PART TWO
LOCAL ELITES IN TRANSITION

Four
From Comprador to County Magnate: Bourgeois Practice in the Wuxi County Silk Industry

Lynda S. Bell

One of the most difficult problems concerning local-elite roles in China during the late nineteenth and early twentieth century is the analysis of new industrial elites. Writing in the Cambridge History of China , Marie-Claire Bergère has presented one prevalent point of view on such men in Shanghai in the early decades of the twentieth century. Bergère argues that members of this group had the motivation necessary to develop a strong industrial order but lacked the government support necessary to accomplish their end. Chinese governments seemed more interested in taxing and controlling industrial enterprises than in developing sufficiently strong fiscal policies to support them. Moreover, as relative latecomers to industrial development, Chinese entrepreneurs depended on the world market to sell their new products, but they were unable to operate on an equal footing with their more advanced international competitors. As a result, China remained weak "as a nation" and failed to generate the necessary stage of "economic take-off."^[1]

Although these last points suggest there may have been larger developmental dilemmas at work, the overwhelming impression left by Bergère's analysis is that the problems of China's modern industrialists were primarily political, as they dealt constantly with inept governments at home and aggressively dominant competitors abroad. These are significant issues, but to fully evaluate industrial elites I would argue that there is also a second set of problems to consider. These concern the socioeconomic position of such men, including their relationship to China's predominantly rural order.

Perhaps the easiest way to explore these equally important issues is to look at lower levels in China's economic hierarchy. The research I present here deals with events in a place of this type—Wuxi county, a semirural, semi-urban district located inland from Shanghai, in the heart of the Jiangnan region.^[2] Although there was no single critical event, the nineteenth and twen-

tieth centuries were cumulatively a time of great change in Wuxi. Elites and peasants alike took part in changes in the local economy, including the introduction of modern technologies and participation in new international trade networks. Elsewhere I explore broader issues of economic development associated with these changes, but in this paper I am concerned particularly with new social relationships and stress how these were shaped by what I call "bourgeois practice."^[3]

By using this concept, I try to pinpoint precisely what it meant to be bourgeois in early twentieth-century Wuxi. Industrialists' behavior included most activities that we associate with the bourgeoisie as defined in European or American contexts. They invested capital in industry, managed factories that employed wage labor, sold products in impersonal national and international markets, organized to promote their economic interests, and sought to influence officials in drawing up favorable economic regulations. But this was not the whole story of their bourgeois behavior. As we will see, new industrialists also built upon old patterns of economic and social dominance developed during Qing times by gentry-merchants of the Jiangnan region.

Given these observations, the twentieth-century industrialists of Wuxi may, in one sense, be considered hybrid types, somewhere between precapitalist merchants and the bourgeoisie of an advanced industrial society. Yet in another sense, they were also actors in a process that had no exact parallel in other national settings. This is why "practice" is also an integral part of my discussion. I use practice here in the sense that Pierre Bourdieu and Marshall Sahlins, among others, have used it. Bourdieu defines practice as social strategies that move through time, creating, recreating, or transforming hierarchical human relationships. Moreover, symbolic behavior is an important part of practice, constantly interacting with material life in different settings to shape unique patterns of dominance. Sahlins works from the same general assumptions but emphasizes that even in one culture practice does not always repeat itself. When historical circumstances change, evolving social strategies may fundamentally transform old patterns of dominance.^[4]

Drawing on these discussions, I emphasize the unique historical process through which silk industrialists emerged in Wuxi and the ways in which they built their strategies of social and economic dominance. Members of this group did not accumulate wealth, power, and prestige through industrial investment alone; they also relied on elite-management activities and political styles developed in an earlier era. Moreover, behind relationships between new industrialists and peasants were significant changes in landlord-ism and the increasing importance of gentry-merchant roles within Jiangnan society. Because these developments had their roots in Qing times, I will begin by looking at the evolving nature of elite dominance in Jiangnan during that dynasty.

