1 / Separate and Unequal

Anti-Manchuism, narrowly defined as ethnic opposition among China’s Han majority toward the “alien” Manchus, figured prominently in the critique of the Manchus’ Qing regime in the years immediately following the antiforeign Boxer uprising of 1899–1900.¹ As Zou Rong (1885–1905), for example, complained in his pamphlet The Revolutionary Army (Gemingjun) of 1903, “Unjust! Unjust! What is most unjust and bitter in China today is to have to put up with this inferior race of nomads with wolfish ambitions, these thievish Manchus [Manzhouren], as our rulers.”² The revolutionaries, to be sure, were never narrowly or exclusively anti-Manchu. While “expulsion of the Tartar caitiffs” (quzhu Dalu) was the first of its stated goals, the revolutionary Alliance (Tongmenghui) in 1905 had three other, no-less-important objectives: revival of China, establishment of a republic, and equalization of land rights. Furthermore, the revolutionaries advanced numerous reasons other than ethnic opposition to justify ousting the ruling Manchus: they were incompetent; corrupt; oppressive and arbitrary; and, above all, incapable of defending China’s national interests against the rapacious foreign powers. Nevertheless, apart from and in addition to such criticisms, the revolutionaries time and time again asserted simply that the Manchus had to go because they were not “Chinese.”

But who, exactly, were the Manchus? They are often described as the descendants of the scattered Jurchen tribes in the northeastern frontier of China, bordering Korea, whom Nurhaci (1559–1626) began in the 1590s to unify and whom his son and successor Hong Taiji (1592–1643) renamed in 1635 “Manju” (“Manzhou” in Chinese). They were, in fact, a much broader, more heterogeneous group; they encompassed the total membership of the Eight Banners (Baqi), of whom the Manzhou were only one (though, admittedly, the core) group. Members of the system were known, collectively as well as individually, as “banner people” (qiren). Thus, the issue of Manchu identity is inextricably bound up with the Eight Banner system.

How valid were the revolutionaries’ criticisms of the Manchus as an alien people? To what extent and in what ways were the Manchus in the late Qing dynasty (1644–1912) separate and distinct from the Han, particularly in view of the 1865 edict cited by Mary Wright in The Last Stand of Chinese Conservatism that purportedly had abolished most remaining differences between the two groups?

THE CASE AGAINST THE MANCHUS

Anti-Manchuism was a loud chorus among the young, revolutionary-minded intellectuals who gathered to study in Japan in the early post-Boxer years. Their immediate concern was the menace of imperialism. Following the multinational invasion of north China to suppress the Boxers and the Russian occupation of Manchuria in 1900, China seemed to be on the verge of being partitioned among the foreign powers. The perilous condition of their homeland could easily be blamed upon the Qing government, the Manchu court, and the Manchu people as a whole. An article in Enlightenment Journal (Kaizhilu) in 1900–1901 spelled out the link between anti-imperialism and anti-Manchuism: “People of our country speak daily of the shame of becoming the slaves of the foreigners, but they overlook the shame of being the slaves of the Manchus. They speak daily of expelling the foreign race but ignore the expelling of the foreign Manchu race [Manzhou zhi waizhong].”³

The most outspoken of the anti-Manchu critics in the early post-Boxer years were Zhang Binglin (1868–1936), Zou Rong, and Chen Tianhua (1875–1905). Zhang Binglin, the oldest and most persistent of the three, had been loosely associated with the radical reformers of 1898 and, like them, had had to flee China after the Hundred Days of Reform. He did not become a revolutionary until the summer of 1900, when, distressed by the empress dowager’s mishandling of the Boxer crisis, he published his first attack on the Manchus. In one of two statements written for China Daily (Zhongguo ribao), the newly founded organ of the Revive China Society (Xing-Zhong Hui) in Hong Kong, he announced that he had cut off his queue—thus severing his ties to the Qing dynasty—and explained why he had done so. Later on, as part of his assault on the monarchical reformers, he published other criticisms of the Manchus, such as a 1901 article on Liang Qichao (1873–1929) in the Tokyo Citizens’ News (Guominbao) and a 1903 Shanghai pamphlet on Kang Youwei (1858–1927). For his anti-Manchu writings, Zhang was arrested in 1903 in Shanghai as part of the Jiangsu News (Subao) censorship case. Following his release from prison in 1906, he joined forces with Sun Yat-sen (1866–1925) in Tokyo and became editor of The People’s Journal (Minbao), the organ of Sun’s year-old revolutionary Alliance.⁴ Zou Rong, the second of the three chief anti-Manchu critics, was the author of The Revolutionary Army, which was published in 1903 in Shanghai as a pamphlet, with a laudatory preface by Zhang Binglin. Zou had written the tirade a year or two earlier while a student in Japan. Because of his unrestrained criticism of the Manchus, he too was arrested as part of the Jiangsu News case. He died in 1905, at twenty years of age, in prison. In part because of his youthful martyrdom, in the post-Boxer decade Zou’s The Revolutionary Army became the most widely circulated of the anti-Manchu writings.⁵ Finally, Chen Tianhua, who was Zou’s elder by ten years, was the author of two pamphlets, A Sudden Look Back (Meng huitou) and An Alarm to Awaken the Age (Jingshi zhong), both published in 1903 in Tokyo, where Chen had been studying for about a year. He subsequently, in 1905, joined Sun Yat-sen’s Alliance in Tokyo and wrote extensively for the first issue of The People’s Journal. Shortly afterward Chen committed suicide in protest against Japanese press reports that denigrated the patriotism of the Chinese students in Japan. His unfinished novel Lion’s Roar (Shizi hou) was published posthumously.⁶

The revolutionaries complained a great deal about the Manchus qua Manchus but seldom in a systematic fashion. Their condemnation of the Manchus, even in a tract so unremittingly hostile as The Revolutionary Army, was usually scattered throughout a broader condemnation of the Qing regime, one that criticized the Qing, for example, for being corrupt, oppressive, or incapable of standing up to the imperialists. However, a sampling of the revolutionary literature of the early post-Boxer years shows that their critique of the Manchus may be summarized in a seven-point indictment. The specifics come largely from the writings of Zou Rong, Chen Tianhua, and Zhang Binglin, supplemented by other publications of the period. Except for Chen’s novel Lion’s Roar, all of these sources appeared prior to the formation of the Alliance in mid-1905.⁷

The Seven-Point Indictment against the Manchus

First, the Manchus were an alien, barbarian group who were different from the Chinese and did not belong in China. “Why do I find fault with the Manchu people [Manzhouren]?” Zhang Ji asked bluntly in the Jiangsu News in 1903. “Because China belongs to the Chinese people [Zhongguoren].”⁸ This stance was what distinguished the anti-Manchu revolutionaries from the anti-imperialist reformers, whose foremost concern was to exclude the imperialist powers (and not necessarily the Manchus) from China; so far as the reformers were concerned, the Manchu rulers were no less “Chinese” than their subjects. For revolutionaries such as Zhang Ji, the Manchus most definitely were not “Chinese” and they had no more right to be in China than the imperialists had.

The revolutionaries differentiated the Manchus from themselves terminologically in two ways. One was to refer to them by epithets traditionally applied to the various “barbarians” on the periphery surrounding China’s civilized core. Thus, Zou Rong asserted,

What our compatriots today call court, government, and emperor, we used to call Yi, Man, Rong, and Di [barbarians of the east, south, west, and north] as well as Xiongnu and Dada. Their tribes lived beyond Shanhaiguan and fundamentally are of a different race from our illustrious descendants of the Yellow Emperor. Their land is barren; their people, furry; their minds, bestial; their customs, savage.⁹

Other labels of a similar sort that were commonly used to designate and denigrate the Manchus were “Donghu” (Eastern Barbarians), which Zhang Binglin often used, and “Dalu,” as in the Alliance membership oath, which combined “Dada” (Tartar) with lu (caitiffs), an archaic term used in the Northern Song period to refer to the Khitan.¹⁰ As for the corresponding terms of self-reference, those revolutionaries who equated the Manchu “other” with barbarians generally identified themselves as “Chinese,” which they usually rendered, as Zhang Ji did, as “Zhongguoren” or, as Liang Qichao had done in his 1898 China Discussion commentary, as “Zhina” or “Zhinaren,” from the Chinese reading of “Shina,” the prevalent Japanese term in the Meiji era for China.¹¹ Interestingly, they hardly ever referred to themselves as “Hua” and/or “Xia,” the traditional terms for the people of the civilized central core when they were being contrasted with the barbarians of the periphery.

The other way the revolutionaries differentiated the Manchus from themselves was by drawing upon the imported doctrine of social Darwinism, which Liang Qichao had helped to introduce. According to their understanding of those teachings, as exemplified in Chen Tianhua’s Lion’s Roar, the world’s population was divided, on the basis of location and skin color, among five large racial groups (zhongzu), or peoples: yellow, white, black, brown, and red; and each of these five peoples was in turn divided into a number of races (zu) and further subdivided into yet smaller groupings such as ethnic groups (minzu) and tribes. Thus, for example, the white people, living in Europe and North America, were made up of three races: Aryans, Teutons, and Slays. Similarly, according to Zou Rong, the yellow people, in East and Southeast Asia, were composed of two races (renzhong): those of China (Zhongguo) and those of Siberia. The Chinese race included the Han (Hanzu)—with Koreans and Japanese as subgroups of the Han—as well as Tibetans and Vietnamese; the Siberian race included the Mongols, Turks, and Tungus, with the Manchus (Manzhouren) as a subgroup of the Tungus.¹² Chen Tianhua generally referred to the Manchus as “Manzhouren,” but for Zou Rong and many others, the usual term was “Manren,” which Liang Qichao had popularized. The most widely used corresponding term of self-reference in this instance was, again following Liang Qichao, “Hanren.” Chen Tianhua, in Lion’s Roar, thus rephrased Zhang Ji’s slogan: “China belongs to the Han people [Hanren].”¹³ The term “Hanzu,” which is how the Han are classified today in the People’s Republic, was seldom used; even more rare was the correlative term “Manzu” for the Manchus. The range of terms that were used to identify Manchus and Han and their relative frequency of use will become clear in the course of this study.

Second, the Manchus had committed a number of heinous crimes against the Chinese people, particularly in the course of their conquest in the mid-seventeenth century. Their barbarous actions marked the Manchus as the ancestral enemies of the Han, and though those deeds happened a long time ago, they demanded to be avenged. The worst of such crimes were the savage massacres of defenseless civilians at various cities in central and south China as the Manchus marched through in 1645. Citing the recently reprinted chronicles of the slaughter at Yangzhou and Jiading (both in Jiangsu), Zou Rong claimed that the Manchu troops had been “let loose, burning and plundering” and that “wherever the cavalry of the thievish Manchus [Manren] reached, there was murder and pillage.” Nor, according to Zhang Binglin in 1903, were these atrocities attributable to only a few individual Manchu commanders; instead, every Manchu person had been responsible and thus culpable. Therefore, “when the Han race [Hanzu] wants revenge against the Manchus [Manzhou], they want revenge against their entire group.”¹⁴

Third, the Manchus had barbarized China by imposing their savage customs upon their Han subjects. Unlike previous foreign conquerors of China who had assimilated the ways of the Chinese, the Qing had forced the Han to adopt their alien Manchu customs, notably their male hairstyle and their official dress. As Zou Rong noted indignantly,

When a man with a braid and wearing barbarian clothes loiters about in London, why do all the passersby cry out [in English], “Pig-tail” or “Savage”? And if he loiters about in Tokyo, why do all the passersby say, “Chanchanbotsu” [lit., “a slave with a tail”]? Alas, the dignified appearance of the Han official has vanished utterly; the dress instituted by the Tang has gone without a trace! When I touch the clothes I wear, the hair on my head, my heart aches! . . . Ah, these queues, these barbarian clothes, these banner gowns [qipao], these peacock feathers, these red hat buttons, these necklaces. Are they the costume of China’s cultural tradition, or are they the loathsome dress of the nomadic and thievish Manchus [Manren]?¹⁵

Fourth, the Manchus had set themselves up as a privileged minority separate from and superior to the Han. According to Zou Rong, “Although it has been over two hundred years, the Manchus stick with the Manchus and the Han stick with the Han; they have not mingled. Clearly there is a feeling that a lower race does not rank with a noble one.” That is, the Manchus did not consider the Han their equal. As an example of the continuing failure of Manchus and Han to intermix, Zou referred to the provincial garrisons, where Manchus stationed in various major cities lived in their own quarters and were residentially segregated from the Han. Chen Tianhua, in Lion’s Roar, cited a ban on intermarriage as another device by which Manchus kept apart from Han. The revolutionaries, furthermore, claimed that the Manchus, from their own separate world, lorded over and indeed lived off the Han. Echoing Liang Qichao’s earlier criticism, Chen Tianhua, in A Sudden Look Back, charged, “They require that the inhabitants of the eighteen provinces collectively provide for their five million people. But up to now they themselves have not farmed or labored. All they do is sit and feed off the Han people [Hanren]. Is this not absolutely hateful?”¹⁶

Fifth, the Manchus subjugated the Han in the manner of a foreign military occupation. They maintained their domination over the Han by keeping their banner soldiers separate and concentrating them in a few strategic places around the country. Chen Tianhua, in Lion’s Roar, commented on the careful thought the early Qing rulers had given to the placement of their troops:

Aware that the Jurchen, by being dispersed, had opened themselves to be killed by the Han [Hanren], they took the several million Manchus [Manzhouren] that they had brought to China and stationed one-half of them in Beijing, where they are called the “palace guard” and the other half in the provinces, where they are called “provincial garrisons” [zhufang].

Zou Rong drew attention to the term used for these provincial encampments:

Suppose we try to explain the meaning of the term zhufang. It is as if they are terrified and are constantly fearful lest the Han people [Hanren] rebel against them, and so they hold them in check like bandits. Otherwise, whom are they defending [fang] against? And why do they need to be stationed [zhu] somewhere?

The obvious intent of these provincial garrisons, as a Jiangsu News article in 1903 summed up, was “to suppress the slaves.”¹⁷

Sixth, the Manchus practiced political discrimination against their Han subjects in at least three ways. They were a numerical minority ruling over the Han majority. According to Zou Rong, “The world recognizes only the principle that a minority submits to a majority. . . . If only the thievish Manchus [Manren] were a majority, but they number merely five million, scarcely the population of a single department or county.” The Manchus discriminated against the Han also by their monopoly of the highest governmental posts, in contravention of the Qing court’s own professed commitment to Manchu-Han equality. According to Chen Tianhua, in Lion’s Roar, “Official posts are supposed to be evenly divided between Manchus and Han, but all the important responsibilities are held by Manchus [Manren].” Following a detailed analysis of the metropolitan administration, Zou Rong concluded, “Opportunities for an official career for a Manchu in comparison with those for a Han are hardly less far apart than the sky is from the ground or clouds are from mire.” The Manchus discriminated against the Han in promotions as well. According to Zou again,

One may often find Manchus and Han of similar grades, graduates of the same year and employed in the same office. The Han [Hanren], however, may be held back for decades, without being transferred to a higher post, whereas the Manchu [Manren] in a twinkling becomes first a board vice-president, and then a board president, and finally a grand secretary. . . . If, by good fortune and against all odds, some [Han] officials do finally rise to the position of grand secretary or board president or vice-president, they are all white-haired and toothless, old and weak, and they share whatever is left over from the hands of the Manchus.¹⁸

Seventh, and last, the Manchus, despite their pretense at accommodation, were fundamentally at odds with and hostile toward the Han people. As evidence, the revolutionaries repeatedly cited a remark that Liang Qichao, in an essay of 1900, attributed to Gangyi (1837–1900), the reactionary Manchu leader at Cixi’s court after 1898: “If the Han get strong, the Manchus are doomed; if the Han grow weak, the Manchus get plump.”¹⁹ Zou Rong claimed that Gangyi’s statement embodied the single underlying rationale behind the court’s successive policies toward the Taiping Rebellion (1851–64), the Sino-Japanese War (1894–95), and the Boxer Rebellion, for in every instance the Manchus derived all the rewards while the Han made all the sacrifices. As for the court’s recent efforts to achieve “wealth and power” (fuqiang): “Today’s strengthening [qiang] is the strengthening of the Manchus [Manren]; it has nothing to do with us Han [Hanren]. Today’s enrichment [fu] is the enrichment of the Manchus; it has nothing to do with us Han.”²⁰ It was, according to the revolutionaries, the Manchus’ anti-Han posture that explained the court’s inability, indeed unwillingness, to defend China’s national sovereignty and territory from the foreign powers. Their policy was epitomized by another of Gangyi’s alleged sayings: “Rather than hand over our land to household slaves [i.e., the Han], let us present it to neighboring friends.”²¹ Having thus sold out to their foreign “friends,” the Qing had become, in the words of Chen Tianhua, “the foreigners’ court,” and the Han had become, in Zou’s words, “slaves of barbarous slaves.”²²

