4 / Strategies of Resistance
Integration by Violence
FIRST BLOOD IN THE 1780S
After the morning prayers, as I passed Ma Laichi’s street, oh God! A crowd gathered against me. Taking long poles and short sticks and whips, they beat me, and the women stood in the doors and threw garbage. With God’s help, not knowing how I found the courage and strength, I beat them one by one, broke their weapons, and defeated them. Thus the vengefulness grew deeper, for Ma Laichi’s fourth son took a crowd to the magistrate and said we had set up a new teaching to deceive the people. The court decided against us, and I was dragged in and beaten forty strokes, while Ma Mingxin received three. As they beat him, the bludgeon split in two! . . . The next day after dawn prayers, I again went by Ma Laichi’s street and cursed the people from one end to the other. The revengers said, “How can he be so bold?” and again went to court. The magistrate decided, “All should return to their native places.”1
Khafīya vs. Jahrīya at Xunhua
After Ma Mingxin’s return from “the west” and his successful campaign of conversion, Khafīya and Jahrīya adherents in “Salar country” (Huangzhong)—and elsewhere in Gansu—became rivals. The Muslims themselves and the Qing state called their altercations “religious disputes,” and official documents record religious justifications for violence to bear out this judgment—vocalization of the dhikr, wearing of shoes at funerals, the length and cut of beards or mustaches, and other ritual minutiae, as well as the more generic but more serious charges of heterodoxy and deception. But these “religious” conflicts clearly had secular causes as well. A wide variety of hot and divisive issues beyond the vocal dhikr compelled the two orders first to confront one another, then to seek redress from the local authorities: rivalry over the profits of religious institutions, establishment of control over converts to Islam, personal conflicts between lineages and individual leaders, and, as the Jahrīya history recorded, the boldness of young men bearding their “enemies” in their lairs and brawling in the streets.2
The confrontation between Ma Mingxin’s and Ma Laichi’s followers might well have been an ordinary event, just another gang war or feud in the complex shuffle of frontier history, had it not taken place among the Salars. The officials of Gansu feared Salar violence, as did many non-Salars who lived within range of Huangzhong, for their reputation as raiders had spread from the Gansu corridor to Shaanxi. This labeling of the Salars as inherently, genetically violent and ferocious had been done so often, in so many official documents and local discussions, that it came to be known as true (see plate 13).
The Salars’ ancestors had migrated from Central Asia to China, probably during the fourteenth century.3 The Chinese accounts of their journey, undoubtedly influenced by centuries of antagonistic hindsight, maintain that the Salars were cast out of their homeland for their violent behavior.4 Isolated by the rough terrain from state power, the Salars earned a reputation for violence that belied their numbers, fewer than fifty thousand in the early twentieth century.5 With a comparative eye on the Montagnards of Vietnam, the hillbillies of Appalachia, the Karen of Burma, and many others, we may surmise that we have received the flatlanders’ view of the Salars, one thoroughly prejudiced against them. The Salars engaged in dangerous, low-status occupations such as long-distance trade by raft and caravan. Many of them settled far from the Xunhua region, and many joined the military. In sedentary, civil China, such professions and mobility sufficed to earn a people the reputation of being congenital bandits. Some Salars did prey on merchants, but their mobile, mercantile professions would have marked them as vagrants, if not bandits, whatever they might have done. We cannot trust the Chinese imagery, which would automatically type the Salars as uncivilized.6
As elsewhere among the Chinese Muslims, the local mosque constituted the core of the Salar communities, which were known as gong. The eight gong on the south side of the Yellow River and the five on the north all supported numerous religious institutions despite relatively small population and poverty. Xunhua County alone counted sixty-two mosques of various sizes in the early twentieth century.7 Salars intermarried with other groups, including Chinese and Tibetans, and adapted in a variety of ways to their frontier environment. They wore the queue under Manchu rule, for example, but many of them successfully resisted footbinding. Physically they appeared to be a combination of all the northwestern peoples, as Joseph Trippner pointed out in the passage translated on page 18. To maintain their connections with their Islamic roots, and to tap wider commercial markets, the Salars allegedly established and maintained connections with Muslim Central Asia that were stronger than those of their Sinophone Muslim neighbors. Their religious tradition gave them motivation, their Turkic language gave them the necessary means of communication, and their far-ranging trade gave them greater opportunity.8
The Xunhua region, which most Salars called home, was a backwater from a Chinese point of view, a frontier town with few eminent literati in its gazetteer and little but conflict to mark its history. Lying amid steep hills on the south bank of the Yellow River, the town had a subprefectural yamen (ting) and a small garrison. These represented the only state authority for miles around, authority over a mixed hill-country population of Tibetans, Salars and other Muslims, and Monguors, but very few Chinese-speakers.9 What seemed to any sensible Chinese to be the back of beyond constituted a cultural homeland for the Salars, who had by the mid-Qing lived there longer than Europeans have lived in North America. When conflict arose among its diverse peoples, however, Xunhua became an unstable social, military, and cultural environment.
In harsh natural conditions, with potential enemies on all sides, none of the frontier peoples could have survived without a martial tradition, without both weapons and the skill to use them, and survive the Salars did.10 We can hardly blame them, but we can also understand the development of stereotypes and fears among non-Salars who lived nearby. An anxious twentieth-century missionary summarized the local stereotype of the Salars, one he shared wholeheartedly:
The geographic situation of [Salar] territory, adjoining as it does upon Tibet, enables them to engage in constant warfare with that people, and thus to nurture within them the fierce spirit of their forefathers. Word that the Salars are out upon the warpath will throw the largest Chinese trading community into a panic. It is commonly said that the Salars can only be governed by a Salar, and even for him the task is not always easy.11
We need not accept this stereotype at face value to understand its evolution or its power. The non-Muslim Chinese and other local peoples formed rigid images of some of their neighbors as brutal, warlike, uncivilized and predatory—in short, barbarous beyond rescue. The appearance of feuds, apparently based on ideological differences, among the Salars created a volatile atmosphere all over the province, for their very name evoked fearful images, and their armed affrays evoked a violent response from the state.
The Sufi combatants in Xunhua or Hezhou were certainly not planning the “great enterprise” of dynastic replacement or a political secession from Gansu, much less a jihād to convert pieces of the Qing empire into the territory of Islamic government. Rather, they sought local revenge on local enemies. Lacking any effective central authority within their own communities—a qādī to whom one side had recourse would have been rejected as heterodox by the other—both Old and New Teaching, as the Khafīya and Jahrīya were called in the reports of Qing officials and the diaries of local gentry, took their cases to the secular officials. They also took matters into their own hands, with young men beating each other in the streets, which called their conflicts clearly to the attention of the authorities.12
From Feud to “Rebellion”
The Khafīya as a religious order had not caused much violent feuding in Gansu until the establishment of the Jahrīya, Ma Mingxin’s revivalist followers, who came to be seen as the destroyers of a once-stable social order. As demonstrated above, the Sufi paths were not merely religious collectives devoted to meditation but rather loci of loyalty for many types of social and political activity. Their internal cohesion and devotion to a single leader made it possible for them to mobilize passion and armed force against their enemies, in this case other Sufis. Twenty years of escalation culminated in 1781, when law cases and street violence between the competing orders brought the Qing authorities to Xunhua to investigate.13 Learning more of the Jahrīya leadership and its apparently subversive activities, they arrested Ma Mingxin, who was not anywhere near Xunhua at the time, as the chief troublemaker, imprisoning him in Lanzhou. Xinzhu, adjutant general at Hezhou, was sent to Xunhua with Yang Shiji, prefect of Lanzhou, to round up Ma’s cohort and end their threat to social harmony.
Some of the Salar Jahrīya Muslims, under the military leadership of Su Forty-three, met the two officials and their small company of troops at Baizhuangzi. Obviously prepared for trouble, the Muslims had concealed their weapons and presented themselves as a welcoming party of Old Teaching adherents. When Xinzhu told them he intended to eliminate the New Teaching, Su and his men overpowered the military escort and killed both officials.14 With this violent act, the internecine strife among Muslims became secondary, and “rebellion,” as defined in Qing law, began. Not only had feuding among Muslims created intolerable social disorder, but even worse, officials—the local representatives of imperial majesty—had been murdered. The state’s theory, and its own claim to legitimacy, demanded that Qing troops intervene quickly and violently to punish the miscreants who appeared to challenge central power.
The Muslims responded with violence of their own, and more of them became “rebels.” This legal charge, combined with the widespread perception of the Jahrīya as heterodox, placed these particular Muslims in direct confrontation with the legitimacy of the Qing state, though the Salars had not initially intended to question that legitimacy. Their actions up to the attack on the government officials, and much of their strategy thereafter, aimed primarily at other Muslims, at the Khafīya, at their local enemies. In their own perceptions, to the extent that we can deduce them from their actions, by killing Yang Shiji and Xinzhu they were reacting to an outside threat—military intervention in their internal affairs and the threat of massacre against their tarīqa. In Qing eyes, however, the Salars under Su took a critical step when they killed officials—from being potentially subversive or heterodox Muslims to being rebels—and the echoes of that act lasted over a century.
To illustrate the power of folk memory in perpetuating fear and violence, and the extent to which the Qing state could engage to keep the peace on its frontiers, let us consider “The Ballad of Su Forty-three,” collected by the ethnographer Wang Shumin, who accompanied Gu Jiegang around the northwest in 1937. The ballad evokes local attitudes toward the 1781 violence from a distance of over a century:
Hush up, hush up, listen to me,
From Xitou here comes Su Forty-three.
Su ahong has Salars in tow,
Ready to fight all the way to Lanzhou.
Su ahong, he hasn’t much brain,
So he follows a teacher named Ma Mingxin.
Ma Mingxin, he’s a really smart guy,
Has thousands of Salars ready to die.
Thirty-six hundred are ready for war,
But Su can’t move without something more.
When cherry-tree leaves roll up like beans,
Su found Hann Number Two, it seems.15
Hann Number Two was smart as hell,
Going to help Su beat old Qianlong well.
Hann Two’s wife was a very bright dame,
And she wanted her man to give up that game:
Husband, you hear!
And Su hear, too!
You bug the Emperor,
And he’ll get you!
You bug the man,
That’s a heavy crime!
The moment he’s mad,
That’s the end of your time!
The emperor’s men are as hot as the sun,
Your Salars are a cloud on the run.
As soon as the sun shines, the cloud is done.
How many Salars are left there? None!
As soon as the sun shines, the cloud is dead.
Can’t you get that into your head?!16
Hann’s wife appears to have been more aware of the consequences of attacking government forces than were her man or Su Forty-three, at least in this ballad.
Despite the threat of Qing retaliation, Su intended to punish the Khafīya for bringing down official wrath on the Jahrīya and to rescue Ma Mingxin from unjust imprisonment in Lanzhou. From Xunhua, he led his two thousand Muslim troops rapidly toward Hezhou, where they invested the town, killing a number of Old Teaching adherents, Qing soldiers, and government officials. Not allowing the provincial military leaders enough time to organize effective resistance, the Jahrīya Muslims secretly crossed the Tao River on hide rafts (Ch. pifazi) to besiege Lanzhou.17 The authorities inside the city, reacting in haste and fear, brought Ma Mingxin onto the Lanzhou wall to show him to the besiegers. The Jahrīya adherents showed Ma immediate devotion and respect, despite his chains, so the Lanzhou officials recognized in him a subversive, a destroyer of the dynasty’s sacred order. They took him down from the wall and beheaded him immediately. In their perception of the danger they faced, this action could be justified despite Ma Mingxin’s lack of direct involvement with Su Forty-three’s attack on Xinzhu and his men (see plate 14).
The “New Teaching bandits,” as Qing officials had begun to call them, attacked the Lanzhou city walls but failed to make any headway. They lacked even rudimentary siege equipment or experience, and all of their military successes had depended on rapid movement rather than careful planning or positional battles. Determined on revenge against the evil officials who had murdered their shaikh, they barricaded themselves, perhaps one or two thousand strong, on the mountains south of the city. The Qing court, thoroughly alarmed at this sudden assault on a provincial capital, sent Imperial Commissioner Agui from the capital to “pacify” them. Unable to reduce the hillside stockades with his regular troops and burdened by the military incompetence of his colleague, Heshen, whom he sent back to Beijing, Agui called in the armed militia of the southern Gansu Tibetans, the Alashan Mongols, and the local Chinese garrisons, all nearby and endowed with a reputation for valor. Agui’s multicultural army cut off the Muslims’ water supply and eliminated them after a three-month siege, killing Su with all his Jahrīya followers in the final battle.18 This success against the Jahrīya continued a longstanding Qing policy of using frontier people to control other frontier people.
This violence in 1781 provided motivation for further armed affrays, stimulating hatred against cultural or religious Others and desire for revenge on all sides. It also vividly illustrated real contradictions within the Gansu Muslim world. This war did not simply pit Muslims against the Qing state, against the Manchus or the non-Muslim Chinese, but also Muslims against other Muslims. Some Khafīya Muslims fought on the side of the government against Su Forty-three. Evidence of similar divisions among Muslims may be found in the annals of every subsequent conflict, an important advantage for the officials enforcing social order and an important caution for us to be very wary of the words Chinese scholars use to describe such wars: “nationality struggle” (minzu douzheng), “nationality righteous uprising” (minzu qiyi), and so forth. By essentializing peoples as minzu categories (“the Salar, Dongxiang, and Hui minzu battled the feudal Qing oppressors,” etc.), such descriptions distort more than they explain, depriving the events of their contextual richness.
The Jahrīya’s Revenge in 1784
After Agui’s victory near Lanzhou and the elimination of Su Forty-three, Qing officials exerted themselves in locating and punishing New Teaching Muslims. Mistakes were made, including the transportation of Ma Laichi’s third-generation successor, Ma Wuyi, and other Old Teaching adherents to the southwest as “New Teaching rebels.” The “pacification” process made more enemies for the Qing, as local officials tried to impress the throne and their superiors with their assiduous prosecution of rebels and their allies. It should be no surprise that some Jahrīya Muslims who escaped the dragnet immediately began to plan revenge for the death of their shaikh (brutally murdered in Lanzhou), their comrades slain in battle, and their fellow Jahrīya adherents persecuted in the pacification process.
Under the leadership of Tian Wu, a Jahrīya ahong but not a Salar, they stockpiled arms, built forts, prepared banners, and gathered substantial food supplies to resist siege.19 Tian’s preparations centered in the highlands on the north side of the Wei River watershed and at Guyuan, in the barren valleys of northeastern Gansu, both far from Xunhua. Jahrīya Sufism, anti-Qing sentiment (shared by some non-Jahrīya Muslims), and the desire to avenge Ma Mingxin combined to evoke a wide response in these particular Muslim areas, but not in others. Tian Wu’s religious vocation and base area indicate that the Jahrīya organization extended well beyond Salar country by 1781. The Sinophone Muslim communities of eastern Gansu, more culturally Chinese than were the Salars, had also been deeply affected by the coming of Sufism and the fear that Qing reprisals against some Muslims would spread to all.
In 1784, three years after Ma Mingxin’s death, Tian Wu and Zhang Wenqing, possibly a relative of Ma Mingxin’s wife,20 called for revenge and attacked local garrisons in eastern Gansu, geographically and culturally distant from the Salar country of the earlier uprising.21 Tian was killed early in the fighting, but this upheaval grew even larger than Su Forty-three’s and took the Qing several months and much expense to quell. Despite pious declarations that only the rebellious miscreants themselves would be punished, Li Shiyao nonetheless executed over a thousand women and children among the eastern Gansu Jahrīya adherents, earning yet another measure of long-lasting Muslim enmity.22 Even the local chroniclers—staunch supporters of the state and the Qing-endorsed superiority of Chinese culture—noted that this massacre of the innocent stimulated the fighting spirit of the remaining rebels.
In the aftermath of this second victory over local Muslims, some Qing commanders realized that this outbreak of violence had been due, at least in part, to the ferocity of anti-Jahrīya violence three years earlier. Fukang’an noted that neither he nor the other Qing generals could actually exterminate the New Teaching, as they had been ordered to do, and that some Old Teaching Muslims, hearing rumors of anti-Muslim massacres, had indeed joined the rebels. To attempt a wide-ranging slaughter of Jahrīya adherents, or forced reconversion to the Old Teaching, would permanently alienate all the Muslims of Gansu, among whom were both good and evil people. He recommended a more lenient pacification than had been practiced in 1781.23
Despite Fukang’an’s plea, the government again proscribed the New Teaching, by which it still meant the Jahrīya, and instituted heavy penalties against any Muslim participating in religious disputation. Not limiting pacification to the Jahrīya alone, the official pronouncements forbade all Muslims to build new mosques, convert non-Muslims to Islam, or adopt non-Muslim children. Ahong could no longer preach outside their own locales, a prohibition aimed at peripatetic Sufi missionaries.24 Those crimes and their penalties, of course, applied equally to non-Jahrīya Muslims, but Khafīya adherents nonetheless continued to serve the Qing in many local capacities, and they obtained some of the lands confiscated from the families of the New Teaching “criminals.” The Qing tried to suppress not Sufism itself, which officials could not comprehend, but the New Teaching, which had been identified as the instigator of violence.
Contrary to their intentions, Qing officials encouraged the Jahrīya by linking it so directly to subversion. Muslims dissatisfied with Qing rule gravitated to the militant suborder as to an underground movement. Its propagation continued, and its anti-Qing character, though no more inherent in its rituals than in those of the Khafīya, was reinforced by its proscription. The ahong who felt themselves responsible for the Jahrīya’s cohesion and continued activity began to write covert collections of miracle stories in Arabic and Persian, describing and praising the karāmāt of Ma Mingxin and his successors. One such book by a Gansu ahong known as Guanli Ye, recently translated into Chinese by two young Jahrīya ahong and a well-known Muslim novelist, was used to transmit not only the suborder’s traditions but also its silsila, which had to be kept entirely secret to avoid Qing persecution.25 Despite the ban on their organization, Ma Mingxin’s initiates found converts from Turkestan to Manchuria, and as far south as Yunnan, where many Gansu adherents were exiled (see plate 15).26 This contrasts sharply with the other Gansu Sufi orders and suborders, which restricted themselves almost entirely to local activity and, as institutions, remained respectable in Qing eyes. The Jahrīya’s uniqueness as a transregional Muslim organization provided even greater evidence to Qing authorities of its menace to social order. Not until the founding of “national” Muslim institutions by Beijing, Nanjing, and other eastern Muslim “new intellectuals” in the twentieth century did Sino-Muslims have any noncommercial structures with so wide a range of activity.