Gentry Dominance in Jiangnan During the Qing Dynasty

The term Jiangnan means "south of the river," and it refers to the vast, rich expanse of alluvial plain situated on the central China coast at the mouth of the Yangzi River. This area, encompassing portions of southern Jiangsu and northern Zhejiang provinces, has a long history as one of China's most significant agricultural regions. Extensive development of Jiangnan began as early as the third century A.D. , as the Han dynasty collapsed and nomadic tribes from the northern steppes began to push into China proper, setting off waves of southward migration that lasted for several centuries. In time, indigenous population growth combined with these patterns of immigration to make Jiangnan a densely settled and highly productive agricultural region. Accumulating wealth from agriculture fostered the development of commerce, cities, and intellectual life. Since the tenth century, a substantial proportion of successful examination candidates to staff China's sophisticated civil service administration came from the ranks of the Jiangnan elite.^[5]

Although the agrarian development of Jiangnan was an ongoing, long-term process, crucial watersheds can also be identified. For the purposes of our analysis, we should pay careful attention to such a period in the seventeenth century, at the time of the Ming-Qing transition. The principal methods of achieving and maintaining elite status in Jiangnan in the late nineteenth century originated during this time. A brief review of the major findings of recent research on these issues will more sharply focus the interrelationships among changes in land tenure, patterns of commercialization, and new strategies for building local dominance.

A large literature deals with landholding patterns and changing land-tenure relationships in the late Ming and early Qing periods in Jiangnan. All the theoretical debates on the significance of these issues cannot be summarized here, so I will concentrate on the general contours of the process itself, something on which most scholars generally agree.^[6] Landholding patterns during this important transitional period generally moved away from large estate-type holdings under the management of a single resident landlord toward small, fragmented holdings, farmed by relatively independent peasant households. Landlordism certainly did not disappear during the Qing period in Jiangnan, but it took on a significantly new form. There were several distinctive features of this new system: a high level of absentee proprietorship, with many of the largest landowning families moving permanently to urban settings; the collection of rent by bailiffs or later by organizations representing collective landlord interests known as "rent bursaries"; and finally, the shift from sharecropping practices to fixed rents, rent deposits, and permanent tenancy rights.^[7]

On the one hand, it is possible to argue that the overall position of the

peasantry improved during the development of this new pattern of land tenure and social relationships in Jiangnan. Growing commercialization reinforced the struggles of peasant farmers against various forms of legally defined servileness through which they won commoner status. Peasants also fought for owning "surface" rights to the land, a kind of permanent tenancy arrangement. By the eighteenth century, most land in Jiangnan had both "surface" and "subsoil" ownership rights. Landlords held the latter and hence had the right to collect rent on the land; tenants held a permanent claim to the former, in reality, a privilege to continuously use the land as long as rent payments were met. Both surface and subsoil rights could be freely bought and sold, so that, in any given locale, a complex and constantly changing pattern of land ownership emerged in which wealthier peasants benefited most. However, this system also made it increasingly difficult to dislodge any tenant farmer from his position as cultivator of a given parcel of land, and tenants as a group were now in a better position to resist rent payment when they judged a landlord's demands unreasonable. As a result, incidents of rent resistance increased as the period wore on.^[8]

On the other hand, landlords responded to this situation with new methods of tenant coercion and control. For example, they increasingly required rent deposits, large lump-sum payments made at the beginning of a tenancy agreement. With the receipt of these deposits, landlords protected themselves from the possibility of rent default at a future date. As sharecrop-ping gave way to fixed rental agreements, rent deposits also helped reestablish landlords' claims to shares of increasing surpluses generated by tenant farmers. Second, landlords benefited from a series of national tax reforms. Beginning in Ming times, these tax reforms involved the eventual merger of the land tax and converted cash corvée payments on individual, working-age males. By the eighteenth century, all taxes came only from landowners, causing local officials and landlords alike to view taxes as ultimately dependent on landlords' continuing ability to collect rent. The state thus had a clear interest in helping landlords resist rent default, using such means as sanctioning arrests of defaulters by the landlords' rent bursary organizations and incarcerating defaulting tenants in county-operated prisons.^[9]