MANCHUS AND BANNER PEOPLE

Among the numerous terms that the revolutionaries of the early post-Boxer era applied to the Manchus, one—“banner people” (qiren)—was strikingly absent. This was odd because it is clear that four of the seven counts in their indictment of the Manchus clearly equated them with the membership of the Eight Banner system. It was soldiers of the banner army who committed the atrocities against the Han during the Qing conquest of China proper (no. 2) and who then kept the conquered Han under continuous subjugation from their various garrisons (no. 5), and it was the broad masses of the banner population who lived a privileged existence segregated from the Han (no. 4) and who benefited from the Qing dynasty’s policy of political discrimination (no. 6). Thus, in the eyes of the revolutionaries, the Manchus were identical to the banner people, and so it was in the eyes of the Manchus, too. As a descendant of the Qing imperial clan recalls, “The Manchus [Manren] called themselves, and were called by others, ‘banner people’ [qiren].” Or, as stated by James Lee and Robert Eng, “The banner system was the principal institution which unified the Manchu people and defined Manchu identity.”²³

The Eight Banner system as it existed in the late nineteenth century was fundamentally little changed from when it was first set up by the founders of the dynasty, Nurhaci and Hong Taiji, prior to the invasion of Ming dynasty (1368–1644) China. Nurhaci created the system in 1615 when he divided all the scattered Jurchen tribes in the mountains of eastern and northern “Manchuria” that he had dominated into eight groups called “banners” (qi), each with a number of companies (zuoling). Twenty years later, when Hong Taiji bestowed the new name “Manzhou” (from the Manchu “Manju”) upon his people, they became the “Eight Banner Manchus” (Baqi Manzhou). Meanwhile, as Nurhaci and Hong Taiji extended their authority into the plains of southern Manchuria and began overrunning Han settlements and enslaving their inhabitants, they apportioned these captives among the leaders of the original Manchu Eight Banners. Such household slaves of Han origin were known as “bondservants” (baoyi, from the Manchu booi, “of the household”), some of whom were organized into separate banner companies of their own that became a part of their master’s Manchu banner. Still later, in 1642, as more and more Han in southern Manchuria either defected to or were defeated by him, Hong Taiji stopped enslaving them and organized them instead into their own set of banners; these Eight Banners composed of Han personnel were known as the Hanjun. Meanwhile, in 1635, Hong Taiji similarly had established another separate Eight Banner organization for those Mongols of western Manchuria and eastern Mongolia who had submitted to his rule. Although other peoples were incorporated into the banner system later on, its basic framework was thus already in place by the time Hong Taiji’s successors “entered through the pass” into China proper in 1644. It had three ethnic components—Manchu, Mongol, and Hanjun—each with its own set of eight banners, for a total of twenty-four banners in the entire banner force.²⁴

The Eight Banner system, though simple in structure, was extremely heterogeneous in composition. In addition to the tripartite division among the Manchu banners, Mongol banners, and Hanjun, there were numerous other significant internal differences: between the Upper Three and the Lower Five Banners, between Old and New banner people, between the imperial lineage and ordinary banner people, between the regular banner companies and the bondservant companies, between the core banners and the affiliated banners, between the Metropolitan Banners (Jingqi) and the provincial garrisons, and between the banner soldiers and their dependents.

In all three ethnic components, each of the Eight Banners was identified by the color (yellow, white, red, or blue) of its flag, which was either “plain” (solid) or “bordered” with a red or white fringe. The Eight Banners were ranked in descending order of social importance: Bordered Yellow, Plain Yellow, Plain White, Plain Red, Bordered White, Bordered Red, Plain Blue, and Bordered Blue. The first three—Bordered Yellow, Plain Yellow, and Plain White—constituted the Upper Three Banners, which from an early date had been under the direct command of the emperor; the Lower Five Banners were originally commanded by various imperial princes and did not come under direct imperial command until the Yongzheng reign (1722–35). The Upper Three Banners were more prestigious than the Lower Five.²⁵

Of the three ethnic components, the Manchu banners were in the late Qing the largest by far. According to The Draft History of the Qing (Qingshigao), compiled after the revolution, 53.5 percent of the officers and soldiers in the Metropolitan Banners belonged to the Manchu banners, 31.8 percent to the Hanjun, and 14.7 percent to the Mongol banners.²⁶ However, when the banner troops in the provinces are taken into account, the Hanjun’s share was considerably less, because the partial demobilization of the Hanjun that took place during the Qianlong reign (1735–96) had a significantly greater effect upon the provincial garrisons than among the Metropolitan Banners. Therefore, it may be roughly estimated that 60 percent of the entire banner force was made up of Manchu banners and the other 40 percent was divided roughly equally between the Mongol banners and the Hanjun.

The Manchu banners were not only the oldest and the most numerous but, not surprisingly, also the most prestigious. They outranked the Mongol banners, who in turn outranked the Hanjun.²⁷ However, just as the entire banner force was not a homogeneous body, neither were the Manchu banners. For example, apart from the Jurchen who formed the founding core, the Eight Banner Manchus also included thirty-seven companies of Mongols, six of Koreans (Chaoxian), one of Russians (Eluosi), and one of Tibetans (fanzi). These “foreign” units in the Manchu banners were in addition to the many individual Mongols and Han Chinese who were scattered among the Manchu banner companies.²⁸ The major difference within the Manchu banners was that between the Old and the New Manchus, depending on the date of their adherence to the Qing cause. The Old Manchus (“Fo Manzhou,” from the Manchu fe, “old”) were those whose ancestors had joined up and been organized into banners during the time of Nurhaci and Hong Taiji; they were principally descended from the Jianzhou and other Jurchen tribal groups. On the other hand, the New Manchus (“Yiche Manzhou,” from the Manchu ice, “new”) were principally the descendants of those northeastern tribes who submitted only after the Qing had invaded China proper; they were relocated southward into the Amur River basin and brought into the banner system during the Shunzhi (1644–61) and Kangxi (1661–1722) reigns to defend against Czarist Russia’s expansionism in the region. These New Manchus included two non-Jurchen Tungusic groups—the Hezhe and Kiakar (Kuyala)—who lived by hunting and fishing amidst the mountains and streams of northeastern Manchuria and adjacent parts of Siberia; and four groups of intermixed Mongols and Tungus—the Daur (Dawoer), Solun (Suolun), Oroqen (Elunchun), and Xibe (Xibo)—who lived on either side of the Greater Xing’an Mountains separating the Mongolian steppe to the west and the Manchurian forests to the east. Because their association with the Qing rulers was more recent as well as more distant, the New Manchus were less prestigious than the Old, but they were allowed a degree of organizational autonomy denied the Old Manchus; even after they had been incorporated into the banner system, they remained under the leadership of their own tribal and clan chiefs.²⁹

One notable subset of the Old Manchus was the imperial lineage (huangzu), which consisted of two large categories of relatives. One was the imperial clan (zongshi), the Aisin Gioro (“Aixin Jueluo” in Chinese), who traced their descent directly from Nurhaci’s father, Taksi. The rest of the imperial lineage, consisting of the collateral lines descended from Nurhaci’s uncles and brothers, were known as the gioro (Ch. jueluo). Members of these two groups multiplied rapidly. At the end of the dynasty, there were twenty-nine thousand members of the imperial lineage in the main line and another twenty thousand in the collateral lines. Although each of the two categories of the imperial lineage were assigned to their own separate banner companies, all members, whether main line or collateral, were subject to the jurisdiction of the Imperial Clan Court (Zongrenfu) rather than that of the regular banner authorities.³⁰ The most exalted of the imperial clan were the titled princes and nobles, who were divided into twelve ranks, the top four of which were qinwang, junwang, beile (a Manchu term), and beizi (from the Manchu beise) respectively. With some exceptions, notably the eight “iron-capped princes” (tiemaozi wang) who had the right of perpetual inheritance, these ranks were inheritable only on a descending scale. The eldest son of a third-rank beile, for example, became a fourth-rank beizi, while all of the younger sons dropped down to “nobles of the ninth rank” (zhenguo jiangjun).³¹ Consequently, only a minuscule number of imperial clan members held a title.

In the post-Boxer era the most prominent member of the imperial clan, aside from the emperor and his immediate family, was unquestionably Yikuang (Prince Qing; see chart 1.1.). Born in 1836, Yikuang was a great-grandson of the Qianlong emperor and belonged to the Manchu Plain Blue Banner, one of the Five Lower Banners to which direct descendants of emperors were assigned. His grandfather, Yonglin, was the first Prince Qing, an honor conferred on him in 1799 after his brother had become the Jiaqing emperor. Yikuang inherited the family estate in 1850 when his uncle, Mianti, died without a male heir. By then, however, due to the impact of the Qing inheritance rules, Yikuang was a mere noble of the tenth rank (fuguo jiangjun). His subsequent rise to prominence came about as a result of his long association with Empress Dowager Cixi (1835–1908). When in 1884 he succeeded Yixin (1832–98), the first Prince Gong, as head of the Zongli Yamen (or proto–Foreign Office), he was made a prince of the second rank, thus reclaiming the family title of Prince Qing. Ten years later, on the occasion of Cixi’s sixtieth birthday, his position was raised to that of prince of the first rank. Finally, in December 1908, at the beginning of the Xuantong reign, he reached the top of the princely ladder and was given the rare right of perpetual inheritance. When a cabinet was formed in early 1911, Yikuang became China’s first prime minister.³² Among the collateral lines of the imperial lineage, the most notable figure in the post-Boxer period was probably Liangbi (1877–1912), a member of the Bordered Yellow Banner, whose ancestor was a younger brother of Nurhaci and whose grandfather, Yilibu, negotiated and signed the Treaty of Nanjing, ending the First Opium War (1839–42). At the time of the 1911 Revolution, Liangbi was a leading proponent of resistance to the Republicans. It was his assassination in late January 1912 that, as much as anything, convinced the court to abdicate.³³

CHART 1.1. Prominent Members of the Imperial Clan in the Late Qing

SOURCE: Li Zhiting, Aixin Jueluo jiazu, vols. 2–3.

NOTE: Dotted line = descent by adoption. Reign titles are set in italic type.

The second ethnic component of the banner system was the Mongol Eight Banners. They were drawn mostly from the horse-riding, sheep-herding peoples of eastern Mongolia bordering on Manchuria. The Mongol banners were also divided between Old and New, depending on whether they had submitted to the Qing before or after the invasion of China proper. The Old Mongols included early adherents such as the Khorchin (Keerqin) and Chahar (Chahaer); the New Mongols were late joiners such as the Barga (Baerhu) and Olot (Elute). The most prominent Mongol bannerman in the post-Boxer era was Xiliang (1853–1917), of the Bordered Blue Banner, who in early 1911 was the governor-general of the three northeastern provinces.³⁴

The third ethnic component of the Eight Banners was the Hanjun (often translated as “Chinese banners”).³⁵ Depending again on when they were enrolled into the system, they too were differentiated between Old and New. The Old Hanjun were descended from the Han frontiersmen living in southern Manchuria who had been captured by Hong Taiji. The New Hanjun were descended from captives of later wars, principally the invasion of China proper in 1644 and the suppression of the Rebellion of the Three Feudatories (1673–81). (Other Han prisoners of war after 1644 were put into the non-banner Army of the Green Standard [Laüying].) The Hanjun were the least well regarded of the three ethnic components of the banner system. It was they, particularly the New Hanjun, who bore the brunt of the Qianlong emperor’s reorganization of the banner system. Despite their relatively low status, the Hanjun were no less a part of the system than its other components.³⁶ The highest-ranking Hanjun among the officials of the post-Boxer era were the two Zhao brothers, Erxun and Erfeng, both members of the Plain Blue Banner. Zhao Erxun (1844–1927) became in 1911 the governor-general of the three northeastern provinces, succeeding the Mongol bannerman Xiliang; after the revolution, he supervised the compilation of The Draft History of the Qing. His younger brother, Erfeng (1846–1911), was the governor-general of Sichuan at the time of the revolution and was killed shortly afterward.³⁷

The last element of the banner system, ranking lower than even the Hanjun, comprised the bondservants. Bondservants owned by the emperor belonged to the Upper Three Banners and were under the jurisdiction of the Imperial Household Department (Neiwufu); they generally served in the palace. Those owned by various imperial princes belonged to the Lower Five Banners. Bondservants were not, strictly speaking, a separate part of the Eight Banner system, but were part of their master’s banner. Because they were hereditarily servile, bondservants as a group, even those in the Upper Three Banners, did not rank with the regular banners. While bondservants had played a crucial role as intermediaries for the Qing rulers during the early part of the dynasty, they were politically insignificant in the late Qing.³⁸

The Eight Banner organization provided the institutional framework within which its members, the banner people, lived out their lives. All banner people—men, women, and children alike—belonged to one of the twenty-four banners. Except for members of the imperial lineage and the bondservants of the Upper Three Banners, who were governed by the Imperial Clan Court and the Imperial Household Department respectively, all were subject to the jurisdiction of their banner. Membership in a particular banner was hereditary and, except when women changed banner affiliation on marriage, was largely immutable.³⁹ To the banner people, the banner and its subunits, particularly the company, were only slightly less important than the lineage and the family.

The Eight Banner system was organized along military lines. Each of the twenty-four banners was commanded by a lieutenant-general (dutong, rank 1B [within a nine-rank system] ) and two deputy lieutenant-generals (fudutong, rank 2A), who were almost always themselves bannermen, often imperial princes and high officials, though not necessarily members of that particular banner or even that ethnic component within the banner system.⁴⁰ Each banner looked after its own affairs, while the Banner Duty Office (Zhinianqi), composed of representatives from the Eight Banners, provided overall coordination.⁴¹ The banner commanders were located in Beijing.

For the banner people, the banner was one of two basic social units. It determined, in part, the social standing of individuals because the different banners varied in prestige depending on ethnicity and color. The color and ethnicity of a banner also determined where its members lived, as each banner was assigned a residential area according to the Manchu system of correlation between color and compass direction. Under this system, which differed from that of both the Chinese and the Mongols, yellow was associated with north, white with east, red with west, and blue with south. Thus, within Beijing’s Inner City surrounding the Forbidden City (the emperor’s Winter Palace), the two yellow banners (the most prestigious) were assigned land in the north, the white banners in the east, the red banners in the west, and the blue banners (the least prestigious) in the south.⁴² This explains why, for example, the Manchu novelist Lao She (1899–1966), whose family belonged to the Plain Red Banner, was born and grew up in the west-northwest sector of Beijing.⁴³ The distribution of the banners by color provided the basis for the further division of the Eight Banners into two “wings” (yi), with the Bordered Yellow, the two Whites, and the Plain Blue forming the east (or, when facing south, left) wing, and the Plain Yellow, the two Reds, and the Bordered Blue forming the west (or right) wing.⁴⁴ A banner’s ethnicity was another, though secondary, factor in determining its residential location. In Beijing, each of the eight sectors into which the Inner City was divided by the color of the banner was further divided into three districts, one for each of the ethnic components of that banner. The result was the creation in Beijing’s Inner City of twenty-four residential zones, one for each of the banners, with the Manchu, Mongol, and Hanjun banners interspersed throughout the city but each living in their own separate quarters.⁴⁵ This spatial arrangement by the color and ethnicity of banners was replicated in most other places where banner soldiers were garrisoned.

Each banner was divided into battalions (jiala, from the Manchu jalan), then subdivided into companies (zuoling). In a Manchu or Hanjun banner, there were five battalions; in a Mongol banner, only two. Each battalion was headed by a colonel (canling, rank 3A), who was appointed from within the banner, as were all subordinate officials. The number of companies within a battalion varied from fourteen to nineteen in the Manchu banners, eleven to fifteen in the Mongol banners, and five to nine in the Hanjun.⁴⁶ Altogether, as we shall see, there were over two thousand banner companies.

