In the mid-nineteenth century, in the city of Changsha, Hunan, a local deity burst onto the stage of Qing imperial history. That deity was Dingxiang Wang, and he was known as the city god (chenghuang ) of Shanhua county. Typically, a city god such as Dingxiang Wang was an object of veneration only within its limited domain, in this case a county that covered one half of the prefectural city of Changsha. However, in 1852, Dingxiang Wang received significant credit for breaking the city's three-month-long siege at the hands of the Taiping, the millennialist Christian sect that sought to topple the Qing and nearly did. Forty-five years later, Dingxiang Wang's power and fame reached from one end of the Qing Empire to the other, as he was worshipped from Kashgar to Fuzhou. This "king who pacifies the Xiang River," as we may translate his name, became known for calming more distant waters, in the oases of Chinese Central Asia.
The exceptional geographical reach of Dingxiang Wang's cult resulted from the involvement of his worshippers in a series of conflicts that raged across the Qing in the latter half of the nineteenth century. This city god became the common object of worship of the soldiers of the Xiang Army, the celebrated fighting force usually known in English as the Hunan [End Page 113] Army. What became the Xiang Army had likewise emerged in Hunan in the early 1850s, as activist gentry organized local militias (tuanlian) into larger and more sophisticated fighting forces meant to combat the Taiping. 1 The army's formation and maintenance thus depended on local social networks and collocal identity. In 1853, when Zeng Guofan (1811–72) reformed these militias as the Xiang Army ( Xiangjun), it was not only a fighting force, but a community rooted in common experiences and practices, among them the worship of this deity. The name "Xiang Army" also emphasized its specific association with Hunan, metonymized by the name of the river Xiang that ran through Changsha. After all, the army's members came overwhelmingly from a small area around Changsha and had grown up near the Xiang River. Moreover, the name carried a natural resonance with the name Dingxiang Wang, which included the same character "Xiang." Nevertheless, soldiers and deity roved far beyond those provincial boundaries: As the Xiang Army fought the Taiping (1850–64) and Nian (1851–68), marched northwest into Shaanxi, Gansu, and Xinjiang during the Muslim uprisings (1862–77), and fought foreign invaders in the Sino-French War (1884–85), its soldiers carried his statue with them.
Where Dingxiang Wang's Hunanese worshippers settled far from home, his story and identity changed along with the new community's estrangement from its homeland. This deity's story is thus one of war, memory, and reconstruction in late Qing China. In the following pages, I explore the case of Dingxiang Wang and his worshippers in Xinjiang. I argue that demobilized Xiang Army soldiers and their descendants retold this deity's story in ways that naturalized and justified their presence as settlers. Dingxiang Wang—in some places renamed "Fangshen"—marked Xinjiang as a Hunanese space. His legends demonstrated that common soldiers had transformed that space through self-sacrifice, both on the battlefield and through irreversible separation from their homeland. Dingxiang Wang became the deity that mediated across that [End Page 114] distance, making life in "exile" intelligible and creating a persistent link with the homeland that also served as a reminder of its unbridgeable distance.
Few have taken note of this storied protector, probably because the leadership of the Xiang Army itself wrote Dingxiang Wang out of history. His presence and story undermined the leadership's own heroic narrative of Xinjiang's reconquest by robbing them of some degree of agency. The way that his legends articulated the soldiers' claim to Xinjiang, as a horizontal movement of heroic commoners, threatened the leadership's project to transform this Inner Asian borderland into a Chinese province through the imposition of a top-down hierarchy of temples.
That tension surrounding Dingxiang Wang reflected the fact that he, as a city god, was worshipped by many constituencies, and was both a member of the imperial hierarchy of spirits and a "popular" deity. In that sense, Dingxiang Wang embodied two different propositions about the relationship between time and space, or history and territory. The anthropologist Safet HadžiMuhamedović, borrowing a category of Bakhtinian literary analysis, characterizes such a proposition as a "chronotope": a "discernible alliance of time and space, a time 'thickened' in a landscape and a set of practices and relationships, a space articulated through a time of specific quality." 2 Put differently, people create, maintain, and negotiate relationships between history, place, and identity through ritual. In the Chinese context, worship of a deity means telling and retelling that deity's story, which in turn indicates the nature of the connection between a community of worshippers and the places where they live.
Such an idea is hardly unfamiliar to Chinese history. Steven Sangren describes two different ways in which deities are mobilized in the creation of territory: The first is the "bounding of uncivilized territory," a historical act for which many Chinese living in former frontier zones venerated their deities. 3 The narrative of a deity's movement into a new space represented the inclusion of that space within the imperial pale. It implied not just the horizontal movement of imperial borders, but the vertical integration of frontier deities into a hierarchy that paralleled the bureaucracy. While we know this "imperial metaphor" to [End Page 115] be an incomplete representation of the world of Chinese belief, it was nevertheless a powerful way of thinking about the relationship between spirits and worldly power, and a means to talk about absorbing and conquering new or unruly elements at the margins. 4 As we will see, Xiang Army veterans reinvented their god as Fangshen, a boundary-expanding deity, and tried but failed to have him recognized as a part of that hierarchy.
The second way in which deities territorialize is by marking the boundaries of the community of worshippers, in part through the deity's processional tour (youjing). During such tours, the deity's image is carried through the space in which its community of worshippers lives. Its movements and stops reinscribe those boundaries on the landscape. Along the way, the deity, as it inhabits the image, often watches performances of dramas, some of which entertain the worshippers, while others are meant for the spirit's eyes only. Dramas recount the deity's legend and reassert the intimate connection between spirit, community, and territory through history. Such practices resemble how Rian Thum characterizes pilgrimage in his study of history and memory in Muslim Xinjiang. 5 In Thum's case, a shrine marking the real or imagined body of an Islamic saint serves as a "fulcrum of history," a place where the social constructions of memory (history) and space (geography) connect at a fixed point through the recitation of the saint's legend. Oral performance reasserts a common historical memory of the arrival of Islam in a place through the saint's travel and death. In Pierre Nora's terms, a shrine is a kind of lieu de mémoire in which the social discourse of memory is mediated by the sacred. 6 Dingxiang Wang served the same purpose for Xiang Army soldiers as they crossed China and settled in Xinjiang.
Dingxiang Wang's accomplishment, as it was retold in sacred legend and at his temples in Xinjiang, was dissonant with the dominant politics of commemoration, which favored the celebration of heroism at "manifest loyalty shrines" (zhaozhong ci). 7 In the Taiping era and after, control of those shrines to imperial martyrs passed from the bureaucracy into the hands of provincial elites such as those who led the Xiang Army. [End Page 116] In Xinjiang, the army's leadership first became the region's provisional government and staffed the "reconstruction agencies" (shanhou ju) until 1884, when Xinjiang became a province, and the same officers moved into its new government. Those officers-turned-officials staked their own claims on the landscape by memorializing its leaders in such shrines and even casting army leader Zuo Zongtang (1812–85) as the province's protecting spirit. The leadership in Xinjiang did not simply promote its own narrative of heroism and territorial recovery but erased those of others.
The first part of this article tells the story of Dingxiang Wang from his adoption as the deity of the Xiang Army, to the spread of his cult to Xinjiang, and finally to his abandonment by the elite but persistent worship by Chinese in the borderland. The second part plunges into history to discern the origins of Dingxiang Wang's legend and how it was adapted in his community's self-fashioning as heroic exiles. The third part draws out the contradictions between his territorialization and that which the army leadership attempted to impose.
This elite process of producing the frontier as an object of knowledge and political project—including phenomena that Laura Newby has described as the "literati conquest of Xinjiang" and James Millward as "coming onto the map"—obscured the social processes of settlement and placemaking among ordinary Chinese. 8 I will argue for greater attention to Qing empire as a multilayered process in which differences of class and belief remained significant even as those of ethnicity came to the fore. After all, Albert Memmi reminds us that colonialism does violence not only to the colonized, but to the colonials, those agents who are allied with the colonial project but remain its instruments. 9 Attending to such experiences requires us to consider sources and perspectives that fit awkwardly into present historiographical paradigms but that ultimately serve to fulfill the promise of Qing imperial history.
The Ever-Securing: From City God to God of Exile
Dingxiang Wang was worshipped first in Changsha, then in the Xiang Army, and later in Xinjiang and across China. The following history of Dingxiang Wang connects his social and cultural role in Xinjiang [End Page 117] to his origins in Hunan. This connection has surprisingly remained underexplored in the historiography, which tends to focus on either end of the god's journey. 10 Moreover, this section places his worship in the context of the "Xiang Army community" ( Xiangjun jituan ). 11 This understanding of the Xiang Army as a sociocultural formation bound together by common origins, practices, and beliefs draws our attention from the lives of the army's famous leaders to those of ordinary soldiers. This shift of perspective will illuminate the interrelationship between the community, the deity, and the sites of worship that joined them together.
