Soulstealers: The Chinese Sorcery Scare of 1768

by Philip A. Kuhn

  The apparent rigor of this system seems less impressive when we examine the actual documents used in it. To begin with, the format was extremely stereotyped. The registers, sometimes known as "fourcolumn books" (ssu-chu-ts'e), contained, for each man, a single page with four headings: "integrity" (ts'ao-shou), "executive performance" (cheng-shih), "native talent" (ts'ai-chii), and "physical fitness" (nien-li), listed in that order. Under each heading, one of three standard ideographs would be filled in:

  Based on their ratings, officials would be grouped into three ranks. The criterion for ranking was the number of categories in which an official received better-than-average ratings. For instance, an official who received ratings of "incorrupt (ch'ing)," "assiduous (ch'in)," and "exceptional (yu)" in the first three categories was ranked in group one. ("Physical fitness" seems not to have played a part in the group rankings. If age or illness made the official unfit, he was impeached in a separate procedure.) Those with two above-average ratings were grouped in group two; and those with one or none comprised group three.23 All three groups, however, were considered fit for duty. Those in group one might be recommended for promotion, which was done in separate memorials attesting to their "outstanding" (cho-i) qualities. Also in separate memorials, those whose general fitness was below standard were impeached (chic-ho). The provincial triennial evaluation (ta-chi) used substantially the same format but added, for each official, a four- or eight-ideograph evaluation (k'ao-yii) that offered an overall assessment of performance.

  How little latitude these fitness reports permitted the evaluating officer! The scale of qualities was hardly fine enough to make careful distinctions among officials. Hardly more revealing were the four- or eight-ideograph evaluations on each man's file in the "Grand Accounting." An examination of numerous eighteenth-century yellow registers suggests that evaluators were choosing their comments from standard phrasebooks. The specificity is still crude, the result bland. Here are a few examples from a 1751 list of magistrates from Chihli ranked in the middle grade (erh-teng). One is reminded of a third-grader's report card, prepared by a teacher who is strug gling for something special to say about each of her charges ("participates actively in class, written work neat").

  One would expect that in the subsequent recommendations for promotion there would be more to say. Indeed the ratings are more complimentary, but the format is just as confining and stilted:

  Although one finds minor differences of vocabulary among provinces (suggesting that each provincial yamen had its own handbook of such stock phrases), the impression left by these registers is of officials who were struggling to differentiate subordinates whose records seemed generally acceptable but of whom they had little or no personal knowledge.

  Such stilted, conceptually cramped procedures grew naturally from bureaucratic life and reflected the mentality of the men who applied them. First, there was the need to avoid risks. Recommendation of a man who later turned out to be disappointing (or worse) could incur penalties for the recommender. Perhaps the more closely the criteria of merit hewed to a narrow, unexpressive format, the more likely were officials to risk making recommendations, on the principle that the less said, the better. Furthermore, descriptions of acts rather than analyses of character were more easily defensible, should anything go wrong. Second, the evaluations probably were adequate to describe what bureaucrats themselves considered a "good" official. In a ruleridden environment, the best official was the one who caused the fewest problems-that is, who exemplified largely negative virtues by avoiding trouble. In any bureaucratic system, to excel can be risky. Nor are whistle-blowers and boat-rockers appreciated. The overzealous official trips over rules more often than does the cautious plodder. Hence prudence, circumspection, and diligence were prominent values in the routine evaluations.