More so than the banner, the company was the primary social organization for the banner people. Membership in a company, as in a banner, was hereditary, and members of a company “resided together within the area designated for their banner.”⁴⁷ The company was the banner person’s primary focus of identification, in the same way that the county was for most other Chinese. Whereas a Han would be identified as “So-and-so of X county in Y province,” a banner person would be known by his or her company and its parent banner. Each company was headed by a captain (also known as zuoling, rank 4A), assisted by a lieutenant (xiaoqixiao, rank 6A). Just as the magistrate was the “father and mother official” to the inhabitants of his county, so the captain was to the members of his banner company. So closely was a banner company associated with its leader, who unlike the county magistrate was not subject to rotation, that it was more likely to be identified by the name of its captain than by its unit designation; for example, the Fifteenth Company in the Third Battalion of the Manchu Plain Red Banner was, in 1904, better known as “Yizhen’s company.”⁴⁸

Finally, the basic unit of the banner organization was the banner soldier (qibing). Each company was assigned a quota of soldiers. In the late Qing, according to Zhang Deze, a company in a Manchu banner had about 260 soldiers; in a Mongol company, more than 100; and in a Hanjun company, about 140.⁴⁹ One of the main functions of a company captain was to maintain a register of all the members of his company, which he submitted to his banner headquarters and to the Board of Revenue every three years. The register recorded the name, age, occupational status, and family relationship of every member of the company. In particular, it classified the men according to their availability for military service: those who were not yet sixteen years (sui) in age were classified as “young males” (youding); those who were between sixteen and sixty as “able-bodied males” (zhuangding); those who were older than sixty as “retired” (tuiding); and those who were ill as “handicapped” (feiding). The purpose of the register was to keep track of those who were eligible for selection as new banner soldiers in the company when old ones retired or died.⁵⁰ Except in the Hanjun, another purpose of the register was to assist the captain in recommending young women from his company who might be eligible for service in the emperor’s palace.⁵¹

All banner personnel, apart from their membership in one of the twenty-four banners, were assigned to a specific geographical location, usually a military garrison, depending on which type of banner company they belonged to. Also, all banner soldiers were enrolled in one of several service branches. These additional affiliations served to further differentiate the banner population. For example, Lao She’s father was a member not only of the Manchu Plain Red Banner but also of the Metropolitan Banners in Beijing and the Guards Division (Hujunying). Just as the statutory strength of the banner system had long been fixed by the late Qing, so had the geographical distribution of the companies, the number and size of individual garrisons, and (except for a few creations of the Self-Strengthening Movement in the late nineteenth century) the size and duties of the service branches. Consequently, the descriptions of the banner system in the 1899 edition of The Collected Statutes of the Great Qing (Da-Qing huidian) and in The Draft History of the Qing differ hardly at all from Thomas Wade’s detailed account in the 1851 volume of the English-language Chinese Repository, which was based on the 1812 edition of The Collected Statutes.

Banner companies were customarily divided among three types—outer, inner, and garrison—but there was a fourth type as well, which may be called “affiliated.” Outer companies (wai zuoling) consisted of the regular banner units located in and around Beijing; they were commonly called the Metropolitan Banners. They were the most numerous of the four types, totaling 1,147 companies. Inner companies (nei zuoling) were composed of bondservants and hence were also known as bondservant companies. They, too, were located in Beijing as well as in Zhili and Fengtian. The Draft History of the Qing suggests that they numbered 115 companies.⁵² Garrison companies (zhufang zuoling) were regular banner units stationed outside Beijing at various locations in the provinces; they added up to 817. These three types of banner companies, which altogether totaled 2,079, are usually thought of as constituting the totality of the banner force. However, as The Collected Statutes also indicates, there was a fourth type of banner company, which consisted of various frontier peoples who belonged to the banner system but were not formally counted among the provincial garrisons. There is no complete accounting of the strength of such affiliated banner companies.

Soldiers of the Metropolitan Banners were divided principally among five long-established service branches (listed in order of prestige): the Escorts (Qinjunying), Vanguard (Qianfengying), Guards (Hujunying), Light Cavalry (Xiaoqiying), and Infantry or Gendarmerie (Bujunying), all of which date from the days of Nurhaci and Hong Taiji. Subsequently, in the early Qing, several new units were created: the Firearms Division (Huoqiying), Yuanmingyuan Guards, and Scouts (Jianruiying). During the Self-Strengthening Movement in the late Qing, yet another new unit, which contemporary Westerners called the Peking Field Force (Shenjiying), was added to the Metropolitan Banners. All of these units, including the Peking Field Force, served principally in one capacity or another as palace guards for the emperor and his residences. Thus, nine-tenths of the Metropolitan Banners were stationed in Beijing’s Inner City. The remainder were quartered near the emperor’s Summer Palace in the northwestern suburb.

The total strength of the Metropolitan Banners, according to different estimates, ranged between 125,000 and 150,000 soldiers and officers (see table 1.1). T. F. Wade’s statistics for the various divisions indicate that the Metropolitan Banners consisted of 7,305 officers, 90,333 regular soldiers, 31,064 reservists, and 2,497 artisans and others, for a total of 131,199. An 1849 source cited by Wang Zhonghan said that the total was 149,425. Yang Du, writing in 1907, quoted figures for the banner units in Beijing that add up to 139,430. The Draft History of the Qing asserts that the Metropolitan Banners numbered 126,989.⁵³ On the basis of these sources, it may be estimated that the total strength of the Metropolitan Banners in the late Qing was about 130,000 soldiers and officers. If the roughly 20,000 bondservants (derived from Wade and The Draft History of the Qing) are added to this figure, then altogether the banner personnel in and around Beijing totaled about 150,000.

Banner personnel who were not in the Metropolitan Banners or in bondservant companies belonged to either garrison companies or, in some instances, affiliated banner detachments. According to The Collected Statutes, the garrison companies numbered 817, divided among ninety-one garrisons, with each garrison assigned a certain number of companies and a quota of soldiers and officers.⁵⁴ As was previously noted, there is no complete accounting of the affiliated companies. The ninety-one provincial garrisons varied greatly in size. The largest ones—namely, Shengjing (7,031), Xi’an (6,588), and Jingzhou (5,668)—had more than five thousand soldiers each; the smallest ones, such as Shunyi in Zhili, had only fifty soldiers. The larger or more important provincial garrisons typically were commanded by a general (jiangjun, often translated into English as “Tartar-general”; rank 1B) or a brigade-general (also fudutong; rank 2A); the smaller garrisons, by a military commandant (chengshouwei, rank 3A; or fangshouwei, rank 4A) or a platoon captain (fangyu, rank 5A).⁵⁵ The commander of a larger garrison often had one or more smaller nearby garrisons within his overall jurisdiction. In the most extreme case, the general at Shengjing was responsible for twenty-four other garrisons in Fengtian. The garrison commander was usually appointed from outside the garrison; his subordinates came from within the garrison.

TABLE 1.1. Composition and Strength of the Banner Troops in Beijing

SOURCES: Wade, “Army of the Chinese Empire,” 254, 308; Qingshigao, 130:3886–89; Yang Du, Yang Du ji, 425; Liu Fenghan, “Qingji ziqiang yundong,” 346–47.

As with the Metropolitan Banners, the banner soldiers in the provinces were divided among several service branches, principally the Light Cavalry, the Vanguard, and the Infantry, with the Light Cavalry as the principal unit. Unlike the Metropolitan Banners, the provincial garrisons did not include Escorts and Guards, and they also largely excluded the Hanjun. Following the large-scale demobilization of the Hanjun during the Qianlong era, almost all of the provincial garrisons were staffed solely by Manchu and Mongol banner soldiers. Only the garrisons in the three northeastern territories (particularly Fengtian) and in Guangzhou and Fuzhou had any Hanjun.⁵⁶

The garrison at Jingzhou (present-day Jiangling), the historically strategic city in Hubei up the Yangzi River from Wuhan, was, except for its large size, a representative provincial garrison. Established by the Kangxi emperor in 1683 following the suppression of the Rebellion of the Three Feudatories, it was headed by a general and consisted of fifty-six companies: forty from the Manchu banners, sixteen from the Mongol banners, and none from the Hanjun. Assisting the general were two brigade-generals, each responsible for one of the two wings into which the companies were divided by banner color. Below the two brigade-generals were ten regimental colonels (xieling, rank 3B), one for each of the eight Manchu banners and one for each of the two wings of Mongol banners. Each of the fifty-six companies was headed by a captain, with ten of the captains serving concurrently as regimental colonels. Each company also had one platoon captain and one lieutenant. Aside from the officers, the garrison was made up of—according to The Collected Statutes of the Great Qing—5,668 soldiers, two-thirds of whom (3,800) were either privates (majia) or corporals (lingcui) in the Light Cavalry. The remaining soldiers consisted of 200 vanguards, 700 infantrymen, 80 artillerymen, 168 artisans, and 720 reservists (yangyubing).⁵⁷ Insofar as they were members of one of the Eight Banners, they were under the jurisdiction of the commander of their respective banner in Beijing. At the same time, because they were stationed in the provinces, they were also under the jurisdiction of the general of their garrison. Thus, no less than their peers in Beijing, the banner soldiers in the provinces possessed multiple identities, depending on their banner, geographical location, and service branch. During the post-Boxer era, probably the most prominent bannerman who came from the provinces was Tieliang (1863–1938), a member of the Manchu Bordered White Banner from the Jingzhou garrison. He rose to the position of grand councilor in 1905 and was both a bitter rival of Yuan Shikai (1859–1916) and a hated enemy of the republican revolutionaries.⁵⁸

By the late Qing, the assignment of soldiers to a garrison was hereditary and permanent.⁵⁹ The last large-scale uprooting of banner soldiers was after the Taiping Rebellion, when the ranks of such decimated garrisons as Hangzhou and Nanjing were replenished by recruits from other garrisons. The Hangzhou garrison, whose statutory strength was 2,216 soldiers, was virtually reconstituted between 1865 and 1880 by 1,461 transfers from Zhapu, Fuzhou, Dezhou, Qingzhou, Jingzhou, and Chengdu. The destroyed garrisons, however, were not always restored to full strength. The Nanjing garrison originally had 4,666 soldiers in 40 companies, but when it was restocked by transfers from Jingzhou following the Taiping massacre, it had only 2,424 soldiers in 24 companies.⁶⁰

The provincial garrisons were widely distributed among four geographical regions: the “capital region” (jifu), northeast China, northwest China, and the rest of China proper (see table 1.2).⁶¹ According to The Collected Statutes of the Great Qing, the banner soldiers in the capital region, which encompassed most of Zhili Province outside Beijing, were divided among 25 garrisons, with a total of 14,238 soldiers organized into 54 banner companies. The largest of these garrisons, with 2,100 soldiers and headed by a brigade-general, was at Miyun, north of the capital. Also within the region were two large affiliated banner detachments, both stationed outside the Great Wall. One was a detachment of 1,000 soldiers at the Mulan imperial hunting grounds, north of Rehe, commanded by the Rehe lieutenant-general. (It is unclear from The Collected Statutes how many banner companies these 1,000 soldiers constituted.) The other, consisting of 10,800 Chahar, Olot, Barga, and other Mongols organized into 120 companies, was under the jurisdiction of the Chahar lieutenant-general at Zhangjiakou. Finally, although they were not formally counted among the capital region contingent of garrisons or among the affiliated detachments, a group of 1,460 banner soldiers guarded the two sets of Qing imperial tombs in Zhili: the Eastern Tombs near Zunhua (where the Shunzhi, Kangxi, Qianlong, Xianfeng, and Tongzhi emperors are buried) and the Western Tombs near Yixian (where the Yongzheng, Jiaqing, and Daoguang emperors are buried).⁶²

TABLE 1.2. Disposition and Strength of the Provincial Banners

SOURCES: Da-Qing huidian, 84: 4b-8a; 86: 4b-8a; Wade, “Army of the Chinese Empire,” 315, 321, 324–25.

The banner contingent in the northeast (or Manchuria) consisted of 338 companies, totaling 35,361 soldiers in forty-four garrisons, six of which were attached “water forces” (shuishiying). They were divided among the three territories of Fengtian, Jilin, and Heilongjiang, each headed by a general. Slightly more than half of the banner troops in Manchuria were located in the southernmost territory, Fengtian, whose capital, Shengjing (also known as Mukden, now Shenyang), was the site of the largest provincial garrison in the country. Yet other banner soldiers not counted among the regular banner companies were located in the northeast. These included three large affiliated detachments. One was the Buteha [Hunting] Eight Banners, made up of Solun, Daur, and Oroqen; another was a “herding” detachment composed of Barga and Olot. Both were located in the Greater Xing’an Mountains in what is now Inner Mongolia, and both were under the jurisdiction of the Heilongjiang general at Qiqihar. The third affiliated detachment was made up of the Hezhe and Kiakar along the coast of the Japan Sea, who were under the authority of the brigade-general at Sanxing in Jilin (now Yilan in present-day Heilongjiang). Also counted as part of the banner force in the northeast were two other groups: the seven hundred soldiers stationed in detachments of twenty to fifty under platoon captains at nineteen gates (bianmen) along the length of the Willow Palisade (Liutiaobian), which enclosed southern Manchuria; and the 539 soldiers guarding the three early Qing imperial tombs in Fengtian (including those of Nurhaci and Hong Taiji).⁶³

The banner contingent in the northwest consisted of eight garrisons, with 15,642 soldiers organized into 119 companies, stretched out along a vast arc from Ningxia in the east to the Yili River valley on the Xinjiang-Russian border in the west. These eight garrisons were distributed among three subregions: Ningxia and the Gansu corridor, eastern Xinjiang, and western Xinjiang. The largest of the northwestern garrisons was at Huiyuan (also Kuldja, now Yining), on the Yili River, with forty companies under the command of a general. Also under the authority of the general at Yili were two tribal detachments, one Solun and one Daur, relocated from northern Manchuria in the mid-eighteenth century; they were encamped along the south bank of the Yili River and were considered as part of the Yili garrison rather than as affiliated banner soldiers.⁶⁴

Finally, the banner contingent in China proper (exclusive of Zhili) consisted of 39,879 soldiers, organized into 306 companies and divided among fourteen garrisons and two attached water forces. Except for Yunnan-Guizhou in the southwest, there was at least one banner garrison in each of G. William Skinner’s “macro-regions,” usually (but not always) in the core city of the region. Each of the three regions closest to the capital had more than one garrison. In the plains of north China, excluding the Metropolitan Banners in Beijing and the capital-region contingents in Zhili, the banner forces were stationed in three cities: Qingzhou (now Yidu), in east-central Shandong; Dezhou, on the Grand Canal in western Shandong; and Kaifeng. In the northwest highlands, the banner forces were stationed in three cities also: Xi’an, Suiyuan (now Hohhot), and Taiyuan. In the lower Yangzi region of east-central China, the banner troops were divided among four garrisons: Nanjing (or Jiangning), Zhenjiang (or Jingkou), Hangzhou, and Zhapu, on the Zhejiang coast halfway between Hangzhou and Shanghai. Each of the other four macro-regions had only one garrison. In the middle Yangzi region, the garrison was at Jingzhou; in the upper Yangzi region, at Chengdu; along the southeast coast, at Fuzhou; and in the West River region of south China, at Guangzhou. At both Fuzhou and Guangzhou the garrison included an attached water force.⁶⁵

The total strength of the ninety-one provincial garrisons was about 110,000 (see table 1.2). Together with the 150,000 soldiers in the Metropolitan Banners and the bondservant companies, the total banner force (exclusive of the affiliated detachments) adds up to 260,000 soldiers. Of these, roughly 55 percent were concentrated in Beijing and 45 percent were scattered among the provinces, with 5 percent in the capital region, 19 percent in the northeast, 6 percent in the northwest, and 15 percent in the rest of China proper. This finding accords well with Wang Zhonghan’s estimates for 1849: 149,425 (52 percent) in Beijing, 52,552 (18 percent) in the northeast, and 85,219 (30 percent) everywhere else in China. If an estimated forty thousand soldiers in the affiliated detachments are added to the others to produce a grand total of three hundred thousand banner soldiers, then the banner troops would have been evenly divided between the capital and the provinces, which is what Zheng Tianting and numerous other scholars have concluded.⁶⁶

These three hundred thousand banner soldiers constituted the core personnel of the Eight Banner system, but they were by no means all of the banner people, for the Eight Banners were not only a military organization made up of banner soldiers but also an administrative system that incorporated the banner soldiers’ numerous dependents. In addition to bannermen’s wives and daughters, the system also included many men who were not soldiers. It is a common misconception that every able-bodied adult bannerman was a soldier. As the banner population proliferated over the course of the Qing period while the size of the banner army remained more or less fixed, it was not even the case (contrary to the recollection of Aixin Jueluo Zongkui) that every banner family produced a banner soldier. Those bannermen who failed to be selected as soldiers and who were not otherwise employed (e.g., as officials) were known as xiansan or (from the Manchu) sula, that is, unsalaried or “idle” bannermen.⁶⁷

These idle bannermen loomed large among the banner soldiers’ dependents, but how numerous were they in the late Qing? Different sets of information yield very different answers. At one extreme, several banner company records from the Beijing area in 1904 suggest that the unsalaried may have constituted no more than one-third or one-half of all adult bannermen. Thus, in one Manchu banner company (perhaps of the Plain White Banner) headed by the gioro Juntong, only 49 out of 128 adult men (38 percent) were xiansan; in another company of unspecified ethnicity commanded by Songhua, 100 out of 216 (46 percent) were xiansan. In a third company, Yizhen’s in the Manchu Plain Red Banner, of 168 able-bodied adult men, 79 were soldiers; the other 89 (53 percent) were presumably xiansan. Viewed from a different angle, these company rosters suggest that the ratio of banner soldiers to adult bannermen was roughly one to two or two to three.⁶⁸ At the other extreme, some authorities suggest, though without much supporting evidence, that this ratio of no fewer than one banner soldier to two adult bannermen is much too narrow. A Short History of the Manchus (Manzu jianshi), for example, asserts that already in the days of Hong Taiji only one in three adult males was selected as a banner soldier and that by the Qianlong reign in the mid-eighteenth century, only one in eight was so selected. Wang Zhonghan similarly estimates that by the late 1840s the ratio of banner soldiers to adult bannermen had widened to one to fifteen.⁶⁹

Yet another type of source, family genealogies from the northeast, yields statistics that fall between these two extremes. According to the records of three banner families in eastern Fengtian—the Guan (Guaerjia, or Gūwalgiya) and the Wu (Wushu) in Xinbin and the Wang (Wanyan, or Wanggiya) in Xiuyan—only about 10 percent of the males in each of the three families ever qualified as soldiers or officials. Among the Guan, for example, over the course of eight generations stretching from before 1644 and to 1825, there were altogether 230 males in the family, of whom only 24 were soldiers and 7 were officials; the other 199 (87 percent) were xiansan. In the eighth generation alone, living in the early nineteenth century, there were 87 Guan males, of whom only 3 were soldiers; 84 were xiansan. Among the Wu, of 226 males going back twelve generations, only 23 were soldiers or officials; among the Wang, of 123 males over four generations, all but 12 were xiansan. Similarly, in Daoyi, a banner village in Fengtian near Shengjing, according to a detailed analysis of its banner registers in the eighteenth century, only about 10 percent of the males were or had been in military or administrative service. In short, the three northeastern family genealogies and the Daoyi registers agree that the ratio of banner soldiers to bannermen was about one to ten.⁷⁰

If one were to accept one to ten as the approximate ratio of banner soldiers to bannermen and then factor in the female half of the system, the banner soldiers would have accounted for about one-twentieth (5 percent) of the entire population of banner people. Indeed, records of the Banner Affairs Office (Qiwuchu) in Fengtian in 1910 confirm that banner soldiers and officials made up 4.9 percent of all the banner people of the province.⁷¹ Consequently, if the total number of banner soldiers was three hundred thousand and the ratio of soldiers to banner people was one to twenty, then the total banner population in the late Qing would have been six million. This is roughly comparable to the figure of five million that Liang Qichao, Zou Rong, and other anti-Manchu critics of the early post-Boxer era regularly gave as the number of Manchus. If so, the banner people constituted slightly more than 1 percent of the total population of China at the time.