The history of Dingxiang Wang may begin in September 1852, when the Taiping army besieged Changsha. 12 At the time, Changsha was home to a number of deities who could have rendered aid to its overwhelmed defenders. They included three different city gods (chenghuang): one for Changsha Prefecture, and one each for the two counties—Changsha and Shanhua—that shared jurisdiction of the prefectural city. According to a theory then espoused by Chinese scholars and repeated in many gazetteers, a place or territory—such as a county, prefecture, or province—is the "double" or "repetition" (fu) of its city god. 13 For this reason, the first emperor of the Ming dynasty (1368–1644) had mandated that city god temples be established for every administrative unit, along with official rites for their worship. This theory of doubling was included in the canon of the so-called statecraft (jingshi) school of Neo-Confucianism, which was compiled in Changsha by a community of scholars who also formed the core group of the Xiang Army leadership. However, other stories present an alternative image of city gods as more active forces who are made manifest during moments of crisis, often [End Page 118] as ghostly warriors defending their city from atop its walls. Such a bifurcation between official and popular conceptions and worship of the city god reflected a widespread understanding of the complementarity of earthly and spiritual power. The chief administrator of a territory, for example the magistrate of a county, was the city god's partner in the visible world: where the magistrate tended to his flock in life, the city god would guide them in the afterlife through the courts of the underworld.
The least among Changsha's city gods was Dingxiang Wang, of Shanhua County, to whom the defenders of Changsha turned when all else failed. 14 According to the Shanhua Gazetteer, when the Taiping had besieged Changsha for several weeks, Magistrate Wang Baosheng led a group of gentry to Dingxiang Wang's temple, which was at the base of the city wall. They then carried his statue to the top of the wall and faced it southward to view the battle. Suddenly, the tide of battle turned: the Taiping troops, who had been using explosives in an attempt to undermine the walls, instead burned themselves up, while their rockets and cannonballs fell all but harmlessly onto empty houses. Those scaling the walls were pushed away from it by some unseen force and tumbled through the air like so many drops of rain. The siege was broken. While historians may attribute the lifting of the siege of Changsha to the resilience of its defenders and a poor tactical choice on the part of the Taiping commanders, the people of Shanhua celebrated their god's timely intervention and reported sightings of him keeping watch on the walls or planning the defense. On May 2, 1853, Governor Luo Bingzhang (1793–1867) officially requested a title for the city god in recognition of his contribution. The Board of Rites named him the "Ever-Securing" (yongzhen ) Dingxiang Wang.
The Xiang Army was formally created that year, and Dingxiang Wang probably entered it along with worshippers from Changsha. The army was not merely a military unit, but rather a social formation constructed around collocal identity, ties of blood and patronage, oaths of loyalty, a moral code, and the common experience of battle against the Taiping. 15 Worship of Dingxiang Wang was a further source of social cohesion, and one that seems to have strengthened following the army's reorganization [End Page 119] in 1864 under the leadership of Zuo Zongtang. While Zuo's new, smaller army was technically named the Chu Army (Chujun), its members consistently referred to it as the Xiang Army, and I follow their self-description here.
For a decade, as the campaign dragged on across China, Zuo Zongtang ordered the worship of Dingxiang Wang before every battle. Worship resembled that given to him in his role as a city god: "Whenever the Hunanese go to fight in other provinces," declared a memorial in 1888, "they always carry with them a statue of Dingxiang Wang… which they worship at a 'temporary residence' (xinggong )." 16 The term "temporary residence" is normally used for the places where city gods stopped overnight during tours of their territory when they could not shelter in a shrine, and it reflects the word used for the mobile encampments where earthly rulers stayed in similar circumstances. In these residences, the city gods often watch dramas while facing south, much as Dingxiang Wang had watched the battlefield in Changsha. Perhaps he continued to observe combat along the march. That a local deity was operating far beyond his usual territory, and yet retained his identity as a city god, attracted the notice of Chinese observers. "The city god of our Changsha does not only orient inward, but can also orient outward," wrote the Hunanese literatus Yi Baisha (1886–1921). "Not only does he stay in Hunan—he goes on campaign to other provinces, as well!" 17 Dingxiang Wang was simultaneously the inward-facing "double" of Shanhua and an outward-facing deity who reclaimed and expanded territory.
This dual identity was embodied in Dingxiang Wang's image. A block-printed book from 1885 depicts him as a scholar-official with a serene expression and long beard. 18 His long-sleeved robe is covered in cloud [End Page 120]
Click for larger view
View full resolution
[End Page 121]
patterns, pointing to his place in the celestial realm. In his left hand, he holds a hu scepter. His right hand rests in his lap, covered by a sleeve. About his waist is a girdle. Overall, Dingxiang Wang appears the picture of an upright scholar-official who has crossed into the spirit world, and he resembles in dress and comportment certain members of the heavenly bureaucracy, including the headdress of a king (wang) with tassels hanging down, which marks his rank. His image is similar to those of the guardian gods of rivers, also enfeoffed as "kings," as they were depicted in nineteenth-century Jiangnan; however, the throne on which he sits is draped with the skin of a tiger, pointing to his newfound identity as a god who accompanied an army on a mission of reconquest. In the 1980s, a team of researchers investigating the remains of Dingxiang Wang's temple in Manas, Xinjiang, which had been destroyed in the 1950s, ran across a statue and noticed the same curious mixture of imagery. 19 Members of the community, when asked about the martial aspects of this civil figure, responded that Dingxiang Wang had been a soldier, and then an official. That same duality of upright official and brave soldier appears in the legends of Dingxiang Wang, which we will discuss in the following section.
The temporary residences of Dingxiang Wang served a range of functions for Xiang Army soldiers. 20 Soldiers suffering from illness sought healing from him, while commanders confirmed their strategic decisions by consulting the god through spirit writing, much as nervous young scholars had done in Shanhua on their way to the provincial examinations. When Zuo's lieutenant Liu Songshan (1833–70) died in battle, Zuo selected his young nephew Liu Jintang (1844–94), the future governor of Xinjiang, to take his place. Dingxiang Wang, it was later written, appeared to Liu Jintang in a dream to give him courage as he took on his uncle's duties. Meanwhile, on the nights before the army met the enemy in battle, soldiers heard the mysterious sounds of the ghostly men and horses whom Dingxiang Wang rallied to their aid the next day.
In 1877, the Xiang Army entered Xinjiang carrying the statue of Dingxiang Wang and in less than two years had retaken the entire region. As soon as the war was over, however, the army's leadership recast [End Page 122] their erstwhile soldiers, who had been recruited into the army through personal connections and marched across all of China, as suspicious members of an unstable underclass. Roughly half left Xinjiang and sought to return home, while the rest stayed and became farmers, traders, and laborers. Few found their fortunes. In these circumstances, packaging the bones of the dead to send home for burial became an important duty shared by Hunanese collocals and their native-place associations. However, that could be an expensive and complicated endeavor. 21 Where Chinese migrants in distant but gold-rich California could afford to dispatch their remains for burial at home across the vast Pacific—so that it became a lucrative trade that outgrew the native-place associations—their counterparts who had made the arduous overland journey into the Northwest did not share that luxury. 22
One of the primary roles of Dingxiang Wang was therefore to connect Hunanese soldiers and settlers with their homeland. The first known temple to him in Xinjiang was founded in Qumul (Hami), which the Xiang Army used as a staging ground for the campaign. In 1992, more than a century later, a worshipper there described the god thus: "The Xiang Braves may have died in Xinjiang, but Dingxiang Wang sent the souls of the dead back home." 23 Indeed, Dingxiang Wang's temple in Qumul did not merely ensure the passage of deceased Hunanese who could not be buried in their homeland, but came to act as a spiritual intermediary for all Han Chinese. In 1939, the writer Mao Dun (1896–1981) visited Dihua, where he saw a Dingxiang Wang temple buzzing with activity. By this time, the temple's priests (daoshi) were Gansunese, rather than Hunanese, and within it they ran a kind of "spiritual post office." Worshippers far from their homes in China proper could send letters to their distant deities, if they paid a fee for a special stamp labeled "Dingxiang Wang's Palace" ( Dingxiang Wang fu ). Thereupon the letters would be burnt so that Dingxiang Wang could spirit them to their recipients. The temple did especially brisk business around the Lunar New Year, when migrants were too far from home for a return visit to be feasible. [End Page 123]
As Wang Penghui argues, this example reflects the role of deities on frontiers in blurring distinctions between Chinese of otherwise diverse origins, or what other scholars of Chinese religion might describe as a deity's role in an ongoing tension and negotiation between localizing and universalizing identifications. 24 Mao Dun even called Dingxiang Wang the "city god of the Han ethnic group" (Hanzu zhi chenghuang), and as Wang notes, his temple was the only city god temple in Dihua, the capital of Xinjiang. The metaphor of the city god as the "god of walls and moats" recalls Dingxiang Wang's ancient function as the protector of a territory, or in this case the boundaries around a community. This seems to suggest that Dingxiang Wang had somehow become recognized both as the city god of Shanhua and, at least from a popular perspective, as that of the provincial capital. At the same time, even though the Dihua Dingxiang Wang temple served a variety of Chinese communities, it was later attached to a Hunanese-Hubeinese native place association (Liang Hu huiguan) founded in 1904. It must be emphasized that the reconquest and reconstruction of Xinjiang was a distinctly Hunanese project: an overwhelming majority of the provincial leadership were former Xiang Army officers from a small area around Changsha, and, as many travelers observed, the Hunanese Xiang dialect remained more useful than the speech of Beijing into the mid-1910s, when the Xiang Army clique lost control of the government. 25 The region became known as "little Hunan" (xiao Hunan), and so it is not surprising that Chineseness as a collective identity should be articulated in part through Hunaneseness. To draw a comparison with the articulation of European identities in colonial contexts, the Dihua temple points to the role of institutions in blurring distinctions among settlers who might have had diverse origins in the homeland but developed a common identity in their new home. Despite the articulation of vertical boundaries between peoples, however, horizontal boundaries of class nevertheless persisted, as we will see. 26
The construction of Dingxiang Wang's further "temporary residences" followed the path of the army and of the merchant networks that formed [End Page 124] around it. 27 Dingxiang Wang temporary residences were built in Turpan's New City sometime before 1893; in Gucheng (Qitai), which was the major hub of trade in North Xinjiang; in Manas (Suilai); in Suiding in Ili; in Wusu, west of Dihua, in 1907; in Ürümchi (Dihua); and, across the Tianshan mountains to the south, in Chaqiliq (Ruoqiang). 28
Of these, the Manas temporary residence provides the most complete picture of Dingxiang Wang in Xinjiang. Manas was the site of the Xiang Army's final victory in North Xinjiang over the forces of Ya'qūb Beg (1820–77), and, before the Muslim uprisings, it had hosted a significant Chinese settlement and trading post. It was also where Xiang Army leaders chose to encamp during the planning of the campaign into the Tarim Basin to the south. Later, some Hunanese settlers decided to build this new residence out of "contrition" for nearly twenty years of missed prayers and sacrifices. Local support was strong: the magistrate solicited funds and labor in the early spring, and, on the summer solstice, the temple was already complete and ready for dedication. A stela erected to memorialize the event reads:
Dingxiang Wang is the city god of Shanhua. … After the tragedy of the Muslim uprisings, the Great Qing pacified Shaanxi, Gansu, and Xinjiang, one by one. The Wang's incense fires then spread north and south of the Tianshan Mountains.