These eighteenth-century wars should not be described as “ethnic conflict” or “Muslim rebellions” but rather as local feuds over both religious and secular issues, inflamed and polarized by heavy-handed state violence wielded by local officials terrified of the Salars’ reputation, men who worried about their careers and their dossiers in the capital more than the lives of the subjects over whom they ruled. Since they would be blamed for disorder in the territory under their temporary control, the spread of the outlawed New Teaching beyond the Xunhua district caused them considerable anxiety, but neither they nor their non-Muslim subjects could distinguish clearly between “good” Muslims and “bad” ones except by membership in a “teaching.” Some Muslims also had a stake in labeling some of their coreligionists as “the wrong kind of Muslims,” and they regularly did so in court after Ma Mingxin’s return from his pilgrimage in 1761. We can see how the complex of national policy decisions regarding the New Teaching, provincial maladministration, local religious and political rivalries, military officials overzealous in their obedience to unenforcable orders, and currents from the Muslim west combined to begin a sanguinary history in Gansu, one that was to last more than a century and a half.
The Qianlong emperor himself was baffled by the northwestern rebellions. In a poignant letter to Fukang’an, he wrote:
In this instance of the Muslim rebels under Tian Wu—how could they manage, without cause or reason, to collect a crowd, set a date, and rebel? Why would Muslims from far and near join up and follow them like sheep? I have pondered this over and over, and I examine myself to find the answer: In the decades since I ascended the throne, I have acted with great caution, not daring to allow myself the slightest arrogance or pretense. I have constantly attended to the people’s sufferings. . . . As for Gansu, I have given exceptional relief—for the past many years I have not heard of floods or drought, famine or shortages, and never of the poor losing their homes, nor of bandits stirring up trouble, nor of local officials extorting or harshly accusing, thereby causing thieves to run amok and disturb the peace. Or did news of Li Shiyao’s investigations of Muslims leak out, so rebels could start rumors flying of [a government campaign to] “exterminate the Muslims” as an excuse to incite riots? I have thought of all these things, but none seems to be the true reason. In the end, why did they rebel? We must get to the bottom of this!27
EVERY SIXTY YEARS A BIG REBELLION
These differences of religious faith led to constant conflicts between Chinese and Moslems. (Chu Wen-djang)
The Muslim revolt in Shensi and Kansu was, however, not just a religious movement. Like the revolt in Yunnan, the Tungan Rebellion was a large-scale community conflict—the coalescence of the persecuted mosque-centered communities for the purpose of survival. (K. C. Liu and Richard J. Smith)
The Shaanxi Muslim uprisings were the inevitable results of the Qing’s reactionary oppression, religious prejudice, and harsh economic exactions, all policies of feudal oppression; they were the explosion of Shaanxi’s social contradictions. (Chen Chongkai)
The Shaanxi-Gansu Sino-Muslim uprising was a struggle of the people of the Hui minzu to resist Qing reactionary control, which included opposition to the systems of feudal control and class oppression and exploitation; most important, it was a struggle to resist the Qing rulers’ minzu prejudice and minzu oppression. (Wu Wanshan)
This new Islam [the xinjiao], which started by disseminating Islamic puritanism among Muslims, partly through the use of jihād, turned into an extremist Mahdi movement which attempted to use jihād (mainly against non-Muslims) to bring the millennium, and ended up as a multifarious range of sects and sub-sects which, in effect, set themselves apart from Islam and turned to ritual introspection. (Raphael Israeli)28
MAP 3. Area of the mid-nineteenth-century Muslim rebellions. (Map by Philip M. Mobley)
Historiography of the Great Rebellion
The contemporary Chinese historical literature on the Muslims of the northwest pays close attention to the violence of the 1860s and 1870s, using two evocative and revealing vocabularies, one in the People’s Republic, another in Taiwan. Self-consciously reversing the Qing officials’ language, The Righteous Uprising of the Shaanxi Hui People, 1862–1875, recently published in Xi’an, employs the unassailably virtuous term qiyi (righteous uprising) to define the moral quality of the Muslims’ military action.29 In its analysis of conflict between local Muslims and non-Muslims, the latter bear the heavy, judgmental appelation dizhu tuanlian, “landlords’ militia.” The anti-Qing armies of the mid-nineteenth century all receive approbation as antifeudal forces, but the Muslim rebels of the northwest (and the southwest as well) carry the additional positive cachet of being minzu heroes, fighting for the independence and unity of all of the diverse minzu of China against Manchu hegemony. In this way of thinking, the Manchus are defined as evil not because they are Manchus—in the dominant minzu paradigm, that minority status should carry a positive nuance—but because they are feudal, oppressive, or incompetent.
In Taiwan, however, scholars employ a vocabulary entirely lacking class and minzu terminology, as in the title of a recent monograph by the late Gao Wenyuan, The Anti-Qing Movement of the Northwestern Hui People in the Late Qing.30 His list of “immediate causes” of the violence includes a wide variety of political, economic, and social shortcomings of Qing government, all covered by the PRC literature as well, but not antifeudal class struggle or the depredations of “landlord” militias inspired by minzu hatred. In Gao’s view Sino-Muslims acted against the Qing as good Chinese, for morally sound reasons—the state’s lack of virtue is clearly established—and the local anti-Muslim forces receive a negative judgment in terms reminiscent of Confucian historiography.
None of this is new, of course, in Chinese historical studies, for we have become accustomed (perhaps deadened) over the course of more than four decades to this contradictory cacophony of dueling paradigms presented in the secondary literature. In this particular case, however, neither set of historians highlights what might seem to Euro-Americans to be a crucial fact: the rebels were Muslims, and Islam may have played a role in motivating them. After all, we have been taught by years of Orientalist discourse to expect violence from Muslims, and all of the Chinese judgments of Islam and Muslims recounted in the previous chapters have included a proclivity for antisocial behavior. Raphael Israeli’s book on the subject is subtitled “A Study in Cultural Confrontation,” and in it he claims almost exclusively religious rather than contextual causation for Ma Hualong’s uprising.31 What role did religion actually play in this context? Are the historians of the People’s Republic correct in seeing Islam as nothing more than one factor unifying the Hui minzu with its brother Muslim minzu against the feudal Qing? Are the Taiwan scholars justified in presenting nineteenth-century Muslim opposition to the Qing as a problem of imperial political virtue? Is Israeli correct to see Islam and the Confucian ideology of the Chinese order as inevitably and permanently incompatible and Islam itself as the primary cause of violence? We will pursue these issues here as we examine the most widespread violence in Sino-Muslim history—twelve years of bloodletting that devastated several provinces, left millions dead and homeless, and advanced the process of frontier incorporation that Gu Jiegang lamented as still unfinished in 1939.
Multifocal Rebellions: Shaanxi
The decades after the 1780s had seen no major Muslim actions against the Qing in Gansu, but during this time the widespread networks of Muslim traders and wandering religious professionals brought news of increased violence and agitation all over China, as well as specific intelligence regarding the Qing military’s incapacities.32 As we have seen, this period also saw subtle but marked discriminatory precedents emerge in the Qing criminal statutes. Dealing primarily with cases in Shandong, Anhui, Henan, and other central provinces, the Board of Punishments affirmed the general perception that Muslims are inherently more violent people than non-Muslims and thus should be punished more severely for the same acts. We have no evidence of Sino-Muslim reactions to these changes, but Muslims certainly knew that their “Huiness” would be a negative, perhaps even dangerous, attribute if they came before the Qing officials’ bench.
After 1850, with the outbreak of the Taiping wars in the south and then hostilities between Muslims and the Qing in Yunnan, the atmosphere in the northwest grew very tense. Violence on a small scale—market brawls or intervillage feuds—became more and more frequent.33 Qing civil administration, which had been shaky since before the 1781–84 uprisings, had not recovered. Corruption and official malfeasance, overwhelming taxation, neglect of the military, and confiscation of food supplies by local troops continued to plague both Muslim and non-Muslim residents of Shaanxi and Gansu.
Muslims suffered from social discrimination as well, enforced not only by the state but also by local security organizations, often organized by local gentry specifically to confront them. The militarization of society had proceeded apace all over the Qing empire, but in the northwest the non-Muslim tuanlian (militias) had an obvious target in the large local Muslim communities, who responded by forming their own armed bands. This process advanced most rapidly in the Wei valley of southern Shaanxi Province, where agriculture and trade supported a dense population that may have included as many as a million Muslims by the mid-nineteenth century. Most Shaanxi Muslims had not been persuaded to join Sufi orders, so their villages and urban neighborhoods remained Gedimu, mosque-based local solidarities connected to one another by the emotional ties of shared Islamic identity but not by any institutional bonds. Faced with hostile, armed non-Muslims, however, these Muslims did form multivillage associations for self-defense, and they constructed several series of stockades and fortified strong points on the hillsides north of the Wei River. The forts of the Fengxiang Muslims, for example, stretched “over a hundred li” in southwestern Shaanxi. To no one’s surprise, these forts and their armed young men came into regular, violent confrontation with non-Muslim tuanlian in the vicinity, but the Muslims never created any provincial or even regional leadership. Elites and their militias, like most people’s concerns, remained almost entirely local, rarely extending beyond a few counties.
The most vicious battles took place in the southeastern part of the province in the early 1860s. Attested in local sources, these feuds alarmed local society in a number of counties and involved thousands of armed men. In addition, the years 1861 and 1862 brought outside forces to play a role in stimulating local conflict. Taiping armies moved north from Sichuan under Shi Dakai and from Henan under Chen Decai, and the western Nian threatened from the east, so friction increased in Shaanxi. Knowing the Qing’s military ineptitude and inadequacy, Muslims and non-Muslims reinforced their stockades, and some Muslims made contact with the invaders as potential allies against the local non-Muslim tuanlian.34
Shaanxi was not the only locus of confrontation. All over Gansu as well, local leaders provided weapons to their followers and organized militias. Mosques affiliated with menhuan had especially effective networks of intelligence and supply, since their members could easily move between communities and their religious leaders owed allegiance to a single shaikh. The Sufi suborders could also concentrate economic resources effectively, for their members regularly contributed to the menhuan’s treasury. Without any central coordination, the Gansu Muslims created no fewer than four large military regions, each controlled by charismatic religious and military leaders. Across the Yellow River from Ningxia, at the town of Jinjipu, the Jahrīya leader Ma Hualong made his headquarters. At Hezhou, Ma Zhan’ao of the Khafīya, both an ahong and a tuanlian commander, became the primary leader. Ma Guiyuan led the Muslim militias at Xining; Suzhou, out in the western Gansu corridor, saw an independent Muslim force take shape under Ma Wenlu. The connections among these centers consisted primarily of information networks, not any coordinating structure, and no historian has been able to document effective decision-making discussions among them, though Jahrīya adherents in all four owed loyalty to Ma Hualong. As we shall see, they met very different fates as local xiedou violence escalated into full-fledged military confrontation in the 1860s.
The initial local cause of the 1862–73 violence, according to Wendjang Chu, was a brawl over the sale of some bamboo poles at a Weinan County market. Other historians have discovered other accounts, all describing events that increased communal tensions. One author, for example, claims that a group of Muslims attempted to break into a non-Muslim theatrical performance toward which they had not contributed.35 Reports from elsewhere in the northwest indicate that localized warfare, including conflicts between competing Muslim groups, occurred in many regions during those same years. News of these conflicts traveled rapidly via the Muslim trading networks as well as Qing official channels, increasing tensions all over Shaanxi and Gansu. The Xining area, for example, suffered from serious xiedou clashes that spring between competing Khafīya suborders, who had been feuding for several years.36 In sum, the “rebellion” was an escalation of existing tensions in a variety of settings rather than a centralized, planned uprising caused by an explosion of new conflicts or contradictions from a single flash point.
All agree that violence of an unusually general nature erupted in the Wei valley and spread with great rapidity in the spring of 1862.37 The fighting did not follow any prearranged plan but rather ran along the roads and rivers that connected the Muslim communities, all of which reacted against the threat of attack by tuanlian, withdrawing to their forts or attacking non-Muslim villages. Southeastern Shaanxi has become known as the origin of rebellion both because of the arrival of the Taiping and Nian armies and because the Muslims of that area were able to kill high officials and gentry. Zhang Fei, a Hanlin scholar and formerly a high-ranking censor, came from Xi’an to Lintong to organize tuanlian and to persuade both Muslims and non-Muslims that the “national enemy,” the Taiping army, was the main threat to social order. After ordering the local non-Muslim tuanlian to leave him unguarded and parlaying unsuccessfully with the two sides, Zhang and his party were captured by Muslims, taken to Cangtouzhen, and killed, making the local Muslims “rebels” in Qing perception and law.38 Elsewhere the local non-Muslim tuanlian mobilized and struck first.39 Generally the Muslim violence, not that of the semiofficial tuanlian, was held to constitute “rebellion,” and local officials certainly perceived it within that category of crimes.40
Faced with epidemic violence and murdered upper gentry, a series of inept Qing commanders, including Yingqi and Shengbao, wavered as to general policy: should they take a hard line and treat all Muslims as rebels, or should they try to separate “good” from “weed” Muslims as their theory demanded? Military contingency argued for the latter solution; given the heavy demands on their military resources for fighting the Taiping and other insurgents, the Qing could put only a few thousand poorly armed troops into the field against an enemy potentially numbering hundreds of thousands. But the local non-Muslims, terrified of the hideous violence of which they believed the Muslims to be capable, demanded the former, an option known as “washing away the Muslims” (Ch. xi Hui).41 One source reports that unofficial “proclamations appeared declaring that all Muslims were ‘to be killed without further inquiry.’ ”42 In these Shaanxi non-Muslims’ eyes, their government could protect them only by ending the Muslim threat permanently—by killing all the Muslims or allowing the tuanlian to do so.
Policy options also divided the Muslims. Some commanders, fearing the worst from the state and knowing the worst of the tuanlian, argued for a full-scale strike against the provincial capital at Xi’an, to neutralize resistance and force the Qing to pardon them and remove the potentially fatal “rebel” label. Others rejected this radical course and pleaded for more diplomatic methods. In an atmosphere already rife with rumors of sanguinary deeds (some of them true) and intensified by the alarms of many armed bands, the bellicose arguments carried the day. Some dissenting Muslims were killed by their coreligionists, while others committed suicide in despair, and in late June the organized Muslim tuanlian besieged Xi’an.43 For over a year Qing troops from Shanxi to the east and Sichuan to the south tried to relieve the city. Fierce fighting between Muslims and Qing armies, with their non-Muslim tuanlian allies, took a terrible toll on both fighters and civilians. With the relief of Xi’an in the fall of 1863, and the relatively competent Dolongga leading troops westward toward Fengxiang, Qing victory seemed assured. But government troops remained too thinly spread across the province, and the anti-Qing Muslims counterattacked and remained in control of substantial parts of Shaanxi for five more years. Not until the defeat of the Taipings and the arrival of Zuo Zongtang as supreme commander, bringing his veteran Hunan troops and methodical logistics, could the Qing finally retake all of Shaanxi in late 1868.
In the meantime, the fighting had created a vast number of homeless Muslim refugees, of whom a majority fled westward to Gansu, where their “Eighteen Great Battalions,” organized according to their native places and leaders, concentrated in eastern Gansu and sought to rebuild their strength until they could fight their way home. These military units remained the Shaanxi Muslims’ institutional form and focus of loyalty for the remainder of the period of violence. In each of the Gansu centers, local Muslim commanders had to decide what to do about the Shaanxi refugees, who were both “us” (Muslims) and “them” (nonlocal, and often members of different religious solidarities). The Gansu leaders were unwilling to integrate the Shaanxi fighters and their many dependents entirely into their communities, so the Shaanxi Muslims came to constitute not only nuclei of resistance to Qing “pacification” in Gansu but also a contentious issue among Muslims. Their horrifying experience in Shaanxi and their vengeful, homeless presence in Gansu contributed to violence in many parts of that province, and their dispersal at the end of hostilities troubled the northwest for years. They never succeeded in forming any effective central command, remaining in eighteen units, whose manpower and then number gradually decreased as disease, warfare, and desperate conditions took their toll.
What began as a brawl, or series of brawls, became a succession of tragic massacres and flights. Literally decimated, the once large and wealthy Shaanxi Muslim communities never regained their position. The 1953 census found only fifty-four thousand Muslims in the entire province, far less than 10 percent of what the population had been a century earlier. The Xi’an community, a few thousand strong, escaped unscathed from the rebellion only by promising again and again, under close military scrutiny, that they would not under any circumstances join the rebels.
Multifocal Rebellions: Gansu
In Gansu, with its more dispersed and much larger Muslim population, the Qing could not hope to inflict so crushing a defeat, nor had the Gansu Muslims any place to which they might flee. The rest of China offered no refuge, nor did Turkic-speaking Xinjiang appear an attractive alternative to most. Qing commanders and policy makers differed on the extent of violence necessary to pacify the Muslims, and in the end they treated each center differently, depending on the resistance offered, the military options available, and their perception of the Muslims’ motivation. After “pacifying” Shaanxi in 1868, Zuo Zongtang turned his attention first to Jinjipu, the well-defended headquarters of the Jahrīya under Ma Hualong, spiritual descendant of Ma Mingxin.44 Ma Hualong had engaged not only in battle but also in negotiation with provincial and national government forces over the years since 1862; he had surrendered on at least one occasion and taken a new name, Ma Chaoqing (One Who Attends upon the Qing). On that occasion, Mutushan, the Qing general at Ningxia, had beaten his more militantly anti-Muslim rival general Duxinga in a policy debate over accommodating the Muslims, and he had accepted Ma’s surrender. The court, hoping to avoid the high costs and difficult logistics of a frontier war, believed Mutushan when he claimed that Ma’s proposed surrender was sincere. Zuo Zongtang, among others, sided with Duxinga. On the Muslim side, rather than quietly returning his troops to their farms, Ma Hualong had instead continued to strengthen his base, building a wider circle of defensive fortifications around Jinjipu and actively aiding the Shaanxi Muslims in their attempts to return to their homes by force.