  Systemic Blocks to Routine Control

  Hungli's Despair at the Routine System

  Hungli had nothing against routine evaluations as such. On the contrary, he recognized that they "affected the fundamental institutions of state" (ta-tien). The trouble was, he believed, that they were applied carelessly and dishonestly. Seven years into his reign, he complained that the Grand Accounting in the provinces was "an empty letter" and was administered "sloppily or perfunctorily." Good officials were not recommended, bad ones not impeached. Personnel evaluations hinged on affairs of the moment and said little about long-term conduct. Judgment of an incumbent official rested on whether the governor-general and governor happened to like him, and might not accord with the man's general reputation. Officials were recommended for promotion through personal favoritism, and scoundrels were shown leniency. "Only the faults of educational officials or petty functionaries are casually noted, so as to make up the necessary numbers." By such chicanery was the autocrat undone: "Our governors-general and governors are Our arms and legs, heart and backbone. If they treat such a crucial government function as mere routine, on what can We rely?"26 Evaluation in the Capital Investigation he considered just as perfunctory.27

  In the hands of the review commission, the triennial yellow registers got the same off-hand treatment. Hungli complained that unworthy candidates for promotion or reassignment were not screened out but were passed right through to imperial audience. "It should be possible at a glance to see [from the registers] whether a candidate is up to the job. But the Nine Chief Ministers (chiu-ch'ing) [that is, the review commission responsible for inspecting the registers] pass the buck to the Board of Civil Office, the Board of Civil Office passes it to the Nine Chief Ministers; and the result is that they all pass it to Us!"28 The monarch knew that, rather than face up to the task of making personnel judgments that might arouse resentment, the bureaucracy passed the job to the one man who could always take the heat.

  The routine inflation of evaluative phrases (k'ao-yii) was all too apparent to Hungli when a candidate for appointment received radically disparate ratings from two evaluators, or when a man he knew to be a bumbler came before him with glowing recommendations. The Yunnan governor recommended one of his circuit-intendants with a kao-yu of "experienced and of solid character."29 But Hungli reviewed the man's record and found that the preceding governor had characterized him as "aged and infirm," defects that the passage of time would hardly have remedied. The new governor had, he complained,

  said nothing about his age and infirmity. And the k'ao-yii of all his other subordinates show only superior traits and no deficiencies. Now human talent is variously distributed. Some men are of decent character but not very able; or they are efficient administrators but insincere in their inner natures ... but in [the new governor's] memorial, the evaluations are very general and there is absolutely no differentiation among them. It seems that in the whole province there is not a single official who can be criticized!

  The k'ao-yu of a certain Taiwan circuit-intendant read "youthful and vigorous, of fine character, able and intelligent, apt at administra- tion."30 Hungli: "But We know him very well. He has a bit of cleverness, but his character is definitely not sincere, and his administrative behavior merely consists of complying with the formal requirements of his job in order to dispose of his public duty. He never had solid abilities, and his physical vigor is not that great either ... This shows how provincial officials do not take seriously the job of evaluating personnel." Although he was aware that laziness and laxity were part of the trouble, Hungli also knew that reliable routine personnel control was hampered by certain systemic problems.

  Patronage versus Discipline

  The provincial governors were line administrators as well as evaluators. Hence they acted under compulsions that went with their office.;' One of these was a powerful desire for a certain kind of personal image, which can be roughly translated as "magnanimity" (k'uan-ta). The essence of a good patron, this quality warmed what might otherwise be a cold, objective bureaucratic relationship between a province chief and his subordinates. "Magnanimity" in a patron meant a concern for the human needs of his clients. Though
the patron's practical payoff was loyalty, his symbolic reward was a certain personal image, one that was sullied every time he had to throw the book at a subordinate and thus treat him as a misaligned cog in a machine rather than as a human dependent. The dignity of the superior was hurt along with the career of the subordinate.

  In what respect could local officials be considered dependents or clients of governors? Though the general answer is surely the paternalism that tinctured the entire Chinese bureaucracy, the specific answer lies in the governor's power to recommend his subordinates for appointment, transfer, or promotion. Except for the small number of local posts that were rated highest in administrative difficulty and were filled by recommendation of the Grand Council32 and except for a portion of posts set aside to be filled directly by the Board of Civil Office, governors had the privilege of recommending men for particular posts within their provinces. If a governor certified that there was no suitable candidate among bureaucrats of his own jurisdiction, he could recommend someone from outside. Even in categories of appointment that the administrative code specifically placed outside their reach, governors pushed relentlessly to expand their appointive power, a power basic to personal patronage networks. Given the speed at which provincial personnel rotated from post to post, such networks could rapidly grow to national scale. Information in the Collected Statutes indicates that at least 30 percent of all posts, from circuit-intendancies down to county magistracies, were in categories that could be filled through a governor's recommendation.33

