An unusual debate on what may seem an arcane topic—China’s imperial civil service examinations—recently took place on the op-ed page of the The New York Times. The argument centered on the question of whether or not China during the past 1000 years or so was a meritocracy. Prominent mainland intellectual Zhang Weiwei, first argued that “[m]eritocratic governance is deeply-rooted in China’s Confucian political tradition.” A few days later, Mark Elliott, a Harvard historian, penned a sharp rebuttal, attacking Zhang for mischaracterizing the historical record. This was not, needless to say, just a disagreement about history. Equally at stake were the claims China’s current leaders make about their own legitimacy when they and their academic supporters—in this case Zhang—argue that they are heirs to a meritocratic system that stretches back into antiquity. This is a highly popular argument among mainland Chinese scholars with a “leftist” or nationalist political bent, who seem eager these days to draw similarities and connections between the current political regime and imperial Chinese traditions.
No historian can disagree with Professor Elliott’s argument that history should not be misused to support contemporary political arguments. He is much less persuasive, however, when he categorically denies that selection into the late imperial Chinese bureaucracy was basically meritocratic. A better way of challenging Zhang Weiwei’s highly problematic use of imperial history is to point out that, even if the imperial civil service examinations were reasonably meritocratic, they bear almost no resemblance to the far more opaque recruitment and promotion mechanisms of the contemporary party-state. China’s historical traditions can lend little support or legitimacy to a political system that shares so little in common with them.
From the Song Dynasty (960-1279) onwards, most high-level government posts in China were held only by those who had excelled in the keju (科举) examination system. This was a sequence of examinations that—with some variance over time—tested aptitude in the Confucian canon, poetry, fiscal and economic policy-making, and military strategy, or some combination of the above. The vast majority of males were allowed to take the lowest-level examination and would gain access to higher levels through superior performance, as marked by the conferral of a degree. The highest-level examination, generally held every three years in the capital, conferred the jinshi (进士) degree, along with a government post of county magistrate or above.
Professor Elliot makes two central claims about this elaborate system: First, despite being theoretically open to all men, it failed to create substantial social mobility. Quite the opposite: because it demanded high fluency in classical written Chinese, it excluded anyone who could not afford the often expensive education required to obtain such fluency—essentially, most commoners. Second, the grading of the examinations was riddled with corruption, and often subjective or outright unfair. Elliott concludes, therefore, that any claim that the imperial Chinese state was meritocratic is “outmoded.”
The basic historical facts Elliot cites are, of course, accurate. Anyone who has read seriously on this subject—particularly the seminal works of historians Ho Ping-ti and Benjamin Elman—would agree that lack of educational access excluded some 90 percent of the population during China’s Ming and Qing dynasties (1368-1644 and 1644-1911, respectively) from any realistic chance of examination success. Probably more than 85 percent of Qing jinshi had at least one ancestor or relative in their extended family who was either a degree holder or very wealthy. Few came from “true” commoner families with no political or economic connections whatsoever. Moreover, cheating was widespread. In the 260-year history of Qing keju examinations, there were at least nine cases—five in the particularly unfortunate year of 1657—in which exam administrators were sentenced to death for their failure to prevent serious cheating.
But does this mean that there was no “meritocracy” in late imperial China? Here Elliott overstates his case. Most importantly, when he points to the lack of social mobility through the examination system as evidence that it was not meritocratic, he confuses meritocracy with social fairness: Even in an ideal examination-based meritocracy, it is completely natural—if undesirable—for wealthy or politically connected families to have considerable innate advantages in money, educational resources, and so on. This is true of the current Chinese college admissions examinations and the SATs, and it was clearly true of the imperial examinations as well.
Whether an examination system is meritocratic depends on whether the examination itself is institutionally open and objectively graded, not on whether state or society provided equal educational resources to the poor or, for that matter, whether x percent of degree-holders were commoners. In fact, this is precisely how Elliott himself defines “meritocratic:” “judged superior according to an objective standard of ability.”
Of course, the charge that the examination system was heavily corrupt is relevant to whether it was meritocratic. But here the question becomes one of degree, and Elliott, like everyone else who has addressed the issue, can point to no systematic evidence that demonstrates precisely how corrupt the system was. Probably no examination system in human history has escaped the taint of cheating, but whether that cheating is so severe that the system fails to be basically meritocratic is another story.
Even taking into consideration the fairly dramatic corruption behind the scenes, I would argue that, for the most part, the men who wrote the best essays in the imperial examinations won the degrees. There is no evidence—including in the Lawrence Zhang dissertation that Elliott cites—that cheating was institutionally or even regularly tolerated. Quite the opposite—it was a severely punished and socially scorned crime. For every case of cheating that we know of, there were numerous men of extraordinary literary talent who made it to the top. A couple dozen cases of corruption per degree class does not mean that the other two hundred sixty men were also selected unfairly.