When Zuo Zongtang finally solved his financial and logistical problems, through a combination of careful planning and foreign loans, he sent three columns into eastern Gansu in 1869, converging on Ma Hualong’s headquarters. Overcoming strong Muslim resistance and two mutinies within Qing ranks, Zuo’s generals reached Jinjipu with their Krupp siege guns in September of 1870 and forced Ma Hualong’s surrender in January 1871.45 At Zuo’s insistence, the entire leadership of the Ningxia Jahrīya, including Ma Hualong and as many of his family as could be captured, were executed following the siege, and thousands of ordinary Jahrīya adherents were massacred during “pacification.” As we will see, Zuo never succeeded in exterminating the New Teaching, but he certainly did try, by strenuous argument in letters and memorials and by one-sided violence after Ma Hualong’s surrender.46
Neither Muslim defense nor Qing suppression followed this same course at Zuo’s next target, the frontier entrepot of Hezhou, a Muslim-dominated trading town lying between China and Tibet.47 The most important rebel leader at Hezhou, Ma Zhan’ao, was certainly not an anti-Qing holy warrior. Like Ma Hualong both an ahong and a general, he was head of the Huasi menhuan, associated with Ma Laichi’s Khafīya. Ma Zhan’ao managed to preserve his political prerogatives and his territory by astutely handling Zuo Zongtang. After seizing Hezhou in 1862, he had established a stable base for antigovernment activity or, as the Muslims saw it, a haven for protection against the hostile Qing order.48 Though Zuo prepared his campaign against Hezhou in 1872 with great care and stockpiling of supplies, he could not overcome Ma Zhan’ao’s carefully positioned troops in the Taoxi triangle.49 The Qing armies crossed the Tao to fight the crucial battle around the town of Taizisi late in that year, and the Muslims drove the government troops back on the river, then out of Taoxi, killing several commanders and retaining complete control of their territory and supply lines. Though the waters of the Tao may not have run red with blood, as one missionary reported, Zuo’s legions fled before the Muslim counterattack.50
Had Ma Zhan’ao been a separatist fanatic, as the typical Muslim leader is often portrayed, he surely would not have behaved as he did after this battle. While still mopping up the Qing remnants in Taoxi, Ma sent his own son to Zuo’s field headquarters at Anding to propose immediate surrender of Hezhou.51 He declared his loyalty to the Qing and his willingness to aid the imperial armies in further campaigns against any rebels, including Muslims. Ma’s effectiveness in dealing with the Qing and protecting his own community may be demonstrated by his influence on Qing policy in the wake of the fighting. As Zuo Zongtang shuffled surrendered Muslims from place to place in Gansu, following He Bi’s seventeenth-century recommendation that the Muslims be isolated, he undertook the reduction of conflict near Hezhou by moving the non-Muslims away. This happened nowhere else in northwest China.52 After the surrender Ma Zhan’ao executed local Muslims foolish enough to disagree publicly with his decision to ally himself with the Qing state.53 For his loyalty and subsequent success in campaigns against “rebellious Muslims” at Xunhua and Xining, he received the feathered cap of the fifth rank (Ch. hualing wupin dingdai), a token of prestige and merit requested for him by Zuo Zongtang.54
His army buttressed by the newly recruited Hezhou Muslims, Zuo Zongtang turned his commanders toward Xining. Though his ultimate objective lay in opening the road to Xinjiang through the Gansu corridor—which meant dealing with the Muslim-held fortress at Suzhou—Zuo calculated that Xining’s strategic position south of the road, and the large number of Shaanxi Muslims who sheltered there, argued for taking that major city first. Well protected by its mountains and rivers, Xining held out against Zuo’s commander Liu Jintang for three months but fell in the late fall of 1872. The gentryman Ma Guiyuan, who commanded the Muslims, was captured in Salar country, and thousands of armed Muslims were killed. But Zuo did not mete out severe justice to the Xining Muslim community itself. Rather, he arranged for the resettlement of the surviving Shaanxi refugees, as he had done after Jinjipu and Hezhou, on arable lands distant from other Muslim centers, mostly in eastern and southern Gansu.
Finally free to recapture Xinjiang for the Qing, Zuo’s forces faced the last Gansu obstacle at Suzhou, a heavily walled fortress astride the main highway through the Gansu corridor, defended by a large number of Muslim commanders from all over the northwest. Ma Wenlu, who held highest local authority, came originally from Xining, and many Shaanxi Muslims had joined his banner. As in his previous campaigns, Zuo moved cautiously, assuring his supply lines and sufficient ammunition for his German siege guns by reinforcing Lanzhou and establishing an arsenal there. By September of 1873 Suzhou was entirely cut off from the east and surrounded by fifteen thousand Qing troops under Zuo’s personal command. Ma Wenlu surrendered on October 24, his walls battered by heavy shells and undermined by tunnels and explosives. The next month Zuo ordered and oversaw the execution of seven thousand Muslims, most local but 1,500 of them from other Gansu and Shaanxi communities. This number surpassed those executed when Meng Qiaofang took the same city from Ding Guodong in 1649. To guarantee that the Gansu corridor would remain open and unthreatened, Zuo removed the few surviving Muslims from all the cities and towns in the corridor, from Lanzhou to Suzhou, and resettled them in southern Gansu: “Their seed will no longer remain in these three prefectures, and one need not worry about collusion between Muslims inside and outside the Chia-yü Pass [Jiayuguan].”55
Thus ended the “great rebellion” of the northwestern Muslims, in total defeat. As we have seen, it actually consisted of multiple rebellions, local in both personnel and outcome, and its objectives, as far as we know, had never included toppling the government or setting up an independent Muslim state.56 The flight of the Shaanxi Muslims, the only connective tissue among the various rebellious centers, did not create or enable any unified leadership. The Shaanxi Eighteen Great Battalions were forced to be outsiders in Gansu, and they never found a home until Zuo Zongtang resettled their remnant survivors by force after his victories.57 One Shaanxi Muslim commander, Bai Yanhu, held his refugee force together and retreated from one Gansu Muslim refuge to another, fighting all the while, then fled after the Xining debacle to Xinjiang, where he joined with several Turkic-speaking Muslim leaders to continue the anti-Qing fight. He ended his days under Russian rule, west of the Pamirs, and his followers are numbered among the ancestors of the Dungan “minority nationality” of Kyrgyzstan.58
Muslim Rebellion or “Muslim Rebellions”?
The above narrative, schematic and rapid, can nonetheless inform our discussion of Muslims in China by challenging a number of paradigmatic constructions: the unified minzu, the fanatical Muslims, the rational and statecraft-practicing Qing officials, even notions such as “rebellion” itself. For none of these is a natural, inherent quality of persons or events; rather, we must use the evidence to describe what people did—that is, we must construct narratives—and modify our paradigms and vocabulary accordingly. To take one dramatic example from this complex series of events, let us consider the Muslim militias of the Wei River valley. Conscripted by their communities for self-defense, armed and drilled, provided with fortifications both physical and ideological, these young men could have had a wide variety of motivations to attack a nearby non-Muslim village. Were they simply ordered to do so by manipulative, fanatical ahong? Did they lust after loot or violence for its own sake? Or had they heard reports (true or not) of massacred Muslim villages down the road, of tuanlian on the rampage, of Qing armies raping and pillaging? Were they predatory or protective, rapacious or terrified, or did they change over time? These questions can be answered only through the construction of focused narratives on the basis of local evidence, not by wholesale, essentialized characterization of religious groups, minzu, or other collectivities.59
These conflicts, like those of 1781 and 1784, pitted Muslim and non-Muslim Qing loyalists against Muslim (and occasionally non-Muslim) insurgents, righteous rebels, or rioters (depending on the historian’s vocabulary).60 “Muslim unity,” a concept dear to scholars, missionaries, and many northwesterners when describing the Muslims, did not prevent numbers of Gansu Muslims from joining the Qing forces or siding with them. Such people found greater advantage in alliance with the state, and they did not hesitate to kill coreligionists. Wang Dagui of the Haicheng region, for example, fought effectively for the Qing and was awarded a cap button of the sixth rank (Ch. liupin dingdai). Local anti-Qing Muslims caught up with him during a counterattack in 1863 and killed him, with his entire family.61 It would be comfortable to say that Old Teaching Muslims were loyalists and New Teaching Muslims rebels, but the evidence does not support such a contention. Though anti-Qing activism in northern Gansu (now Ningxia) centered in the Jahrīya, some of the most successful resistance to the Qing came from within the Khafīya and from the refugee Shaanxi Muslims, very few (if any) of whom were Sufis at all, not to mention adherents of the New Teaching.
Nor did Ma Hualong, fifth-generation holder of Ma Mingxin’s position as head of the Jahrīya, oppose the Qing with the fanatic’s lust attributed to him by most accounts.62 Branded by the Qing as the main leader of this rebellion, Ma was not a secessionist, and he sought to protect his followers and his territory around Jinjipu, suing for peace when possible until forced to last-ditch military resistance. He often requested amnesties for various groups of Muslims and was reported to be treating non-Muslims well in the territory under his control.63 Even according to official accounts of the campaigns against him, Ma initially had no desire to establish an antidynastic enterprise. He took no royal titles of office within his movement, though one of his Jahrīya subordinates did, under the influence of Taiping and Yunnanese rebellious styles.64
Jihād, Islamic “holy war,” has held pride of place in Euro-American thinking about Islam since the Crusades. A vision of fanatical hordes unafraid of death has dominated our perceptions and prevented a clear understanding of wars fought by Muslims. Literally “effort directed toward perfection,” jihād became a general term for military action by Muslims to expand Islamic territory or, if need be, to defend it. The idea of a war to convert Qing territory into Islamic territory could almost never be entertained by a Sino-Muslim leader, as compared to Turkic-speaking Muslims in Xinjiang, who often did declare jihād against the Qing. Indeed, virtually all of the Shaanxi and Gansu Sino-Muslims, as we have seen, shared a strong sense of belonging in China and of the Qing state’s legitimacy. Shengzhan (lit., “sacred or holy war”), the Chinese translation of jihād, came to mean self-sacrifice for Islam, a call to heroism rather than political domination.65 Muslims in Shaanxi and Gansu engaged in shengzhan not to topple the state but to attack local officials, to take revenge on local enemies, or to defend themselves.
The bloodshed and sacrifices of Ma Hualong’s stubborn defense of his main base at Jinjipu might well have been avoided had Zuo Zongtang or his superiors in Beijing been willing to talk peace with the “fanatical Muslim rebel,” as Mutushan had done. But Zuo was pressed for victory, imbued with the monolithic Confucian conception of imperial authority, and convinced that the New Teaching lay at the root of Gansu’s troubles. He foiled Ma’s attempt to surrender to Mutushan and threw army after army at Jinjipu, determined to set an example for all rebellious frontier peoples. In light of this evidence, Zuo Zongtang appears as much a fanatical defender of his faith, the Qing imperium and its bureaucratic Confucianism, as Ma Hualong was of his Jahrīya and his territory.66
Neither the creation of negative images of Muslims nor the usefulness of such images to the Qing state should surprise Euro-American historians. Lacking texts from the Muslim side, we cannot rely on evidence manufactured by self-serving military officials to deduce Muslim fanaticism. Such tactics have long been a mainstay of official Chinese historiography when describing those who failed to achieve their goals in opposing the state—“A winner a king, a loser a bandit.” We can read the sources carefully to comprehend the discriminatory, prejudiced filters through which Qing officials and local gentry viewed the Sino-Muslims and also search outside the official correspondence and military compendia for evidence regarding Muslim motivation and behavior.67
We must also remember that the flood of Shaanxi refugees into Gansu, driven from their homes amid slaughter and pillage, helped to keep anti-Qing sentiment alive in communities within which they settled. Jinjipu, Xining, and Suzhou sheltered large numbers of Shaanxi Muslims, and those areas became strongholds of anti-Qing Muslim bands. But even as they worked together to resist the Qing, the Shaanxi Muslims and the Gansu Muslims—including Ma Hualong’s Jahrīya—viewed the conflict differently. According to one source, Ma Hualong tried to use the Shaanxi Eighteen Great Battalions as a front line of defense, protecting his northern Gansu troops and forts. Many Gansu Muslims mistrusted the outsiders, and Zuo Zongtang’s lieutenants later summarized: “Ma Hualong was a New Teaching Muslim, holding to different religious objectives than those of the Shaanxi Old Teaching [in this case, Gedimu] Muslims, and they were mutually incompatible [lit., ‘fire and water’].”68 Hezhou’s Muslims, led by Ma Zhan’ao, successfully controlled the Shaanxi refugees’ local potential for violence and thus placed themselves in an advantageous position to hold their territory after “pacification.” Thus, the Shaanxi Muslims and the Jahrīya—not Muslims in general—provided crucial nuclei of resistance, and those two violently anti-Qing groups differed considerably in motivation. The Shaanxi refugees wanted to fight their way back home, while the Jahrīya adherents seem to have been moved deeply by commitment to their shaikh and to Islamic revival, by the Qing’s policy of proscribing their tarīqa and exterminating them, and certainly by the desire to protect their homes.
Zuo Zongtang’s lenient, even generous acceptance of Ma Zhan’ao’s surrender stands in strong contrast to the brutality meted out at Jinjipu and to the massacre at Suzhou. Nakada Yoshinobu explains Zuo’s benevolence as the result of several lines of reasoning. A campaign against Hezhou would have taken a long time, and Zuo’s superiors looked to him for a rapid, inexpensive victory after the protracted battles in Shaanxi and northern Gansu. It had, after all, been almost a decade since the beginning of the troubles. Appeasement of the Hezhou Muslims, which could be made to look like victory, produced success and fame for Zuo without incurring high costs.
In addition, Ma Zhan’ao had volunteered to serve the throne, allowing Zuo to “use the Muslims to control the Muslims,” a conventional and attractive solution to frontier problems. This variation on an ancient Chinese political theme saved lives and expense for the Qing commanders while preventing unified Muslim resistance to Qing power, a unity inaccurately predicted by Qing ethnography to be inevitable and permanent. Zuo exacerbated already-existing divisions among Sino-Muslims by supporting some groups and persecuting others, and by physically separating communities of Muslims from one another by large-scale resettlement. This strategy proved particularly effective in dealing with the Shaanxi refugees, who had become so intractable wherever they lodged. Zuo moved them to isolated Muslim enclaves in eastern Gansu—at Zhangjiachuan and Pingliang, especially—to prevent them from returning to their Wei valley villages. Their homes and fields in Shaanxi had already been taken over by local non-Muslims, and Zuo knew the refugees would demand satisfaction if they returned.69
Ma Zhan’ao had been willing to negotiate and had demonstrated no antistate, dynastic, or jihādist ambitions, so his “rebel” status might be negated rather easily and made to appear temporary. Unlike Ma Hualong, who had been associated with the heterodoxy and long-term violence of the Jahrīya, Ma Zhan’ao led one part of what the Qing called the Old Teaching, which at this point included both the Khafīya, with its many subgroups, and the Gedimu congregations of both Shaanxi and Gansu. The Qing had blamed the entire northwestern debacle on the New Teaching for almost a century. The active participation of an Old Teaching leader in anti-Muslim campaigns served to divide the Muslims further, to isolate the New Teaching more effectively, and to establish more firmly the “legitimacy” of the Old Teaching.70 First as a rebel leader, and then as a “returned” loyalist, Ma Zhan’ao built an alliance between Muslim local elites and Qing authority in order to control local violence while retaining a measure of local power in the face of Qing officials and armies. Events after 1873 bear out the ironic observation that Ma Zhan’ao, by surrendering Hezhou, had preserved it as a territorial base for himself and his descendants, for the entire Hezhou region remained under the control of his family for another half century.
To Hezhou’s northwest, the Xining Muslims had been engaged in local feuding with one another before 1862, and many Shaanxi refugees made their way across Gansu to this multicultural outpost to intensify already-existing tensions. Local Muslims knew that the Qing armies would follow the Shaanxi refugees as soon as they could mount a campaign, and controversy arose over the appropriate response to Qing oppression and violence. Despite feelings of loyalty and solidarity with their coreligionaries from Shaanxi, fear of the army and of local non-Muslim tuanlian hindered organization until the attack finally came. Once Liu Jintang’s army started killing in the Muslim suburbs, whole communities lay open to massacre, however law-abiding they may have been. Official pronouncements to the contrary notwithstanding, Qing soldiers could not easily distinguish “good” Muslims from “weed” Muslims—troops do not take the time to make such subtle distinctions among civilians assumed to be armed and hostile. From immediate experience, the Muslims knew that they would be treated as enemies by the Qing military. Ma Guiyuan’s small army held off the enemy for a few months, but he was defeated and executed by a coalition of government troops and local militia, some of them Muslims. Ma Zhan’ao’s troops participated in this campaign on the Qing side, ensuring that the Salars at Xunhua did not join Ma Guiyuan’s army. With the execution of Ma Guiyuan, the final massacre at Suzhou, and the flight of the remnant Shaanxi refugees under Bai Yanhu to Xinjiang and then Russia, the long violence in Gansu came to an end. The triumph of Qing arms has been attributed to the abilities of Zuo and his commanders, to the harsh terrain and scattered Muslim communities, and to the logistical and technical superiority of the Qing armies. We may now add one more crucial factor: the Muslims did not act from any unified strategy, nor did they unanimously oppose the Qing authorities.
This interpretation of the 1862–73 rebellion helps us toward a new understanding of Islam in China, especially the Old Teaching–New Teaching conflict, and of Muslim relations with non-Muslims in Gansu. We find mutual fear, economic rivalry, and influence from the Taiping and Nian rebellions to have been important stimuli to violence, added to the state’s corrupt, incompetent fiscal administration and its endemic suspicion of closed voluntary associations within Chinese society. This last motivation found expression in a continued demonization of the Jahrīya. Zuo Zongtang argued strenuously in government councils for complete proscription:
The reason why the New Teaching must be prohibited is that it claims to be from God and makes ridiculous prophecies. This group’s behavior is very strange and often lures foolish Muslims into willing slavery. The victims often are trapped into conspiracy without knowing how and are even willing to face execution without the slightest regret. . . . This makes the New Teaching a potential danger to the Empire. . . . [Ma Hualong] healed the sick and granted children to those who prayed for the birth of children. . . . When the New Teaching is eliminated . . . then Shaanxi and Gansu can expect to be safe for a hundred years.71
Zuo lost that political battle, and the Jahrīya recovered from the terrible devastation of Jinjipu, but not without changing both its behavior and its organization, as we will see in chapter 5. Some modern scholars have concluded that the New Teaching represented a decadent, corrupt form of Islam. In fact, as we have seen, it was part of an Islamic revival, a movement to return to Koranic purity through mystical unity with the divine and a purging of Chinese-influenced customs, but neither its doctrine nor its behavior were inherently anti-Qing.