The metaphor of "incense fires" (xianghuo) stands for the popularity and geographical reach of a deity's cult, as worshippers placed ashes from the incense burner of the senior temple into that of a newly founded one. 29 Incense implied pilgrimage and continued connection with Dingxiang Wang's primary temple in Shanhua, which worshippers recognized as his place of origin.
The timing of the temple's erection is telling: roughly eighteen years after the reconquest, the Xiang Army community was reproducing [End Page 125] itself in the borderlands. The army's leading officers simply became the civil officials of the new provincial administration, and they remained entrenched in the government until the end of the Qing. 30 Hunanese officials' children who were born in the Northwest married each other, and their sons often joined the administration, reinforcing the Hunanese dominance of institutions. More Hunanese were recruited from the Xiang Army's home counties, as well. Former soldiers, in contrast, usually lived as unmarried men or in groups of bachelors, as marriageable Han women were rare. 31 Some were assigned wives by the Refugees Agency (Nanmin ju), while others acquired Muslim wives through brokers. Those unions resulted in many children who belonged neither to the Chinese nor Muslim communities. These circumstances heightened anxieties about descent and belonging.
This temple construction was thus accompanied by a reimagining of the place of Hunanese people in Xinjiang and a renaming of Dingxiang Wang, who in South Xinjiang was called "Fangshen." Temples to Fangshen were established, also in the mid-1890s, in Kucha, Wensu, Kashgar, Yarkand, and Qarghiliq. 32 Fangshen, despite being recognized as the same deity under a different name, was later worshipped separately from Dingxiang Wang. Indeed, some places in the North had a Fangshen temple as well as one to Dingxiang Wang: Wusu, Ghulja (Ningyuan), and Qumul, where in 1908 Fangshen's statue was installed alongside [End Page 126] that of Dingxiang Wang in the same temple. 33 The fact that Fangshen and Dingxiang Wang could be worshipped side-by-side indicated that Fangshen had rapidly gained an independent identity that was nevertheless still tied to the Hunanese presence. Given that the Dihua temple served as a "spiritual post office" for all kinds of Han Chinese, it is possible that Dingxiang Wang became more commonly accepted, while Fangshen stood specifically for the Hunanese community.
While Dingxiang Wang continued to receive adoration in Xinjiang, his star in Changsha fell. In 1885, a group of 101 leading Hunanese members of the Xiang Army who were then officials in Xinjiang province—formally founded a year earlier in 1884—collectively sponsored the reprinting of a didactic work, the Deyi lu of Yu Zhi (1809–74), at a press run by another Xiang Army veteran and Xinjiang returnee, Zhou Han (1841–1911). 34 In it, they singled out Dingxiang Wang for praise, noting his special significance to Hunanese everywhere and key role in postwar reconstruction. However, it appears that Dingxiang Wang was becoming associated more specifically with Zuo Zongtang's Xiang Army and its veterans in Xinjiang, while the army's earlier leaders no longer favored him. Wang Kaiyun, who completed his chronicle of the Xiang Army's exploits in 1881, retained the story of how the Shanhua magistrate placed the statue of Dingxiang Wang on the walls of Changsha, but omitted any statement about the god's concrete contributions. Wang Kaiyun's work was soon superseded by Wang Dingan, who was more openly critical: "The people of Xiang and Chu believe in spirits and superstition, and of old they believed in the Changsha city god." 35 Wang Dingan's account dates to 1889, only one year after Dingxiang Wang had received honors from the empire a second time for his role in the Sino-French War, and yet it rejects as mere superstition the importance of a deity who was still worshipped not only in Changsha, but in a new transregional community formed by veterans.
Although a thorough study remains to be written, a document produced at the Dingxiang Wang temporary residence founded in [End Page 127] Nanjing in 1883 attests to the existence of this transregional community. 36 In 1935, the temple's association leaders compiled a history of their own community and the network of temples with which they communicated. Apart from Xinjiang and Hunan, the gazetteer lists centers of worship in Fujian, Zhejiang, Shaanxi, and Gansu, as well as their own. The gazetteer was written with some sense of urgency, as Dingxiang Wang's worship was fading elsewhere. After the Xinhai Revolution, his temple in Changsha was an object of violence in "anti-superstition" campaigns. Ultimately, the Changsha temple burned in the fire of 1938, and the Cultural Revolution destroyed most of what remained. Until 2015, a piece of stone that had formed part of the temple remained on display in a restaurant that later stood on the site. Today, even the restaurant is gone, and while a new city god temple has been erected nearby as part of a history-themed shopping street, Dingxiang Wang is not slated to be honored within it. 37
Soldier's Sacrifice: The Evolution of the Dingxiang Wang Legend
The erasure of Dingxiang Wang from elite accounts of the Xiang Army reflects a competition over history and territory. To return to the idea of the chronotope: The changing legend of Dingxiang Wang and its expressions of when and how Hunanese people arrived in Xinjiang reflected shifts in how Hunanese settlers constructed their relationship with the borderland. The horizontal territory defined by the legend connected the contemporary Hunanese presence to pre-uprisings history through a story of exile. This section will trace the legends of Dingxiang Wang from Changsha to Khotan, showing how the same cluster of narratives traveled with the deity and were adapted in his worshippers' new circumstances.
Dingxiang Wang's rise to prominence drew the curiosity of contemporary literati. In 1877, one expressed the persistent sense of mystery surrounding the deity. 38 He speculated that Dingxiang Wang [End Page 128] originated either as a water god worshipped locally in Hunan, or perhaps as the spirit of the local anti-Manchu hero He Tengjiao (1592–1649). He Tengjiao was a supreme commander for the Ming when the Qing armies crossed into Hunan, and he led his embattled forces in resistance all along the Xiang River. According to several accounts, when Changsha was all but lost, He Tengjiao entrusted his seal of office to a relative before the Qing captured him and placed him on a boat. Somewhere downriver, He seized the opportunity to jump deep into the river, where he evaded his captors and floated ten li to safety, crawling ashore at a Guandi temple. He Tengjiao and his relation were reunited, along with the seal, and the fight continued. Later, he was captured but refused food for seven days before the Qing executed him. According to Wang Fuzhi (1619–92), when the retreating Ming learned of He's self-sacrifice, they honored him posthumously as Zhongxiang Wang, "King of the Middle Xiang," which title obviously resembled Dingxiang Wang. Indeed, in He Tengjiao's hometown of Liping, Guizhou, it was his spirit who was believed to be Changsha's city god, and who frightened the Taiping armies away. 39 Wang Fuzhi was an anti-Manchu intellectual who briefly led a guerilla resistance against the Qing invasion, and his writings later profoundly influenced the core community of the Xiang Army, who construed Wang as part of a distinctly Hunanese scholarly lineage. 40 However, there is no clear association between He Tengjiao and Dingxiang Wang in other known Hunanese sources.