SEPARATE . . .

On 23 July 1865 an edict issued in the name of the nine-year-old Tongzhi emperor (r. 1861–75) endorsed a memorial from Governor Shen Guifen (1817–81) of Shanxi urging a fundamental reform of the Eight Banner system. The governor had proposed that

hereafter all banner people [qiren] who wish to go out and make their own living . . . be permitted by the lieutenant-general [of their banner] to venture forth. If they wish to settle down and work in the provinces, they should be permitted to notify the local county magistrate and be registered as banner people. . . . In all litigation, they should then come under the jurisdiction of the county magistrate. If some do not attend to their own affairs but stir up trouble, they should be punished by the local officials on the same basis as civilians [minren]. Those who wish to be reclassified as civilians should be entered in the local registers as civilians.

It is on the basis of this “important” edict that Mary Wright concludes, “Most of the last restrictions separating the Manchus from the [Han] Chinese were removed in 1865.”⁷² The Tongzhi emperor’s decree did not, however, have the far-reaching impact that Wright ascribes to it, simply because it was never carried out. Consequently, even after 1865, the Manchus continued, as before, to live lives that were separate from and unequal to those of the Han.⁷³

Manchus were segregated from Han in four important respects: administratively, occupationally, residentially, and socially. First, they were kept apart from Han administratively by a system known as “separate governance of banner people and civilians” (qimin fenzhi). As the 1865 edict indicates, Manchus were classified differently from Han. They were registered as “banner people,” whereas the non-banner people, who were nearly all Han, were generally registered as “civilian.” These classifications were hereditary and essentially permanent. Not since the seventeenth century had Han been enrolled into the banner system. As for the Manchus, they had rarely been permitted to leave the system, and this had occurred only on the explicit instruction of the ruler, such as when the Qianlong emperor ordered the demobilization of a large number of Hanjun and their reclassification as civilians. The 1865 edict cited by Wright admittedly made it easier than before for a banner person to resign from the system, because it gave blanket authorization to the commanders of the twenty-four banners to approve individual requests to get out. The decree, however, merely permitted—it did not order—the reclassification of the banner population. There is no evidence that many members asked or were allowed to resign from their banners.

Due to their different classifications, Manchus were governed separately from Han. As civilians, the Han were under the jurisdiction of the regular local administrative hierarchy, which ranged upward from the county magistrate to the provincial governor; as banner people, the Manchus were subject to their own set of officials, from the captain of their banner company to the lieutenant-general of their banner in Beijing. Generally speaking, civilian officials had no jurisdiction over banner people, nor did banner officials have authority over civilians. For example, at Xiuyan, a garrison town in eastern Fengtian with between five hundred and six hundred banner soldiers, the banner people were under the command of their garrison’s senior military commandant and ultimately the general at Shengjing; the civilian population, on the other hand, was governed by a second-class subprefect (tongpan) locally and by the metropolitan prefect of Fengtian at Shengjing. Legal cases involving members of both populations in Xiuyan required joint adjudication by the garrison’s commandant and the subprefect.⁷⁴ On the face of it, the 1865 edict did away with at least part of this administrative segregation when it extended the judicial authority of local civil officials to cover those banner people who had opted to settle in the provinces but had kept their banner classification. It appears, however, that few banner members exercised this option either. Consequently, both the separate classification and the separate governance of banner people and civilians continued in force down to the end of the century.

Second, Manchus were segregated from Han occupationally in that they, in essence, constituted a hereditary military caste from which non-banner civilians were excluded. As members of the Eight Banner system, their primary responsibility was to fight, when needed, as soldiers in the banner army, and when they were not soldiering, they were permitted only to serve as officials in the Qing government or to farm. They were specifically barred from engaging in trade or other occupations.⁷⁵ These prohibitions applied not only to the banner soldiers but to other members of the banner system as well. Although the 1865 edict explicitly allowed the banner people “to go out and make their own living,” it does not appear to have had any practical consequences. Manchus in the late nineteenth century still labored under the old occupational restrictions, which by and large set them apart from Han.

Third, Manchus were segregated from Han residentially. Half of the banner population were hereditarily assigned to the Metropolitan Banners in Beijing, and most of the rest to one or another of the ninety-one provincial garrisons that were unevenly distributed throughout the Qing empire. Whether in Beijing or in the provinces, all banner people were tightly bound to their respective garrisons. In particular, they were not allowed to travel beyond a certain distance of their garrison headquarters without the explicit approval of their superiors. In Beijing they were restricted to within forty li (about thirteen miles) of the capital, and in the provinces to within twenty li of their garrison (although in the northeast it was one hundred li, or thirty-three miles). These limits applied as well to the location of banner settlements away from a garrison city. For example, Sanjiazitun (Three Family Village, in present-day Fuyu County), whose banner personnel were under the jurisdiction of the Heilongjiang general, was located ninety-five li north of Qiqihar, near the outer limit of the permissible range for Manchuria. Any banner person who ventured beyond these specified distances without permission was considered a deserter. Manchus thus lacked the Han’s freedom of movement. When the expectant prefect Jitai, of the Hanjun Bordered Red Banner, died in Shandong in 1888, his son had to petition his banner commander for a travel permit to accompany the coffin back to the capital. As the Shandong governor observed in a covering letter forwarding the petition,

If the family of the deceased had been classified as civilian, there would have been no need to request permission to begin their journey. This youth, however, belongs to a banner. Heretofore, each banner person, when leaving the capital on official business, has been required to report in writing to his lieutenant-general’s headquarters before he can proceed⁷⁶

In the urban centers where they were garrisoned, except in Manchuria, the banner soldiers and their dependents lived in special quarters known as “Manchu cities” (Mancheng), which may be classified into four types depending on their spatial relationship to the indigenous population.⁷⁷ One type was the “attached twin city,” where the Manchus lived in a walled enclosure that was external to but contiguous with another walled city that the non-banner people inhabited. Beijing’s Inner (or “Tartar”) City is the sole example of this type. When the Qing forces captured the Ming capital in the seventeenth century, they expelled all of the residents of the Inner City, which they took over entirely for their own people, and they relegated the Han and other non-banner people to the old walled addition to the south known as the Outer (or “Chinese”) City⁷⁸ A second, and related, type was the “detached twin city,” where the banner people were quartered in a new, specially built walled city that was near to but separate from an existing walled city where the other people lived. In Shanxi, for example, the Manchus were garrisoned at Suiyuan, built in the 1730s, while Han, Mongols, and others lived about a mile and a half to the northeast in the Ming-era city of Guihua; known locally as the New and Old Cities respectively, these two settlements are now districts of the city of Hohhot in Inner Mongolia. Other Manchu cities of this second type were at Qingzhou (Shandong), Ningxia, and, until it was destroyed in the 1860s, Urumqi (Xinjiang).⁷⁹

Two other types of Manchu cities, which Mark Elliott aptly calls “intramural,” were located within an existing walled city. One type was a walled compound entirely enclosed within the larger Han city, such as in Kaifeng and Zhenjiang.⁸⁰ The other type, by far the most common, was not fully surrounded by its own set of walls but instead shared a part of the larger Han city’s external walls. The amount of land occupied by the banner people in these intramural Manchu cities varied greatly. As depicted on maps, the Manchu city in Jingzhou took up the entire eastern half of the walled city; in Xi’an, the northeastern third; in Guangzhou, the western fourth; in Nanjing, a substantial portion of the southeastern sector within the old Ming imperial city; and in Taiyuan, merely the southwestern corner.⁸¹ As in Beijing, wherever the banner people occupied portions of an existing city, they had first cleared the area of its former residents; they then built a set of internal walls, which were not so high nor so massive as the external ramparts but which nevertheless clearly demarcated one group of people from the other. Fuzhou and Guangzhou were exceptional in that their Manchu cities were not walled.⁸²

In principle, Manchu cities were the exclusive residential quarters of the banner people. In Beijing, non-banner people originally could visit the Inner City during the day but not dwell there. At Hohhot, non-banner people, who lived at the Old City of Guihua, could not enter the new Manchu city of Suiyuan at will. Similarly, at Zhenjiang, the Manchu garrison “was off limits to Han.”⁸³ In reality, with the passage of time, a growing number of non-banner civilians did live in the Manchu cities. In the capital, some high-ranking Han officials were, by imperial grace, allowed to reside within the Inner City. For example, Zhang Zhidong (1837–1909), who served on the Grand Council during 1907–9, lived on Baimixie Street just outside Di’anmen, the northern entrance into the Imperial City.⁸⁴ More commonly, shopkeepers managed in time to gain admission into the Manchu cities, especially since the banner people themselves were prohibited from engaging in commerce. In Beijing, by the mid-nineteenth century the population of the Inner City included some fifteen thousand non-Manchu shopkeeper households, in addition to the seventy-six thousand regular banner households.⁸⁵ Nevertheless, despite this Han influx, the Manchu cities remained down to the late Qing predominantly Manchu in population, even if they were no longer exclusively so, as the scattered census returns from 1908 and 1910 reveal. In Beijing, as late as 1908, the banner people still constituted 54 percent of the 414,528 individuals in the Inner City. In Chengdu, they were 53 percent of the households in the Manchu City; and in Xi’an, 64 percent.⁸⁶ Non-banner civilian immigration, therefore, had eroded but did not destroy the Manchu cities’ fundamental characteristic as ethnic ghettoes.

Another characteristic of Manchu cities was that even with their Han inhabitants, they were sparsely populated, particularly when contrasted with the rest of their host city. (Beijing was an exception: the population of the Inner City slightly exceeded that of the Outer City.)⁸⁷ Intramural Manchu cities, as noted, sometimes took up as much as one-quarter or even one-half of an entire city, yet proportionately their resident population was considerably less. In Guangzhou, where they occupied about one-fourth of the city’s area within the walls, the banner people probably numbered no more than one-tenth, or perhaps even one-twentieth, of the city’s estimated total population of six hundred thousand. Foreign visitors in the late nineteenth century often commented on the parklike atmosphere of the spacious yet sparsely populated Manchu cities. One lauded the Manchu City in Xi’an for “its wide, healthy spaces, its lovely gardens, its grand old trees.” Of Chengdu, where the banner people took up one-fifth of the city, another wrote, “The Manchu quarter is one of the most picturesque parts of the city, with the charm of a dilapidated village set in untidy gardens and groves of fine trees.”⁸⁸ The sharp difference in population density and in the pace of life between the Manchu and Han quarters in Chengdu was captured well in a novel by Li Jieren (1891–1962), Ripples across Stagnant Water (Sishui weilan), written in 1935 but set at the end of the Qing. Separated by a low wall, the two areas were known locally as the Small City (Shaocheng) and the Big City (Dacheng):

The low wall created two entirely different worlds. In the Big City, there were houses and shops everywhere and numerous pedestrians would be moving to and fro on the cobblestone streets. Not a drop of green could be seen in the whole district. By contrast, as soon as one entered the Manchu City, there were trees in abundance. . . . Everywhere was one vast expanse of green. A broad unpaved road [probably present-day Changshun Street] flanked by low yellow earthen walls extended itself among the shadows far into the distance. Inside the walls, flowers and trees complemented the low houses, interspersed with ponds covered with lotus blossoms. And human beings were rarely to be seen. Unlike, for example, the Big City where pedestrians would be everywhere and on some of the downtown streets you would have to squeeze through shoulder to shoulder. In the Manchu City, you could walk down whole streets without encountering other pedestrians. In any case, those whom you did meet were utterly different from people in the Big City where, except for some cultured older types, everyone was invariably in a hurry. Here [in the Small City] most of the male pedestrians would be strolling with a bird cage in their hands or a fishing-rod on their shoulders. As for the females, dressed in long gowns, waists cinched and hair tied back, they would walk along slowly in their slippers smoking long bamboo pipes. The Manchu City was indeed a different world.⁸⁹

In sum, in an urban environment such as Chengdu, Manchus and Han lived out their lives in separate worlds and generally had little to do with one another. Because the Manchu City had few stores and workshops, it was economically dependent on the rest of the host city. Aside from commerce, however, the two populations did not interact much. As one longtime resident of Chengdu recalls, “Before the Xuantong era [1908–12], the Han people very rarely entered the Small City for recreation; similarly, the banner people hardly ever went to the Big City for activity. The line between the two was very strict.”⁹⁰

As for the sizable minority of banner people who did not live in urban garrisons, they too were residentially segregated. Those in Zhili Province outside Beijing lived on land that long ago was cleared of its Han inhabitants, for in the mid-seventeenth century, the newly established Qing, in a move that paralleled the expulsion of previous residents from Beijing’s Inner City, had “enclosed” (quanzhan) much of the land within a five-hundred-li (170–mile) radius of the capital, evicting the Han farmers and reallocating the land to its banner soldiers and bondservants to farm.⁹¹ Although many Han farmers subsequently managed to filter back into the region and, despite a prohibition on the sale of banner land, even regain control of the land from their Manchu proprietors, Manchus and Han lived apart from each other in separate settlements. Thus, near the Summer Palace, soldiers of the Outer Detachment of the Firearms Division resided in walled villages whose inhabitants were entirely banner people. The bondservants working on imperial or princely estates in various parts of Zhili, such as Xiaoyingzi in Qinglong County and Zhoujiazhuang in present-day Yi County, likewise lived in separate communities composed of banner people. Such ethnic enclaves were also widespread in Fengtian. For example, the village of Daoyi, north of Shengjing, was composed predominantly, if not exclusively, of Hanjun.⁹²

The most extensive example of the Manchus’ residential segregation was central and northern Manchuria, which, even more so than the five-hundred-li zone around Beijing, was supposed to be an exclusively banner domain from which non-banner civilians were barred. This was the region north and east of the Willow Palisade. The palisade, an earthen levee about a yard high planted with a row of willow trees with a parallel trench about three yards deep, was composed of three sections radiating outward from the vicinity of Kaiyuan, north of Shengjing, dividing Manchuria into three distinct regions. To the south was the Liao River plain of Fengtian, which the Han had long cultivated. To the west was the arid steppe occupied by Mongols. Finally, to the east and northeast was the forested region of Jilin and Heilongjiang, from which the original (or Old) Manchus had come. The Qing rulers were determined to maintain this third region of Manchuria, in Robert Lee’s words, as “a fountainhead of ancestral virtues and a reservoir of military power,” by prohibiting Han immigration beyond the Willow Palisade.⁹³

The Qing’s efforts at keeping northern Manchuria a Manchu preserve free of the Han were a mixed success. On the one hand, the Willow Palisade was no Great Wall; it was lightly guarded, poorly maintained, and easily penetrated. There were along the entire length of the barrier only nineteen gates, through which all traffic into and out of the region was supposed to be funneled, and each was guarded by fifty or fewer banner soldiers. Furthermore, by the nineteenth century, the palisade itself was run-down at so many points as to be all but useless. The British consular official Alexander Hosie (1853–1925) observed in 1896 that “little but the gates, each with its guard of a few soldiers, remains.” Long before then, Robert Lee says, the palisade had been “breached at will,” as Han from overcrowded north China ignored the ban in search of land to farm. On the other hand, notwithstanding its porosity and deterioration, the Willow Palisade could not be, and was not, ignored with impunity. Thus, even though the barrier had supposedly been breached at will for over a century, there were still very few Han immigrants living in Jilin and Heilongjiang in the 1860s.⁹⁴ Indeed, it was the very success of the Qing in keeping the Han out of the northernmost parts of Manchuria that allowed the Russians in 1860 to walk off with the extensive territory north of the Amur River and east of the Ussuri.⁹⁵ Even when afterward the Qing began to allow Han immigration into northern Manchuria to forestall the Russians from further encroachments, the effect was geographically limited. The opening-up started with the Lalin district along the middle reaches of the Songhua River (in Jilin) in 1860 and the Hulan basin (in Heilongjiang) in 1862; it was subsequently extended to the upper Mudan valley, the Tumen valley, and the upper Ussuri (all in Jilin) between 1878 and 1882 and the Tongken valley (in Heilongjiang) in 1896.⁹⁶ However, aside from these specific areas where Han immigration was now unrestricted, much of northern Manchuria beyond the Willow Palisade was still in the late nineteenth century populated largely, though no longer exclusively, by banner people.