An essential part of this Sufi faith lay in obedience to the leader of the order, in whose hands lay the power of initiation into the order and whose tomb provided individual and communal inspiration to his descendants and followers. Such leaders may be found all over the Muslim world, including the darwish of Turkey, the pir in India, the serigne of the Wolof, and China’s jiaozhu. The extraordinary loyalty they inspired in their followers constituted an important element in their potential threat to Qing state interests. With China’s long history of peasant uprisings, rebellious sects, and millenarian movements very much in mind, officials reacted with extreme hostility to the activities of “heterodox” religious groups in times of social unrest. Voluntary association, always problematic in totalitarian states, became identified with subversion in the eyes of the Qing authorities.72 In the nineteenth century the Jahrīya embodied structural and ideological characteristics that, seen from the Qing point of view, made it subversive even in times of peace and made Ma Hualong an intolerable menace, despite his willingness to negotiate and to protect non-Muslims under his control.73
We may thus conclude that religion did indeed play a role in the violence of mid-nineteenth-century northwest China, but it was neither a consistent nor a dominant role. Local economic, social, and political conditions certainly loomed large in the calculations of all protagonists, as did state policy toward the Muslims and Qing decisions on taxation, military preparedness, and local control. Like ethnic identity, religious identity can never be an absolute predictor of behavior, and in these alarms and clashes we see Muslims making a wide variety of decisions—some active, some reactive—in the face of a majority society and state they correctly perceived to be potentially hostile.
Islam did not determine the actions of the Shaanxi Muslim battalions, Ma Zhan’ao, or even Ma Hualong, but all of them acted inside frameworks of Muslim community and identity, which were different from those of their non-Muslim neighbors. Sometimes led by ahong, engaged in what they called shengzhan, answering a shaikh’s call to unified resistance, going into battle prepared for shexide (the Sino-Muslim transliteration of the Arabic shahīd, “martyrdom”), these Muslims do not seem to have been very Chinese. Yet, as we have seen, they were Chinese, like other anti-Qing rebels—fighting for the safety of their villages and families, reacting to a state authority to which they granted legal legitimacy as long as it did not attack them, fighting against the Qing or making local peace deals as they judged most expeditious or noble or safe. Therein lies the historical puzzle of Muslim “rebellion” in China, a puzzle we confront again in the 1890s.
EVERY THIRTY YEARS A SMALL REBELLION
[Hann] Nuri was a Xunhua Salar Muslim, leader of the Old Teaching at Gaizigong. He and the New Teaching adherent Hann Si accused each other over an old grudge and started a feud [Ch. xiedou].74 Their contention was most turbulent, and could not be settled rationally. Governor-General Yang Changjun drew up an indictment and commanded the Xining prefectural office to hear the case, for Xunhua was in Xining Prefecture. Woshikeng’e presided, and both litigants followed correct procedure.
Nuri came to the prefect’s court and deposed that the Old Teaching esteemed and loved the Koran and saw the New Teaching as a heretical doctrine. He cited many precedents as proof, talking with confidence and composure. . . . [He said that] the former governor-general, Zuo Zongtang, had held that Gansu Muslim rebellions’ origins lay in religious disputes [between Muslims]. If religious litigations do not cease, armed conflict will be born. The roots of conflict are surely in the New Teaching.
[At the end of Nuri’s deposition] Hann Si sensibly stayed silent, not saying a word. Woshikeng’e held that Muslim religious disputes were commonplace, and sent both litigants back to Xunhua, ordering them to settle their quarrel and feud no more.75
The New Old and New New Teachings
Is the above case not history repeating itself, as the Jahrīya and Khafīya go at it again?76 On the surface, the case of Hann Nuri and Hann Si, who is conflated with Hann Musa in some sources and called Hann Laosi in others, could have been adjudicated in the 1760s or 1860s. We find the same sorts of people (feuding Muslims) in the same location (Xunhua) suing one another over the same sorts of things (accusations of heterodoxy). But it was not repetition, for this case came before the Xining judge Woshikeng’e in 1894, and northwest China had undergone many dramatic changes since Ma Mingxin’s time, as had the rest of the Qing empire.
MAP 4. Center of the 1895–96 Muslim rebellions. (Map by Philip M. Mobley)
In the almost two centuries since Sufism had made its first inroads among the Gansu Muslims, the menhuan had arisen as a dominant institution in many centers, especially Hezhou, Xining, and Ningxia, but also in smaller towns such as Didao and Xunhua. Saintly lineages that led Sufi suborders, the menhuan also constituted crucial loci of secular power, bearing strong structural resemblance to the corporate lineages that controlled local systems all over the Qing empire.77 The division between “Old” and “New” at Xunhua stemmed from a power struggle within a single menhuan, for both litigious Hanns belonged to the Huasi menhuan of the Khafīya, which the Qing had called part of the Old Teaching. Nuri and Musa followed different leaders within their solidarity, and their xiedou involved not only currents of revivalism from the Muslim world to the west, as Ma Mingxin’s and Ma Hualong’s had, but also Zuo Zongtang’s pacification of Gansu in the 1870s.
When Ma Zhan’ao surrendered after defeating Zuo Zongtang in 1872, his supporters included two brothers, Ma Yonglin and Ma Yongrui, who were ahong from the Huasi menhuan’s leading family, the descendants of Ma Laichi. Ma Yongrui had played an important military role in the 1860s, while his brother Yonglin had provided wealth to the cause through his commercial success. After “pacification” had been completed through the elimination of Muslim dissenters, Ma Yongrui retired in Hezhou, while Ma Yonglin continued as a leader in local politics and in the menhuan. Ma Yongrui’s son, Ma Rubiao, followed his father and uncle into religious studies, and the story of the new New Teaching revolves around him.
Either through the influence of an Arab missionary named Selim who came to Gansu, or during his pilgrimage to Mecca, or both, Ma Rubiao became involved in the Shādhilīya, a Sufi order popular throughout the Muslim world.78 Though not very different from the Khafīya of his ancestors, this new affiliation gave Ma Rubiao the impulse to reform his own menhuan, to purify religious practice and bring it into conformity with what he had come to regard as “pure” Islam. Returning from his pilgrimage, Ma Rubiao followed a familiar pattern; he gathered disciples, drawn by his charisma and that of his texts and reformist ideas, and split the Huasi menhuan. While Ma Zhan’ao lived, he was able to mediate between the reformist and conservative elements within the Huasi, but after his death in the late 1880s internal conflict over leadership, ritual practice, and finances spread from Hezhou to other Khafīya centers, especially Xunhua.79 As the latest innovators, for so they were branded by their opponents, Ma Rubiao’s reformist followers bore the condemnatory appelation of New Teaching, while his uncle Ma Yonglin led the Old Teaching resistance to change.
By 1887 xiedou had broken out between the rival Huasi factions at Xunhua, and the escalation of fear and violence began again. Given their ferocious reputation, religious feuding among the Xunhua Salars always represented danger to Qing officials, though in this case the initial impetus to conflict did not lie among the Salars but among the Sino-Muslims at Hezhou. As the troubles continued and lawsuits multiplied, the authorities all over the province grew more concerned about widespread violence. In the fall of 1894, as the Xunhua litigants argued in court, their followers fought it out in the streets of Xunhua, and Muslims were killed. The Hezhou commander Tang Yanhe decided to send Ma Yonglin and Ma Yongrui to mediate, since they held high status in the menhuan within which the xiedou were taking place. Eager to forward his antireformist cause within the Huasi, Ma Yonglin secretly encouraged Hann Nuri to attack the New Teaching, which he did, killing at least two ahong. Khafīya adherents also believe that Ma Yonglin knew that the Qing had severely depleted their northwestern defense forces in order to fight the Sino-Japanese War, and that he told Hann Nuri that there were no Qing armies west of Tongguan, in Shaanxi.80 Following long-standing Khafīya custom, the New Teaching group sought relief from the Qing law courts, sending representatives to Lanzhou, where they accused both Hann Nuri and Ma Yonglin before Governor-General Yang Changjun.
Qing Action, Muslim Reaction
Xunhua. In the spring of 1895, Yang Changjun sent Chen Jiaji, a Xining military official, to Xunhua with a small military force and orders to end the troubles. By the time he reached Xunhua, Chen seems to have decided that Ma Yonglin and Hann Nuri’s Old Teaching, the antireformist group within the Huasi menhuan, was responsible for the xiedou. He locked the Xunhua town gates against the Muslims who had gathered to present their cases to him, arrested eleven Old Teaching leaders, and had them executed, displaying their severed heads to awe the Salars into submission.81 His troops fired on the protesting crowds from the walls, and the Old Teaching Muslims reacted by surrounding Xunhua, several thousand strong, on April 2, 1895. Thus began a year of warfare that devastated southern Gansu from the Tao River to Xining (see plate 15).
Chen Jiaji had acted in what he perceived to be a productive manner, punishing the destroyers of public order and parading the grisly results as a deterrent to further crime. From the Old Teaching Salars’ perspective, their leaders had been unjustly slaughtered by an evil Qing official, and, for all they knew, they were next. From this point on, Chinese sources limit themselves to descriptions of troop movements, battles, strategies, and massacres. We must infer the personal and collective motivations of the belligerents (and the bystanders) from the extent of their participation.
Fearing that the violence might spread to other Muslim centers, Governor-General Yang planned a rapid military campaign to relieve Chen Jiaji at Xunhua. Within a few days of Chen’s debacle, General Deng Zeng, who had participated in the wars of the 1860s, moved east from his base at Xining, and Tang Yanhe was to cross Jishi Pass from Hezhou to catch the Xunhua besiegers between two Qing armies. Before Tang even reached the mountains, however, he attacked several Muslim villages near Hezhou and announced that he would kill all Muslims, regardless of their menhuan affiliation or behavior.82 Muslims all over the province heard immediately and reacted to Tang’s threat of massacre, so the Old vs. New Teaching controversy within the Huasi menhuan disappeared rapidly in the fear inspired by this brutal, publicly stated intention. Thereafter, the violence was no longer limited to the Old Teaching or the Salars, though it remained geographically confined.83 Muslims in areas attacked by Qing troops had little choice but to oppose them. As in the 1860s, soldiers operating in hostile territory could hardly be expected to inquire as to the “good” or “weed” status of Muslims, all of whom they perceived as potential enemies. Yang Changjun tried to quell the fear with proclamations declaring Qing benevolence and demanding an end to the fighting:
Initially the Muslims of Gansu were without religious conflict. From the early Qianlong period, the “Eat Befores” and “Eat Afters” arose [referring to the qiankai-houkai conflict]. With such theories came menhu [menhuan] and from menhu came conflict over power and profit. The religion divided Old from New, who accuse each other and take murderous revenge. . . . What good do all these rebellions do? . . . Immediately turn back and peacefully pursue your occupations!84
Once Muslim leaders had been declared “rebels” by the state, it would be hard to imagine them believing any promises of Qing benevolence, with the memory of the 1870s still fresh and raw and Tang Yanhe’s extermination order being enforced in local fact and general rumor.
Yang’s proclamation encapsulated the perceptions of both the state and the local non-Muslims: Muslim violence, which they called rebellion, stemmed from conflict within the Muslim communities. Violence in society may not be tolerated, so Muslim internecine strife must be condemned, along with all institutions that appear to promote it, such as New Teaching and menhuan. If conflict occurs, the state must intervene, despite the antistate unity created by direct military interference in Muslim affairs. This dogmatic, straightforward analysis, reinforced by centuries of precedent, dominated Qing policy toward the northwestern Muslims, contributing to violence and tragedy.
The local non-Muslims, as frightened of the infamous Muslims as the Muslims were of them, locked themselves in their walled cities and stockades. They called for protection from the state and for extermination of the Salars. Over a century of violent confrontations and their subsequent exaggeration in oral tales had taught Gansu people that Muslims on the rampage, especially Salars, killed non-Muslims wholesale. The governor-general blamed the menhuan; solidarities that he perceived as subversive to the all-encompassing legitimacy of the Qing state. Christian missionaries, with their own particular vision of Islam and Muslims derived from the Middle East and Orientalism, theorized a jihād in Gansu. Seeing the Near East’s holy wars mirrored in this distant corner of what they called the Celestial Empire, they misunderstood the workings of communal hatred and community memory in northwest China, the inertia of local feuding and thirst for revenge, and the ineffectiveness of the Qing government in preventing violence or stemming it once it began.85
Hann Nuri, knowing that he faced superior forces once the government mobilized, used the Muslim domination of the Yellow River to cross from Xunhua to the north side in secret. He surprised and defeated Deng Zeng’s relief force moving south from the Bayanrong garrison, while Tang Yanhe’s column from Hezhou was halted at Jishi Pass. Successful in blunting the first Qing attempts to subdue them in the spring of 1895, the Xunhua Muslims alone had neither the numbers nor the weapons to resist an onslaught by concentrated armies.86 The diverse Muslims of southwestern Gansu all feared the ethnocentric wrath of the majority and the state, violently expressed by local militia and government troops, so the insurgents worked to convince other Muslims of the immediacy of the danger.
Hezhou. The largest concentration of Muslims lay at Hezhou, where Tang Yanhe commanded the garrison inside the walled city, while Ma Yonglin and the Khafīya held sway in the Muslim suburbs, with their headquarters in Bafang, south of the city wall. Within Hezhou Prefecture only large settlements, such as the market towns of Hanjiaji and Tangwangchuan, had mixed populations; the rural villages were exclusively Muslim or non-Muslim. In peacetime travelers remarked on the delicious produce, delightful climate, and bustling markets of Hezhou, especially its Muslim suburbs. A wealthy borderland entrepot, the town attracted non-Muslims from all over China as well as Muslims of many languages and cultures.87 Hezhou’s Muslims paid great attention and devotion to their religious institutions. Like Naqshabandīs all over the world, Hezhou Sufis built tombs, some of them elaborate and large, to venerate their shaikhs and to seek their intercession. The various solidarities supported at least seven large tomb complexes, each with its prayer halls, as well as a dozen major mosques and many small ones. Madrasas in Hezhou sent graduates to Muslim communities all over China, and hundreds of religious professionals led the liturgy and read the sacred books.88
The Hezhou city wall was strong and easy to defend, but the suburbs, without walls, often fell victim to manmade disaster. If contemporary reports are to be believed, each victory for Muslim or non-Muslim was accompanied by slaughter among the non-Muslim, Muslim, Tibetan, and other civilians, for communal war never excludes noncombatants from its casualties. Despite commercial symbiosis and the intercommunity cooperation that it entailed, Hezhou had a sinister reputation as a center for heterodox Muslim movements and as a headquarters for Muslim bandits. The Hezhou Muslims served Gansu mothers as a bogeyman for disobedient children: “When I was a baby, whenever I wouldn’t stop crying, my mother would say to me sternly, ‘Don’t cry any more! If the Hezhou Muslims hear you, they’ll come and kill you!’”89 By the late nineteenth century, as we have seen above, those Gansu mothers could draw on a century of violent memories to terrify their children and verify their fear of the Muslims.
The story of Chen Jiaji’s murderous attempt at pacification of Xunhua quickly reached Hezhou, as Tang Yanhe’s local garrison moved out to break the Xunhua siege, so some Hezhou Muslims organized to defend themselves and to seek revenge against the state, especially after receiving news (or rumors) of Tang’s intemperate proclamation that he would “wash away the Muslims.” The sources tell us little of the recruiting procedures of these groups, but we know that local anti-Qing leaders rose immediately to the occasion. Chief among them was Ma Yonglin, who actively agitated for violence, resorting to terror against Muslims who resisted his appeals. Reluctant Muslims lost their homes to arson or their families to assassins.
Ma Zhan’ao had trained his son, Ma Anliang, who had gone to negotiate the surrender in 1872, to be both an administrator and a soldier, and Ma Anliang had gone to Xinjiang as a cavalry commander in a Gansu unit of the Qing military. In his role as a Qing military official, Ma Anliang earned the enmity of many Muslims. He led only one faction of the Huasi menhuan, and the others had good reason to complain of his access to state authority. Familiar with this local history, the anti-Qing Ma Yonglin sought and obtained the support of a leading Muslim named Min Fuying. Min’s grandfather and father had opposed Ma Zhan’ao and rebelled against Ma Zhan’ao’s local ascendancy in 1874, two years after the surrender. Their campaign had been brutally quelled, all of its leaders executed. Min Fuying, with many followers, joined Ma Yonglin in the hope of taking revenge on the weakened Qing and Ma Anliang for his ancestors’ deaths.90
From his headquarters in Bafang, Hezhou’s largest Muslim suburb, Ma Yonglin could persuade or intimidate the local people more effectively than could the distant Ma Anliang, clearly an officer of the hostile Qing armies. Perceptions of history, both personally experienced and orally transmitted, taught some Hezhou Muslims that violence did not discriminate clearly between friends and enemies, so they joined the rebels in self-defense. Others had learned, from the same history, that violent opposition to the Qing state brought only sword and fire down upon their homes, so they opposed Ma Yonglin. All Gansu Muslims, whether inclined to violence against the state or not, knew that local officials could attack them without warning. In an atmosphere of such insecurity and fear, being a Muslim could not in and of itself determine the choices any anxious resident of Hezhou might make. People decided to fight for one side or the other, not to fight, or to flee for many reasons—personal relations with followers or leaders, personal histories of family loyalty, guesses as to who might win, and more. Ma Yonglin, for example, was apparently not yet ready to declare open rebellion against the state, so he continued to play the mediator in negotiations with the officials, a stance reminiscent of Ma Hualong’s thirty years before. When Qing commander Lei Zhengguan arrived at Hezhou with his forces from eastern Gansu, Ma Yonglin offered to be intermediary between the apprehensive local Muslims and the new garrison.
Despite superficial resemblance to earlier conflicts, the 1895 violence did not herald a much wider war, as the Wei River valley battles had in the 1860s. No flood of bitter, homeless Shaanxi refugees urged the Gansu Muslims to fight. No Muslim communities were uprooted, nor, with only one exception, did large numbers of Muslims move across the province to join the struggle. The bellicose Muslims of Xunhua, Hezhou, and Xining remained isolated from other Muslim centers. Though communication certainly did exist between Hezhou and Zhangjiachuan, the Haicheng/Guyuan region, and Ningxia, those communities did not share the immediacy of the threat felt in southern Gansu after Chen Jiaji’s brutal warning to the Salars and Tang Yanhe’s and Yang Changjun’s proclamations. As we will see below, even the Taozhou Muslim community, southeast of Hezhou within the Tao River bend, had interests sufficiently different from those of the Xunhua-Hezhou-Xining militants that it armed itself in defense against them. In addition, the Jahrīya, the New Teaching of the 1860s, had changed to the point that its formerly militant centers in Ningxia, Zhangjiachuan, and elsewhere in eastern Gansu took no part in the fighting.