Rather, the stories that accumulated around He seem to reflect two of the tropes surrounding Dingxiang Wang: self-sacrifice in the struggle against an invading force and diving into the river. Therefore, the deity's story may be far older, in the sense of emerging from a collection of traditions. "Dingxiang Wang" literally means "the king who pacifies the Xiang River," the wide waterway that flows through Changsha, and it should not surprise us that his known origin legends associate him with its fickle waters. Perhaps Dingxiang Wang was worshipped already in the Song as the city god of Tanzhou, as mentioned in the 1224 Bintui lu. 41 In [End Page 129] the meantime, Dingxiang Wang fell into obscurity until the nineteenth century. 42
Indeed, the Jiaqing-era Changsha County Gazetteer describes Dingxiang Wang as a forgotten god, a former city god of Changsha known by different names in different places. 43 In Shanhua, it relates, he was instead called Chengxiang Wang, "the king of clarifying the Xiang," or perhaps of "bathing in the Xiang." While variation between "Ding" and "Cheng" may be attributed to dialectal differences, it also indicates an action that the deity took in many of his legends: submerging himself in roiling waters in order to calm them. Regardless, as of 1818, Dingxiang Wang was not the city god of Shanhua or anywhere else, nor were his temples labeled explicitly in local gazetteers. 44 Nevertheless, some of Dingxiang Wang's worshippers later referred to him simply as the "noble" ( wangye), and earlier Qing gazetteers do note the presence of smaller temples ( dian) near the Xiang River where locals worshipped a "noble." It is possible that this indicated Dingxiang Wang, then playing the role of a river god.
Instead, Shanhua's city god at the time was considered to be the spirit of Li Fengsheng, who had come to Shanhua as an acting magistrate in [End Page 130] 1743. 45 Li died by dark magic, but according to legend acted from beyond the grave to guide his flock back to orthodoxy. His spirit became the city god. In 1838, however, a crisis emerged that Li Fengsheng could not address: a plague of locusts descended on Changsha and stripped the fields bare. Fortunately, the story goes, the magistrate was an upright gentleman named Han Dunwu, known alternately by a given name Han Yuan, or by a style name "Baowu." (There is no record of a Han Yuan, Baowu, or Dunwu serving as magistrate of Shanhua.) According to legend, Han Dunwu was a direct descendant of the Tang scholar Han Yu (768–824), and he possessed exceptional abilities to detect and control malevolent spirits. Nevertheless, the locust crisis proved an insurmountable challenge. In a last and desperate act, Han Dunwu dressed in his formal robes, strode out into the fields, and swallowed handfuls of locusts, swearing that he would consume them all himself if necessary. Magistrate Han died from his efforts, as though drowning in locusts instead of water, but he was not gone forever.
Han Dunwu reappeared as a spirit in July 1847, much as He Tengjiao had reemerged from the river. 46 One day, a water seller was gathering water from the Xiang when the spirit of Han Dunwu possessed his body, apparently having traveled through the waters. God and man proceeded to the magistrate's office to explain the situation: Han Dunwu, since his death, had served as the tutelary god of Yanggu County, Sichuan. 47 He was then reassigned to Xiangtan County, Hunan—just downriver—as its city god. Now he had been transferred back to Shanhua and wished to report for duty. The magistrate recognized Han Dunwu as Dingxiang Wang and accepted his appointment as Shanhua's city god.
Dingxiang Wang now possessed a treble identity: as an upright magistrate turned city god, a defender of Changsha, and a god who pacified dangerous waters. The former two of those aspects were on display during the siege of Changsha in 1852, but the third remained salient in the retelling of his legend in Xinjiang. In Qumul, Dingxiang Wang was identified in an oral source from 1991 as the magistrate of [End Page 131] Xiangxiang County, which neighbors Shanhua and Xiangtan. 48 When the Xiang River flooded, it is said, this magistrate donned his official robes and immersed himself in the river in order to calm the waters. The story is plausible, as magistrates were known to act out roles in local rituals of self-sacrifice. 49 Indeed, magistrates in such rituals took on the roles of legendary figures—and so, if and when Han Dunwu or someone like him waded into the river, perhaps it was not the origin of the Dingxiang Wang story, but an enactment of it. When the flood did not retreat, however, he leapt further into the river, and only when he drowned did the water calm. According to the story, the Qing court posthumously granted him the name "Dingxiang Wang," with the title of king (wang), to celebrate his pacification (ding) of the Xiang. Unfortunately, no records appear to verify the story, but it does reinforce the intimate connection between Dingxiang Wang the city god, water, and self-sacrifice.
All of those elements survived Dingxiang Wang's transformation into "Fangshen." Yet Fangshen's story had one critical difference: Fangshen was not a magistrate, but an exiled soldier. 50 The Wensu County Gazetteer (documenting a county in the northwestern Tarim Basin, north of [End Page 132] Kashgar) provides an origin legend for Fangshen that ties him both to Hunan and to Xinjiang: 51
The Fangshen Temple is outside the east gate of the county seat. According to the Fangshen Biography, the former Khotan Prefect Liu Shinan once found it recorded in the Daoguang-era Posthumous Record of the Garrison Soldier: 52
The god's surname was Huang, and his style name [zi] was Dingxiang. He was from Changsha County, Hunan, and his family had lived for generations in Millet Dike Village beside the Liyin Temple. He was born in the early evening on Jiaqing 6.5.6 [June 16, 1801]. In Daoguang 1 , he fought with his neighbors over some water. His elder brother, in a fit of rage, accidentally killed someone. The god took his brother's place in acknowledging the crime without concealing anything, accepting the punishment. The high official appreciated his honesty and affection [for his brother] and sent him as a garrison soldier to Gansu.
In Daoguang 6 , the god was transferred to [the garrison at] Kashgar. At that time, it was the Jahāngīr uprising, and they had diked up the city, flooding it with water. The god magnanimously swam into the water—he opened the dike, and the water rushed out. In the besieged city, soldiers and civilians alike celebrated their new lease on life. So he sacrificed himself. In life, he was upright; in death, he was made a deity.… His temples have spread across Xinjiang, and he is called "Fangshen." This is probably from the Book of Odes: "They will come and offer pure sacrifices to the Spirits of the four quarters [fang ]." 53 [End Page 133]
It was found that, if one is to make a man into a god, then he must be worthy to have an imperial stela recording his deeds, and then in the present day he may be offered ceremonial sacrifices. In Guangxu wuxu , the officials of Southern Xinjiang memorialized requesting him to be sacrificed to… and they attached a copy of his Biography to await his receipt of honors.
Thus was Fangshen made both a Hunanese and a loyal servant of the empire in Xinjiang. The writer, Pan Zongyue (b. 1858), was a native of Ningxiang, Hunan, and member of the Xiang Army community, and he had come up through the ranks as a young man fighting in Gansu. Pan was surely aware of Dingxiang Wang. He implausibly claims that this humble villager had a "style name," used for signing poetry and artistic prose, that was "Dingxiang," and thus identical to the name of the Shanhua city god, as well as a surname, "Huang," that was a near-homophone with "Wang." The simplest conclusion to draw would be that Pan Zongyue recorded or created a story that intentionally altered the details of the Dingxiang Wang legend to adapt it to new circumstances.
The story about Jahāngīr flooding Kashgar does not seem to appear in any accounts of the war in Chinese or in Chaghatay, and we may reasonably doubt its authenticity. Rather, it was an adaptation of the Dingxiang Wang story also recorded in Qumul: the young soldier, rather than the official, sacrificed himself to calm the dangerous waters by diving into them and thereby saved the city from foreign invasion. Moreover, it provides a satisfying symmetry to the deity's narrative: he was exiled to Xinjiang for a dispute over water in which he saved his biological brother, and later died saving his brothers-in-arms by diving into water during a conflict. That narrative symmetry is not merely a matter of foreshadowing, but an allegory for the experience of the Xiang Army itself and its emerging identity. Dingxiang Wang stood for the interests of common soldiers, many of whom joined to secure livelihoods during a time of hardship, and so the magistrate is reimagined as a farm boy. Moreover, because the Xiang Army recruited through familial connections, actual brothers in many cases marched together. 54 They [End Page 134] sacrificed themselves on the battle field, and yet, their reward was to be stranded on the frontier.
Yet the name "Fangshen" is connected here to a passage from the Book of Odes. The original text referred to human sacrifices performed in ancient China in different quarters to give thanks for a bountiful harvest. A Qing literatus such as the gazetteer-compiler Pan would probably have understood the reference. However, it is also possible that the writer provided this gloss as a way to tie the deity to antiquity and thus suggest its primordial origins, or else to demonstrate his own erudition. Although many gods of one or all of the four directions (fang) are referred to throughout the Chinese historical record, nevertheless I am unable to locate any deity called simply "Fangshen." Perhaps a Hunanese literatus in Xinjiang first misinterpreted the passage to mean "They will come and offer sacrifices to 'local god,'" taking fang as a shorthand, and then reinterpreted the phrase again to indicate a specific deity, the "Fangshen." However, such a contortion seems unlikely. The assertion would appear to be an attempt to locate Fangshen not just in the history of Xinjiang or Hunan, but of ancient China, tying the interconnection between the two places not to a reconquest in recent memory, but to a point in the distant past that was itself grounded in the authority of the Classics. "Fangshen" is thus presented not as an invention, but as a rediscovery or recovery.