Last, Manchus were segregated socially from Han. In particular, they were forbidden to marry Han. For example, the Manchus at Xiuyan, the garrison town in eastern Fengtian, observed four “big taboos” concerning marriage. Along with prohibitions against marrying someone with the same surname, someone from the same generation, or a minor was a prohibition against marrying a Han. According to local informants, the ban on Manchu-Han intermarriage was “a regulation of the dynasty and could never, ever, be contravened.” The prohibition does not appear in the Qing code as such, but successive Qing rulers, including the Yongzheng emperor and Empress Dowager Cixi, declared it to be imperial policy. As Ding Yizhuang asserts, “The prohibition against marriage between banner people and civilians was indeed extraordinarily severe.”⁹⁷

As a cumulative consequence of the various aspects of segregation, Manchus and Han generally lived in separate communities and kept apart from each other. Indeed, the banner garrisons were so isolated from the local population that they often became “language islands,” where, as in Jingzhou, the banner people, after two and a half centuries, still spoke the standard Chinese (Mandarin; Putonghua) dialect of north China, from which their ancestors had been dispatched, rather than the dialect of the people among whom they were living. In Guangzhou, according to a 1959 field survey, “prior to the 1911 Revolution the Manchus all spoke standard Chinese. Although they could understand and talk Cantonese, they normally did not speak it.” Similarly, according to the journalist Harry A. Franck in the early 1920s, the Manchus at Xi’an were “recognizable to the others by their Peking dialect.”⁹⁸ The persistence of such linguistic distinctions is additional evidence that the two groups had rarely interacted.

. . . AND UNEQUAL

In addition to keeping Manchus and Han separate, the Qing rulers dealt with the two populations unequally. Their policy was not so much anti-Han as pro-Manchu, but their favoritism toward the Manchus occurred unavoidably at the expense of the Han. Thus, down to the end of the nineteenth century, the banner people received preferential treatment over the non-banner civilians in three broad areas: legal, political, and economic.

Legally, Manchus (as banner people) were treated differently as well as better than Han (as civilians). If, for example, a Manchu and a Han were brought before a court on a legal matter, the Han was required to kneel before the magistrate, whereas the Manchu was permitted to stand.⁹⁹ More important, if found guilty of a crime, a Manchu was subject to a lesser punishment than a Han. According to the Qing code, banner people could opt for a beating with a whip instead of a bamboo rod, and they could substitute wearing the cangue in place of penal servitude or even military exile. In one early nineteenth-century case, a Manchu in Heilongjiang was found guilty in the accidental death of a boy. If the defendant had been a Han, he would have been beaten with one hundred blows of the heavy bamboo rod and exiled for life at a distance of three thousand li (one thousand miles), but because he was a bannerman, his sentence was reduced to an equal number of blows of the whip and the wearing of the cangue for sixty days.¹⁰⁰ The 1865 edict permitting banner members to leave the system would have greatly reduced such forms of discrimination between Manchus and Han, because it specified that those banner people who had asked to be reclassified as civilians, as well as those who had resettled in the provinces but had kept their banner status, would be subject to the same laws as the civilians. However, since the edict was not widely implemented, the special consideration that Manchus received in legal matters remained in force down to the end of the century.

Politically, Manchus benefited from preferential treatment in recruitment, appointment, and tenure. In recruitment, there were four methods by which Manchus could enter the civil bureaucracy much more easily than could Han. First, when bannermen participated in the regular (or “literary”) examinations, they were assigned separate quotas that were more generous than those for the Han. Thus, among the 4,457 who passed the triennial metropolitan examinations and were awarded the jinshi degree between 1862 and 1894 were 242 bannermen, who constituted 5.4 percent of all successful candidates, which was several times greater than their share of China’s total population. They included the previously noted Xiliang and Zhao Erxun, both of the class of 1874. Two other bannermen officials of the post-Boxer era who entered the civil service by way of the regular examinations were Duanfang (1861–1911), governor-general of Zhili in 1909, and Natong, a grand councilor during the Zaifeng regency (1908–11) at the end of the dynasty; they gained their provincial, or juren, degrees in 1882 and 1885 respectively.¹⁰¹

Second, Manchus were allowed to take a separate set of examinations that were less demanding than the literary ones and from which the Han were excluded. These “translation” (fanyi) examinations led to degrees that were identical to those awarded under the literary examinations except for the addition of the prefix fanyi (e.g., fanyi jinshi). As their name suggests, they were an exercise in translation: from Chinese to Manchu for Manchu bannermen and Hanjun, and from Manchu to Mongol for Mongol bannermen. They typically asked for a commentary on and a translation from the Four Books of the Confucian classics. Compared with their literary counterpart, the translation examinations were shorter, consisting of, for example, two rather than three sessions for the metropolitan examination; moreover, their success rate was several times higher. In the early nineteenth century, about one in fifteen bannermen passed the translation examination for the provincial degree, and about one in five or six passed the metropolitan examination. Because they were less difficult, the translation examinations lacked prestige and did not attract the most capable of bannermen candidates. Nevertheless, they were no less effective a way for bannermen to enter government service than were the literary examinations. For example, the Hanjun Fengshan (d. 1911), who in 1907 replaced Yuan Shikai as commander of four Beiyang New Army divisions, acquired his provincial degree via a translation examination.¹⁰²

Third, Manchus were also allowed to enter the government by taking yet another translation test that qualified them for employment as low-ranking metropolitan officials known as Manchu-language scribes (bitieshi, from the Manchu bithesi). The test itself was quite easy, and while it was open to holders of the translation examination degree, it could be taken by other bannermen as well. It was, however, closed to Han. Among the officials in the post-Boxer era who obtained their first posts as Manchu-language scribes were Fuqi, the acting general of the Guangzhou garrison when he was assassinated in 1911, and Songshou, governor-general of Fujian-Zhejiang in 1911.¹⁰³

Finally, yet other Manchus entered government service without having been certified by any sort of examination; they did so by taking advantage of their father’s hereditary yin privilege (granted for meritorious service) or by purchasing a degree or position of their own. Thus, Empress Dowager Cixi’s confidant Ronglu (1836–1903) started his career as an honorary licentiate (yinsheng), a degree bestowed upon him in 1852 in recognition of his father’s death while resisting the Taiping rebels; this led to an initial appointment to the Board of Public Works as a second-class secretary and culminated in his appointment as a grand councilor in 1898. Similarly, Zhao Erfeng, younger brother of the metropolitan degree holder Zhao Erxun, began his career by purchasing a post in the salt administration in Guangdong.¹⁰⁴ The purchase of degrees and offices in order to enter the civil service was not, of course, a Manchu monopoly. Yuan Shikai, for example, started by buying the title of secretary of the Imperial Patent Office attached to the Grand Secretariat.¹⁰⁵ Yuan’s “irregular” beginning, however, was an exception among high-ranking Han officials in the late Qing. Such was not the case among their Manchu peers, for whom an “irregular” beginning was more often the norm than the exception.

Because they had more numerous as well as less demanding ways of entering the government, Manchu officeholders were, not surprisingly, less academically qualified than Han. Over the course of the Qing period, 72.1 percent of Han governors-general held either the metropolitan or the provincial examination degree, whereas only 33.9 percent of Manchu bannermen and even fewer Mongol bannermen and Hanjun did. For the same reasons, Manchus also began their official careers four to six years sooner than did Han. Whereas the average age at first appointment of a Han as governor-general was just over fifty-six, that of a bannerman was between 49.5 and 52.1 years.¹⁰⁶

In appointment, as in recruitment, Manchus had an advantage over Han because many posts in the government, particularly in the capital, were reserved for them. In the metropolitan administration, the Qing allocated most positions among members of six ethnic or status categories, namely, imperial clansmen, Manchu bannermen, Mongol bannermen, Hanjun, bondservants of the Upper Three Banners, and Han.¹⁰⁷ These might be called “ethnic slots.” In most instances, half of all of the top posts in the metropolitan bureaucracy were reserved for members of the first five categories, and the other half were allocated to the Han. This was the well-known institution of Manchu-Han dyarchy, with the five various groups of bannermen collectively making up the Manchu half of the structure. On the basis of a detailed analysis of the top personnel of the Grand Council, the Grand Secretariat, and the Six Boards in the mid-nineteenth century, John Fairbank confirmed that “a roughly equal balance was in fact maintained between Manchus and [Han] Chinese in the official posts at the capital.” The principle of dyarchy was observed among the top ranks of the Censorate as well.¹⁰⁸ Insofar as an equal number of Han were allowed to serve with Manchus, dyarchy can be considered nondiscriminatory. Indeed, with regard to official appointments, successive Qing emperors prided themselves on their alleged impartiality; their guiding thought, they claimed, was that “Manchus and Han were as one family” (Man-Han yijia).¹⁰⁹ But since the Manchus were only 1 percent of China’s population, their half-share of the top metropolitan posts was obviously disproportionate to their numbers.

In any case, the principle of dyarchy applied only to a few posts at or near the top of the metropolitan administration. It did not, for example, extend to the shadowy post of board supervisor (zongli), who outranked a board’s two presidents and oversaw the operations of their ministry; when the appointment (which was more common in the eighteenth than the nineteenth century) was made, the board supervisor was almost always a Manchu.¹¹⁰ More commonly, dyarchy did not apply to the far more numerous positions at the middle and lower echelons of the Qing’s nine-rank administrative system, where Manchus greatly outnumbered Han both absolutely and proportionately. At the Grand Secretariat, for example, twenty of the twenty-four mid-level positions (83 percent) were reserved for Manchus (fourteen Manchu bannermen, four Mongol bannermen, two Hanjun) and only four were for Han. The situation at the Six Boards was similar. At the Board of Revenue, of the 141 statutory positions in the three middle ranks, ninety-seven (69 percent) were reserved for Manchus (four imperial clansmen, ninety-one Manchu bannermen, two Mongol bannermen); the remaining forty-four slots (31 percent) were to be filled by Han. The Censorate and the Hanlin Academy were the only metropolitan agencies where at the middle ranks Manchus were not more numerous than Han.¹¹¹ The Manchus’ preponderance was most egregious among the bottom three ranks, where in particular the post of Manchu-language scribe was set aside exclusively for bannermen. The scribe’s primary duty was to translate documents, including memorials addressed to the throne, from Chinese into Manchu; he also served as copyist and archivist. Though lowly, the post existed in large numbers in every metropolitan agency. At the Board of Revenue, Manchu scribes numbered 141, almost half of the ministry’s statutory officials; they comprised one imperial clansman, 120 Manchu bannermen, four Mongol bannermen, and sixteen Hanjun.¹¹² Manchus thus monopolized the bottom of the metropolitan bureaucracy to the almost total exclusion of Han.

Not only did the principle of dyarchy not extend below the top ranks of most organs of the metropolitan government, but it did not apply to certain agencies at all. Except for some clerical posts requiring a knowledge of Chinese that were held by Han, only imperial clansmen were appointed to the Imperial Clan Court, which had jurisdiction over members of the imperial lineage, and only bondservants of the Upper Three Banners were appointed to the Imperial Household Department, which served the needs of the court. The broader group of bannermen similarly filled all 161 posts at the Court of Colonial Affairs (Lifanyuan), which was responsible for dealing with the Mongols, Turks, and Tibetans living in the outlying realms of the Qing empire.¹¹³ Of course, no Han was appointed to the Eight Banners either. Overall, according to Chen Wenshi’s exhaustive analysis of 2,277 posts in the fourteen most important metropolitan agencies, 1,559 (68.4 percent) were earmarked for Manchus—1,255 imperial clansmen and Manchu bannermen, 196 Mongol bannermen, 108 Hanjun—and only 416 (18.3 percent) for Han, with the remaining 302 (13.3 percent) unspecified.¹¹⁴

Away from the capital, there were also some (though not nearly so many) posts that were set aside for Manchus and from which Han were excluded. These were typically located along the frontiers of the empire, especially the three regions of Manchuria—Fengtian, Jilin, and Heilongjiang—which were governed as territories, outside the regular provincial administration, down to the end of the nineteenth century. Each of the three territories was ruled by its respective banner general, each in command of a large banner force. With the Qing rulers determined to maintain Manchuria, especially its northern part, as a Manchu preserve, civil administration was largely dominated by bannermen as well. For example, the prefect of Fengtian, the chief civilian official in the entire northeast, was always a Manchu. At the garrison town of Xiuyan, in eastern Fengtian, of the forty-six second-class subprefects who had jurisdiction over the non-banner population from 1772 to 1856, thirty-six were Manchus and only ten were Han. Even when the civilian administrative apparatus was extended into northern Manchuria in the late nineteenth century to keep pace with the influx of Han immigrants, most of the local officials continued to be Manchus. At Hulan, Heilongjiang, the subprefect—from the time the post was created in 1863 to oversee the affairs of the Han settlers, until its abolition in 1905–was never a Han.¹¹⁵ Of course, the banner administration in the provinces, as in the capital, was another exclusive preserve of the Manchus.¹¹⁶

Because most provincial and local posts were not earmarked as “ethnic slots,” Manchus were not nearly so dominant there as they were in the capital; still, their presence was disproportionately large. It was greatest among the top two ranks of the provincial administration, where dyarchy prevailed informally. Thus, over the course of the Qing dynasty, Manchus were 57 percent of all governors-general and 48.4 percent of all governors; in terms of length of tenure, they were in office 61.6 percent of the time as governors-general and 51 percent as governors. The Manchu presence diminished below the level of governor, particularly as the post declined in importance. Thus, among provincial financial commissioners and judges, 28 to 29 percent were bannermen; among prefects, 21 percent; and among county magistrates, only 6 percent. However, even at 6 percent, Manchus as local administrators exceeded their share of the population.¹¹⁷

Once they were appointed, Manchus also received preferential treatment politically, in that they were likely to have longer tenure than Han. Among governors-general, for example, the average length of service of a bannerman was five years five months, while that of a Han was eleven months shorter. Among county magistrates, 26.4 percent of Manchus (as against 20.9 percent of Han) served longer than three years.¹¹⁸ Probably the main reason why Manchu officials had greater longevity than Han was the difference in their mourning requirement. Whereas Han officials had to give up their post and return home for twenty-seven months when a parent died, bannermen mourned for only one hundred days and could then resume their duties or receive immediate reassignment. Thus, out of 1,052 commanders of provincial banner garrisons (i.e., generals and brigade-generals) spanning 113 years from 1796 to 1908, only four left office because of mourning.¹¹⁹

It is widely, if mistakenly, assumed that the preferential treatment of Manchus at the expense of Han in the political realm greatly diminished during and after the Tongzhi Restoration, the period of dynastic recovery in the 1860s. The belief rests, in part, on the indisputable fact that far fewer bannermen were appointed as provincial rulers after 1851, as the Qing court began to recognize the power of the new Han “regional” armies and their commanders that arose during the suppression of the Taiping Rebellion. Thus, whereas Manchus constituted 57 percent of all governors-general over the entire Qing period, they were only 34.6 percent during 1851–1912; similarly, the percentage of Manchus as governors declined from 48.4 percent for all of the Qing to 22.2 percent for the last sixty years of the dynasty. There was, however, no parallel shift from Manchus to Han at the metropolitan level of government, where appointments were still made on the basis of dyarchy. When, for example, the Zongli Yamen was established in 1860 as a subcommittee of the Grand Council to handle China’s burgeoning foreign relations, its staff of forty-two secretaries was evenly divided between Manchus and Han.¹²⁰