Haicheng. Shortly after the outbreak of violence at Xunhua and Hezhou, a jail break and flurry of military activity did occur at Haicheng, in eastern Gansu about 250 kilometers northeast of Hezhou.91 On May 30, 1895, a band of one hundred Muslims attacked the magistrate’s yamen in Haicheng; killed the magistrate, his clerk, at least two women, and a servant; plundered the storehouses; burned the official buildings; and, perhaps most important, opened the jail to free Muslim prisoners.92 The alarm spread rapidly, for the “rebels” (as they were immediately branded) had not bothered to cut the telegraph lines, so nearby officials heard immediately of this outbreak and declared martial law in several counties.93
Local militia under local leadership made quick work of this minor incident.94 Not staying to hold the town for long, the escapees and their rescuers fled to Hezhou before the attacks of local tuanlian and Qing troops. Yang Changjun reported that the incident had lasted two weeks.95 At least eight Muslims were executed, and many Muslim civilians lost their lives in the ensuing cleanup by Qing troops, undertaken with brutal inefficiency. The Haicheng incident revealed not a widespread Muslim conspiracy but rather the isolation of the Taoxi and Huangzhong fighting from other Muslim centers. There were no threats in other areas, no reactions even in nearby, heavily Muslim districts such as Zhangjiachuan or Guyuan. The response by Qing authorities in eastern Gansu indicates a fear of Muslim violence greater than actual Muslim capacity to act violently against the state or against local enemies outside of Taoxi-Huangzhong.
After reinforcing Tang Yanhe’s garrison, Lei Zhengguan closed the gates of Hezhou in the face of Ma Yonglin’s agitation in the suburbs outside the walls. His guards checked all travelers on the road through the east gate, the only one left open, while strong patrols held the walls opposite Bafang. The tension rose, and Ma Yonglin, Zhou Shixiang (also called Zhou Qishi), and Min Fuying finally led their armed followers to besiege the city on June 8, 1895. They wanted to join their banners to those of the Xunhua Old Teaching, but as a Muslim force approached Jishi Pass from Xunhua, Muslim troops, led by Ma Fulu and Ma Guoliang, stopped them.96 These commanders, brothers of Ma Fuxiang and Ma Anliang respectively, clansmen and forerunners of the Ma family warlords of the Republican period, clearly felt no sense of solidarity with the antistate activity of Ma Yonglin or the rebellious Xunhua Muslims. Their interests and loyalty lay with the dynasty, a choice their fathers had made that had served their families well.
Didao. Unable to unite with coreligionists to the west, Ma Yonglin and his allies looked eastward, to an alliance with the sizable Muslim community at Didao. Located on the east bank of the Tao, Didao controlled a vital ferry and floating bridge crossing between predominantly Muslim Taoxi and non-Muslim Taodong. The Muslims of Didao constituted about 10–15 percent of the town’s people, and several suborders and menhuan had their headquarters there. The Hezhou Muslims did not attack Didao themselves, but rather persuaded some of the Didao Muslims to violence. The Mufti, a Didao menhuan that traced its leaders back to Hidāyat Allāh, responded by taking the tuanlian forts west of the Tao, crossing the river, and besieging the town. For Ma Yonglin, Didao constituted a crucial military objective. Its floating bridge gave eastward access not only to its own substantial hinterland, but also to the entire upper Wei valley. No forts guarded the mountain road from Didao to Weiyuan, so a successful crossing of the Tao would leave the valleys all the way to Tianshui open to Muslim attack.97
In addition to the Mufti militants, Ma Yonglin’s supporters enlisted the aid of armed bands among the Mongolic-speaking Muslims of Hezhou’s hilly eastern prefecture, the people now called the Dongxiang minzu by the Chinese government. Those Muslims numbered 150,000 in the 1953 census, far more than the Salars, but their villages lay scattered over a wider territory.98 That as late as the 1940s many of these Muslims (who call themselves Santa) had not yet learned Chinese confirms their relative cultural isolation. Of their folkways we know little, except that they treated Muslim guests with extreme politeness and had, according to Chinese sources, a fanatical loyalty to their local leaders.99 A Japanese traveler noted that, unlike most Mongols, they did not drink any alcohol, indicating strong adherence at least to that prescription of Islamic law.100
Both widespread fear of violence from the Muslims of the Dongxiang and their reputation for lawlessness derived, at least in part, from their physical setting, in the poverty-stricken hill country east of Hezhou, west of the Tao. Their markets lay either at Hezhou or across the Tao on the “non-Muslim” east side. So in order to survive, they depended on the river crossings at Tangwangchuan and Didao. If the non-Muslims who occupied the east bank of the Tao closed the ferries and floating bridges, the people of the Dongxiang could not travel eastward for marketing or for seasonal work, on which they depended for income to supplement their meager hillside agriculture.101 Much of the violence associated with these Muslims revolves around those crossings and the roads leading to them.102 The extreme poverty of the hill country also played an important part in their “antisocial” inclinations.103 We do not know to what extent their Mongolian cultural heritage or language divided them from Chinese-speaking or Turkic-speaking Muslims.
The gazetteer account of the battles around Didao contains further evidence that local Muslims fought local battles in 1895 and confirms the importance of the Qing monopoly of substantial walled cities in defeating Muslim attempts to seize the military initiative. Besieging a fortified city is never easy without artillery; the mobile, often mounted Muslim soldiery found it well-nigh impossible.104 The inhabitants of Didao town could thus resist a Muslim attack, but the bulk of the non-Muslim population lived in smaller, less protected places, so the prefect requested funds from Lanzhou to raise local braves for resistance. The local gazetteer describes the Didao war, which lasted from mid-June into August, 1895:
The Mufti rebels . . . privately made banners and uniforms.105 They arranged with Tongzhi period ex-rebels who had escaped punishment to entice thieves and bandits from the Dongxiang to attack Siwashan fort. Altogether they killed and wounded over six hundred Han. They united to force the Shikangcha and Songmingyai stockades, then they surrounded and attacked Jinggucheng [all on the west side of the Tao opposite Didao]. Xining and Hezhou were thoroughly alarmed at this time.
On June 29 they crossed the river, gathering at the lumber mill. They surrounded the city for more than a month, not resting day or night and killing many. The prefect, . . . the clerk, . . . and local gentry . . . did their utmost in defense, so the rebels failed. The rebels heard that General He Jianwei was coming to the rescue, so they cut the floating bridge and escaped in the night. When [He] Jianwei arrived, he ordered his army to advance and relieve Jinggu. The rebels besieging the town congregated at Gaojiaji. Then Imperial Commissioner Dong Fuxiang was ordered to advance over the river and attack the rebel horde at Gaojiaji and annihilate them. . . . [Qing troops] destroyed the gongbei and moved the miscreants to Yongdaojiazhuang, ordering that they not be allowed to resist the Han or again establish menhuan. The case was closed.106
The composition and outcome of the Didao uprising strengthened the Qing and local perception that menhuan and gongbei caused the Muslim troubles in Gansu. Both local and national Qing officials had numerous precedents for considering sectarian solidarities to be inherently dangerous, especially when they espoused a religious ideology that could be branded subversive. Though Islam was not in itself anti-Qing, as the Qianlong emperor had declared, some of its forms could be and were construed in that way by hostile local authorities.107
Old Taozhou. Blocked to the east and north, the rebellion also could not spread to the south, where the large and wealthy Muslim community of Old Taozhou reacted to the violence by uniting with local non-Muslims, both Chinese and Tibetan, in defense. Old Taozhou lies inside the big bend of the Tao River, separated from the rest of Taoxi by mountains and a wide expanse of grassland occupied by seminomadic Tibetans. The town’s economic life depended on the roads that led northward to Lanzhou, via Labrang and Hezhou, and southward to Sichuan and Yunnan. To the south and east, the Tao was unnavigable by manned craft, though some commodities such as lumber could be floated downstream, and only a tenuous floating bridge connected the north and south banks at Choni. To the west lay the towering Tibetan massif, culminating in Amnye Machin, a range twice calculated (incorrectly) to be higher than Everest.108
Like Hezhou, Xunhua, Xining, and other towns in southwestern Gansu, Old Taozhou marks the Chinese-Tibetan ethnic frontier with a Muslim-dominated market, its Muslim community vulnerable between potentially hostile non-Muslims. Community-minded Taozhou Muslims supported thirteen mosques, thirty-one ahong, and five gongbei in the early twentieth century.109 The various Taozhou ethnic groups felt Tibetan political influence much more than Hezhou did, for the headquarters of the Choni tusi, a Qing-appointed, hereditary Tibetan leader, lay only twenty-five kilometers away, on the bank of the Tao. Direct Qing power, exercised from Minzhou and the “new city” of Taozhou, also played a considerable role, so in Taozhou we find a truly tripartite world, one in which non-Muslim Chinese, Sino-Muslim, and Tibetan political and social forces coexisted or conflicted more or less in parallel, without complete domination by one group or another. Certainly both the Qing state and the Choni tusi could exert coercive power, but from a considerable distance, so the Taozhou peoples governed themselves by cooperation among local elites.
In general terms, the Old Taozhou communities performed different socioeconomic functions. The agricultural and pastoral economy lay primarily in Tibetan hands, though Sino-Muslim and non-Muslim Chinese farmers clustered around the market towns.110 Like coreligionists all over Gansu, the Muslims of Taozhou undertook distribution of goods and marketing, including interregional trade. The local non-Muslim Chinese farmed and produced artisanal goods such as metalware and pottery. The Muslim gentry (Hui shen) of Taozhou, like Muslim elites of other Gansu towns, performed both secular and religious functions. Ahong, merchants, successful examination candidates, and military men all partook of this status. By the late nineteenth century, ahong were often landlords, merchants, or even local tuanlian commanders. Their interests thus included both loyalty to Islamic exclusivity and the very real benefits they derived from association with the Qing state and non-Muslim society.111
Physically and socially separated from Muslim centers to the north, delicately balanced between Qing authority to the east and the Tibetans to the west and south, with nomads and mountains on three sides, the Old Taozhou Muslims had a strong vested interest in peace. Located in an ideal commercial middleman’s position, they were correspondingly vulnerable in times of trouble. Their numbers were never sufficient to threaten the Tibetans, nor could they escape the Qing presence. Accommodations, not confrontation, marked their reaction to the violence at Xunhua and Hezhou in 1895. Banditry remained in check that summer, a sign of stable control by local gentry:
In the fifth month of year 21 of the Guangxu reign period [June 1895], the Hezhou and Huangzhong Muslims rebelled. Dissatisfied elements among the Taozhou Muslims had started rumors, stirring up trouble and inciting disorder, taking a challenging position and wildly rampaging around. Wild stories circulated among the Han . . . who panicked and fled. People’s spirits were agitated, and the markets were empty. But the Han and Muslim gentry . . . held to the correct way and did not fear.112
The multicultural gentry managed to keep control, despite the provocations of local rowdies. When a Hezhou Muslim led his fighters across the grasslands in pursuit of fleeing Qing troops, the gentry-led locals captured and executed him. Set in its defensive posture, Taozhou saw no major outbreaks of violence in 1895.
The Limits of Violence
We have now seen the spark of violence fall on Xunhua, Hezhou, Haicheng, Didao, and Taozhou in 1895. At Haicheng local troops quickly ended any threat from a large-scale jailbreak. At Didao the non-Muslim majority and Qing officials acted expeditiously, with the help of outside troops, to blunt a rebellious thrust by a local Muslim group (Mufti) and its Dongxiang-based allies. At Taozhou Old City nothing happened at all beyond rumors and local agitation, put down by the gentry. Only in Taoxi and Huangzhong, centered on Hezhou, Xunhua, and Xining, could the anti-Qing Muslims undertake concerted violence. Unable to take major cities, they either kept their forces moving or based themselves in smaller fortresses surrounding the Qing garrisons.
After his subordinates failed to relieve Xunhua, Deng Zeng himself led a column from Xining, via Bayanrong, in early June. The Salars could not concentrate enough force on the north or west sides of their siegeworks around Xunhua to stop his column, which reached the town on June 12, ending the two-month siege.113 Hann Wenxiu, who had emerged as a central Salar leader during the siege, fled across the river to the northern gong. Had Deng been able to continue his advance, catching the remaining Salar forces between Lei Zhengguan’s Hezhou garrison and his own troops, he might have ended the Xunhua threat. But Deng realized that the violence could easily spread to his base at Xining, so he returned to Bayanrong rather than penetrate more deeply into rebel territory.
Back at Hezhou, Ma Yonglin and other anti-Qing commanders ordered continuous assaults on the city wall and on non-Muslim villages in the immediate hinterland. Throughout June and July small forts and stockades south of the city fell to the Muslims, who slaughtered the defenders in thousands and took their land. The Chinese chronicles extol the heroism of the defenders of Ninghepu (now Hezheng County), where seven corps of militia drove off the Muslim attackers, made some successful forays beyond their walls, and neutralized a large besieging force.114 During July the Xunhua leader Hann Wenxiu brought his fighters to the Xining region and led them, with local Muslims, to take forts in the northern and southern valleys that converge on that city. From Didao to north of Xining, Gansu burned, but the rest of the province lay relatively quiet, except for the movement of Qing troops and supplies toward the disturbances.
Governor-General Yang Changjun planned to relieve Hezhou by using Tang Yanhe’s army, still stationed south of the city, in conjunction with a sortie by Lei Zhengguan’s garrison inside the walls. Tang, whose Muslim subordinates Ma Fulu and Ma Guoliang had held Jishi Pass against the Salars, moved north via Laoyaguan in early August. According to several accounts, Tang was persuaded by a promise to surrender and a bribe from Muslim leaders to remain at Shuangcheng for the night, where his army suffered an ambush and scattered in the face of the enemy.115 Lei Zhengguan, shut up in Hezhou, could not move against the besiegers without the other half of the pincers, so the slaughter continued in the countryside.
Ma Yonglin and the other Hezhou Muslim leaders found themselves with two conflicting objectives. They needed to take Hezhou, consolidate their base, and eliminate the Qing armies and mintuan stockades that still threatened their local security. But they also needed to break out of Taoxi, either toward Xining or north and east, to find allies in other Muslim communities before larger Qing armies arrived to “pacify” them. In the end they achieved neither objective. Ma Yonglin did traverse the Dongxiang to cross the Tao in early August, but his invasion was easily turned back by Qing cavalry under the redoubtable Ma Anliang, who remained in the field to lead the Qing forces in counterattacking toward Hezhou. However deeply Ma Anliang may have felt solidarity with other Muslims, it certainly did not extend to Ma Yonglin, who had plundered Ma Anliang’s home base.116
In August 1895 Dong Fuxiang was in Beijing as a Xinjiang representative at the empress dowager’s sixtieth birthday party. Concerned about the Gansu violence and impressed by the powerful Manchu minister Ronglu’s recommendation of the young soldier, the court sent him into the field once more against the Gansu Muslims. A central figure in northwestern politics and internationally known for his role in the attack on the Beijing legations in 1900, Dong Fuxiang has been confused with Deng Zeng,117 erroneously called a Muslim,118 and maligned by Europeans as an arch-criminal (see plate 16). He first gained prominence during the Muslim rebellion of the 1860s, when he and his father formed a group of local partisans to defend their home county, the son finishing the war as a colonel in the Qing army commanded by Liu Jintang.119 Ma Anliang, Ma Fulu, and other important Hezhou Muslim officers enlisted in Dong Fuxiang’s army after Ma Zhan’ao’s surrender, hoping for advancement within the Qing military and for social order with a measure of local autonomy for the Sino-Muslim elite.120
Coming from Beijing to eastern Gansu in 1895, Dong brought something new to the field against the Muslims, the greatly increased firepower of European small arms. Though Dong’s troops had not been trained in the European fashion and still resembled the anti-Taiping armies of the mid-nineteenth century, they were armed with Remingtons and Mausers. Dong used these new weapons, vastly superior to the Muslims’ muzzle-loaders, halberds, and swords, to great effect.121 Never before had the Hezhou Muslims been forced to recognize the superiority of Qing arms. Their experience in the 1860s had led them to believe that the Tao River, the mountains of the Dongxiang, and their own mobile, stout-hearted cavalry would serve to block any invasion.
Dong’s successes were not only military. He also used localist politics to halt a movement of Muslims from Ningxia to southwestern Gansu. This episode provides a measure of evidence for contact and cooperation among Muslims all over Gansu, but it must be examined carefully. It took place in the fall of 1895, around harvest time.
When [Dong] Fuxiang arrived at Pingliang, . . . [two local literati] of Jinjipu came to welcome him there. Fuxiang asked them, “How are the fall crops?” They replied, “All harvested, with only the Muslims’ fields yet unharvested.” Fuxiang was surprised and asked, “Where have they all gone?” “Most of them have gone to Hezhou.” Fuxiang said, “I shall send people to urge them to return home.”
After the gentry left, he commanded his clerk quickly to prepare a proclamation to be printed in thousands of copies. . . . In brief, it stated: “At the moment, Hezhou has rebelled and Xining reacted. I, the commander, have received the emperor’s command to take forty battalions of horse and foot [soldiers] to attack Hezhou and Xining. You have all listened to and been misled by rumors and run off, leaving your ripened crops to the birds. Hezhou is in the south, Ningxia in the north, as different as horses and cows. The court clearly differentiates good [people] from evil [people]. Could it possibly exterminate you all without exception? Within one month you must return, each to his farm, never again to be frightened out of your wits in this way,” etc., etc.