  Quite apart from the danger of cliques and factions, however, Hungli had constantly to deal with governors' desire "to acquire a reputation for magnanimity," which made it hard for them to evaluate personnel honestly.34 Hungli scolded G'aojin (then serving as Anhwei governor) for submitting an implausible evaluation of a previously dismissed subordinate who was now due for reassignment: "In the case of a man who has already left office, there is nothing to prevent [his superior] from skirting the truth about him in order to elicit the admiration of his subordinates. This bad practice among governors-general and governors of purchasing reputation (ku-ming) is wholly inappropriate."35 Of course, Hungli considered "magnanimity" of this sort perfectly appropriate when exercised at the top of the system, namely by himself. Indeed, it may be a law of bureaucratic practice that every official tries to reserve this genial quality to himself (or those above him), while holding his subordinates to strict application of the rules.

  Once a governor had recommended an official for a post, it was awkward for him to admit a mistake. Hence evaluations tended toward a certain consistency: no governor was likely to change his opinion of a subordinate he had recommended until the man's per- fornmance, either good or bad, was so egregious that he had no choice. Hungli complained that, if it man of modest ability had been rec- onmmended for an easy post, his superior would seldom report any outstanding accomplishments; if 'a promising man had been recommended for it hard post, then he would seldom report faults.36

  Reading the Leader's Mind

  Another systemic evil that stymied personnel evaluation was called "seeking to accord [with the desires of one's superiors]" (ying-ho): that is, currying favor by molding one's judgments to conform to what one believed the boss wanted. This produced ludicrous distortions in the routine evaluation system, as officials trimmed their standards to the winds of imperial preference. The problem was endemic to the highest officials in both capital and provinces. "When We are lenient in one or two cases in which leniency is appropriate, then all the officials scurry to be lenient. When We are strict in one or two cases that require strictness, then they all scurry to be strict." On the surface this might seem to be a case of "the grass bending before the wind"a commendable deference to royal leadership. But in reality, Hungli warned, "it stems from self-interest and self-aggrandizement." Ignoring the "great principles," acting without proper estimation of right and wrong-were these really the right way to advance in honor and rank? When officials in the capital need correction, "We can personally admonish them." In the provinces, however, "the governors-general and governors have sole charge of their jurisdictions. If they are determined to conform themselves to whatever they think We want, without any firm views of their own, then their subordinates will flock to do the same in order to curry favor with their superiors," and national affairs will really be in parlous straits.37

  Obstacles to Impeachment

  Like any system of accountability, the personnel evaluation process included one self-defeating mechanism that nobody knew how to cope with: reporting misconduct was dangerous to the reporter, but so was failure to report it. The administrative code included a range of penalties for "failure to investigate" (shih-ch'a). Failure to report a subordinate for dereliction of duty made one liable, oneself, for administrative discipline. But if one did report him, a whole range of other embarrassments might surface (including tales he might bear about his colleagues, whom one had also "failed to investigate") that might have even worse results. Hungli knew that penalties for "failure to investigate" hindered his access to information. Reporting heterodox sects, for instance, was dangerous to an official's career. "Because it arose in his jurisdiction, and hence would adversely affect his fitness report, he might consider covering it up."3"

  Here are some penalties visited upon officials who "failed to investigate," from the administrative code of the Board of Civil Office:

  In the case of an official who receives bribes in the course of duty, and whose misconduct is not reported, [when the affair finally comes to light] a prefect stationed in the same city will be charged with failure to investigate and will be demoted one grade and retained in his post. A circuit-intendant will be fined a year's [nominal] salary.3`' A prefect not in the same city but stationed within a hundred li will be fined one year's [nominal] salary. A circuit-intendant will be fined nine months [nominal] salary ... When a governor-general or governor impeaches an official in a "failure to investigate" case, let him clearly state in his memorial the distance [from the scene], to serve as data for investigation. If there is a misstatement of the distance, any official through whose hands the memorial passes will be demoted two grades and transferred.'"'