Indeed, not only did the post-Taiping court continue to favor Manchus at the expense of Han by making metropolitan (if not provincial) appointments on the basis of ethnic slots, it exhibited even greater favoritism toward Manchus when it excluded practically all Han from several of its Self-Strengthening reforms. Thus, both the Peking Field Force (formed in 1862) and the Kunming Lake Naval School (Kunminghu Shuishi Xuetang, opened in 1888) drew their personnel entirely from the ranks of the Metropolitan Banners, while the Translators College (Tongwenguan, founded in Beijing in 1862 and in Guangzhou two years later) recruited most, though not all, of its students from the local banner population. The court’s exclusion of Han from its efforts at modernization was also evident at the Navy Yamen, founded in 1885 after the naval debacle of the Sino-French War (1883–85). Unlike the Zongli Yamen, where the staff was evenly split, the Navy Yamen’s administration was practically all Manchu. While the top five commissioners were divided roughly equally (in the spirit of dyarchy) between Manchus and Han, with the Manchus usually holding an edge of three to two, only one of the thirty-one secretaries who served during the yamen’s ten-year history was a Han.¹²¹

Finally, economically, too, Manchus were treated preferentially over Han. The clearest example of economic favoritism was the subsidy paid to the banner soldiers, which went to support not only the soldiers but the rest of the banner population. The Manchus were thus hereditary stipendiaries of the state. Most banner soldiers, particularly those stationed in China proper, received two kinds of compensation: a monthly payment in silver and a semiannual distribution of grain. Compensation varied with service branch and rank. According to the published pay scale for a normal year (i.e., without an extra intercalary month), an Escort, the most prestigious among the Metropolitan Banner soldiers, received forty-eight taels of silver (at a rate of four taels a month) and 22.2 piculs of grain; a private in the Light Cavalry, the banners’ common soldier, received thirty-six taels and also 22.2 piculs. The pay of an officer was, of course, higher. A lieutenant in the Metropolitan Banners, for example, received sixty taels of silver in regular salary (probably with an additional amount of “integrity nourishing” money) and seventy-two piculs of grain. Compared with the Army of the Green Standard, a Han unit, the pay scale of the banner force was at least twice as high.¹²²

A minority of banner personnel, particularly those in less populated areas such as Manchuria, received a part of their pay in land, in lieu of the grain distribution. This was known as “emolument land” (suiquedi) and “soldiers land” (wutiandi), which they could farm themselves or, more likely, rent out to others. The land was theirs so long as they remained in active service, but when they retired or died, it reverted to their garrison for redistribution to others. Like the monetary payment, which they also received, the land allocation was graduated according to rank. At the Baishizui Gate along the Willow Palisade in the Liaoxi corridor, the garrison commander (a platoon captain) received eighty taels annually in silver and nineteen acres in emolument land; an ordinary banner soldier, twenty-four taels and thirteen acres.¹²³ As another example of economic favoritism, Manchu landholders were generally exempt from the land tax. All banner land in China proper as well as in central and northern Manchuria was entirely tax free, while banner land in southern Manchuria was taxed only nominally and at a lower rate than non-banner land.¹²⁴ In an attempt to ensure that Manchus remained economically independent, the Qing prohibited the transfer of property from banner people to civilians (qimin bujiao chan). Although widely ignored and even occasionally revoked within China proper, where banner land was relatively scarce, the prohibition was enforced in Manchuria, where banner land was much more plentiful. It seems to have kept much of the land in the hands of the Manchus; thus, in Shuangcheng (then in Jilin but now in Heilongjiang), 83 percent of the arable land was still owned by Manchus well past the 1911 Revolution.¹²⁵

Apart from his stipend, whether paid in silver, grain, or land, the banner soldier enjoyed other economic benefits as well. He was given monetary grants to help defray the cost of weddings and funerals. If he had distinguished himself in battle and then retired because of illness, he was entitled to disability benefits of one tael a month for the rest of his life. When he died, his widow received half of his stipend. His orphans, too, received a small monthly stipend.¹²⁶

In yet another expression of the regime’s policy of economic favoritism, when the Manchus, despite their stipends, fell into economic difficulties, they could look to the Qing state to come to their rescue, or at least to try to help them out. The banner people, particularly the majority who were garrisoned in urban centers, became impoverished in different ways. They lived beyond their means, spending more than their stipends permitted and going into debt to grain merchants and other lenders. They could not keep up with inflation, as their stipends remained fixed while the cost of living increased. They, ignoring the ban on the transfer of property to civilians, sold off their allotted banner land to their Han tenants. And, perhaps most commonly, they proliferated in numbers far beyond the capacity of the banner system to accommodate. Because the banners did not expand in size to keep pace with the increased population, employment opportunities within the system became increasingly scarce. By the late Qing, as previously explained, perhaps only one in ten bannermen could become a banner soldier. Since only the banner soldier was compensated, and since others in the banner system were barred from any occupation other than soldiering, officeholding, and farming, the economic situation of the banner people became increasingly difficult over time.¹²⁷

Successive Qing emperors, from the late seventeenth to the late nineteenth century, undertook a variety of measures to try to solve what came to be called the “problem of the Eight Banners’ livelihood” (Baqi shengji wenti). Early Qing emperors canceled monetary debts to the central treasury that impoverished Manchus had incurred. They redeemed banner land that had been mortgaged to non-banner civilians. They demobilized a large part of the Hanjun in the provinces and replaced them with Manchu and Mongol bannermen from Beijing. They created new positions in the banner system, such as that of reservist, to provide additional employment and income opportunities. They relocated unsalaried bannermen and their families from Beijing to virgin lands in Manchuria.¹²⁸ At best, these early measures slowed down but did not reverse the Manchus’ long-term trend toward impoverishment. The economic plight of the banner people worsened in the late nineteenth century. The post-Taiping court itself was in such financial straits that it was obliged to cut back on its payments to the banner soldiers. The outright grants of money for weddings and funerals were replaced by loans, which were deducted from the soldiers’ subsequent pay. The stipends themselves were reduced by 30 to 40 percent.¹²⁹ The 1865 edict permitting the banner people to leave the system and find work on their own was one more attempt, ultimately ineffective, to solve the problem of the Eight Banners’ livelihood.

In Beijing and other urban garrisons of China proper, where they had no other source of income than the banner soldiers’ stipends, the Manchus sank deeper and deeper into the state of genteel poverty that Lao She, in Beneath the Red Banner (Zhenghongqi xia), recalls so graphically from his Beijing childhood. As a member of the Guards Division, Lao She’s father received a stipend of three taels a month and two semiannual distributions of rice. The land the family had been given, amounting to twenty to thirty mu (three to five acres) north of the city, “had been sold generations ago, leaving only a little more than one mu on which there now stood a few gravestones.” Entirely dependent on his father’s meager stipend, they lived continually on credit: “Buying on credit and running up debts had become our way of life. . . . As soon as the monthly allowances came in, we paid our debts, which left us so little cash that we needed credit again.”¹³⁰ The poor but proud bannerman became a stock figure in late Qing fiction.¹³¹

Eyewitness accounts of China’s Manchu cities in the late nineteenth century dwell not only on their spacious, parklike qualities but also on the poverty and indolence of their inhabitants. The British consular official Alexander Hosie, after a visit to the banner garrison in Chengdu in the early 1880s, observed that its residents were “slip-shod, down-at-heel, lazy-looking” and that “the people, especially the women, were badly, even slovenly, dressed; everything announced the presence of parasites battening on Government pay, without affording any adequate return.” Seasoned traveler Elizabeth Kendall, speaking in 1911 of Chengdu also, concurred with Hosie’s observations: “Loafing in the streets and doorways are tall, well-built men and women, but they had a rather down-at-heel air, for their fortunes were at a low ebb when I was at Chengtu.” Regarding the Manchu quarters in Guangzhou, the American missionary B. C. Henry wrote, “Their part of the city shows a marked contrast to the purely Chinese portion. Their houses are smaller and poorer, and an air of neglect, thriftlessness, and decay spreads over all.”¹³²

Banner people living in rural areas were not so stricken as their urban peers. On the one hand, they were spared the seductive extravagances of urban life. On the other, they could fend for themselves; they were not absolutely dependent on the soldiers’ stipend and the largess of the state. In Manchuria, where the population was sparse and land was plentiful, many banner soldiers lived not in the garrison town to which they were assigned but rather in outlying villages, which was permissible so long as they were within one hundred li of the garrison. There, the soldiers and their families could farm or, as local circumstances allowed, make their living in yet other ways. For example, the Manchus of Sanjiazitun, who belonged to the Qiqihar garrison of Heilongjiang, supplemented their soldiers’ income of twenty-four taels a year by trapping fish in the Nen River. Other Manchus farmed. In Daoyi, the Hanjun community north of Shengjing, according to a reconstruction of its population in the late eighteenth century, only 4 percent of all the men were soldiers, and 3 percent were hereditary craftsmen. “The vast majority of the men . . . were clearly farmers who tilled plots of Banner land.” Bondservants on imperial and princely estates, too, made their living from farming; ironically, though their social status among Manchus was low, they may have been better off economically than the more prestigious, but sedentary, banner soldiers in the cities.¹³³

TWO PEOPLES, TWO CULTURES

Manchus and Han were, at the beginning of the Qing period, two very different peoples with very different cultures. During the more than two hundred years that the Qing ruled over China, each group necessarily affected the other. It is well known that the Manchus over time underwent a process of “Sinicization” (or, more accurately, “Hanification”); it is less well known that the Han were subject to the reverse process of “Manchufication.” Such mutual acculturation, however, was never complete. As a result, down to the end of the nineteenth century, Manchus and Han remained, to a greater extent than is generally realized, culturally different from each other.

The Qing rulers, from the beginning, were much worried about the possibility of becoming culturally absorbed by their far more numerous Han subjects. They were keenly aware of what had happened to previous conquest dynasties, including the Jurchen Jin (1127–1234), from whom they claimed descent. When Hong Taiji in 1636 lectured his officials on the perils of acculturation, he admonished them to heed the warnings of the fifth Jurchen emperor, Shizong (r. 1161–90) and to learn from the ultimate fate of the Jin:

Fearing that his descendants would imitate the customs of the Han, he [Shizong] prohibited this, and many times urged his people not to forget Taizu and Taizong [the dynastic founders]. He ordered them to follow the old customs in clothing and language, to practice horsemanship and archery regularly so they would be ready for warfare. In spite of these imperial admonitions, later generations disregarded him and forgot their horsemanship and archery. Thus we come to Aizong, where the gods of grain were threatened, and subsequently the dynasty was wiped out.¹³⁴

The Qing therefore attempted, on the one hand, to inculcate the ways of the original (or Old) Manchus among the rest of the banner population and, on the other, to preserve this distinctive culture of the banner people from being eroded by the Han.

The Manchu way of life was summed up in the Chinese phrase guoyu qishe, “national speech and mounted archery.”¹³⁵ “National speech” during the Qing period was Manchu. It differed greatly from Chinese in grammar and orthography. Manchu, an Altaic language, is inflected, agglutinative, and atonal; Chinese, a Sino-Tibetan language, is none of these. Manchu is written with a modified form of the Mongolian alphabet; Chinese, with characters. The Qing rulers were concerned that every bannerman know Manchu—not just every Old Manchu whose ancestral speech it was, but all other banner people as well. Thus, schools for banner youths were set up to teach the language. At the Jingzhou garrison in Hubei, it was required that “all bannermen, regardless of their personal background or status, study Manchu writing in order to look after the fundamentals.”¹³⁶ Examinations in translation were instituted to screen bannermen who were prospective officials. Those who excelled at the language earned bonuses, rewards, advancement, and other material benefits. As the Yongzheng emperor once explained, “If some special encouragement as this is not offered, the ancestral language will not be passed on and learned.” Conversely, those who performed poorly on such tests were penalized. Every five years, soldiers and officers in the banner force were supposed to take an oral examination in Manchu. Those who failed were given until the next review to achieve competence; if they failed again, they faced dismissal.¹³⁷ The Xibe, a New Manchu group of Mongol origin, were so adept at learning and absorbing Manchu that it eventually replaced Mongolian as their “native” language.¹³⁸

The effort to inculcate and preserve the Manchu language among the banner population at large was generally not so successful as among the Xibe. The declining use of Manchu among the banner people was already marked in the early eighteenth century and only became worse thereafter, as successive emperors lamented. The Jiaqing emperor, for example, complained that “officials in the Imperial Household are simply not versed in the Manchu language, and when copying they must write stroke by stroke, and letter by letter; not only do they not comprehend the sense, but they do not even recognize the very graphs.” The extant memorials in the Palace Archives (in Taiwan) confirm the steady decline of the Manchu language among bannermen officials, most of whom wrote only in Manchu in the early Qing, then in both Manchu and Chinese in the mid-Qing, and finally only in Chinese in the late Qing. The archives of the Hulan banner detachment in Heilongjiang, far from the center of Han cultural influence, exhibit a nearly identical pattern of change. By the end of the nineteenth century, only about 1 percent of the Manchus in Hulan could read Manchu, and no more than two-tenths of 1 percent could speak it. At around the same time, at the office of the Shengjing general in southern Manchuria, reportedly the sole occasion for using Manchu was to write the memorial wishing the reigning emperor a long life; all other public business at the yamen was conducted in Chinese. By then, too, banner-family genealogies in Fengtian were no longer written in Manchu but in Chinese. In Beijing, Lao She’s cousin Fuhai “knew only a smattering of Manchu and always spoke Chinese, which he used for any occasional writing he did.”¹³⁹

However, the Manchus’ linguistic acculturation, though far along, was never complete; Fuhai, after all, still knew “a smattering” of Manchu. The banner people, while taking up Chinese, did not stop using Manchu altogether, particularly those living in out-of-the-way places. Manchu was widely spoken in the Yili valley of western Xinjiang among the Xibe; it was also used, though less widely, in Manchuria. In Fengtian in 1870, the Scottish missionary Alexander Williamson observed that “some of the more aged still speak the Manchu language.” He also noted that in “rare” instances some of the young people, if they already knew Chinese, were given additional instruction in the Manchu script. The British consular official Alexander Hosie, after a tour of the northeast in the 1890s, asserted that “the Manchu language . . . is to all intents and purposes a thing of the past” but admitted that it was still spoken in “remote corners” of Jilin and Heilongjiang “where Tartar tribes have kept themselves isolated and beyond the tide of Chinese immigration.” Even within China proper, Manchu was not entirely a dead language in the late nineteenth century. In Beijing, some, though admittedly not many, official documents continued to be produced in Manchu. As Ch’en Chieh-hsien concedes, “The conservative instincts of the Manchus were very strong, and so the official archives and the Historiography Office contain numerous records written in Manchu, dating as late as the last years of the dynasty.” Also, students in banner schools were still taught the Manchu language. Thus, prior study of Manchu writing was one of the principal prerequisites when the Beijing Translators College started recruiting banner students for European language study in 1862. The Guangzhou Translators College similarly stipulated that its banner students, in addition to learning English, should also study, as time allowed, written and spoken Manchu (Qingzi Qingyu). Because banner soldiers were routinely tested in spoken Manchu, minimum proficiency was sufficiently widespread in the Peking Field Force, the semimodern banner army organized in the 1860s, for its officers to be able to issue voice commands to their troops in Manchu. Indeed, the Peking Field Force’s promoter, Yixin (Prince Gong), in 1878 commended this practice to other Qing military units of the time as a way of keeping secrets from inquisitive Westerners.¹⁴⁰

Manchu name-giving practices reflected the process as well as the limits of their linguistic acculturation. Whereas Han typically had three-character names made up of a monosyllabic surname and a disyllabic personal name, Manchus originally had surnames and personal names that were generally both polysyllabic. Furthermore, their surnames were made up of two components: a lineage name preceded by a sublineage name. For example, the Aisin Gioro, the imperial clan, were a subset of the larger Gioro lineage. Finally, Manchus seldom used their surnames in public, such as when memorializing the throne; instead, they usually referred to themselves by their personal names, for example, Ronglu, a member of the Suwan Gūwalgiya.¹⁴¹

Other members of the banner population besides the Old Manchus adopted and went by Manchu-style names. The ancestors of the post-Boxer official Duanfang, for example, were Han Chinese from Zhejiang who, when they moved to southern Manchuria in the late Ming, became subjects of the Qing and were enrolled in the Manchu Plain White Banner; they then Manchufied their surname from Tao to Tohoro (Tuohuoluo in Chinese). Members of the Hanjun often altered an originally monosyllabic Han family name by adding to it the two-syllable suffix giya (jia in Chinese) to make it sound Manchu; thus, the monosyllabic surname Li would become (in Chinese) the disyllabic Lijia.¹⁴² Despite such examples, the Qing rulers seemingly did not require all of their bannermen followers who were of Han origin to adopt Manchu-style names. Thus, the bondservant family of Cao Xueqin (1715–63), author of Dream of the Red Chamber (Hongloumeng), evidently never changed their surname. Neither did the Hanjun family of the late-Qing officials Zhao Erxun and Zhao Erfeng. Nevertheless, many members of the Hanjun did go by Manchu names. For example, six of the seven Hanjun who were sent to study police matters in Japan in 1901 had Manchu-style two-syllable names, and so did two of the six students who were identified as Hanjun in the school directory of the Metropolitan University (Jingshi Daxuetang) for 1906. And the Hanjun graduate of the Beijing Translators College who was China’s minister to the United Kingdom in 1902–5 had both a Han-style name (Zhang Deyi) and a Manchu name (Deming [1847–1918]).¹⁴³