The Khotan Prefecture Gazetteer also records an effort to have Fangshen recognized officially by the Qing: "Fangshen's temples and statues are all over Southern Xinjiang. He is esteemed for warding off peril on the same day that it appears. It would appear he ought to be inducted into the Record of Sacrifices." 55 Yet, there is no record of the request, nor of the textual sources mentioned in the Wensu County Gazetteer: the Fangshen Biography ( Fangshen shilüe ) and Posthumous Record of the Garrison Soldiers of the Daoguang Era ( Daoguang shuzu yimo). I have yet to locate either of these works, despite extensive searching and consultation with specialists in Beijing, Taipei, Hunan, and Xinjiang. They may be fictive, or else hidden in the not-yet-digitized archives of the Board of Rites. Unless there is evidence that at least one of these two sources exists, or more sources emerge to attest to beliefs and practices surrounding the deity, then it would then appear that Fangshen was an intentional fabrication. [End Page 135]
Nevertheless, subtle differences with a second Fangshen legend indicate instead the presence of an active tradition, and that divergent narratives may have emerged organically in the reconstruction era. This one was recorded in the Kucha gazetteer, and apparently in near-identical form in a now-lost Kashgar gazetteer. 56 It reads:
The Fangshen Shrine is north of the prefectural seat. We find that the god was originally from Shanhua, Hunan. His surname was Huang, and his name was Guifang. In the Daoguang reign, he was sent to Kashgar as an official, where he assisted the general at the Black Water Encirclement. When the bandit Jahāngīr flooded the city, in a panic, [Huang] alone, magnanimously, and in righteous rage jumped off the city walls and into the water, pledging to sacrifice his life. In a flash, the water streamed out of the gap, opening the city, and everything was all right. The people recognized his virtue and so sacrificed to him, and when they do pray, none is without a numinous response. All across the South [of Xinjiang], Chinese and Muslims enthusiastically burn incense to him.
This account again plays fast and loose with history in its construction of an origin legend. The Black Water Encirclement, also called the Battle of Tongguzluq, actually took place in 1758 along the Qarasu River during the Junghar War, not in the nineteenth-century conflict with Jahāngīr. However, the story adapts the same tropes that persisted in the deity's legend from its early days in Changsha: self-sacrifice during a flood and invasion, this time by an upright official.
The assertion that even the Muslims of Xinjiang burn incense to Fangshen draws on another trope that became popular in the Northwest during and after the Muslim uprisings. That is the story of sacrifice and transformation, in which a heroic figure journeys into Xinjiang to reunite a family torn apart by war. 57 In one example, a boy's father sojourned beyond the Pass into Xinjiang before the outbreak of the uprisings. When the Xiang Army reconquered the Northwest, the boy learned [End Page 136] that his father still lived. The self-sacrificing child suffered hardships but eventually located his father. In their meeting-place, a tree grew that symbolized not only the fulfillment of the parent-child relationship, but the planting of a seed of Confucian values in the borderlands. The story asserts that Chinese and Muslims alike worship the tree, showing how the goodness of filiality is self-evident. Meanwhile, the reclamation of a lost parent stands not only for the reclamation of lost territory, naturalized through the metaphor of family, but the creation of an even more permanent bond through the affirmation of universal values. Fangshen would therefore appear to be an agent of this same transformation.
As for Huang Guifang, this name does not appear in accounts from Xinjiang before 1864, nor in the official compilation of Jahāngīr War dead. 58 Nevertheless, there was one well-regarded official with exactly this name in the Xinjiang administration in the late 1890s. This army major from Yangzhou was praised for having moved to Xinjiang from afar and gained respect and notoriety among the people of the new settlement of Xinping for planting trees and grasses in the desert. His inclusion thus points to another symbol and method of Xinjiang's territorialization, land reclamation and the transformation of the environment. 59 Given that this adulation for Huang Guifang in 1908 comes from Hunanese officials working in Kucha Prefecture, of which Xinping was a part in the 1890s, it is likely that they intended to honor their colleague by writing him into the Fangshen legend. The trees that he planted were seen not only as sources of much-needed lumber, but as symbols of ideological triumph, like that which grew where the boy met his father—or those which Zuo Zongtang's armies planted along the road into Xinjiang, and which gained a special meaning for those who praised the Xiang Army in poetry. 60
Regardless of how the name came about, the transformation of Dingxiang Wang into Fangshen coincided with a reimagining of Xiang Army soldiers as heroic agents of the restoration of empire. 61 Novels and dramas cast Zuo Zongtang as the protagonist of an epic struggle to reclaim Qing territory, while in the post-Taiping era, others saw the [End Page 137] army's march into Xinjiang as the common man's sacrifice for the greater good. This discourse reflected the idea of the Pass, that gateway between China and Inner Asia at Jiayuguan through which exiles passed on their way to life in the frontier. To go beyond the Pass meant separation from home, typically for conviction for a crime. Nevertheless, some officials sent into exile in Xinjiang, such as Lin Zexu (1785–1850), were now considered heroes to the Chinese people. Chinese and other non-Muslims who died in the Muslim uprisings were regarded as martyrs and celebrated in the increasingly popular "manifest loyalty shrines," as were the exiles who defended Kashgar against Jahāngīr. Xinjiang was becoming a topos not of shame and exile, but of heroism.
This heroic reimagining of Xinjiang construed it as the object of a civilizing mission, which the Xiang Army took up as an ideological program for the creation of Xinjiang as a province. 62 Yet the Fangshen legend places the agency of that civilizing mission in the hands of an ordinary exile. According to the Kucha legend, the death of Huang Guifang organically created a new site of memory, the temple to him, where members of every community remembered his sacrifice. The Hunanese soldier in the borderland, who in life certainly experienced violence and deprivation, thereby becomes a natural and positive part of the borderland, and one that is accepted even by reconquered Turkic Muslim people.
The rapidity of Dingxiang Wang's transformation into Fangshen—from obscure city god to the deity of the Hunanese in Xinjiang in less than fifty years—seemingly points to an intentional effort to rewrite his story. The notion of the heroic Hunanese exiled soldier is too perfect a reflection of the experiences of Xiang Army soldiers and too neat a legitimization of their settlement. Yet the differences between the Wensu and Kucha legends would indicate the opposite, that a cluster of ideas about this deity persisted in the broader Xiang Army community: Dingxiang Wang as tamer of waters, defender of the city, and upright official. Moreover, photographic evidence demonstrates that the worship of Fangshen as a separate deity persisted in some places through at least 1927. 63 The image of Fangshen, known only from a single example in Qarghiliq, resembled that of Dingxiang Wang, and he apparently played [End Page 138] a similar role as a curer of disease. The continued relevance of both deities suggests that Fangshen was not simply an invention, but something that emerged from the Hunanese community and adapted to shifting circumstances, even as it remained tied to the same cluster of narratives and images.
Prasenjit Duara has characterized this path dependency in the imagination of spirits as "radically continuous" and "radically discontinuous." 64 That is, the latitude for reimagining a deity is limited—instead, a community inscribes new meanings on it, and those that might be sublimated under certain circumstances tend to reemerge in others. The identity and memory of a spirit is always contested, and the flexibility in that contestation helps to ensure its survival. In the case of Dingxiang Wang, a major discontinuity in the tradition, the inscription of a soldierly identity, coincided with a movement first beyond the Pass, and then across the Tianshan Mountains. That discontinuity also reflected a cutting-off of communication between the settler community and the settlers' homeland of the kind that necessitated the use of Dingxiang Wang's temple as a spiritual "post office." Under these conditions, the metaphor of the god's incense fires stood not just for the spread of his cult to Xinjiang, but for the separation that its transportation implied. The exiled soldier, the new image of Dingxiang Wang or Fangshen, appeared in the same moment.
At the same time, the Fangshen legend embodied a sacred-historical chronotope in which the death of the exile indicated the defense of territorial boundaries, but also a transplantation of the values of the community whose experience was reflected in his story. The Hunanese population of Xinjiang before the Muslim uprisings had been small and little noted, especially compared to merchants and settlers from Shanxi. Indeed, the whole configuration of Chinese communities in the region was radically different before and after the uprisings, the only common factor being the persistent presence of sojourners from Shaanxi and Gansu. However, Fangshen's legend demonstrated that Hunanese settlement was not an innovation, but a restoration, and that the reconquest was not accomplished by mere force of arms. Rather, the Xiang Army and its soldiers had been guided by the spirit of an exiled Hunanese commoner and followed his path again into the Northwest, [End Page 139] beyond the Pass to Gansu, and then to Kashgar. Their journey was his journey, and the conquest of Kashgar and Khotan was not the end of the war, but a destiny to be fulfilled.
The Apotheosis of Zuo Zongtang
Yet such a politics of recovery squared awkwardly against the Xiang Army leaders' vision of Xinjiang as territory. Their government prioritized establishing the temples required of each administrative unit as described in the imperial register of sacrifices, so that the new province would also comprise a stable ritual formation. During reconstruction, they erased existing popular Chinese deities and instead favored their imagined revival of Xinjiang's ancient Chinese past. Nonetheless, the army leadership's own favored deity soon appeared in the form of Zuo Zongtang.