An important conclusion to be drawn from these observations on land tenure patterns is that both landlords and tenants developed new ways to assert their rights in the changing Jiangnan countryside. In the face of growing peasant independence, landlords struggled to assure their continued hold on elite status and, with state support, devised various new methods for doing so. A useful way to view this process has been suggested by Kathryn Bernhardt, who characterizes this trend as movement away from power built through tight patron-client relationships between individual landlords and their own tenants to a new more broadly based form of elite dominance that she calls "gentry hegemony." She uses this term not only to characterize the

new situation in which the Qing state supported landlord economic rights vis-à-vis defaulting tenants but also to refer to a more generalized pattern of gentry dominance through active community leadership, in contrast with the previous more private and direct control of resident landlords over dependent peasants on their estate-type holdings. Under these new conditions, individual gentry members often took on new roles as organizers of communitywide irrigation projects and rural relief efforts. These activities further strengthened the elite's growing tendency to undertake multiple, nonparticularistic leadership roles, designed to enhance both community development and local social control.^[10]

We have thus seen that new styles of gentry leadership emerged in Jiangnan against the backdrop of important changes in land tenure and shifts in relationships among elites, the state, and rural society at large. As a final step, we must also consider two other equally significant trends—the commercialization of Jiangnan's rural economy during Qing times and the continued expansion of elite functions via gentry-merchant roles.

The general contours of late imperial commercialization in Jiangnan are well known. As population increased, land productivity also improved, a process fueled by development of a wide range of labor intensive agricultural techniques. Peasants further enhanced land productivity through localized crop specialization; for example, they converted portions of land from grain production to cash crops, such as cotton or mulberry trees to feed silkworms. These developments also spurred the growth of peasant household-based cottage industries, especially cotton and silk yarns and textiles. By the eighteenth century, these converging trends in population growth and commercialization resulted in Jiangsu and Zhejiang provinces becoming grain deficient. To pay for grain imports, Jiangnan sold its handicraft products outside the region, contributing to the nationwide escalation of long-distance, interregional trade.^[11]

Within this evolving economic context, the possibility for individuals to accumulate private fortunes through commerce multiplied. In addition to long-distance trade, commercialization led to widely dispersed marketing within the Jiangnan countryside, with permanent, intermediate-level marketing centers emerging.^[12] At this lower level of the economic hierarchy, locally important gentry families began to assume new merchant roles. Pingti Ho has presented evidence, for example, that local "official" families in Jiangnan engaged increasingly in commerce throughout the Qing period.^[13] As Mary Rankin has further demonstrated, by the late nineteenth century, this strategy was widely used in many Jiangnan locales, as commercial and manufacturing opportunities grew at an even more dizzying pace.^[14] Moreover, this volume's essays by William Rowe and Madeleine Zelin suggest that such gentry strategizing was an empirewide phenomenon, emerging during Qing times in commercialized areas of Hubei and Sichuan as well.

Overall, then, the changing nature of landlordism, the proliferation of state-sanctioned elite management of community-oriented public institutions, and the development of gentry-merchant roles created a complex and highly diversified pattern of expanding elite functions during the late imperial period. However, we should also underscore that throughout this process achieving degree-holding rank—the legitimate claim to gentry status— remained the ultimate mark of social pedigree within the system as a whole. Perhaps the best evidence to support this claim is that men who made their initial fortunes through commerce either continued to purchase degrees or provided classical educations for family members who might eventually succeed in the more regular examination route to degree-holding status.^[15] Thus commercial wealth was one way to bolster family fortunes but was insufficient in and of itself to legitimate elite status. Although it was now possible to enter the informally defined, lower ranks of elite society through activities geared toward local development and social control, these patterns of social change had not produced an alternative to gentry degree holding as the principal criterion for upper-elite status. For instance, the type of gentry networks and associational activities that Timothy Brook describes in this volume for Ningbo persisted throughout Jiangnan during Qing times, despite the growing interest in trade among many gentry members.

