (shih ta-fu) at court.” This preference for administration through clerks rather than through graded officials pervades descriptions of Southern Sung government. From the viewpoint of the monarch and his sole minister, a clerical administration was more reliable, stable, and malleable than its official alternative. But, from the viewpoint of the regular bureaucracy, this predilection for “clerks and petty men” frustrated their careers and so instigated contemporary invective against Ch’in Kuei and other sole councilors. As we have seen above, career success under the Sung personnel system depended on regular advance through established channels of promotion. By not filling mid-level billets in the Six Ministries, Ch’in Kuei effectively blocked the advance of potential critics and possible adversaries into higher office.³⁶⁵

This Southern Sung reliance on clerical administration had origins in Northern Sung attempts to reduce clerical corruption and inefficiency by removing the boundaries between lower-level officials and the higher orders of clerks. Officials as intellectually divergent as Wang An-shih and Su Shih offered suggestions on how to achieve this goal, but it was Wang who was credited – or blamed – with the first concrete steps in this direction. He maintained that in antiquity there had been no distinction between the two groups and therefore “officials and clerks should be as one.”³⁶⁶ In a move to effect this union, Wang created in 1070 the position of examiner (chien-cheng) to monitor the document flow in and out of the Secretariat–Chancellery subunits and to act as liaison between the chief councilor and the clerical staff.³⁶⁷ Abolished in 1130, the examiners were restored by Ch’in Kuei during his first tenure as chief councilor in 1132. Chang Tun, the first of the sole councilors in the 1090s, served as examiner in the early 1070s, as did Tseng Pu.³⁶⁸ Although official biographies rarely speak to such issues, one may surmise that these years of apprenticeship under Wang An-shih afforded Chang and other leaders of the next generation a detailed, working knowledge of the clerical world and how it could be mobilized.

The dividing line between the clerical and the official worlds was in fact more permeable than many secondary sources suggest. In the early Sung, senior Secretariat clerks (t’ang-hou kuan, literally, “officials behind the hall”) had been graded positions.³⁶⁹ Sung personnel statutes contained provisions for the transfer of such clerks to graded status, although promotion caps known as the “laws of halt” limited their advance into the highest ranks of officialdom.³⁷⁰ Yeh Shih observed that, although the Yüan-feng reforms had produced a surge in the number of central-government clerks, the real transformation resulted from the liberal policies adopted toward clerical transfers into graded status for senior clerks during the Hui-tsung years.³⁷¹ Ts’ai Ching not only expedited such transfers but ignored the “laws of halt” of the Yüan-feng period that prohibited these officials from advancing into graded positions high enough to confer yin privilege.³⁷² In essence, he allowed clerks to become officials and then to confer official status upon their sons. This policy contributed to the oversupply of officials in the late Northern Sung. More importantly, it produced many graded officials in early Southern Sung who had not entered officialdom through the examination system but were “clerks and petty men.” This fact is fundamental to understanding the difference between Northern and Southern Sung government and the propensity of the latter for sole councilors.

The bureaucratic chaos that resulted from the loss of court archives and relocation of the capital to Hang-chou during the late 1120s and early 1130s enhanced the power of the clerks. Their collective memory and their private copies of statute books and personnel records were major sources for rebuilding the documentary base of the restored dynasty.³⁷³ In 1127, when the new Emperor Kao-tsung established his temporary court at Yang-chou (in Huai-nan-tung circuit), only 258 clerks reported for duty. He promoted them, and, within two years, their numbers swelled quickly. The reconfiguration of the central government in 1129 established clerical quotas as follows: Secretariat–Chancellery 238, central State Affairs 204, the Six Ministries 920, and the Military Affairs Commission 327.³⁷⁴ As soon as he began his second term as chief councilor in 1138, Ch’in Kuei authorized the addition of “above-quota” clerks in central-government agencies and in the provinces.³⁷⁵ When he died in 1155, there were 4,000 clerks alone in the seven prefectures of Che-tung circuit (Liang-che-tung circuit). Steps were taken to reduce this number by half, but “private clerks,” off-quota personnel in the private employ of local officials, multiplied tenfold and “were a horrible bane to the people.”³⁷⁶

As soon as he ascended the throne, Emperor Hsiao-tsung ordered a 20 percent reduction in clerical staffing at central-government agencies and forbade the practice of adding “above-quota” clerks.³⁷⁷ By 1168, the numbers of clerks in the central offices had fallen 28 percent below the quotas of 1131.³⁷⁸ A contemporary memorial blamed Ch’in Kuei for the excess: “He violated the statutes because he disliked having graded officials close to him, and he preferred clerks, whom he rewarded beyond all normal standards.”³⁷⁹ This reduction in clerical staffing reflects Hsiao-tsung’s desire to restore elements of the traditional Northern Sung model of government. His rejuvenation of the outer court audience protocols in the 1160s, mentioned above, also reflects this trend. The vitality and status of these rituals in the Southern Sung co-ordinate closely with the rise and fall of sole councilors. Largely moribund under Kao-tsung and Ch’in Kuei, they returned in the 1160s and 1170s, only to fade once again under Han T’o-chou and Shih Mi-yüan.³⁸⁰ Ritual gatherings of diffident officials benefited neither the emperor nor the sole councilor. As Hung Mai implies, there were, anyway, few officials to attend. And for an emperor and a councilor who were in accord, the open-court memorial process was little more than a distraction.