The provincial government planned city god temples for every county and prefecture and for the province itself, although budgetary restrictions meant that many were never finished. Educated Turkic Muslim observers found these "idol-temples" (Chaghatay butkhāna) to be lively centers of activity, complete with regular dramatic performances that attracted common Muslims to the unholy spectacle. 65 Muslims were aware that those performances reenacted the deeds of "past emperors," and those at the Dingxiang Wang temples were surely among them. City god processions took the deity's statue on tour through its territory, both within and beyond the city walls, stopping at shrine and temples—and in Kashgar, the stops included major Islamic shrines. The procession functioned just as it did in China proper, as a territorial statement, yet one that Muslims did not necessarily find convincing. As one described these rituals: "Sometimes they carry the idol through the city streets and around to every market, making a round of the city. They have this stupid idea that it's beneficial, and that it will keep them from getting sick, or the city from catching fire." 66 Muslims could be very familiar with the procedure and evident meanings of worship, and it was clear that the [End Page 140] provincial government prioritized constructing religious institutions. In Turpan, Muslims were obligated to donate to the construction of the city god temple, although they were permitted to do so anonymously, and witnessed Chinese settlers transform the city's Eastern Mosque into a temple by performing rituals there with the local government's support. 67 Yet the Xiang Army apparently did not engage with the Prophet Mu.hammad or Islamic saints as threatening deities, but as historical figures beyond the realm of numinous competition.
Indeed, the Xiang Army's imposition of a ritual hierarchy related not only to the project of creating Xinjiang as a province but also to their self-presentation as the revivers of an imagined era of Chinese supremacy over the Western Regions that dated to the Han (206 B.C.E.—220 C.E.) and Tang (618–907) dynasties. 68 Kashgar's city god did not belong to "Kashgar," as the place was known in the earlier Qing, but to "Shule," a Tang-era kingdom, the name of which the Xiang Army assigned to the new prefecture. The same was true across the province, as antiquarianism and golden-age revivalism guided the leadership's vision of the new society, reflecting Xinjiang's "literary conquest" and "coming-onto-the-map." Indeed, this project did not simply inscribe the new territory with antique meanings—it served to elevate the position of Xiang Army leaders and to eliminate other Han Chinese experiences.
The provincial administration did not recognize any local deities, save one: the city god of Barköl (Balikun). 69 Barköl was a mountain redoubt, a strategically positioned town between Turpan and Qumul, that in 1864 was home to a Qing garrison and a sizable Han Chinese population. The people of Barköl had remarkably managed to survive the Muslim uprisings and were alive to greet the Xiang Army on its entry into Xinjiang. They attributed this miracle to the numinous action of their own city god, who during numerous assaults appeared at night atop the walls and drove the Muslim forces away. The deity was joined, although not as visibly, by Guandi, whose temple dated to 1772 but was destroyed in the uprisings; by the Dragon King (longwang) of nearby Pulei Lake; and by a popular deity called General Yang Si. Yang Si's peculiar history is worth mentioning here, as his own legend and history have some [End Page 141] resonance with those of Dingxiang Wang: He was thought to be a young warrior from Song-dynasty Changsha who in death was called upon to control floods. 70 In the last years of the Ming, Hunanese migrants brought Yang Si into the loess plateau of Southern Shaanxi and Gansu, where over the course of the first half of the Qing he gained a new grandfatherly identity, as his cult diminished in Hunan. This latter Yang Si was the one worshipped in Barköl—Hunanese by origin, but Northwestern in identity, brought to the town by Shaanxi traders and settlers.
The Barkölese esteemed all of these four deities for providing them with over a decade of protection. "If not for the temples of Guandi and the gods," Barköl's literati asked in 1882, "if not for the protection of the spirits, how could human effort have managed this?" 71 In response to their request, the Xiang Army leadership requested honors for Barköl's city god. The Board of Rites named him "Numinous Aid" (lingji). Their story was one of connection with the earlier Qing and miraculous survival, and the recognition of the city god reflected the continuity of the community across the rupture of the Muslim uprisings.
Guandi, however, saw his temple incorporated into the Xiang Army's own narrative of heroism and recovery. The temple was rebuilt in 1882 as a home for two stelae, one from the Han and one from the Tang, that celebrated early Chinese victories in the Western Regions. The Han stela, dating to 137 C.E., celebrated a victory at Dunhuang, and it was known from the earlier Qing. 72 The 640 Tang stela recorded the arrival of Tang general Jiang Xingben (d. 645) during a conflict with the king of Gaochang, and it had been rediscovered during the Xiang Army's arrival in 1876. 73 In 1901, a new inscription provided a series of temporal signposts that placed the temple and Barköl into the Xiang Army leadership's preferred historical trajectory: the Han victory, the Tang expansion, the temple's destruction in 1866, the arrival of the Xiang Army, and reconstruction. Barköl was not even called by its contemporary name, but instead "Yiwu," indicating a kingdom that existed there during the Tang. Unsurprisingly, literati came to believe that the Guandi temple was itself an artifact of the Tang, and so it was [End Page 142] a particular object of interest in a region where few artifacts that the Xiang Army recognized as Chinese actually survived from that idealized dynasty.
Soon the Barkölese story of resistance and numinous aid faded from the historical record. The 1908 gazetteer explicitly denies the role of its deities:
When, fortunately, the Dynasty regained its power, and the great army marched to Barköl, they peered in the city and saw that it was weak… and its people were poor, there was none who was not speechless at how strange this was. They thought, to hold out for ten years, they must have had the help of Heaven! They did not know that it was truly thanks to human effort. 74
Here the agency of the deities is denied, but the efforts of the people themselves are also minimized, making the Xiang Army the only heroes of Barköl's liberation. A story of continuity and resistance was incompatible with one of loss and heroic reclamation. Staying and surviving presented a claim to an authentic belonging in Xinjiang that dated to the Qing, while the Xiang Army preferred to present itself as retaking lands lost to China since the Tang.
Meanwhile, the Xiang Army leadership offered its own gods, and their efforts included the apotheosis of Zuo Zongtang himself. The Xiang Army had long drawn on the Taiping-era changes in memorial practices to erect "manifest loyalty" shrines to their own fallen heroes, and Zuo himself requested shrines to be built for deceased officers. Later, Zuo Zongtang and Liu Jintang—Zuo's erstwhile lieutenant and the first governor of provincial Xinjiang (1877–91)—were both enshrined as outstanding officials near their homes in Hunan, and small shrines to them sprang up all over Xinjiang. In the late 1890s, however, the Hunanese regard for Zuo transformed from memorializing a meritorious human being into worshipping him as a protecting spirit.
So the story goes: early in the morning on May 3, 1897, a blacksmith surnamed Zhu entered the compound of Governor Rao Yingqi. 75 He went straight to Rao's office and announced that Zuo Zongtang was [End Page 143] coming up the Pass to "exterminate the bandits." Rao knew that Zhu had a reputation for excitability and delusion and calmly explained to him that Zuo's campaigns were twenty years past. Zhu, however, was insistent on his vision of strong soldiers and stout horses. Rao dismissed Zhu, but soon found himself leading soldiers down that very road into battle during the He-Huang Uprising in Gansu. Rao found the rebels frustrating to fight, as they would disappear into the mountains. However, "with divine assistance" (ruo you shenzhu), he wrote, the army could apprehend them in a day or so.
Rao wrote that he had once heard Zuo himself say, "If, in his whole life, a poor provincial graduate should receive the favor of the dynasty and be enfeoffed and respected as a great person, what should that official then plan to do? He's only mortal. He should thus repay the dynasty by dying and becoming a hungry ghost, killing bandits!" By "provincial graduate," Zuo seemed to have meant himself, as he had never passed the metropolitan exams. For Rao, Zuo's pledge of immortal loyalty to the dynasty invoked the experiences of two Tang military heroes, Wang Jun (653–792) and Zhu Ci (742–84), who had once battled Inner Asian forces. Those engagements had seen soldiers return from the dead to strike at the enemy again. Rao proposed that Zuo, who had died in 1885, be enlisted to the Qing cause once again. He invoked the support of scholars and commoners and reported that they had already begun work on a new shrine to Zuo in Dihua. Rao only asked for imperial approval, which he received.
Zuo was enshrined not only in Dihua, but also less formally in settlements with a Hunan-Hubei native place association. 76 His shrines tended to abut Dingxiang Wang temples. In 1927, Sven Hedin visited one such shrine, which was located next door to the combined Dingxiang Wang and Fangshen temple in Qumul, and described how Zuo's image was flanked by placards on which were inscribed his officers' names, forming a kind of posse comitatus in death. Liu Jintang also had his appanage in the afterlife. Wu Aichen describes a similar scene at the temple complex of Dihua in the late 1920s, when the Dingxiang Wang [End Page 144] temple was decorated with placards and full of swirling incense smoke. Nearby, Zuo's shrine held two lifelike statues of him and an adulatory poem. Liu's smaller shrine held only his image, and his poem was simpler.