  In sensitive cases, willful failure to impeach could bring, not a slap on the wrist, but real terror. When one of his trusted province chiefs concealed information in an impeachment case of 1766, Hungli complained that he had been personally betrayed.41 "Chuang Yu-kung has received Our highest favor and has been selected for choice appointments. Yet he has the nerve to strut about, flaunting his favored status . . . This case is one in which he has intentionally deceived Us!" Not only was the ingrate cashiered, but his case was switched from the track of administrative discipline to that of criminal sanctions. Chuang was arrested, haled to Peking for interrogation by the Grand Council, had all his property confiscated, and was jailed to await decapitation. Yet the point was not to destroy but to chasten. Hungli granted him amnesty some months later and appointed him acting governor of Fukien.42

  Hungli, who well knew that bureaucratic culture made impeach ment distasteful to his province chiefs, read memorials carefully and was not easily fooled. He discerned a system of mutual protection by which governors shielded the reputations of their immediate subordinates, the provincial treasurer and provincial judge, from charges of "failure to investigate." The governor's impeachment memorial would say, "just as I was writing this, the reports of the provincial treasurer and provincial judge reached my desk and were not at variance from my own inquiries." Scoffed Hungli: "This might really occur once in a thousand or a hundred cases; how could it turn up every time?" From then on, when a governor was impeaching a subordinate, he was to state whether his information came from his own inquiries or from a report, and exactly how it had been con- veyed.433 Because impeachment for irregularities in impeachment was subject to the same irregularities, could the monarch ever be confident that routine procedures would produce sound evaluations of personnel? He saw himse
lf confronting a system in which vertical clique-formation within the provincial bureaucracy made it hard for that bureaucracy to police itself. Higher officials and their subordinates "form cliques behind the scenes" and "collude by mutually ingratiating and coercing . . . These evil ways must be rigorously stamped out."-'-' They were not, however, likely to be stamped out through the routine procedures of bureaucratic control, and Hungli knew it.''

  Nonroutine Systems of Evaluation

  Confidential Reporting from the Field

  Given what he saw to be the futility of routine evaluation, the monarch naturally made the most of opportunities to inject autocratic power into the system. But to do so he needed reliable, undoctored information. From the beginning of his reign, he tried to get confidential personnel evaluations from the provinces. If governors would not evaluate their subordinates honestly through the open channel for fear of stirring up resentment, perhaps reporting confidentially in palace memorials would make them feel more secure. In the first year of his reign, Hungli had so instructed them.-"' "Now, in the beginning of Our reign, We are not well acquainted with the men serving as circuit-intendants and prefects in the provinces. You [governors] may report on the worthiness and the activities of your sub ordinates through [confidential] palace memorials." Yet even in the confidential channel the province chiefs felt insecure. Three years later, Hungli complained that his original decree was being ignored.47 All governors had "reported once," but not since. He pointed out that governors customarily kept their posts longer than their subordinates, so that the passage of men through various jurisdictions opened a splendid opportunity for fresh evaluations. Now all governors were to send confidential memorials (mi-tsou) "from time to time." Yet to enforce this demand required ceaseless struggle. In 1759 Hungli was shocked, but perhaps not surprised, to find the governor-general of Liangkiang sending in sheer boilerplate through the confidential channel: "Recently We happened to be examining a memorial from Yenjigan evaluating his subordinates. In it, Wei Che- chih still appears as Huai-an prefect, and Dingcang still appears as Hsu-chou prefect. These men held these posts more than ten years ago! How come there were no follow-up evaluations of them? This is a matter for confidential memorials, which can do no harm to one's reputation for magnanimity.14'