Southern Sung texts, as early as 1132, often repeat the phrase “the clerks are strong, the officials are weak.” The formulation implies an official helplessness, as clerks dominated routine government and used their superior knowledge of statutes and bureaucratic procedure to connive against their reputed masters.³⁸¹ Yeh Shih claimed that the graded officials themselves were to blame. They never deigned to master the technicalities in the very statutes they themselves had allowed to proliferate.³⁸² Handbooks for Southern Sung officials warn that clerks do not respect “public opinion.” The official must be vigilant and never delegate authority to them, for the people make no distinctions between who is a clerk and who is an official.³⁸³ In 1141, as the Jurchen delegation approached the capital to sign the peace accords, many officials refused to attend the formal reception ceremony. To solve this dilemma, Ch’in Kuei dressed his clerks in official robes, and they stood at audience in place of the graded officials. The image was prophetic.

Along with the growth of clerical power came the increasing consolidation of separate statutory authorities into the hands of the chief councilors. The Sung founders had insisted upon strict division between the civil authority of the Secretariat–Chancellery and the military authority of the Military Affairs Commission. A separate Finance Commission and eunuch control of imperial finance also excluded chief councilors from access to detailed information on revenue and budgets. Gradually, and in the face of continual opposition from the monarchy, the councilors eroded both dividers. The late Northern Sung is again the key period. The Yüan-feng reforms disbanded the Finance Commission and placed its subunits under the jurisdiction of the chief councilors in the Department of State Affairs.³⁸⁴ In 1122, Wang Fu’s Frontier Defense Office eroded the boundaries between the chief councilors and the Military Affairs Commissioner. By the reign of Hsiao-tsung, chief councilors routinely carried concurrent appointment as Military Affairs Commissioner.³⁸⁵ This combined authority, plus another concurrent appointment as Commissioner of State Revenue (Kuo-yung shih), enabled Han T’o-chou to collect taxes directly from the provinces and pursue his disastrous war against the Jurchen in 1206.³⁸⁶

These consolidations of power paved the way for the longest-ruling sole councilor in Sung history, Shih Mi-yüan, who held the office from 1209 through his death in 1233.³⁸⁷ The authors of the Sung history declined to label Shih a “nefarious minister,” largely because the Tao-hsüeh teachings had spread widely during his administration and were declared state orthodoxy in 1241 under the sole administration of his nephew, Shih Sung-chih (d. 1256). Yet his contemporary critics were not so kind. Wei Liao-weng (1178–1237), a leading Tao-hsüeh scholar of the period, submitted a lengthy memorial on how Emperor Li-tsung should seize the opportunity presented by Shih’s death to reform his government. Wei’s text contained ten points. First, eliminate the Secretariat examiners (Chung-shu chien-cheng) and restore the integrity of the three departments so that “the Secretariat obtains the imperial will, the Chancellery resubmits the memorial, the Department of State Affairs promulgates the action.” Second, end concurrent appointments between the chief councilor and the Military Affairs Commissioner, thus restoring the latter’s independent authority. Third, discontinue the practice of allowing the chief councilors to work at home and re-establish the Hall of Administration as a center for the conduct of official business. Fourth, restore permission for ministers-in-attendance to present memorials. Fifth, revitalize the Classics Mat lectures. Sixth, separate the Censorate and the Bureau of Policy Criticism and remove both from the influence of the chief councilor. Seventh, restore the integrity of the Institute of Academicians and the ability of Secretariat drafters to “return for correction” court documents and appointments. Eighth, restore the ability of “speakers” to submit memorials at audience in open court. The last two concern military issues and are not relevant to this discussion.³⁸⁸

A recent commentary on this memorial rightly concludes, “Wei’s proposals were no hot air: they went to the heart of the problems of imperial and bureaucratic power in Sung.”³⁸⁹ Wei forcefully advocated that the Northern Sung model of government should be restored. His proposals spoke to every major issue in the eleventh-century search for a government of balanced functional units. Somewhat ahistorically, he saw the Yüan-feng reforms as a bulwark against authoritarian control. He inherited this view from twelfth-century political thinkers who looked to strict separation of the three departments, strong exercise of “return-for-correction” authority, and wide avenues for “public opinion” as safeguards against abuse of documentary protocol by monarch or chief councilor.³⁹⁰ As I have argued above, the perceived integration of these features into a smoothly working government was largely a historical chimera. However, Wei’s proposals – if his descriptions of contemporary court practice are accurate – reveal how extensively the sole councilor Shih Mi-yüan had subverted all the major institutions of Sung government. In the end, it was the chief councilors and their coalitions of “clerks and petty men” that won the struggle for power in Sung. Reforms that had begun in the Northern Sung when gentlemen-scholars set out to rid the clerks of petty corruption ended when the clerks and petty men completed the corruption of the gentlemen-scholars.