He chose three of the corporals of the guard who were skilled in rhetoric and gave each a platoon. From Pingliang they went to Guyuan, posting copies of the notice as they went. At large villages, they would exhort people. . . . Those among the Muslim braves who returned to their homes were registered by name, age, and residence and the official list sent first to Ningling subprefecture. Then the provincial capital was notified. When Yang Changjun saw the Ningling report, which said that the Muslims from this district who returned from Hezhou numbered over ten thousand, he said to a provincial official, “Dong Xingwu [Fuxiang] has not even advanced on Hezhou yet, and Hezhou’s flank army has disappeared.”122
Far more effective than Yang Changjun’s earlier warning, Dong’s traveling lecturers actually thwarted the Ningxia reinforcements, returning them to their unharvested fields. They were frightened by Dong’s army and reputation, giving up their expression of Muslim solidarity in the face of threatened Qing repression. The Muslims may also have been persuaded by Dong’s evocation of localism, for he was a local man himself and would have chosen the most convincing arguments. This incident, evidence for Muslim intercommunity cooperation, nonetheless had no effect on the southern Gansu violence. Some combination of Dong Fuxiang’s skill, Muslim fear, self-interest, localism, and perhaps religious rivalry prevented the northern and southern Gansu Muslims from uniting their forces against the Qing. The Ningxia Muslims made no further concerted efforts to aid the anti-Qing Muslims at Hezhou. Dong Fuxiang set up his headquarters at Anding, Zuo Zongtang’s old base east of Lanzhou, and prepared to invade Taoxi.
His first objective was to cross the Tao and break Ma Yonglin’s siege of Hezhou, relieving Lei Zhengguan’s garrison. As Dong moved toward the river, the Hezhou besiegers renewed their efforts to take the city, for they could not concentrate their forces to repel Dong’s invasion if Lei Zhengguan’s Hezhou garrison remained in their rear. They failed. Beginning on November 17, 1895, Dong killed thousands of Muslims at Kangjiayai, crossed the Tao at Didao, relieved the siege of Jinggucheng, and struck across the Dongxiang. Ma Anliang’s cavalry crushed Muslim resistance on the slopes of Oxheart Mountain and broke the Hezhou siege in a week of heavy fighting, relieving the town on December 4. Over four hundred Muslim leaders were captured and executed—Ma Weihan, the leader of the Mufti from Didao, died with Ma Zhan’ao’s old subordinate Zhou Qishi, and Ma Yonglin’s entire family suffered the extreme penalty. The Hezhou uprising, which had cost tens of thousands of lives and the devastation of Bafang, ended in a complete victory for the Qing state. With the Muslim general Ma Anliang in charge, the Qing could be sure of southern Gansu for the remaining fifteen years of its rule.
Xining. Like Lei Zhengguan’s defense of Hezhou, the Xining garrison’s successful resistance against Muslim besiegers kept another large Qing army intact and active deep in “Muslim” territory. Xining, like Old Taozhou and Hezhou, stands on the frontiers of several ecologies, cultural groups, and political-military powers, astride the main road between Tibet and Mongolia, a road crucial to the political and religious histories of both peoples. It lay at the western end of Qing administrative penetration into Amdo, and its Muslim trading community occupied the cultural middle ground between China and Tibet.123 Xining’s city wall, one of the most imposing in premodern China, towered over a wide plain at the confluence of three rivers. Its garrison could dominate both the local hinterland (Huangzhong) and the essential caravan routes that passed nearby.124 If the Qing empire could control Xining, it could maintain its domination over the politics of Amdo, the region’s communication with the Mongols, and the commerce of its highways. But the roads from Lanzhou and the Gansu corridor to Xining, the city’s lifelines to the world of China, passed through several steep, narrow passes, places at which a small defending force could easily prevent passage by a much larger one, rendering the city’s position tenuous in times of trouble.
Many aspects of Xining’s socioeconomic life showed marked blending of Chinese and Tibetan ways. Xining Chinese built Tibetan-style houses, but local wedding customs strongly resembled those of China proper.125 Despite their local accommodations, many Xining Chinese saw, and continue to see, the Tibetans as a primitive, inferior, uncultured people, separated from civilization by their barbaric character but useful as mercenaries and auxiliaries in frontier warfare. Xining floated on a sometimes restless sea of non-Chinese people, who put up with Qing (or any other external) hegemony to the extent that it profited them or the state could preserve its prerogatives with troops. Around the turn of the twentieth century many groups lived in significant numbers both in the suburbs and in the hinterlands—Chinese, Tibetans, Mongols, Monguors, Sinophone Muslims, and Salar Muslims. No simple mapping, however, could do justice to the intricate intermixing of groups throughout the region, for they often did not live in clearly delineated areas but rather in adjacent villages or neighborhoods.
Xining’s high walls faced those of the Muslim suburb, Dongguan, only a few meters distant. The valleys leading to the city bristled with forts held by both Muslim and non-Muslim militias. Datong, Bayanrong, and other county towns could also be easily surrounded by insurgents based in hilltop stockades. In such an atmosphere, the Qing officials both regularly and brutally used military force against the locals. All the communities armed themselves for defense of their lives against their neighbors and/or the state. Through the summer of 1895 local tuanlian and government troops reacted to the escalating violence around Hezhou and Xunhua: Muslim residents of Xining were subjected to searches on the road, exclusion from the city walls for fear of spies, and other official and unofficial discrimination. As the Xunhua Muslim leader Hann Wenxiu raided closer and closer to the city, many suburban Muslims joined him, and Muslim forces attacked stockades and fortified villages throughout Huangzhong.
In these “little wars” around Xining, communal or religious identity functioned very much like lineage identity in Chinese society, creating one vector of unity for villages or groups of villages, a boundary against the dangerous Others. In Philip Kuhn’s terminology, the simplex tuan of many Xining Muslim villages engaged in vendettas (Ch. xiedou) with non-Muslim villages and with Tibetans.126 The funds to buy arms and provisions for such activities (which, in other parts of China often were provided by lineage organizations) in Muslim Gansu came from the communal treasury of the mosque, to which all were expected to contribute, and from the wealth of successful commercial families. They did not, however, attempt to overthrow the Qing state or found any type of autonomous Muslim polity, for their ability to organize rarely exceeded the confines of their villages or multivillage local systems. The spatial arrangement of Muslim and non-Muslim communities, mixed and limited by the harsh terrain, did not encourage regional organizations, nor did the division of Xining Muslim society by the various religious solidarities. Indeed, as we have seen, fighting between Muslims could be as vicious and enduring as Muslim vs. non-Muslim battles.
Tension rose very high as Salars assaulted Deng Zeng’s own camp near Bayanrong and took commander Li Shijia prisoner at Jingningpu. Deng’s army, positioned primarily to fend off Salar violence emanating from Xunhua, could not prevent local Muslim raiding even within sight of Xining’s walls. On September 1 armed Muslims mounted the Dongguan walls to begin a bombardment of Xining that lasted several days.127 Deng Zeng feared the worst—capture of the main citadel—and brought his entire army back to the city, where he met the Muslims head-on. A major battle ensued, with heavy losses on both sides but no clear victory.128 Though the Xining garrison ventured out to take the Muslims in a pincer movement, they failed, so Deng entered the city but could not break the siege. The Xining Muslims’ decision to attack the city thus followed the pattern of Hezhou and Didao. Local Muslims, fearing an assault from apparently bloody-minded local tuanlian or from the Qing authorities, struck first in the hope of securing a quick victory and control over the walled town. At the roots of this violence lay local emotions and fears, not any grandiose dream of Muslim hegemony based on religious ideology. The weakening of Gansu’s Qing garrisons by withdrawals to fight the Sino-Japanese war encouraged some Muslim leaders, and their motivation to fight overrode their knowledge of the local garrisons’ strength and the height of the walls.
Battles continued all over Taoxi-Huangzhong in September, with the Qing garrisons and local tuanlian consistently holding the advantage through their control of stockades such as Hejiapu, across the river from Didao on the western, “Muslim” side of the Tao; Baitasi, at a crucial Yellow River crossing north of Hezhou; and Ninghepu, south of Hezhou. Deng Zeng and Lei Zhengguan held out in stubborn defense of Xining and Hezhou, as Qing reinforcements began gradually to arrive from Shaanxi, Ningxia, and Xinjiang.129
Overwhelming Force at Xining
The Xining Muslims, scattered in dozens of villages and divided by Qing garrisons and tuanlian, established a headquarters and bastion at Duoba, west of Xining City. To this fortified market town, with its large Muslim population and many mosques, came fighters and villagers who feared to remain in the open. The Duoba leaders employed their fortress as a staging area for attacks against Xining and other enemy strong points. When Li Peiying tried to join Deng’s defense, the Muslims blocked his relief force for over a month at the narrow passes leading to the city, despite Deng Zeng’s efforts from within the citadel.130
Deng, waiting for reinforcements and trying to contend with chaotic conditions in Xining’s hinterland, enlisted the aid of the Qinghai Tibetans as Agui had done at Lanzhou in 1781. Eager for loot and a chance to strike at the Muslims, the Tibetans joined with alacrity. This alignment—Qing troops, local non-Muslim tuanlian, and Tibetans all attacking local Muslims—deepened hatred and rivalry between Tibetan and Muslim common folk, despite their ongoing commercial association in the wool trade. The enmity often burst into brutal, sickening violence during the next few decades of disorder. Tibetan participation in these wars demonstrated yet again to local gentry and Qing officials the wisdom and efficacy of using one group of frontier people to fight another.
Chinese people have generally seen Tibetan culture as inferior to their own, but the Tibetans maintained their language, folkways, religion, and economy remarkably well in the face of their self-confident antagonists.131 Only Tibetans could live and produce wealth in the highlands, across which no one could fight as effectively; for this skill they were granted a ferocious reputation as killers and robbers. The experience of Chinese, European, and Muslim travellers who ventured into their territory partially justifies this fame.132 With some exceptions, Amdo Tibetan society, especially its religious institutions, remained relatively independent of outside control even in areas of Qing or Muslim power, or Chinese cultural influence. The lamaseries held much of the local secular and economic power on the grasslands, often greater than that of the Qing state, which bought partial Tibetan compliance by assessing lower taxes than those of nearby non-Tibetans. In addition, some Tibetan leaders—the tusi families—received hereditary patents of office from the Qing.133 This cultural and political combination resulted in many Tibetans’ learning some Chinese and adapting in part to Chinese culture, the while remaining clearly Tibetan.134 In this complex Muslims played the economic middleman, making available the products of agricultural China and the flatland artisans to the Tibetans, who in turn sold them pelts, wool, and other highland products.135
Beyond trade and productive activity, the Tibetans played a role in the military affairs of Gansu. When the tusi called for troops, every family had to respond. The lamaseries, too, had extensive arsenals, and the monks organized militias and stockpiled arms to defend their temples. The brutality and extent of conflicts along this frontier, from the Tao valley to Xining and Kökenōr, illustrate painfully the fear and hatred that sometimes characterized communal relations in Gansu.
The lamaseries concentrated wealth and political power. Contributions flowed in to support the monks and rimpoche, whose prayers and pilgrimages spiritually maintained the Tibetan world. The most important rimpoche at Labrang, for example, sat on a golden throne, surrounded by precious objects and clad in jewel-encrusted robes in stark contrast to the poverty of the pilgrims praying in his courtyards. Scholars went out from the great lamasery at Kumbum, near Xining, to propagate and strengthen religious practice and allegiance all over Tibet. Their missions also maintained the political power and relationships of the rimpoche and their agents, who were often their regents.136 Using their long experience in dealing with the Qing state and many cultural Others, the Tibetan leaders chose friends and enemies with care. In 1895 they correctly allied themselves with the Qing. They called up their cavalry in the Xining area, operating both independently and under Qing command against the Muslim insurgents.
Early in the fall of 1895 Qing victory had seemed imminent. Wei Guangtao was on his way with a large force of the Hunan Army, Zuo Zongtang’s old command; Dong Fuxiang had taken Hezhou and was free to strike through Salar country toward Xining; Niu Shihan had finally left Ningxia to assist in the final victory. But the bitter Gansu winter interfered, as did tenacious Muslim resistance. The Muslims knew that submission meant violent death at the hands of the imperial armies (including Muslim commanders such as Ma Anliang) and their own non-Muslim neighbors. After halting Niu Shihan east of the Xining River gorges, the Salars had offered to surrender on terms favorable to the Qing. But the Qing commanders, many of them Gansu men, felt too strongly about Muslim depredations and the necessity for thorough “pacification.” They intended to make no mistakes like those of Zuo Zongtang, who had left too much power in Muslim hands.
So Qing forces converged on Xining from the north, east, and southeast. Wei Guangtao’s Hunan braves fell heavily on the Muslim villages, and Dong Fuxiang’s troops joined in the despoliation, but the Muslims still held out. As late as mid-February, even with Qing troops stationed in the villages, Deng Zeng and his commanders were not able to prevent non-Muslims from being ambushed and slaughtered by Muslims in broad daylight, within sight of Xining’s walls. Nonetheless, as the weather improved, so did conditions for infantry and cavalry assaults, supported by the few Qing artillery pieces. Finally, in March, the Duoba Muslims, their numbers greatly reduced through unsuccessful sorties and depleted food supplies, surrendered unconditionally. Their leaders lost their heads, and the common soldiers dispersed to their ruined, empty villages. Thousands of Muslims tried to follow the tracks of Bai Yanhu west toward a Muslim refuge. Unlike Bai’s veterans in the 1870s, they never reached Russian territory but were hunted down by pursuing armies, frozen by the late winter of the north Tibetan marches, and starved by the barren wasteland. Finally captured and relocated near Korla, in Xinjiang, fewer than two thousand survived.
Lessons Learned
The violence between some Muslims and Qing loyalists and allies of various cultures had devastated a fourth of a province, killed tens of thousands, and sustained the worst fears of all. The Qing officials retained their conviction that religious disputes among the Muslims, particularly those caused by the New Teaching (whatever that may have been) and the menhuan, led to conflagration and bloodshed.137 The local non-Muslim Chinese verified their tradition, that their Muslim neighbors were bloodthirsty fanatics. The Muslims confirmed that they were to be discriminated against, excluded, even slaughtered, because they were Muslims, and thus different. The total Qing military victory did not bring lasting peace, for even brutal pacification could not eliminate communal mistrust and hatred. As long as the Muslims and non-Muslims remained neighbors and state authority could not maintain local order, the problems would continue to exist.
The events of 1895 also illustrated again the disunity of the Gansu Muslims, their vulnerability to piecemeal attack. Official and local fears still emphasized the ferocious loyalty Muslims supposedly felt for one another, but the Muslims could not actually unify their communities in action. Despite efforts by the Xunhua, Hezhou, and Xining Muslim militants to call for a general uprising, many Gansu Muslims remained aloof or actually opposed their coreligionists. Between the pacification of Gansu in 1872 and that of Hezhou in 1895, a number of Muslim leaders faced realistically the Qing capacity to control its peripheries. Adjusting to that capacity, most of them became Qing commanders or allies in order to preserve their local power and, as they thought, the very existence of their communities.
Poised on the frontier of cultural China, resisting or adjusting to the changes occurring in China, Gansu folk also struggled to remain themselves. Violence from their neighbors or their government often threatened their lives and property, so hatred and fear dominated their memories and their decisions in time of trouble. When Chen Jiaji killed eleven Old Teaching Muslims and displayed their heads at Xunhua, many Muslims felt compelled to act in self-defense and vengeful anger. Their poverty had not changed, nor had their harsh environment, but Qing pressure had grown more insistent, more “modern” in its demand for incorporation and in its enforcing power. The Muslim communities, too, had changed, for most of the menhuan, originating as revivalist Sufi institutions, had evolved into essentially conservative local solidarities, organized to preserve the community power and wealth of important lineages.
Neither Ma Anliang, the Qing cavalry commander, nor Ma Yonglin, the disrupter of social order, fits into the standard Qing categories of Muslim villainy. Neither was a “New Teaching bandit” or an anti-Qing state-building holy warrior. Ma Yonglin failed because his coalition of anti-Qing and anti-reformist Muslims could not organize widespread, protracted resistance to state power. Ma Anliang succeeded because he served the Qing loyally against some of his Muslim neighbors, bringing new weapons to frontier conflicts. His reward, like his father’s, was local authority, albeit closely controlled by Qing officials, an authority that evolved into Gansu’s unique system of Muslim warlordism in the twentieth century (see plate 17).138 The “Ma family warlords” (Ch. Majia junfa) represent a new development in the Muslim worlds of northwest China, the modern nation-state’s incorporative power expressed locally through cooptation of existing elites. That process is evident in the following narrations of the lives of four individual Gansu Sino-Muslims who followed very different paths, all of which led toward a closer relationship with New China.
1. Ma Xuezhi, Zhehelinye daotong shi xiaoji, cited in Yang Huaizhong, “Lun shiba shiji,” 320–21.
2. Elsewhere I have argued (unfortunately without sufficient care regarding words such as “ethnicity”) that local feuding and street violence, almost always undertaken by young males, must be understood at least in part as a stimulating, enjoyable activity (Lipman, “Ethnic Conflict,” 79–81). Michael Ignatieff, traveling in Croatia, found the combination of young males, guns, ethnicity (including religion), and alcohol particularly potent and dangerous in creating an atmosphere of murderous violence (Blood and Belonging, 3–5).
3. For a recent investigation of their origins, see Mi, Salazu, esp. chaps. 1 and 2.
4. Trippner, “Die Salaren,” 241–76, esp. 241–50. According to a legend quite general in the sources, the Salars migrated eastward guided by a white camel and carrying a bag of Central Asian soil. They settled around Xunhua, where the local earth matched that in the bag, and the white camel transformed itself into a prominent white boulder.
5. Population statistics for the Salars, like those of any group in China, must be evaluated with great care. The Xunhua ting zhi of 1844, 4.28a-31b, gives 2,780 households for the eight main Salar communities on the south side of the Yellow River. This would indicate a much smaller population than the thirty thousand Salars counted by the newly established People’s Republic in 1953. But there were five other Salar communities on the north side of the river, outside Xunhua Subprefecture, as well as many Salars residing elsewhere or living on the road. We also have no way of knowing how Salar identity, or any other ethnic identification, was determined in the 1953 census, not to mention the censuses conducted by the Qing and Republican governments.
6. Trippner, “Die Salaren,” 261. The late Rev. Claude Pickens, Jr., a missionary with long experience among Chinese Muslims, traveled to the northwest in 1936. In response to an early draft of this book, he wrote that he deplored the overemphasis on Muslim banditry, calling it a Han perversion of the truth. The Muslims had received this characterization, in his opinion, without earning it, victimized by ethnic prejudice and Han control over the written record.
7. Zhou, Qinghai, 185.
8. “Sara Kai oyobi Mōko Kaikai,” 76. Arienne Dwyer, who has done the most extensive contemporary work on their language, informs me in a personal communication that the Turkic elements in the Salar language most closely resemble the Turkmen language, from west of Samarkand.