The result of this particular trajectory of regional social change in Jiangnan was that, as elite functions diversified, they also tended to concentrate within a relatively small number of degree-holding or potential degree-holding families. Late Qing local histories often refer to a strata of local notables known as shendong , a term we usually translate as "gentry managers." Everyone could identify these men—a small core of local leaders with power to mobilize their communities at large for any and all forms of social, political, or economic activity. Many gentry managers were degree holders, but this was not an absolute prerequisite for entry into this locally defined functional elite. Literacy and wealth seemed far more important, yet most wealthy, literate families also aspired to have, or already had, a member with degree-holding status. Thus, our picture of local-level elite society in late Qing Jiangnan is that of domination by a small, self-contained group of prominent families, engaged not only in landownership and degree seeking but also in assuming a multitude of new local managerial roles.^[16]

Commercialization and the Wuxi County Elite

When industrialists appeared on the local scene in Wuxi in the late nineteenth century, they built upon elite strategies already at work, developing their bourgeois practice by combining leadership of new industrial enterprises with other forms of local social management. A brief look at commer-

cialization in Wuxi will provide a clearer picture of how the local socioeconomic environment fostered this trend.

The eighteenth century was a period of substantial commercial development in Wuxi. The main impetus came from two sources: Wuxi's role as a key marketing center in the growing interregional rice trade and the involvement of local peasants in cotton textile production. We have already seen the general contours of the rice trade for Jiangnan as a whole, so it remains only to point out that Wuxi became a major transshipment point for rice within Jiangnan during the eighteenth century. Strategic positioning made this development possible. Not only did Wuxi sit on the northern shore of Lake Tai, a major transport link between southern Jiangsu and northern Zhejiang, but the Grand Canal also passed directly through the middle of the county, only a few kilometers south of its confluence with the Yangzi. Many smaller canals and rivers also converged in Wuxi, making it one of the best water-based transport hubs in the entire Jiangnan region.^[17] Rice imports came down the Yangzi River to feed Jiangnan's burgeoning population, and requisitions of "tribute grain," a supplementary tax assessment levied on central China's most prosperous rice-producing provinces, also passed through Wuxi on their way north via the Grand Canal to Beijing. The combination of these two trends led to a substantial rice market in Wuxi, where officials responsible for meeting tribute grain quotas as well as merchants involved in the long-distance rice trade congregated.^[18]

Concomitant with these developments in the rice trade, peasants in Wuxi also began to produce cotton textiles for market. Although it is difficult to map this trend with any great precision, population growth probably contributed to this new stage of intensified labor efforts by Wuxi peasant households. Throughout Jiangsu as a whole, population doubled from the early eighteenth through the early nineteenth century, and gazetteer figures suggest that the rate of growth may have been even faster in Wuxi.^[19] In response to both declining land-per-person ratios and the increasing need to purchase additional grain supplies, many peasant families in Wuxi began to spin and weave cotton during slack agricultural periods to make goods that they could sell at market or keep for home use and pawn when money was needed. Although Wuxi peasants did not grow cotton themselve cotton imported from surrounding counties provided the raw materials necessary for this new sideline activity.^[20]

Trade in rice and cotton provided two key components of countywide trends in commercializing Wuxi during the eighteenth century. Not entirely clear, however, is exactly when gentry members began to participate in merchant activity. Jerry Dennerline has gathered evidence in the form of local folklore suggesting that gentry members began to sponsor market-town development in southeastern Wuxi as early as the sixteenth century. Denner-

line also points out that even rather small landlords developed local granaries and sold their surplus rice at market.^[21] In lieu of more comprehensive work on merchant communities in Wuxi during Qing times, Dennerline's observations allow us to suggest that as population grew and land available for purchase declined, even before the eighteenth century some gentry families turned to commerce as a supplement to their land-based income-earning capacity.