Narratives and deities in late Qing Xinjiang posed a set of overlapping claims about history and territory, the sites for the constitution of which sometimes abutted each other. The people of Barköl could claim to have survived the Muslim uprisings, forming a bridge between the earlier and the later Qing, across a period that was increasingly regarded as one of total devastation and loss. That same idea served the self-fashioning of the Xiang Army leadership, who recreated Xinjiang's landscape of worship so as to emphasize their own connection to a long-lost age. Their project to recreate the region as a province meant establishing a hierarchy of temples managed from the provincial center, which were thought to bring stability to this territory, and enshrining Zuo Zongtang himself as a spirit who, in death, would protect what he had supposedly created. However, their own soldiers narrativized their march across China as a journey into "exile" undertaken by a forgotten hero. For them, the relevant comparison appears to have been between the defense of Changsha—in 1852, or perhaps during the Qing conquest—and the defense of Kashgar during the Jahāngīr War. Both spoke to the idea of a city god as a protector of a place and a preserver of territory. Regardless of whether we regard the deities that all of these actors offered as objects of belief, metaphors for experiences, or vehicles for historical claims, they were topoi for the articulation of belonging.
The cases of Dingxiang Wang, Fangshen, and the other deities visited in this article point to the key role that religion played in imagining communities in China at the turn of the century. Dingxiang Wang in Xinjiang supported the project of empire by giving meaning to conquest through narrative, historicization, and personalization, but also revealed its incoherence. Territory could be won, borderlands enclosed, but the question of to whom those spaces belonged remained open. Xinjiang was technically Qing territory, and certainly the sacrifices to the spirits of mountains and rivers carried out before the uprisings were revived under the Xiang Army. However, the Xiang Army leadership represented Xinjiang as a part of a territory that needed to be reclaimed, not simply for the Qing, but for an idea of China that transcended the present [End Page 145] empire. It is telling that the shape of the province was not designed in Beijing, but derived instead from plans made in Changsha, and that the Qing court later attempted to dislodge the Xiang Army from the provincial government. 77 This was the time when, it has been argued, the activists of that community helped articulate a new collective identity, occupying an ambiguous space between Chineseness and Hunaneseness. 78 Xinjiang for them was an object of Hunanese agency, and especially of those enshrined leaders who transformed the borderlands, in the restoration of imperial territory, but in service of something older than the empire. We may understand their project as a top-down process of territorialization that aimed at creating a traditionalistic society held together through the worship of standardized gods.
Yet within that project dwelled unruly spirits, the worship of which recalled other chronotopes—propositions about the relationship between time and space—and therefore other collective subjectivities. Dingxiang Wang and Fangshen point to the Xiang Army community's identity as a "diaspora": the Xiang Army soldiers and their descendants in Xinjiang considered themselves more victims of empire than its perpetrators, and yet worked to valorize that suffering through a narrative of heroic overcoming. 79 That narrative was suffused with longing for a homeland, not over an ocean, but in a remote corner of a vast empire. Moreover, an emerging transregional consciousness at the center of the dispersal, as expressed in the Nanjing temple gazetteer, pointed to a retroactive attempt to understand this scattered people as a community despite separation. It was this politics of identity, as much as its material basis in geographic dispersal, that made the Xiang Army community in Xinjiang diasporic. Moreover, Dingxiang Wang manifested the imagination of locality through communication, the experience of displacement but intimate connection. 80 While the eld of history increasingly recognizes Chinese migration as diasporic in these terms, studies of Chinese diasporas have typically been confined to those that spread across oceans. Dingxiang Wang points the way, I would argue, to comparing people's [End Page 146] movement across land, particularly into the isolation of Xinjiang, to the more familiar cases from distant California or Southeast Asia. Each creates a different kind of Chinese space that links otherwise disparate places.
Dingxiang Wang's "grand tour" and settlement therefore indicates another imagination of Chineseness in the shifting terrain of identity in the late Qing. It was unlike that articulated by a contemporaneous Daur writer in Ili, who advanced an imperial, pan-Manchu identity. 81 Nor was it like the anti-imperial nationalisms that emerged among Han Chinese elites from the stuff of premodern identity systems. Rather, it was a use of the vocabularies and morphologies of local religion to root people in place through narrative and worship—a "territorial cult" inscribed across the landscape of the Qing. There was a class dimension to this cult, which the Xiang Army leadership in China proper wrote out of its own history, and which gained its strongest support among the descendants of the lower-class soldiers whom the leadership regarded with fear.
The invisibility of deities such as Dingxiang Wang, and for that matter of poor Han settlers, in Anglophone literature on Xinjiang indicates a need to reorient scholarly priorities in the historical study of Inner Asian borderlands. The fundamental promise of the New Qing History is to recenter marginalized voices and thus nuance the understanding of the empire—and, increasingly, of its legacies in the twentieth century. 82 Scholarship on Xinjiang naturally tends to favor Turkic-language sources and experiences, much as prioritizing Manchu or Mongol sources is meant to illuminate lifeworlds that were inarticulable in the framework of Sinocentric history. I would argue that historical anthropology as practiced in "China proper" serves a similar purpose, although the experiences and imaginations in question belong not to people marked as ethnically different, but to people whose class differences were obscured by the idea of China as a shared cultural or ethnic identity. Attending to the Chinese subaltern is not only germane to the New Qing History approach but is essential for nuancing the complexity of imperial domination in the borderlands. Doing so, however, may require attention to the perspectives of people who left few records of their own, but for whom deities were very much alive. [End Page 147]
* Earlier versions of this article were presented at the AAS-in-Asia Conference in Kyoto in 2016, the Lockridge Workshop at the University of Montana, and the Montana East Asia Workshop. It builds on initial findings presented in Schluessel, "Cong chenghuang dao shuzu." Thank you to two anonymous reviewers and to William Ma, Steven Miles, Tobin Miller-Shearer, Onuma Takahiro, and Qiu Yuanyuan, who greatly enriched this article with their comments. A very special thanks goes to Katherine Alexander, who very kindly shared not only her great insight, but also the 1885 image of Dingxiang Wang reproduced here.
1. Kuhn, Rebellion and its Enemies, 140–42, 145–48 . Most histories of the Xiang Army end with the departure of Zeng Guofan from its leadership in 1864 following the victory over the Taiping. More recent scholarship takes a more nuanced view of its history. See Platt, Autumn in the Heavenly Kingdom , 24, 357; Wang Jiping, "Lun Xiangjun jituan" ; and Zhao Weixi, Xiangjun jituan, 18–19 . Indeed, modern historians of the Xiang Army place the Northwest and Xinjiang campaigns in the army's "middle period," following the Taiping defeat but preceding its role in the Sino-French War. Wang Dun, Xiangjun shi, 248–49.
10. There are a number of short articles in Chinese on Dingxiang Wang in either Changsha or Xinjiang, including Liu Hua, "Cong 'Dingxiang Wang' xinyang kan zongjiao de shehui gongneng." For a particularly comprehensive and subtle examination of religion in turn-of-the-century Dihua, and one that bears on this argument, see Wang, "Chongjian fenghua." The one work in Chinese to address the phenomenon comprehensively is Tang Juan, "Dingxiang Wang xinyang de qiyuan ji qi zai Xinjiang diqu chuanbo yanjiu."
13. Goossaert, "The Shifting Balance," 7–8; "Managing Chinese Religious Pluralism"; Johnson, "City-God Cults," 365–68, 373, 388–91; Qin Huitian, "Chenghuang kao," in Huangchao jingshi wenbian, 55:14b–15a; Rong Zhen, Zhongguo gudai minjian xinyang yanjiu, 179–86.
16. Chen Zuolin, "Bingzhu li tan," in Jinling suozhi jiu zhong (xia), 311; Sangren, History and Magical Power, 50, 157; Wei Rongguang, GX 14.9.3, "Zou wei Hunan Shanhua xian chenghuang yongzhen Dingxiang Wang, Zhejiang Zhaobaoshan Guanyin dashi xuan zhu ling ying, qing zhi ge banbian jia fenghao shi," FHA 03–5548–108; Dai, Xiyu beiming lu, 496–97; Peng Mingjun et al., comps., Hunan Dingxiang Wang Nanjing xinggong zhilüe, 3–6.
18. Yu Zhi, Deyi lu . I am very grateful to Prof. Katherine Alexander for sharing this work with me, and to Prof. William Ma for advice on interpreting the image. My points of reference are Wang Qi, San cai tu hui , and Doré, Researches into Chinese Superstitions , Vol. X: Boards of Heavenly Administration. Dingxiang Wang's images closely resemble that of the guardian of the Huai River in Doré, 92 (facing).
28. Suiding xian xiangtuzhi, 203 ; Suilai xian xiangtuzhi, 80–81 ; Tulufan zhiliting xiangtuzhi, 133 ; Qitai xian xiangtuzhi, 40 ; Ruoqiang xian xiangtuzhi, 314 ; Wang Penghui, Qingdai Minchu, 78; Dai, Xiyu beiming lu, 496–97; Xuxiu Wusu xian zhi .