9. Monguors, or Tu people (lit. “indigenes, autochthones, people of this earth”), occupy much of the Huangzhong triangle. They live in lineage settlements. Their language, folkways, history, and ethnic interactions have been studied by Louis Schram, who lived among them for many years. Though biased in their favor, his monograph (Monguors) contains a wealth of information on the border region as a whole.
10. Arienne Dwyer’s as-yet unpublished fieldwork indicates that the ethnocultural distinctiveness of the Salars is rapidly disappearing under the acculturative pressures of the Chinese-dominated state and both Chinese-speaking and Tibetan-speaking neighbors. She believes that their Turkic language, already replete with Chinese and Tibetan grammatical and lexical influences, will not survive another generation.
11. Andrew, The Crescent, 17.
12. Yang Huaizhong, “Lun shiba shiji,” 311–16.
13. Nakada, Kaikai minzoku, 88–89.
14. GNQSL 18.47a–b.
15. The surname Hann, belonging to the leading families of the Salars, romanizes a Chinese character different from but homophonous with the ethnonym Han.
16. The ballad may be found in its entirety in Wang Shumin, “Qianlong sishiliu nian.” It has been reprinted in several anthologies, including ZYSCZX, vol. 1, 802–19. Gu Jiegang included an assessment of the Muslim problem, in embryonic form, in his diary of his trip to the northwest, Xibei kaocha riji. His views on non-Chinese minorities within China may be found in Schneider, Ku Chieh-kang, esp. chap. 8.
17. These pifazi, a mode of transportation uniquely suited to the rapidly flowing rivers, bulky cargoes, and easily available materials of the northwest, have been the subject of much curiosity among foreigners and non-northwestern Chinese, nowadays appearing in many travel brochures and videos as a noteworthy tourist attraction of Gansu. For sources on their history, see Iwamura Shinobu, “Kōga jōryū.” Both Moore, “Raft Life,” and Köhler, “Die Bedeutung,” offer excellent pictures of the rafts.
18. The copious memorials and reports regarding this bloody uprising and its repression may be found in Qinding Lanzhou jilue. Agui’s participation is particularly noteworthy, and his own literary and official descriptions may be found in Nayancheng, A Wenchenggung nianpu (1813), from 22.43a, where the first emergency memorial from Lanzhou is reported, through Agui’s final reports, ending at 25.18b. Some Sino-Muslim historians proudly note that none of the “bandits,” as the Qing called them, surrendered, choosing instead to die heroically, while other scholars find this self-sacrifice for religion foolish.
19. Saguchi Tōru, “Chūgoku Isuramu,” 79–80; and Schram, Monguors, vol. 3, 64. For one perception of Tian Wu’s vengeful motivation, see the Ping Hui jilue, an anonymous account of Muslim violence in the northwest written by a minor Gansu official, in HMQY, vol. 3, 12.
20. Local sources in Xinjiang, where Ma Mingxin’s wife and daughters were sent in exile, record that her surname was Zhang, and that she came from the same town as Zhang Wenqing in Tongwei County (Wu Wanshan, Qingdai xibei, 40, n6).
21. The official compendium of campaign documents, authored primarily by Agui and his colleagues, the same group who succeeded against Su Sishisan, was published as the Qinding Shifengpu jilue. A secondary account may be found in Wu Wanshan, Qingdai xibei, 39–52.
22. An Weijun, Gansu xin tongzhi 17.3.
23. Wu Wanshan, Qingdai xibei, 50.
24. Fletcher, “Naqshbandiyya,” 34.
25. As noted above, Guanli Ye’s entire text has now been translated into Chinese. Zhang Chengzhi, “Kakusareta Chūgoku,” is a brief description and analysis of the text and its importance. The book’s description of the Jahrīya silsila agrees with that of Aubin, “En Islam Chinois,” though they give different Arabic names to the second shaikh, Ma Mingxin’s immediate successor.
26. There have been many studies of the Jahrīya and its transmission after Ma Mingxin. All of them cite Shan Huapu, “Shaan-Gan jieyu lu,” an oft-reprinted source, as an authority. The most important Euro-American version is Aubin, “En Islam Chinois.”
27. Wu Wanshan, Qingdai xibei, 48. The emperor disingenuously neglected to mention the financial scandal in Gansu that had only recently been cleaned up. He can hardly have forgotten such a major law case, which resulted in the execution or severe punishment of dozens of provincial and local officials for some of the crimes he mentions here.
28. These epigraphs are taken from Chu Wen-djang, Moslem Rebellion, 4; Liu and Smith, “The Military Challenge,” 217; Chen Chongkai, “Cong Shaanxi Huimin,” 52; Wu Wanshan, Qingdai xibei, 144; and Raphael Israeli, Muslims in China, 204.
29. Shao Hongmo and Han Min, Shaanxi Huimin. This vocabulary, of course, does not originate in the 1980s but descends from the revisionist work on structures of power in traditional society, especially peasant rebellions, done by many twentieth-century Chinese scholars. It finds obvious expression in works such as Bai Shouyi’s Huimin qiyi, the compendium of primary sources compiled in the early 1950s.
30. Gao Wenyuan, Qingmo xibei.
31. Israeli, Muslims in China.
32. C. K. Yang has studied the “mass actions” reported in the Veritable Records of the Qing for the years 1796–1911 and reports more than six thousand of them (“Some Preliminary Statistical Patterns”).
33. Schram, Monguors, vol. 3, 66, notes a number of local disturbances, some of them sizable, in Huangzhong between 1859 and 1862. Communal relations had clearly deteriorated all over the region, and the vast violence taking place further east, with its consequent channeling of Qing coercive resources away from the northwest, had predictably stimulating effects on simmering local conflicts.
34. Wang Zongwei, “Qingdai Shaanxi”; and Zeng Liren, “Xi Nian jun.” The role of the Taiping, Nian, or other non-Muslim rebellions in causing or stimulating or encouraging the northwestern Muslim rebellions has yet to be elucidated in a nonpolemical fashion. Fletcher, “Naqshbandiyya,” 40, argues that the Taiping invasion of Shaanxi in 1862 touched off communal violence. Curwen, Taiping Rebel, 264, finds that the Taipings intended to take advantage of already-existing unrest. Jen, The Taiping Revolutionary Movement, 470–71, states that the Taiping general Chen Decai linked his army with rebels in Hanzhong, southern Shaanxi. The combined army, led by Lan Dashun, was to be responsible for the death of Dolongga in 1864. Li Xiucheng, however, in his deposition to Zeng Guofan’s staff, said that the Taipings had not been in communication with the Shaanxi or Gansu Muslim rebels (Curwen, Taiping Rebel, 311). Work on the Taiping invasion of Shaanxi is now being done by provincial and local historians in China. See, for example, Wu Wanshan’s explanation in Qingdai xibei, 65–66; he finds that the eastern Shaanxi Muslims had already planned to move against the Qing but used the arrival of the Taiping armies as an opportunity rather than part of a self-conscious alliance. His argument that anti-Qing rather than communal elements prevailed among the rebels ignores plenty of evidence of already-existing local conflict.
35. Chu, The Moslem Rebellion, 25, mentions the bamboo poles. See also Mei, “Stronghold.” As part of an important oral history project in southern Shaanxi, Ma Changshou collected many local tales and versions of “the beginning” of this violence (Tongzhi nianjian, esp. chaps. 1–4). The multiple incidents, local conflicts, and lack of Muslim central leadership in these accounts accurately represent the conditions in Gansu as well.
36. Wu Wanshan, Qingdai xibei, 114, reports violence among Muslims at Tankar, Bayanrong, and Xining, especially between adherents of the Huasi and Mufti menhuan, both suborders of the Khafīya.
37. Chu, Moslem Rebellion, is a study of government minority policy and the military strategy of Zuo Zongtang. With a similar perspective, Lanny Fields has studied Zuo’s “statecraft” (jingshi) orientation in a useful book on the mid-century wars, Tso Tsung-t’ang and the Muslims. The best study of the official documents on the rebellions remains Nakada, “Dōchi nenkan.” An English narrative, with some attention to the interaction of Zuo’s policies and the Gansu environment, may be found in Liu and Smith, “The Military Challenge,” esp. 211–34. All of these are solidly based in the primary sources, but Chu’s book in particular contains errors of interpretation due to its national policy perspective, while Liu and Smith claim, without any evidentiary citations, that Ma Hualong’s command network covered all of Gansu. Scholars in China cited above—Feng Zenglie, Wu Wanshan, Ma Changshou, Gao Wenyuan—and many others, including the dean of Sino-Muslim scholars, Bai Shouyi, have produced voluminous and detailed histories of these events. Primary sources include the massive official compendium of military documents, Yixin, Qinding pingding Shaan Gan Xinjiang Huifei fanglue in 320 juan, published during the Guangxu period; local gazetteers; collections of personal writings from Qing commanders and occasional Muslims; Mu Shouqi’s Gan Ning Qing shilue, 20–24; Ma Changshou’s oral history project, cited above; and accounts by local gentrymen in Shaanxi and Gansu. Many of the Qing sources are listed in Chu’s bibliography, Moslem Rebellion, 207–19.
38. Chu, Moslem Rebellion, 25–27. Some of the primary texts may be found in Gao Wenyuan, Qingmo xibei, 191–203.
39. The geographical spread of the violence may be traced in the Qinding pingding Shaan Gan Xinjiang Huifei fanglue (see n37) and other contemporary sources. Shan Huapu has done a preliminary survey (“Shuo Shaan-Gan”). In areas of endemic conflict, we must assume that villages were organized over a wide area, constituting what Phillip Kuhn has called “multiplex tuan,” unless geography made complex structures impossible. Certainly many non-Muslim areas possessed subcounty military units, some with formidable stockades, against what they perceived to be a constant threat of Muslim violence (Kuhn, Rebellion and Its Enemies, 67–69). The possibility of “extended multiplex tuan” seems somewhat more remote, given the lack of high-ranking gentry and the relatively sparse resources of the northwestern countryside.
40. Ping Hui zhi, in HMQY, vol. 3, 61, relates that Zhang Fei threatened to accuse non-Muslim militia of rebellion if they did not disperse, but this seems to have been extremely rare.
41. We should certainly take note of this sinister name’s resemblance to the “ethnic cleansing” practiced in Eastern Europe and the former Soviet Union after 1989, especially in former Yugoslavia. The sickening violence perpetrated in the name of local purity there certainly found its match in northwest China during the 1860s, 1890s, and 1920s.
42. Ma Xiaoshi, Xibei Huizu geming jianshi, cited in Liu and Smith, “The Military Challenge,” 217.
43. Ping Hui zhi, in HMQY, vol. 3, 62.
44. The Qing continued to refer to the Jahrīya as the New Teaching throughout this period, and they saw it as the main disruptive influence in the northwest. Zuo and his superiors debated whether to proscribe it, allow it to exist under strict controls, or treat it like any other Muslim group. A relatively liberal policy prevailed over Zuo’s objections, and the Jahrīya was not forbidden.
45. These mutinies are generally believed to have stemmed from the activities of the Gelaohui, a secret mutual-aid society that played a major sociopolitical role in Shaanxi, Gansu, Sichuan, and Henan throughout the nineteenth and early twentieth centuries. Many of Zuo’s troops were Gelaohui members, so when he garrisoned the northwest, the secret society extended its influence all the way to Xinjiang through migration and troop movements. See Yang Zengxin, Buguo zhai wendu, 2nd (yi) collection, 3.1a–5a, which describes the presence of the Gelaohui from the Xinjiang provincial governor’s point of view. A secondary account has been published by Kataoka Kazutada (“Shinkyō no karōkai”).
46. Zuo’s apparently contradictory analyses of the causes of the rebellion—that it was the fault of the non-Muslims and that it was the fault of the New Teaching—can be reconciled if we recognize that he had different audiences and different purposes as he described the rebellion to his staff, to his superiors in Beijing, to members of his family in letters, and to posterity in reports written for the permanent record.
47. Here again I use these terms not in reference to nation-states or other “legitimate” authorities, but rather to describe cultural zones.
48. Yixin, Qinding pingding Shaan Gan Xinjiang Hufei fanglue, 83.7b, notes that when Ma Zhan’ao took over Hezhou at the beginning of the disturbances in the early 1860s, he aided local Qing officials in their escape from the town, doing his best to protect life and property.
49. On Zuo’s logistical preparations, see Chu, Moslem Rebellion, chap. 4.
50. Andrew, Crescent, relates a complete Muslim victory. Lattimore, Inner Asian Frontiers, cites the Russian scholar V. Shakhmatov, who agrees. Other sources that confirm this account include at least one Republican-period local gazetteer from the region, Chongxin xian zhi of 1926, 4.50b, and Tian, “Longshang qunhao.” For extensive primary source citations, see Gao, Qingmo xibei, 293–309.
51. Zuo’s first meeting with his victorious enemy’s son impressed Zuo sufficiently that he renamed the young man Anliang, an abbreviation of anfen wei liangmin, “making peace for the good people.” Even decades later, Ma Anliang reminded his own violent and rebellious son of the meaning of that name—their family, beginning with Ma Zhan’ao, had determined to build a safe haven for Muslims in Gansu, with themselves leading the elite under a Qing and then a Republican state aegis (GNQSL 29.23b–24b).
52. See Chu, Moslem Rebellion, 150; Liu and Smith, “The Military Challenge,” 234.
53. GNQSL 24.9a–10a.
54. Ma Zhan’ao has not fared so well at the hands of PRC historians. In time-honored fashion, they judge (pingjie) him severely for his collaboration with feudal forces. Though approving of his surrender, for it saved tens of thousands of lives, they find him and his descendants to have become willing servants of the oppressors and the founders of the exploitative, semifeudal Ma family warlords. For a succinct statement of this judgment, see Li Songmao, “Qing Xian-Tong.”
55. Cited in Liu and Smith, “The Military Challenge,” 235. Though Qing policy makers had worried about collusion between Gansu and Xinjiang, the Shaanxi and Gansu anti-state forces and actions were overwhelmingly local throughout the nineteenth century. Xinjiang’s various rebels, most of them Turkic-speaking and well connected to Muslim centers west of the Pamirs, rarely made any connections with their predominantly Chinese-speaking coreligionists east of Hami.
56. That objective, attempted only once in a Chinese-speaking part of China, belonged to Du Wenxiu and his “State that Pacifies the South” (Ch. Pingnan Guo) in Yunnan, an entirely different environment in which Muslims rebelled against the Qing in the 1850s and 1860s. Only a few direct links have been conclusively documented between those two Muslim anti-Qing movements, though a large literature on the Yunnan rebellion has been produced by Chinese, Japanese, and Euro-American scholars. A recent doctoral dissertation from Sweden makes an important contribution to the sociocultural background for a study of the Yunnanese Muslims (Wang Jianping, “Concord and Conflict”).
57. Chu, Moslem Rebellion, 149–56, describes and tabulates the resettlement program.
58. A large literature on Bai Yanhu and his Shaanxi band’s long flight has been produced in both China and the former USSR, and Svetlana Rimsky-Korsakov Dyer has done linguistic and ethnographic work on the Dungan in the Central Asian republics. Bai was not the only anti-Qing Sino-Muslim commander to flee to Russian territory, but he is certainly the best known. In the judgments of contemporary Chinese historians, the Muslim rebel Bai Yanhu and the victorious Qing general Zuo Zongtang share an ironic juxtaposition. Zuo Zongtang’s brutal victory over the “righteous uprising” of the northwestern Muslim peoples certainly makes him a supporter of reactionary, exploitative, corrupt Qing power, but his recovery of Xinjiang from the separatist Muslim leader Yakub Beg earns him positive evaluation as a reunifier, a restorer of China’s legitimate authority over eastern Turkestan. Bai Yanhu, on the other hand, began in Shaanxi as a righteous rebel, stayed that way throughout his legitimate wars in Gansu, then became a much more problematic and negative presence by supporting the separatists in Xinjiang and then—greatest sin of all—fleeing the homeland into Russian territory. Clearly, the Xinjiang Muslims’ separatist movements and the Shaanxi-Gansu Muslim uprisings take very different places in this historiography. Ma Tong, the most senior scholar of Sino-Muslim studies in Gansu, visited Bai’s descendants in Kyrgyzstan in 1991 to investigate the dual charge that Bai had treacherously betrayed China by joining Yakub Beg’s separatist movement in Xinjiang and that he had also intended to help imperial Russia invade China. He found both allegations unfounded and now maintains an entirely positive evaluation of Bai’s anti-Qing movement (Gansu Huizu shi, 95–97). Bai Yanhu presents no problem at all for Taiwan-based Gao Wenyuan, who unambiguously regards Bai’s long-term, uncompromising, peripatetic struggle against the Qing as heroic (Qingmo xibei, 407–24).
59. Elizabeth Perry analyzed armed peasant groups in Huaibei, Anhui Province, using precisely this method, and discovered fundamental differences between bands that were protective and others that were predatory (Rebels and Revolutionaries).
60. Zhu Chongli, “Tongzhi shiqi,” reminds us that though undeniably ethnoreligious conflicts dominated this period, instances did occur in which Muslims and non-Muslims cooperated against the state.
61. Wang Dagui is discussed at some length in Chu, Moslem Rebellion, 58–59, 65–66, and in the Haicheng xian zhi of 1908, 9.1b.
62. Aubin, “En Islam chinois,” 524–28, lists the masters (shaikhs) of the Jahrīya in a chronological table, with their various appelations, dates, and the locations of their tombs. Ma Tong, Zhongguo Yisilan . . . shilue, 491, constructs a less detailed table.
63. Ma Shouqian, a senior historian at the Central Nationalities University (Zhongyang Minzu Daxue) in Beijing, has evaluated Ma Hualong positively, despite his status as an elite Muslim landlord and ahong, because the anti-Qing struggle he led was fundamentally defensive in nature. Had Ma Hualong taken a separatist stand, like that of Yakub Beg in Xinjiang, he would certainly be viewed very differently by contemporary Chinese scholars (“Qing Tongzhi nianjian,” esp. the final section).
64. Nakada, “Dōchi nenkan,” 91–92, 133–34. Mu Shenghua ahong, of northeastern Gansu, chose to call himself the “king who pacifies the south and restores the Ming” (Ch. pingnan fuming wang), recalling the multiple Taiping kings, Du Wenxiu’s Pingnan Muslim state in Yunnan, and the desire to restore the Ming, which had motivated so many anti-Qing rebels (Ma Chen, “Mu Shenghua”).