To enhance their long-term staying power in a rapidly changing economic environment, prominent families also built lineage organizations, a trend that seems to have gained momentum in Wuxi during the eighteenth century. Again, Dennerline's work gives us some insight into this process. Branches of the Hua and Qian families, living in the vicinity of the important market town of Dangkou, each organized lineage-based charitable estates in the eighteenth century, pooling some wealth from several agnatic kinsmen. On the one hand, these estates built their corporate presence by purchasing land that then became the permanent property of the estate. At least in part, this was a response to the problems generated by increasing population, partible inheritance, and a rapidly evolving land market. From the viewpoint of those lineage members who were also part of the local gentry-elite, a corporate estate provided a means to keep ownership of a core of landed property within the control of a kinship-based group that they themselves could lead. On the other hand, these estates provided local charity and relief, enhancing the moral reputation of local lineage leaders.^[22]

In the nineteenth century, as subordinate branches of the original Qian and Hua lineages began planning to organize their own charitable estates, a new strategic twist materialized. A group of Huas in Dangkou who wished to found a new estate had no degree holders among them; instead, they were "local merchants, manufacturers, and purveyors of the wine and soy sauce for which Dangkou was known."^[23] owever, by the time that Hua Hongmo, son of one of the original four brothers who had proposed the estate had formally organized it in 1873, Hongmo himself had passed the prefectural and provincial examinations. Under his directorship, the "New Hua charitable estate" amassed landholdings of nearly four thousand mu (one mu= one-sixth acre) but also developed an important new method for increasing its corporate wealth. Hua Hongmo's own private rice warehouses in Wuxi city were turned over to the charitable estate, thereby "channeling revenues earned from storing the grain of inland merchants en route to Shanghai into the Huas' Dangkou projects."^[24] Thus, in Hua Hongmo, we have a classic case of how multiple roles in social management built elite position. Not only did Hongmo seek and receive degree-holding status, but he also fashioned his local career through managing substantial landholdings on behalf of his lineage organization. Meanwhile, he made a private fortune via the increasingly lucrative Wuxi grain trade. By turning his warehouses

over to the charitable estate, he assured that the source of this fortune would not be divided among his heirs upon his death but would remain intact, a means to promote the economic well-being and local prestige of the Dangkou Huas at large.

This was the tradition of elite strategizing, increasingly merging careers in commerce, scholarship, and gentry management, into which local industrialists entered in late-nineteenth-century Wuxi. We should turn now to that story itself—the world of the local silk industry elite.

The Origins of Bourgeois Practice

The late nineteenth century was a tumultuous time in China. To analyze elite roles in industrial development, at least three major events must be noted. These are the Opium War (1839-42), the Taiping Rebellion (1851-65), and the New Policy reforms (1902-11). Trends set in motion by these events had a substantial impact on the economic activities and political orientation of Jiangnan local elites.

Of these events, perhaps the relationship between our story and the Opium War is most obvious. The Treaty of Nanking, signed at the conclusion of this war, made Shanghai one of five "treaty ports" open to foreign business and trade. This marked the beginning of Shanghai's development as a major new commercial city as well as China's most important center of new-style industry. Early industrial efforts were usually financed with foreign venture capital, but Chinese investors soon became active as well. By the turn of the twentieth century, a large proportion of Shanghai's new industry was Chinese-owned. Moreover, by that time, a kind of "spin-off" effect was also at work: men who began their careers as comprador agents for foreign firms in Shanghai began to return to their homes to develop industry there as well.

The relationship of the Taiping Rebellion to the story of industrial elites in Jiangnan is at first glance, not so apparent. However, for Wuxi, the Taiping period marked a crucial watershed for evolving gentry management in the Jiangnan region. Because wartime devastation in many Jiangnan locales had been so severe, rural rehabilitation efforts led by gentry members were essential in the post-Taiping period. The pace of gentry management escalated, and its scope expanded; thus, in many places, new patterns of commerce and industry also became part of overall plans for the region's economic reconstruction.