32. Kataoka Kazutada in his brief study of temples in late Qing Xinjiang lists Dingxiang Wang and Fangshen as separate deities, but as we will see, an attention to the temples' contexts where they are listed in gazetteers immediately reveals that they involve the same god. ShinchōShinkyō, 291 . For details, see Hetian zhilizhou xiangtuzhi, 394 ; Kuerkala Wusu zhiliting xiangtuzhi, 170 ; Ningyuan xian xiangtuzhi, 209 ; Kuche zhiliting xiangtuzhi, 318 ; Shache fu zhi, 661 ; Wensu xian xiangtuzhi, 259 ; Xuxiu Wusu xian zhi, juan shang, 27; Yecheng xian xiangtuzhi, 372 . Wang Penghui ( Qingdai Minchu Xinjiang Zhendi dao ) indicates that the Wusu temple was established in 1892. Given that Fangshen's statue was actually installed inside a preexisting Huoshen temple, I would suggest that the 1892 date actually refers to the establishment of that temple. Similarly, the Ghulja (Ningyuan) Fangshen temple is meant to date to 1884, which is very early relative to other records of Fangshen. This may indicate that it was actually a different temple.
36. Goossaert, "The Shifting Balance," 25; Peng Mingjun et al., comps., Hunan Dingxiang Wang Nanjing xinggong zhilüe , 1–2, 4–5; "Xiangren fandui chaihui chenghuang miao," Shenbao, MG 2.1.5 (January 5, 1913); Yi Baisha, Diwang chunqiu , 89.
37. Personal communication and observations, Changsha, June 2017.
38. Chen Zuolin, "Bingzhu li tan," in Jinling suozhi jiu zhong (xia), 311; Wang Fuzhi, Yongli shilu, juan 7, "He Duzhang liezhuan," 7b–8b.
41. Zhao Yushi, Bintui lu, 8:14a. There are several known manuscripts and prints of this work, written in 1224, and not all of them include the section on city gods. A Song printing that does include the relevant phrase is reproduced online in the Airusheng database, with a note indicating the print's re-covering in Yuantong 2 (1334, see the final page of the afterword), apparently demonstrating its antiquity. Several other city gods listed bear the title of "king" (wang). A 1646 printing in the Shuofu collectanea is abridged and does not include this section, but it appears in the 1752 printing, which is based on the Song version. The most complete reproduction of the 1752 printing is Qi Zhiping, ed., Bintui lu, 1983. Some modern reproductions have Hunzhou for Tanzhou, which is simply a printing error.
42. None of the intermediate Changsha and Shanhua gazetteers names the city god of Shanhua. Their temples did not move from their locations east of the yamen. Fang, Changsha fu zhi, 6:55a–58a; Su and Tan, Changsha fu zhi, 7:51a–58a. A Ming-era inscription does not name the city god, but indicates that his title was not "king" (wang), but "marquis" ( hou). Huang Qiazhong, "Chongxiu Shanhua chenghuang miao ji," in Chen Yunrong, Xiangcheng fanggu lu, 144.
43. Jiaqing Changsha xian zhi, also reproduced in Liu and Zhang, Changsha xian zhi, 36:2. It should be noted that Dingxiang Wang is distinct from the similarly named Ding Wang. Ding Wang was the son of Emperor Jing of Han (r. 188–141 B.C.E.) who was enfeoffed at Changsha. His worship continued there through the imperial period, and a neighborhood, Ding Wang Tai, is named for him. (Sun and Yang, Changsha fu zhi, zhengsi ji .)
44. Wang Xunxiu and Wang Yunying, comps., Jiaqing Shanhua xian zhi, 11:26b; 20:5b. The 1613 Shanahua County Gazetteer, compiled by Huang Qiazhong, is apparently lost, and not all of it was reproduced in later versions of the gazetteer. If found, it would surely shed light on the religious history of the Changsha area.
45. Peng Mingjun et al., comps., Hunan Dingxiang Wang Nanjing xinggong zhilüe, 1–3; Wu Shenglan, Chu nan xiao ji . Li Fengsheng's identity as an acting magistrate and jinshi is verified in the Jiaqing-era, Wang, Shanhua xian zhi, 15:4a . It indicates that Li actually arrived in 1730.
47. Yanggu County was in Shandong, not Sichuan.
48. Zhongguo minjian gushi jicheng, Xinjiang juan (shang ce), 188–89.
50. A contrary hypothesis, that Fangshen's presence in Xinjiang genuinely predated the Muslim uprisings, has been advanced by Chen Guoguang in "Xinjiang 'Fangshen' ben shi Qingdai Xiang ji aiguo zhi shi"; "Xinjiang 'Fangshen' ben shi shubian aiguo zhi shi." Chen argues that the contrasting geographies of Dingxiang Wang and Fangshen cults suggest the spread of Fangshen northward from Kashgar, where the deity was supposedly martyred, and the intermingling of his cult with that of Dingxiang Wang as it was brought southward. According to this hypothesis, the beliefs merged in the late Qing. I concede that Han Chinese garrison soldiers in Xinjiang before the Muslim uprisings must have worshipped gods, and that, even though Fangshen is nowhere to be found among the temples listed in Chinese sources from that period, it is possible that a self-sacrificing garrison soldier was deified. Chen's argument, however, rests entirely on the Wensu and Kucha narratives, which date to 1907–1909 and therefore provide no direct evidence for beliefs and practices before 1864. That is, to believe that Fangshen existed before 1877 would require us to assume the genuine antiquity of these stories about miraculous deeds performed by valiant Hunanese exiles, that are only known from after 1877, that differ in content, and that were told by Hunanese soldiers far from home. Moreover, the political implications of these stories, which rhetorically bind this specific part of the Qing borderland specifically to Hunan, make sense in the post-1877 context but much less so in the earlier period. It also seems unlikely that Chinese beliefs and practices mostly survived the Muslim uprisings, given how much destruction was visited on the Chinese community that did live in Xinjiang before them. On the strength of the evidence, the hypothesis that Fangshen was Dingxiang Wang, nativized, seems to be the more plausible one.
52. "Liu Shinan" could be Liu Jiade (b. 1838, Huoqiu, Anhui), a Hunan Army veteran who was prefect of Yarkand at this time. Alternatively, it may have been Liu Zhaosong from Xiangxiang, Hunan, then magistrate of Dihua and another Xiang Army veteran and follower of Zuo. Liu Zhaosong's daughter married the son of then-governor Pan Xiaosu (b. 1838) in 1903, suggesting that he was close to the center of the Xiang Army leadership. Numerous archival documents detail their careers, but both are discussed in FHA 03–5403–089, GX 27.6.20, Rao Yingqi, "Zou wei wei ren Liu Zhaosong shuli Shache zhilizhou zhizhou deng yuan que shi." The marriage is documented in FHA 04–01–12–0624–084, GX 29.2.22, Pan Xiaosu, "Zou wei yu Hetian zhilizhou zhizhou Liu Zhaosong shu ernü hunqin guan xiaozhe li ying huibi shi."
53. Translation from James Legge, Chinese Classics, She King , 382. The gazetteer text as reproduced reads lai fang li si, with li "rites" standing for the similarly written yin "sacrifice." I am grateful once again to Onuma Takahiro for urging me to pursue this classical reference.
54. Schluessel, Land of Strangers, chapter 2, provides a more detailed account of the army's composition. See also lists of known officers and soldiers in Wang, Xiangjun shi , and Zhao, Xiangjun jituan .
64. Duara, "Superscribing Symbols," 778.
65. On drama: Jarring Prov. 207, "Butlar üchün ta'ī qilinghan öylärning bayāni," "Butlarning bayāni," and "Chang chiläning bayāni"; British Library, India Office Records, L/P&S/10/825 File 2273/1919 Kashgar: monthly diaries 1912–1920, Reg. No. 924, December 1915; Jarring, "Culture Clash." On processions: SOAS Archives PP MS 8 #57 E. Denison Ross, Kashghar: Dialogues; and Poskami, Kitabi Äbdullah, 147.
68. Chou, "Frontier Studies," 229–44.
69. GX 8.3.1 Liu Jintang, "Shenling xianying ken ci bian'e fenghao zhe," in Liu Xiangqin gong zougao, 3:17a–20b.
71. GX 8.3.1, Liu Jintang, "Shenling xianying ken ci bian'e fenghao zhe," in Liu Xiangqin gong zougao, 3:17a–20b.
75. Rao Yingqi, GX 23.8.8, "You chen Zuo Zongtang zhuanci ban bian'e pian," in Gongzhong dang Guangxu chao zouzhe, vol. 11, 147–48.
76. Hedin, History of the Expedition in Asia, 228; Ningxiang xian zhi, gushi 10, xianmin 34:4a–6b; Qing shi liezhuan, juan 61; Wu Aichen, Xinjiang jiyou, 39; Wu Yinsun, GX 32.4.10 "Zou wei yigu Yi-Ta dao diaoren Xining dao Yinglin zhengji zhuozhu qing zhun fu si shi," FHA 04–01–12–0649–111; Zhou Hong, "Wan Qing Minguo Xinjiang Han ren zhuti," 52.
81. Porter, "Manchu Racial Identity."