Indeed, if corruption really undermined fairness so severely that meritocratic selection was largely nonexistent, contemporary commentators would probably have given the issue far more attention than they actually did. Failed examinees, by far the most vocal and thorough critics of the system, complained most often of the inhumane physical and psychological stress that the examinations inflicted, and of the intellectual rigidity of the bagu (八股) essay format instituted in the Ming and Qing. Corruption and cheating attracted less attention, despite the great moral problems they presented. Moreover, the fact remains that perhaps 10 percent or so of jinshi degree holders had no significant political or economic connections. That is hardly insignificant, especially if we consider how difficult it was for commoners to receive a good education under 18th- and 19th-century socioeconomic conditions in the first place.
Two additional thoughts:
From a comparative perspective, even if the imperial examinations only provided basic meritocracy to 10 percent of the population—the percentage that Elman estimates had adequate educational resources to be competitive—that would still be significantly more meritocracy than what most contemporary Western European, Japanese, Middle Eastern, and Indian governments offered. In much of Western Europe, political rank and title were still largely a matter of direct aristocratic inheritance in the 18th century. Besides, 10 percent of the Qing population in 1800 was some 30 million people, largely equal to the entire population of France.
There is also the problem that the Qing system, the main focus of Elliott’s essay, was arguably far less functional than the Ming and (especially) the Song systems, if only because the population at least doubled, if not tripled, during the dynasty, while the number of degrees issued remained unchanged. In fact, they actually declined quite seriously for Han Chinese, as a significant quota was set aside for Manchu—the ruling ethnicity of the Qing, who conquered “Han” China following the Ming’s collapse—examinees. These conditions increased the difficulty of obtaining a degree and made cheating much more attractive. One plausible theory is that Manchu rulers created these rules so that Han literati would not overwhelm Manchu ranks within the political elite. I find it somewhat surprising that Elliott, who perhaps more than anyone understands the unique “Manchu-ness” of the Qing state, does not mention these circumstances.
Ultimately, although Elliott identifies a number of important problems with the keju system, his argument that it was not basically “meritocratic” is unpersuasive. This does not mean, however, that Zhang Weiwei’s use of imperial history in his original opinion piece is any less problematic.
Zhang’s argument seems to be that the tradition of meritocratic political recruitment in imperial China lends an extra layer of historical legitimacy and cultural authenticity to what he sees as the similarly “meritocratic” system of governance under the current Chinese Communist Party. The problem, however, is that the recruitment mechanisms of the two systems are so fundamentally different that whatever meritocratic legitimacy the imperial state may have had lends almost nothing to the claim that the current party-state is a “meritocracy.”
Most importantly, almost none of China’s current high-level officials—or those that ran the country in previous decades—entered public service on the basis of open, uniform, and written examinations. Although China now selects entry-level bureaucrats through its National Civil Service Examinations, the system was not created until the 1990s, by which time nearly all top-level Party and state officials in service today were already advancing through the political ranks. These people certainly could have been among the “best and brightest,” but the party-state’s opaque recruitment and promotion procedures do not give their selection anything resembling the meritocratic public legitimacy that imperial examinations—to the extent that they worked as designed—conferred upon jinshi.
Moreover, even the National Civil Service Examinations are hardly the equivalent of keju examinations. The current system is designed to recruit entry-level bureaucrats, largely equivalent to county-level clerks and runners in the Qing system. Once in the system, promotion is based on the usual combination of personal connections, seniority, and job performance, and has nothing to do with examination performance. The imperial system, on the other hand, directly recruited jinshi into the upper echelons of the government. How high an official could rise was also tied to his initial examination performance, as many of the most prestigious positions were limited to jinshi of a certain rank or higher.
None of this is to say that the imperial system was necessarily superior. Performance on an academic examination is probably a questionable indicator of actual administrative talent. One could even argue that the imperial state’s over-reliance on examination performance was politically and economically detrimental in the long run. Nonetheless, the point here is that the imperial system and the party-state’s current recruitment procedures have very, very little in common. The former did at least recruit, through somewhat open competition, a sizable share of the country’s top academic talent, whereas it is unclear, at least from the outside, what the latter has recruited. If Zhang Weiwei wishes to argue that the Communist party-state is a “meritocracy,” he should realize that comparison to the keju system does his argument no particular favors. Quite the opposite: it only highlights how comparatively opaque the current regime is.