Manchu men’s names, like the Manchu language in general, underwent a process of acculturation that was substantial but by no means complete. In the beginning, among the first couple of generations, Manchu men had polysyllabic personal names (e.g., Nurhaci) that in their native language may have been meaningful but when transliterated by sound into Chinese characters were gibberish; furthermore, they did not arrange their personal names in generational order, as Han often did.¹⁴⁴ In time, however, Manchu names began to show Han traits. With the imperial clan itself taking the lead, Manchus started to shorten their personal names to disyllabic ones (e.g., Yinzhen, for the future Yongzheng emperor), to adopt names that were meaningful and felicitous in Chinese, and to assign names on a generational basis. By the time of the Guangxu emperor (r. 1875–1908), all the males of his generation in the imperial clan had the character zai in their personal names, such as Zaitian (the emperor), Zaifeng (1883–1952; his brother and future regent), and Zaizhen (1876–1948; his cousin). By contrast, the previous generation had used the character yi (e.g., Yikuang, Yixin, and Yihuan), while the following generation used the character pu (e.g., the future Xuantong emperor Puyi [1906–67] and his brother Pujie [1907–94]). In a further refinement of the generational principle, the second character in the personal name of each person in the direct line of succession contained a radical that distinguished these names from all others of their generation in the imperial clan. Thus, the tian character in “Zaitian” and the feng character in “Zaifeng” shared the “water” radical; however, the zhen in “Zaizhen” (written with the “hand” radical) did not, because as the son of Yikuang (Prince Qing), Zaizhen was not in the direct line of succession.¹⁴⁵

Among nonimperial Manchu families, a similar process was at work. At Xiuyan, in eastern Fengtian, the Manchus in the seventh or eighth generation continued as before to give their sons polysyllabic Manchu personal names that were meaningless when transliterated into Chinese, but at the same time they began to also give them Chinese names that were disyllabic and meaningful and that conformed to the generational principle. Thus, in the seventh generation of the Gūwalgiya lineage were sons with two names, one Manchu and one Chinese, such as Duolunbu/Shiman, Delinbu/Shizhu, and Tehengbu/Shizhen. Within the family and the banner, these boys used their Manchu name, but outside they used their Han-style name. Then, from the eighth or ninth generation on, at the beginning of the nineteenth century, the Gūwalgiya at Xiuyan stopped giving polysyllabic Manchu names to their sons, who thereafter used Chinese names exclusively.¹⁴⁶

Some Manchus, though not the imperial clan, went further toward acculturation by replacing their polysyllabic surname with a Han-type monosyllabic surname to go along with their Chinese disyllabic personal name, as the Hanjun often did.¹⁴⁷ The Qing rulers, somewhat surprisingly, found serious fault with this practice; though they themselves had led in adopting Han-style personal names, they drew the line at family names. The Qianlong emperor repeatedly warned the Manchus “not to succumb to Han surnames, which will cause confusion,” and when the ancient and politically prominent Manchu lineage of Niohuru adopted the Han-style surname Lang, he ridiculed them for having “forgotten their roots.” (The Niohuru, whose name was derived from niohe, Manchu for “wolf,” had chosen Lang as their surname because it was a homophone for the Chinese word for “wolf.”)¹⁴⁸

With regard to name-giving, therefore, Sinicization had proceeded apace for more than two hundred years. Manchu men had abandoned their original polysyllabic personal names in favor of Han-style disyllabic names; they had adopted the Han practice of choosing characters with auspicious meanings for the names; and they had assigned names on a generational basis. For all that, however, there remained down to the end of the nineteenth century one feature of Manchu name-giving that had not changed. Except among some Hanjun such as the two Zhao brothers, bannermen still did not, by and large, use their family name but called themselves only by their personal name—for example, Yikuang, Ronglu, Gangyi, Duanfang, Xiliang, and Tieliang. In this respect, most Manchus remained conspicuously different from Han.

Besides “national speech,” the other core element of the Manchu way of life—one that differentiated them even more from the Han—was mounted archery. Mid-Qing emperors repeatedly insisted that horsemanship and archery constituted the “fundamental essence of us Manchus” (wo Manzhou genben), which had to be preserved. The Jiaqing emperor in 1820 even admonished the banner soldiers in Jilin not to use the musket in hunting but to use the ancestral bow and arrow instead. It was not only the Old Manchus but the rest of the banner system as well—not only banner soldiers but also ordinary bannermen—who were expected to maintain competency in riding a horse and shooting a bow and arrow. Among the Scouts Division of the Metropolitan Banners, young boys began such training at the age of seven. Students at banner schools, such as those in the Jingzhou garrison, practiced mounted archery as part of their regular curriculum. Candidates for the civil service examinations (literary and translation) were required to demonstrate minimum competence in archery, both mounted and on foot, before they were permitted to sit for the tests.¹⁴⁹

For most banner soldiers, military training focused almost exclusively on horsemanship and archery. They practiced those skills on a regular (if rather relaxed) schedule and at different levels of organizational complexity. Among the Metropolitan Banners, for example, the Vanguards were supposed to practice archery on foot every fifth day (or six times a month) and archery from horseback twice a year, in the spring and fall. They also held joint maneuvers with other units once a year. Every third year there was a grand review at the imperial hunting park at Nanyuan, south of the capital, in which nearly twenty thousand banner soldiers participated. Similarly, in Manchuria, the troops of an entire garrison got together in the second and the eighth lunar month for month-long field exercises, and in the winter they went on collective hunting expeditions.¹⁵⁰

However, the effort to preserve the Manchu martial tradition fell victim to the same forces that undermined the concurrent effort to maintain the Manchu language. In the late nineteenth century, the Qing rulers finally overcame their scruples about modern weaponry and allowed the newly organized Peking Field Force in Beijing and similar “foreign-rifle detachments” (yangqiangdui) in various provincial garrisons to substitute the rifle and the cannon for the bow and arrow. In the rest of the banner army, banner soldiers continued down to the end of the century to practice the traditional skills of horse riding and archery, but these exercises had become formalistic and meaningless. Moreover, the emperor himself had long ceased to take his equestrian heritage seriously. Jiaqing was the last ruler to go regularly on the traditional annual or biennial month-long expedition to the imperial hunting grounds at Mulan, one hundred miles north of the Great Wall. The expeditions were terminated altogether in 1821, and the Mulan hunting grounds were thrown open to cultivation in 1863. In Manchuria, the last annual hunts in Jilin and Heilongjiang were held in 1875.¹⁵¹ Bannermen, living off the stipends paid to banner soldiers and barred from most other occupations, began instead to cultivate in their ample leisure time a variety of nonmilitaristic interests, such as raising birds, collecting cricket-feeding vessels and pigeon whistles, and singing Beijing opera. As Lao She observed sardonically of his Manchu elders, “The true meaning of life was to be found in the daily pursuit of their hobbies, in which they sought out what was exquisite, refined, and enchanting.” It was these aesthetic activities, rather than horse riding and archery, with which the Manchus came to be associated in the late Qing.¹⁵²

As the tradition of the Manchu mounted archer declined, so too, as is well known, did the efficacy of the banner army. The conquest of Xinjiang in the 1750s was the last military engagement in which banner soldiers constituted the main fighting force. They were, for example, ineffective in the 1790s against the White Lotus Rebellion (1796–1804). The Qing dynasty thereafter was forced to rely on a succession of non-banner forces: first the Army of the Green Standard, then locally raised militias, and finally regional armies. Nevertheless, the banner soldiers were not so effete or so militarily inept in the mid-nineteenth century as they are usually portrayed. In the two Opium Wars and the Taiping Rebellion, the Manchus unavoidably did much of the fighting, and they often battled bravely against heavy odds. During the First Opium War, the banner garrisons at Guangzhou, Zhapu, and Zhenjiang carried the burden of the defense on the Chinese side. At Zhenjiang, where the garrison had been reinforced by banner soldiers from Qingzhou in Shandong, the banner troops, according to Peter Ward Fay, “fought bitterly and skillfully even after the walls had been scaled and one gate blown in, and would neither surrender nor run but died where they stood, or else managed to avoid the English long enough to regain their own quarter and there make an end to themselves and their families.” In the Second Opium War (1856–60), it was once again the banner troops who led the resistance against the British and French both at Guangzhou and at Beijing. Banner soldiers from Qiqihar, for example, were part of the army of Senggelinqin (d. 1865), which in 1858 repelled the invaders at Tianjin and Dagu.¹⁵³

Meanwhile, other banner garrisons in central China were engaged in a long and even more desperate struggle against the Taiping rebels, whose heaven-sent mission was to slay the Manchu demons. In March 1853 the garrison at Nanjing fought off the Taipings for thirteen days; when it finally succumbed, the banner people paid dearly for their resistance. As described by Jen Yu-wen,

The revolutionaries, in a fury of revenge for the bloody battle, carried out a brutal massacre of all but a few hundred people lucky enough to escape into the Outer City. At one point several thousand Manchu women were surrounded and driven through one of the gates, there to be burned, stabbed, or drowned. According to the later investigations of Tseng Kuo-fan [Zeng Guofan, 1811–72], more than thirty thousand lost their lives in this great massacre of Bannermen.

(The Taipings were to be treated no less terribly when Zeng Guofan recaptured the city for the Qing in 1864.) At Hangzhou, the banner garrison staved off a bitter two-month siege in 1860, only to capitulate a year later. A majority of the banner people in the city, numbering eight to ten thousand, committed suicide rather than fall into Taiping hands. The nearby garrison at Zhapu, already victimized in the First Opium War, was also decimated.¹⁵⁴

Banner troops elsewhere participated in the struggle against the Taiping and other rebels as well, and suffered similarly. The small garrison at Cangzhou, in the capital region southeast of Beijing, lost over two hundred soldiers trying to halt the Taipings’ northern expedition in 1853. Troops from the Manchurian garrisons that were sent south also absorbed tremendous (though improbably high) casualties. In Jilin, according to its general in 1865, “over 10,000 banner troops had died in the campaigns, and of those who returned, almost half were disabled.” Likewise, in Heilongjiang, according to the gazetteer, “the number of troops sent to China proper totaled about 67,730 men; of those, only about 10 to 20 percent survived.” The banner garrisons in the northwest similarly fought against the Chinese Muslim rebels and the invading force of Yacob Beg (1820–77) in the 1860s and early 1870s. Casualties at the Xinjiang garrisons rivaled those at Nanjing and Hangzhou.¹⁵⁵ In sum, the banner soldiers’ role in both Opium Wars and the mid-century rebellions was by no means inconsequential.

Even as they tried to keep their Manchu followers from succumbing to Han ways, the rulers of the Qing dynasty were making several efforts in the opposite direction, that is, to Manchufy (or alternatively, as Mark Elliott puts it, “Manjurify”) their Han subjects. The Manchus made three cultural impositions upon the Han: men’s hairstyle, official dress, and Manchu language. The first was universally and stringently enforced; the other two were more limited in scope and effect.

The queue was the male hairstyle of the original Manchus, a variant of the way men of the northern tribes, including the Jurchen, had traditionally worn their hair; it involved shaving the front and sides of the head, letting the rest of the hair grow long, and braiding it into a plait. The Chinese style current in the Ming period was to let all the hair grow out, coil it up into a topknot on the crown of the head, and hold it in place with a cap made of horsehair. Within a year after they fought their way into China proper, the Qing demanded, on pain of death, that the men among their defeated subjects wear their hair the Manchu way. At first the queue requirement was tremendously unpopular among the Han, who protested that shaving the head was contrary to the “system of rites and music” of ancient China and that it violated the Confucian injunction against harming the body that had been bestowed by one’s parents. The Qing rulers, however, viewed the queue as a visible emblem of submission and would neither withdraw nor modify the regulation. When a former Ming official, Chen Mingxia, in 1654 voiced his disapproval of the queue ordinance and urged a return to the Ming fashion, the Shunzhi emperor had him executed for treason. (Chen was also accused of corruption.) As the Qianlong emperor reaffirmed in 1768, when confronted by a baffling outbreak in Zhejiang and Jiangsu of queue-clipping (not queue-cutting), “Wearing the queue is a fundamental institution of this dynasty.” The only ones exempt were men in mourning, young boys, Buddhist monks (who shaved off all their hair), and Taoist priests (who let their hair grow). All other Han males in Qing China were coerced into abiding by the requirement. (Among the non-Han peoples in the empire, such as the Turks in Xinjiang and the Miao in the southwest, only leaders were obliged to wear the queue, not commoners.)¹⁵⁶ As a result, notwithstanding their initial opposition, most Han came to accept the Manchu hairstyle. The “long-haired” Taiping rebels were an obvious exception; otherwise, by the end of the nineteenth century, as the anti-Manchu critics themselves admitted ruefully, the queue had become a part of the inherited tradition of China.¹⁵⁷

Official dress was the second aspect of Han culture that was Manchufied. Like the hairstyle, the clothing style of the Manchus differed greatly from that of the Chinese during the Ming. In general, the Ming style of men’s clothing, among both officials and scholars, was wide, bulky, and loose, as suited a sedentary population. The Ming official’s formal dress consisted, in part, of a full-length, loose-fitting, wide-sleeved robe, a stiff hooplike belt around the mid-section, a double-crowned hat with two “fins” protruding perpendicularly from the rear, and soft slippers with upturned toes. The scholar’s casual wear lacked the hoop belt and substituted a square cap for the finned hat, but was similarly ample and flowing (see plates 1 and 2).¹⁵⁸ The Manchu style was narrow and tight and reflected its equestrian origins: it was close-fitting on top to conserve body heat; loose-fitting at the bottom and slit in front and back as well as along the sides to facilitate riding a horse; and had long, tapered sleeves flaring into “horse-hoof cuffs” that protected the back of the rider’s hands from the cold. This full-length robe was worn over trousers that were all but concealed, and it was sometimes worn under a half- or three-quarter-length front-fastening surcoat called a “horse jacket” (magua). Qing officials dressed differently from Ming officials in yet other ways. They did not wear the hoop belt. They wore rigid-soled boots, not soft slippers. They wore hats of two seasonal types, neither with protruding fins: one was a cone-shaped, brimless summer hat made of bamboo, the other a round winter hat with an upturned brim made of animal fur. Both types of hat were topped with a glass or jewel ball, the color of which indicated the wearer’s rank. Finally, officials of the fifth rank or higher also wore a long necklace of 108 beads. Both the hat knob and the necklace were Qing innovations. As for the Qing scholar, he wore a gown that was cut like the official robe—with tapered sleeves and slit below the waist—and a round cap rather than the square cap of the Ming (see plates 3 and 4).¹⁵⁹

The Qing rulers demanded that their clothing style, like the queue, be adopted. As with Chen Mingxia a century earlier, Liu Zhenyu was executed during the Qianlong reign for urging that the costume be changed, presumably by returning to the Ming fashion.¹⁶⁰ However, the dress code was required only of the scholar-official elite and not of the entire male population. Therefore, the great majority of Han men were free to continue to dress as they had during the Ming. Nevertheless, in the course of the Qing period, they too, on their own, took to the Manchu style of dress, typified by the long gown with tapered sleeves worn under the three-quarter-length “horse jacket.” As a result, by the late Qing, not only officials and scholars, but a great many commoners as well, had been Manchufied in their attire.¹⁶¹

The third instance of the Manchufication of Han culture was the elevation of Manchu to an equal status as Chinese as one of the two primary official languages of the Qing empire. Thus, the Qing rulers required that many official documents be rendered not only in Chinese but in Manchu as well. Consequently, throughout the Qing period, as summed up by Ch’en Chieh-hsien, “Imperial edicts, memorials to the throne, official documents, and even coins and stone inscriptions all bore Manchu writing.” Furthermore, early Qing rulers had insisted that some of their Han subjects learn the Manchu language. The scope of the language requirement, however, was even more limited than that of the dress code. It applied to only a tiny elite among the officials. When on one occasion it was proposed that “all literati, whether Han Chinese or Manchu,” be ordered to study the Manchu language, the Jiaqing emperor decided otherwise.¹⁶² Only some members of the Hanlin Academy—the Han metropolitan degree holders most likely to reach the highest posts—were obliged to enroll in a three-year course on the Manchu language. This requirement was still in effect in the early nineteenth century, as one of the Hanlin academicians ordered to study Manchu was Lin Zexu (1785–1850), a metropolitan graduate of 1811.¹⁶³ By the late nineteenth century, however, Han officials seemingly were no longer required to study Manchu. Nevertheless, because of the state’s extensive efforts to popularize its use throughout the Qing period, the Manchu language inevitably left its mark on the Chinese language. A number of Manchu words thus found their way into Chinese, particularly the specialized vocabulary of governmental administration, such as zhangjing (from janggin, “secretary”), bitieshi (from bithesi, “Manchu scribe”), jiala (from jalan, “banner battalion”), and niulu (from niru, “banner company”). The Manchu imprint was also evident in the speech of Han Chinese from Beijing and Shengjing, the two cities with the largest banner population, where the local dialects include many Manchu loan-words and their morphology and syntax exhibit some features that, according to Stephen Wadley, “suggest that they were influenced” by Manchu and other Altaic languages.¹⁶⁴

Sinicization of the Manchus thus was not the only process of cultural change occurring during the Qing period; concurrent with it, on a much smaller though not insignificant scale, was the reverse process, the Manchufication of the Han. Furthermore, as of the late nineteenth century, neither culture had triumphed over the other. Each of the two peoples retained cultural elements that made them distinct despite more than two centuries of living together.