65. Encyclopedia of Islam, rev. ed., vol. 2, 538–40. The expansion of Islam’s territory at the expense of non-Islamic territory is seen by Raphael Israeli as a primary motivation for Muslim rebellion in China (“The Muslim Revival,” 121–23), but I find no evidence for this assertion in the Chinese record or in the scattered statements from the rebels. Jihād as a formal legal concept has not been demonstrated in any primary sources from China; rather it seems to be part of a complex of images of Islam—partially Middle Eastern, partially Orientalist—transported to China. Shengzhan, on the other hand, meant a religiously sanctioned war, an “us” vs. “them” communal war, not jihad in its legalistic Islamic sense.
66. For Zuo’s ideology and politics in the context of Qing period jingshi activism, see Fields, Tso Tsung-t’ang, passim.
67. Ma Changshou, Tongzhi nianjian, presents new personal and local data germane to understanding the motivations for violence among both Muslims and non-Muslims in Shaanxi, county by county. Among the yet-to-be-mined sources for this rebellion, the depositions and confessions of captured rebels will certainly give us a much clearer idea of their organization, their motivation (as the rebels themselves presented it, under duress, before execution), and their behavior. Various provincial “literary and historical materials” series have also published reminiscences collected early in this century.
68. Yi Kongzhao, Pingding Guanlong jilue 7, in HMQY vol. 4, 8. For a general summary of conflict within Muslim communities and its relationship to Muslim uprisings, see Gao Zhanfu, “Guanyu jiaopai.”
69. There seems no doubt that the Shaanxi Muslims’ main desire was to go home, and that Zuo Zongtang’s refusal to allow them to do so kept them fighting as long as they did (Luo and Wu, “Lun Qingdai,” 81).
70. Nakada, “Dōchi nenkan,” 150.
71. Chu, Moslem Rebellion, 157–58.
72. This subject, very much à la mode in Chinese studies of the 1960s and 1970s, is reviewed in Wakeman, “Rebellion and Revolution,” which notes the scholarly conflict regarding the Islamic content of Muslim movements (p. 227).
73. The same may be said for the White Lotus, that complex amalgam of Buddhist and other folk organizations falsely conflated by Qing authority into an inherently rebellious, heterodox unity. For the Qing’s reaction to White Lotus activism, see Overmyer, Folk Buddhist Religion, esp. chap. 5. A comprehensive analysis of the transformation of “White Lotus” from a religious to a sociopolitical construct may be founnd in Ter Haar, The White Lotus Teachings. The difficulty of analyzing such religious movements accurately is discussed in Esherick, The Origins.
74. These two Hanns were not related to one another, as far as we know, and neither was a Han, that is, a non-Muslim Chinese. A number of families surnamed Hann held the position of tusi over the Salars under the Qing.
75. GNQSL 24.40a–41a.
76. The title of this section is not intended as a translation of any Chinese terms, though terms such as New-New Teaching (Xinxinjiao) and New-Flourishing Teaching (Xinxingjiao) do appear in Chinese texts. Rather, I intend to indicate that the entities designated as Old and New Teachings in 1895 were not the same as the Old and New Teachings of the 1780s or even the 1860s. That is, the terms have only contextual meaning, no absolute meaning, and must be examined carefully whenever they appear, to discover to what they might refer.
77. Mary Rankin and Joseph Esherick have reviewed the literature and theoretical spectrum on local elites, especially in relationship to state power, in their “Concluding Remarks,” in Esherick and Rankin, Chinese Local Elites, 305–45.
78. Guan Lianji and Liu Cihan (“Yibajiuwu nian”) hold that Selim’s influence caused Ma’s conversion; Gao Wenyuan (Qingmo xibei, 433) claims that Ma became a reformer on the pilgrimage; and Ma Tong (Zhongguo Yisilan . . . shilue, 238) writes that Ma took Selim as his teacher in Gansu, then went to Mecca and received initiation in the Shādhilīya.
79. Ma Tong’s informants told him that Ma Zhan’ao favored the New, forcing Ma Yonglin to forego any public opposition to the reformists until after Ma Zhan’ao’s death (Zhongguo Yisilan . . . shilue, 239).
80. Ibid., 240. Unfortunately for the Muslims and Gansu, this claim was false.
81. GNQSL 24.42b.
82. Guan and Liu, “Yibajiuwu nian,” 48–49. Mu Shouqi also mentions proclamations, but narrates a different, probably apocryphal beginning to the violence. According to Mu’s story, Yang Changjun unintentionally instigated the Hezhou violence by trying to save characters in his announcement. Instead of writing “Sa-la-er Hui,” which unambiguously means “the Salar Muslims,” he wrote “sa Hui,” which might refer to the Salars or might mean “disperse and scatter the Muslims.” Fearing the latter import, the Hezhou Muslims rebelled (GNQSL 24.44a).
83. Rijnhart, With the Tibetans, 58.
84. GNQSL 24.43a.
85. The late nineteenth century brought Euro-American missionaries to Gansu, and their direct observations enrich our understanding of the region, despite their clearly stated and consistent religious biases.
86. Despite the Sino-Japanese War, which was just ending back east, the Qing did have plenty of troops available in the northwest, apart from Deng Zeng, Lei Zhengguan, and Tang Yanhe’s local garrisons. Yang was able to call upon forces under Dong Fuxiang from Xinjiang; the Muslim cavalry under Ma Anliang; Li Peiying from the corridor; He Jianwei from eastern Gansu; Wei Guangtao from Shaanxi; and Niu Shihan from Ningxia, all within the space of a few months (Wu Wanshan, Qingdai xibei, 244–47).
87. Fan Changjiang, Zhongguo de xibei jiao, 90. Ge Chifeng’s account of his visit to Taoxi includes oral evidence that many Hezhou people, Muslim and non-Muslim, came from Nanjing and still spoke a Yangzi dialect (Zangbian caifeng ji, 2). Studies of the wool and hide trade mention merchants from Shandong, Shanxi, and Mongolia purchasing and shipping goods from Hezhou. Li Wujin, “Gansu Linxia,” gives a lengthy list of the Hezhou market’s commodities.
88. Liao, “Gansu zhi minzu wenti,” 86; and Ma Zikuo, “Musilin,” 16.
89. Xu Wanli, “Cong Hezhou shuoqi.”
90. For the Min family’s conflict with Ma Zhan’ao, see Hezheng xian zhi of 1930, 206 ff.
91. Haicheng lies at the south end of a broken, eroded loess plain stretching from Ningxia on the north to the Liupan mountains in the south. The population of this region (currently called Xi-Hai-Gu for its three counties), predominantly Muslim, was widely scattered across a landscape suitable for herding, with some poor agriculture in the loess valleys. The Haicheng xian zhi of 1908 gives the figures 8,489 Han and 39,051 Hui for the county, a total of 47,540.
92. There are several accounts of the jailbreak and the subsequent Qing campaign against the Muslims: Haicheng xian zhi, 9.5b–6a, and the biography of Ma Yanqun at 8.12a; An Weijun, Gansu xin tongzhi of 1909, 17.94b; Guyuan zhou zhi of 1909, 4.37b; and several Western accounts, especially Rockhill, “The Dungan Rebellion,” 414–18. Although the Chinese accounts do not agree on exact details, none claims that the “rebellion” at Haicheng was of more than minor importance. The Europeans inflate its significance. Mu Shouqi, GNQSL 24.47b notes that the yamen was guarded by a ferocious hound, which hated Muslims and killed one of the intruders.
93. Chongxin xian zhi of 1926, 3.3a; and Zhenyuan xian zhi of 1935, 17.37b. Guyuan, an important town near Haicheng, was reported to be in Muslim hands (North China Herald, Sept. 27,1895).
94. The Haicheng gazetteer gives credit for the operation to a local martial arts specialist who had been an antirebel commander under Zuo Zongtang (Haicheng xian zhi, 8.12a).
95. Yang’s victory memorial may be found in the Guyuan zhou zhi of 1909, 8.29b.
96. GNQSL 24.49b.
97. Erich Teichman reported the ease with which the crossing from the Tao to the Wei watershed could be made in the early twentieth century (Travels of a Consular Officer, 124–25). The entire upper Wei valley to Tianshui, or even beyond, would be open to a Muslim army that took Didao. That had happened in the 1860s, when Longxi and several other upper Wei towns were devastated by the Taoxi Muslims.
98. In 1941 the Chinese Communist Party reported only 20,000–30,000 Muslims in the Dongxiang. The reluctance of minorities to be counted, often due to fear that conscription and/or taxes would surely follow, may be responsible for the very low figure, which derives from oral and traditional written sources rather than a census (Zhongguo Gongchandang Xibei Yanjiushe, Kangzhanzhong, 278–79).
99. Ma Zikuo, “Musilin,” 15–16.
100. “Sara Kai oyobi Mōko Kaikai,” 75–78.
101. Andrew, The Crescent, 32, describes the Salars and “Mongol Hwei-Hwei” of the Dongxiang, heading east with their families to work the harvest season on farms in eastern Gansu and Shaanxi.
102. Fu, “Taiping Tianguo,” 7, cites stories from official sources of Muslim complaints against non-Muslims who closed the ferries or harassed Muslims trying to cross.
103. Ma Hetian saw the poverty of the people who lived on Niuxinshan (Oxheart Mountain) in the Dongxiang, commenting that these people were thieves primarily because they were incredibly poor (Gan Qing Zang, 15).
104. Note that the sieges of Xunhua, Hezhou, Xining, and Didao described in this section all failed to take the walled citadels. Herbert Franke concludes that “the comparative safety of towns, together with the wide area enclosed within the walls, always gave the defender an advantage over the attacker.” This certainly held true for late Qing Gansu, where siege artillery was virtually unknown, except for Zuo Zongtang’s Krupp cannon, and where strong fortresses were the rule. Indeed, even the Qing armies found it difficult to take fortified Muslim strongpoints such as Duoba, near Xining. If a city were well defended and provided with food, only rapid breaching or mining of the walls, which had deteriorated in some cases, could prevent a prolonged siege that favored the defenders if rescuing armies were on the way (“Siege and Defense,” 194).
105. The Mufti menhuan, descended via Ma Shouzhen from Hidāyat Allāh within the Khafīya suborder of the Naqshabandīya, did not naturally or invariably oppose the authority of the Qing state or make war on non-Muslims. Only a month before the 1895 violence began in Didao, Mufti adherents captured a non-Muslim bandit and conveyed him to the authorities for punishment (Hezheng xian zhi of 1930, 216).
106. Didao zhou xu zhi of 1909, 12.3b–4a.
107. The comparison of Chinese Muslims, especially those associated with the New Teaching (whatever that may have meant) in the northwest, to secret societies has been used for centuries by both military and civilian officials to describe what they see when they look at menhuan or other Muslim solidarities. See, for example, the Qianlong emperor’s opinion in the Daqing lichao shilu 1208.2. The analogy to the White Lotus sect is particularly germane, for the term New Teaching similarly became an epithet applied rather indiscriminately to Muslims who behaved rebelliously. See also Iwamura Shinobu, “Chūgoku Isuramu.”
108. The botanist-explorer Joseph Rock, whose hyperbole was matched only by his ill-humor, wrote to the National Geographic Magazine (February 1930) that he had found a mountain range over twenty-eight thousand feet in height, surpassing the Himalayas. His measurements, which he later corrected in The Amnye-Ma-chhen Range, were inaccurate, as were those of Leonard Clark. An equally egotistical traveller, Clark repeated the error almost twenty years later while fleeing from the Communists with defeated Muslim and Tibetan troops and printed it in The Marching Wind, especially chap. 11. The mountain is spectacular; for photographs and an imperfect historical account of its exploration, see Rowell, Mountains of the Middle Kingdom, chaps. 11–12.
109. Liao, “Gansu zhi minzu wenti,” 85.
110. Travelers reported that, despite the poor quality of the loessial soil, Taozhou peasants cultivated their terraced hills all the way to the top. Both Rock and Ekvall, especially the latter, describe the gradations of pastoralism and agriculture of the Tao valley and emphasize the cultural and economic symbiosis they found there (see Ekvall, Cultural Relations). During a field research trip to the Taozhou region (now called Gannan, southern Gansu) in May 1996, I gained the same impression of an intensely cultivated arid region. The hillside fields suffer from lack of water during the growing season, though they are eroded by the snowmelt in springtime.
111. Old Taozhou provides excellent examples of these multiple roles within the Muslim elite. The leaders of the Xidaotang solidarity, based in Taozhou, took all of them. See chapter 5. A typology of Muslim gentry in the mid-nineteenth century may be found in Fu, “Taiping Tianguo.”
112. Taozhou ting zhi of 1907, 18.3b.
113. GNQSL 24.48b.
114. The story of Ninghepu’s resistance is in Hezheng xian zhi of 1930, 216–20. Mu Shouqi worried a great deal about the wisdom of arming the common people, for greater harm could come to them if they were armed. In addition, the enemy could provide himself with weapons by taking them away from the inexperienced farmers (GNQSL 25.9a). But the Ninghepu militia, organized by local gentry and defending their own homes with semiprofessional élan, made all the difference in resisting the Muslims.
115. GNQSL 25.6a–7a; Gao Wenyuan, Qingmo xibei, 437; Wu, Qingdai xibei, 243.
116. Mu Shouqi records Ma Anliang’s devotion to his father’s pro-Qing stand, a position that caused him often to be accused of treachery by anti-Qing coreligionists (GNQSL 28.18a–18b). He was, however, successful as a mediator throughout his career, dealing with Muslim and non-Muslim elites all over the province, and he remained a popular leader among the urban Muslims, especially merchants in Lanzhou and Hezhou, who stood to lose by widespread violence.
117. Bonin, “Les Mahometans du Kan-sou.”
118. Broomhall, Islam in China, 161, claims that Dong, though a Muslim, “because of some small personal pique, had sided against his co-religionists.” This thoroughly incorrect information, coming from the Western world’s most-cited authority on Islam in China, must give us pause to consider the quality of our sources! Tian, “Longshang qunhao,” 3–14, is only one of several Chinese sources that unambiguously identify Dong as a non-Muslim. Teichman, Travels, 109, relates that Chinese informants told him that Dong was not a Muslim, but that he refused to believe them. The European press had so universally associated Dong with Muslim troops during the Boxer uprising that he could not credit a different version of the story.
119. d’Ollone, Recherches, 272.
120. Tian, “Longshang qun hao,” 4. While in Beijing, Dong Fuxiang had come to the attention of Ronglu, one of the empress dowager’s favorite ministers. Their relationship led to the old empress’s admiration for the Gansu army and its commander and her order that the northwesterners be brought to the capital region as part of her metropolitan forces during the Boxer crisis. Dong’s connection to that powerful court faction would have enhanced both his reputation and his power in Gansu. See Xue Zhengchang, “Dong Fuxiang.”
121. GNQSL 25.21a.
122. GNQSL 25.12b–13a.
123. Again, I use these as cultural terms, not as the names of sovereign states or territories enclosed by boundaries drawn on maps by negotiators. Marco Polo, who passed through Xining on his way to the Yuan capital in the late thirteenth century, made careful note of the physical characteristics of the Tibetans who controlled the city at that time. He described the yaks and musk deer and reported “no lack of corn.” He observed “idolators,” by which he meant the Tibetans, as well as Muslims and Nestorian Christians in the town, a combination that was once again found in the late Qing when the China Inland Mission established a station there. The able missionary Charles Ridley made excellent studies and maps of the area, as did German Catholic priests who lived nearby. The studies of the latter are summarized in Eberhard, “Notes on the Folklore,” 306–10.
124. The French priest Evariste Huc, with his traveling companion Joseph Gabet, made the journey from Kumbum, a great lamasery near Xining, to Lhasa in 1845–46. Their account, published in English as High Road to Tartary, contains vivid descriptions of Amdo.
125. Rijnhart, With the Tibetans, 27–28. See also Frick, “Hochzeitssitten.” The Filchner expedition observed both the strength of Chinese culture in the area and its attraction and accommodation to Tibetan culture, especially in the material realm (Filchner, Wissenschaftliche Ergebnisse, vol. 2, 124).
126. For a complete discussion of the growth and structure of tuanlian, focusing on south China, see Kuhn, Rebellion, esp. 76–78, on the relationship between militia and lineage organizations.
127. GNQSL 25.8a–9a, 10a. Lacking modern artillery, the Dongguan Muslims relied on homemade cannons and small arms, hopelessly inadequate to the task of breaching Xining’s high walls.
128. Xining dengchu, 1.2b.
129. GNQSL 25.8b, 10a, 12b.
130. GNQSL 25.17a.
131. Ekvall, Cultural Relations, 9–12, 83 in summary, and esp. chaps. 3 and 4.
132. Sutton, In China’s Border Provinces, contains many harrowing stories of the loathing inspired in the European author by his experience with the Tibetan nomads. The North China Herald reported their depredations near Xining and the response by Ma Qi, head of the local Muslim military, in 1924 (see esp. August 13).
133. Zhang Qiyun, Xiahe xian zhi, 3.3a–3b, notes Tibetan fluency in the Chinese language, as does Wang Zhiwen, Gansu sheng xinanbu, 33. Dreyer, China’s Forty Million, 10–11, discusses briefly the development of the tusi system under the Qing. It was primarily intended to govern “semicivilized” non-Chinese peoples by utilizing a hierarchy of titles granted to powerful local lineages, thereby keeping them contented and loyal.
134. A small group of Huangzhong Tibetans around the town of Kargang even converted to Islam under the influence of Khafīya missionaries, but despite some alteration in their clothing styles and food, they remained Amdo Tibetans in language and culture (Li Gengyan and Xu Likui, “Yisilan jiao zai Kaligang”).
135. Ekvall, Cultural Relations, chap. 4.
136. The commander of the Labrang monastic army during the early Republican period was the elder brother of the monastery’s chief rimpoche. The obviously political process of ascendancy to the status of a high rimpoche has always been clear, even to most Tibetans, but it has not necessarily diminished their respect or devotion to those same rimpoche.
137. Among the Qing officials who held this view was Yang Zengxin, later the imperious autocrat of Xinjiang, who served after 1895 as prefect of Hezhou. He wrote often that divisions among the Muslims should be blamed for the violence of northwestern society. Mu Shouqi, Republican period gentryman and compiler of local history, agreed, as did many other Gansu elites. It did not occur to them that the state might respond differently to conflict within Muslim communities (GNQSL 25.37b–39b).
138. That evolutionary process is summarized in Lipman, “Ethnicity and Politics.”