Manchus and Han differed particularly in women’s clothing. Manchu women typically wore a one-piece gown. Also, as described by A. C. Scott, they “tended to build their hair up in elaborate shapes on top of their heads, using wire frames, false hair and ornate decorations.” (The “bat-wing shapes” of this headdress, known colloquially as liangbatou, was, according to Valery Garrett, “characteristic of nomadic headwear from the Eurasian steppe.”) Finally, Manchu women did not bind their feet, and they wore platform shoes that were two or three inches high. The attire of fashionable Han women differed in all three respects. Following Ming custom, they dressed in a two-piece outfit, with a loose three-quarter-length garment worn over a pair of trousers; they combed their hair back and coiled it into a bun at the rear or two buns on the sides; and they bound their feet and wore tiny cloth shoes (see plate 5).¹⁶⁵ Han women were not required to adopt Manchu dress, and, unlike the men, they never did. Nor, by and large, did Manchu women emulate Han fashion. Thus, down to the end of the nineteenth century, Manchu and Han women retained their distinctive costumes.

Yet other cultural differences divided Manchus and Han. For example, they differed in how they styled themselves when addressing the emperor in writing. When memorializing the throne, a bannerman referred to himself, in Chinese, as nucai (your slave), whereas a Han utilized the term chen (your minister). The Qianlong emperor once directed all his officials to call themselves chen, but for some reason the directive never took effect.¹⁶⁶ Manchus and Han also differed in where they placed the position of honor, with Manchus locating it on the right side and Han on the left.¹⁶⁷ And they greeted people differently. Manchus greeted a fellow Manchu (one who was socially superior) by dropping the right knee to the ground and letting the arms hang down; this was known in Chinese as daqian, which David Hawkes, in his translation of Dream of the Red Chamber (also known as Story of the Stone [Shitou ji]), calls the “Manchu salute.” The Han greeting consisted of clasping both hands in front of the chest and bowing.¹⁶⁸

POSTWAR ATTEMPTS AT REFORM

The clearest evidence that down to the end of the nineteenth century Manchus and Han were still separate peoples, who kept apart from each other and who retained some distinctive cultural traits, comes from some of the reforms that were proposed in the wake of China’s loss in the Sino-Japanese War of 1894–95. The defeat, which led to the cession of Taiwan and then the Scramble for Concessions, had opened the floodgates of Japanese and Western imperialism. These alarming events alerted some members of the Chinese elite to the possibility, indeed likelihood, that their country would be partitioned among the imperialist powers. To prevent this from occurring, various scholars and officials began offering suggestions for reform, some of which went far beyond the limited aims of the Self-Strengthening Movement. The reformers soon found a receptive audience in the youthful Guangxu emperor, who had been freed from the constraints of his aunt Cixi’s regency since 1889. On 18 December 1897, after the German seizure of Jiaozhou Bay in Shandong, the emperor issued a rambling edict that, inter alia, called on all officials who were knowledgeable in military affairs to suggest changes.¹⁶⁹ Six months later, on 11 June 1898, he greatly broadened his appeal for reform proposals, thus beginning the Hundred Days.

The postwar reformers, led by the new metropolitan degree holder Kang Youwei, organized political clubs in the guise of “study societies,” such as the Society for the Study of Self-Strengthening (Qiangxuehui) in Beijing in late 1895, and they founded political journals, such as Current Affairs News (Shiwubao) in Shanghai in August 1896. One bannerman who was active in this unprecedented reform effort from below was the imperial clansman Shoufu (1865–1900), who was to pass the regular metropolitan examinations in 1898. In May 1897 Shoufu published an article in Current Affairs News titled “An Address to the Worthies of the Eight Banners Concerning Current Trends,” in which he, like many others at the time, warned of the peril of partition. He focused generally on the threat that partition posed for all Chinese, but at one point he spoke to the self-interest of the banner people. He prophesied, “If the present dynasty is invigorated, then the Eight Banners will continue as the aristocracy.” The unstated alternative, of course, was that if China were partitioned and the Qing dynasty extinguished, then the banner people would lose their privileged status.¹⁷⁰ In Beijing in the same year, Shoufu, together with Kang Youwei, founded the Study Society to Know Shame (Zhichi Xuehui); its purpose was to heighten popular consciousness about the “shame” of China’s defeat by Japan.¹⁷¹

Some of the changes these popular agitators called for verged on what in the past would have been regarded as treasonous. One proposed reform involved the Qing’s sumptuary laws. Previous scholars and officials who had recommended a change in the dynastic dress code had run the risk of severe punishment, including, in the case of Liu Zhenyu in the Qianlong era, even execution. Nevertheless, in the summer of 1896 a certain Jiang Shuzi published a short article in Young J. Allen’s Shanghai reformist monthly A Review of the Times (Wanguo gongbao) that dared to suggest abandoning the Manchu-style official dress. Jiang asserted that it was only after Peter the Great had altered his country’s clothing style to accord with the rest of Europe that Russia became strong, and that the Meiji emperor by following Peter’s lead had achieved similar results in Japan. Such changes, including a change in the calendar (presumably from lunar to solar), would symbolize both an imperial determination to make drastic changes and a willingness to involve the entire population in that reform effort. They would help bring about a “new citizenry.” Jiang then asked rhetorically, “If China wishes to reform, should it not begin with the calendar and the costume?” He admitted, in conclusion, that a commoner such as himself ought not intrude in “matters of institutions and rituals,” which were the sole prerogative of the court; nevertheless, he proceeded to recommend that the Qing learn from the experiences of Petrine Russia and Meiji Japan. While Jiang did not specify how the clothing style should be changed, undoubtedly it would have meant abandoning the Manchu gown and, as Japan had done, adopting European garb.¹⁷²

Kang Youwei went even further and suggested abolishing the dynasty’s queue requirement. In a lengthy account of the Meiji reforms titled A Study of the Governmental Reforms in Japan (Riben bianzheng kao), which he offered to the court in 1898, Kang too alluded to the desirability of the Qing emperor’s emulating the Japanese ruler, but unlike Jiang Shuzi, he referred not only to court dress but also to hairstyle. He wrote that in Japan “the official dress had all been changed to the Western style and the Japanese emperor had personally cut short his hair.” He implied that the Guangxu emperor should follow suit.¹⁷³ (Kang Youwei supposedly submitted three other, much more detailed memorials on Manchu-Han relations in 1898, but they, unlike A Study of the Governmental Reforms in Japan, are thought to be later fabrications.)¹⁷⁴

The most comprehensive reform proposal concerning Manchu-Han relations came in a memorial to the throne presented by Zhang Yuanji (1867–1959) on 5 September 1898. A metropolitan degree holder (class of 1892) from Zhejiang and founder of a foreign-language school (the Tongyi Xuetang) in Beijing in 1897, Zhang Yuanji had only recently been appointed to the Zongli Yamen. He proposed five sets of reforms, the second of which was “to dissolve Manchu-Han differences” (rong Man-Han zhi jian) so that China could unite in common opposition to the aggression of the foreign powers. Zhang, in this section of his memorial, recommended six reforms.¹⁷⁵ First, he urged an end to the administrative separation of Manchus and Han by reclassifying the “various Manchu and Mongol banner people in China proper” as civilians and transferring them to the jurisdiction of local officials. Only the imperial clan were to be exempt because they were too exalted to be treated as ordinary civilians; they would continue to be under the Imperial Clan Court. Second, he called for an end to the social and occupational segregation of Manchus and Han. There was no better way to integrate the two populations than for the emperor to issue a command that “banner people and Han [qi-Han] intermarry.” He also urged the emperor to “lift the prohibition on venturing forth to trade so that the banner people may make their own living, and to permit those who wish to live elsewhere [than the garrison cities] to petition the local authorities to be reclassified as civilians.”

Third, Zhang proposed an end to dyarchy. He conceded that it might have been sensible at the beginning of the dynasty to maintain separate slots for Manchu and Han officials. But the practice had become a source of “mutual discord between banner people and Han.” Furthermore, the system no longer benefited the Manchus (Manren) themselves. Because they generally began their careers earlier, they were not as well educated as the Han (Hanren), and they were often assigned to posts in desolate places beyond the Great Wall. He recommended that henceforth only one qualified person, without regard to whether he was Manchu or Han, be appointed to each position. Fourth, Zhang seemed to call for a phasing out of the banner soldiers’ stipends. Current soldiers would continue to be paid in full, but when vacancies in the banner units occurred, they should be allowed to lapse. Fifth, he urged that industrial schools be set up in Beijing and in the provincial garrisons to give vocational training to unemployed banner people. Last, he recommended that, pending the outcome of the reforms within China proper, the current institutions be left intact in Inner and Outer Mongolia, Xinjiang, and Qinghai (and presumably Manchuria as well).

Zhang Yuanji’s was one of only two known memorials submitted during the Hundred Days that focused on Manchu-Han relations. The other came from Yuan Chang (1846–1900), the Jiangsu provincial treasurer at Nanjing. The text of Yuan’s memorial has not been published, but its existence is known because it elicited an important edict from the emperor on 14 September 1898. Yuan apparently had brought up the vexing problem of the Eight Banners’ livelihood. The emperor, in response, took note of the increase in the banner population and attributed their plight to their being “prohibited from engaging in trade in the provinces.” Officials in previous reigns—notably Songyun (1752–1835) and Fujun (1749–1834) in the Jiaqing and Daoguang (1820–50) periods, and Shen Guifen in the Tongzhi period—had addressed the problem. Now that China was inaugurating a comprehensive renovation, it was time to relax the prohibition in order that the banner people might learn the vocational skills of the four classes (scholars, farmers, artisans, and merchants) so that they could provide for themselves. The emperor, furthermore, called on the Board of Revenue to develop specific programs to resettle the banner population based on the experiences from the Jiaqing and Daoguang eras.¹⁷⁶

The emperor’s edict responding to Yuan Chang’s memorial greatly alarmed the banner people, especially those in Beijing. It seemed to augur an end to the banner stipends and perhaps even—if Zhang Yuanji’s more radical ideas were enacted—an end to the banner system, to dyarchy and ethnic slots, and to the separation of Manchus and Han. According to Lao She, in Beneath the Red Banner,

Several months before I was born [in 1899], my maternal uncle, [my elder sister’s husband] Duofu, and his father were all in a state of great agitation. They fervently opposed the Reforms of 1898. . . . They had heard that when the changes came, bannermen were going to have to go out and earn their own keep. No longer would the Imperial government provide them with allowances.

Similarly, in Lao She’s historical pageant Teahouse (Chaguan), the first act of which is set in Beijing in 1898 right after the Hundred Days, one bannerman observes to another that the reforms would have meant “cutting off our stipends and making us work for a living.”¹⁷⁷

Lao She’s relatives and friends need not have fretted. One week after this edict was issued, the emperor’s reform effort came to an abrupt end. On 21 September, Empress Dowager Cixi, fearful of where all the proposed changes might lead, came out of retirement, put the emperor under house arrest, and resumed her regency. The emperor never acted on Zhang Yuanji’s memorial to reduce Manchu-Han differences. Nor had he, as Jiang Shuzi and Kang Youwei had recommended, shed his queue and abandoned the Manchu-style court dress.

CONCLUSION

The Eight Banner system was, originally, a heterogeneous, multiethnic organization. The founding members were the Jurchen whom Nurhaci united and organized into banners and whom Hong Taiji later renamed Manzhou. Also known as Old Manchus, they were the largest and most prestigious group within the system, but they were not the only group. The Eight Banner Manchus themselves included the New Manchus as well; these were generally either “Mongolized” Tungusic peoples (such as the Xibe) or “Tungusized” Mongols (such as the Daur), but some were Korean, Tibetan, and even Russian. Furthermore, in addition to the Manchu banners, the overall system had two other ethnic components—the Mongol banners and the Hanjun.

Despite these as well as other internal divisions, the Eight Banners by and large presented a united front to the outside world. With some exceptions, such as the addition of the Tibetan company to the Manchu banners in the 1770s, membership in the system was based on birth and was closed to new recruits. Thus, regardless of their ethnic origins, all members of the Eight Banner system shared a common status and a common identity that set them hereditarily apart from the rest of China’s population. Thus, both individually and collectively, they called themselves and were called by others “banner people.” Since their primary function was to hold themselves in readiness to protect and defend the Qing emperor and his court, the banner people, in this regard, constituted a hereditary military caste. They were differentiated from the non-banner people, who were generally known as “civilians.”

Finally, by the end of the nineteenth century, the banner people also came to be called Manchus. These Manchus were not the Manzhou of Hong Taiji or even the Eight Banner Manchus, who were only part of the banner personnel. Rather, they were thought of as constituting the total membership of the banner system and were most often referred to in Chinese as Manren. In short, what was originally a multiethnic military organization (the Eight Banners) was also a hereditary caste (the banner people) that was increasingly being viewed (at least in the eyes of non-banner people) as an ethnic group (the Manchus).¹⁷⁸ By a similar process, the non-banner “civilians” came to be equated with the correlative of the Manchus, the Han. In Zhang Yuanji’s memorial of 1898, for example, the terms “Manchu” and “banner people” were used interchangeably, both in common opposition to “Han.”

The Guangxu emperor’s decree of 14 September 1898 proves conclusively that Mary Wright was very wrong when she asserted that “most of the last restrictions separating the Manchus from the [Han] Chinese were removed in 1865.” If the Tongzhi emperor’s edict permitting banner people to leave the Eight Banner system, be reclassified as civilians, and make their own living had indeed been implemented, it would have been unnecessary for his successor to issue a nearly identical decree thirty years later. Also, if most “restrictions” separating Manchus and Han had been removed in 1865, there would have been no need for Zhang Yuanji’s six proposals to “dissolve Manchu-Han differences.” The Tongzhi edict, however, had never been carried out. As a result, down to the end of the nineteenth century, Manchus and Han remained in many ways separate populations.

The continuing differences between Manchus and Han serve to confirm the essential validity of much of the revolutionaries’ seven-point indictment against the Manchus in the early post-Boxer era. Although they had absorbed much of Han culture, the Manchus were, as charged, an alien people who in some respects were still manifestly different from the Han; their men, for example, did not use Han-style surnames, and their women dressed differently from Han and did not bind their feet. The Manchus had, as charged, barbarized (i.e., Manchufied) the Han when they successfully imposed their hairstyle upon Han men and their costume upon Han officials. The Manchus were, as charged, a privileged minority separate from and superior to the Han; they were administratively and residentially segregated, they were barred from marrying Han, and they were stipendiaries of the Qing state who were prohibited from any employment other than soldiering, serving as officials, and, in some regions, farming. The Manchus did, as charged, constitute a foreign occupying force; they were a hereditary military caste and were garrisoned within their own walled citadels that were strategically distributed throughout the empire. The Manchus did, as charged, receive preferential treatment that was denied to the Han; they were dealt with more leniently under the law, and they had more opportunities to enter and advance in government service. In short, the revolutionaries’ indictment against the Manchus qua Manchus was not a mere propaganda ploy devoid of substance; rather, it did have a basis in contemporary social reality.

It was, however, not the revolutionaries but the postwar reformers such as Jiang Shuzi, Kang Youwei, and Zhang Yuanji who first raised what can be called the “Manchu issue” and brought it out into the open for public discussion. Previously, Manchu-Han relations had been a taboo subject. Chen Mingxia, for example, had been executed in 1654 for criticizing the queue; so had Liu Zhenyu in the Qianlong reign for proposing a change in the Manchu-style official costume. But in the late 1890s, not only did Jiang Shuzi and Kang Youwei suggest abolishing the queue and adopting Western dress, but Zhang Yuanji urged ending dyarchy and encouraging intermarriage between Manchus and Han. In making such proposals, they knew that they were treading on thin ice, as Jiang Shuzi acknowledged. They spoke up cautiously. Thus, Jiang’s and Kang’s separate recommendations to get rid of the queue, though unambiguous, were indirect: the Guangxu emperor should do as the Meiji emperor had done. Furthermore, the postwar reformers held back from the obvious. Thus, no one dared call for the outright disbandment of the banner army, even though it had become patently useless as a military force.¹⁷⁹ The closest that anyone came to suggesting an end to the banner system was when Zhang Yuanji proposed that as vacancies in the banner units occurred they be left unfilled. Nevertheless, despite the cautiousness of their approach and the moderation of their proposals, the postwar reformers presented the most serious challenge to Manchu-Han relations since the anti-Manchu Taiping Rebellion.