To complete our picture of shifts in elite behavior during these years, we must also consider the political impact of the New Policy reforms. These reforms marked a critical period in rethinking political structures, especially at the local level, and had important ramifications for the range of political activity undertaken by local elites. Perhaps the most important events within

the context of the reforms were abolishing the examination system in 1905 and restructuring local-level recruitment into political service. Provincial assemblies and local "self-government" bureaus now served as recruiting grounds for men to move into higher levels of government service. However, locally based, interest group-oriented organizations, including merchants' and industrialists' guilds, which had increased throughout the Qing, and chambers of commerce, which appeared in the New Policies period, also emerged at the local level. As commercialization and industrialization escalated, these sorts of organizations developed in many locales, serving as vehicles of self-regulation and self-protection for local businesses and also acting as collection agencies for commercial and industrial taxes. With the New Policy reforms, these organizations received official sanction to undertake new quasi-governmental roles.

Clearly, these major events of the late-nineteenth and early-twentieth centuries affected local elite behavior, especially in the rapidly developing Jiangnan region. However, it is difficult to assess exactly how these new developments differed from past elite activity. A case study based on one industry in a single locale cannot provide a definitive answer. However, the silk industrialists of Wuxi are an important example because they represent a group at the forefront of the most advanced types of local-level change, both economic and political, during this period.

Based on what we will see of the activities of the silk industry elite in Wuxi, I will argue that even the most progressive industrialist among them, Xue Shouxuan, built upon previous patterns of elite behavior to create a familiar style of local dominance. Because he used commercial links to the peasantry to build a fully integrated silk industry, the continuation of a peasant family-based rural order dominated by local elites was essential to his goals. In the political sphere, although the examination system had long since been eliminated, we will see that Xue also continued to build upon local practices linked to gentry management, constructing his own network of personal political relationships from the county level upward. Despite these continuities, Xue's behavior differed in its content. The kind of commerce he promoted brought the peasantry into an international marketing system for raw silk, and his filatures used the most advanced technology and management practices available. His total repertoire of economic and political activities thus assumed a hybrid form as they grew out of and built upon existing patterns of local social dominance. The results of this combination were powerful. Industrial development resulted in a high degree of commercial dependency among the peasantry, as Xue built a system of economic control that was unusually tight, even by Jiangnan standards. To make these arguments, I must begin with the early stages of silk industry development in Wuxi, namely the establishment of commercial links between local elites and peasantry via the cocoon-marketing system.

Gentry, Peasants, and Compradors in Early Wuxi Silk Production

Two principal activities occurred in the first stages of silk industry development in Wuxi. First, peasants switched from cotton handicraft production and began to raise mulberries and rear cocoons for sale to Shanghai's new steam-powered silk filatures. Second, comprador agents began to establish commercial firms that acted as cocoon purchasing agents for Shanghai filatures. Cocoon marketing was based not only in Wuxi but also in other areas new to sericulture, including several counties to its west and north.^[25] However, because Wuxi was so favorably located in terms of water transport and had such extensively developed commercial networks, it emerged as the most important point of both new cocoon production and trade within the Lower Yangzi region.^[26]

The introduction of sericulture and the development of the compradorbased cocoon-marketing system were inseparably linked in Wuxi; both developed through encouragement and initiative of members of the local gentry elite. To demonstrate these linkages, we must turn now to a step-by-step breakdown of the early years of silk production in Wuxi. Elsewhere, I explore in detail the motivations of the Wuxi peasantry for engaging in sericulture.^[27] Here, I emphasize the local gentry's role at each stage of silk industry development.

Sericulture Promotion in the Post-Taiping Period . In the decades following the Taiping Rebellion, many counties in the Lower Yangzi region required substantial agricultural rehabilitation and development. Wuxi fell within the area hardest hit by depopulation through death, flight, and agricultural devastation.^[28] Local gentry members were concerned that land be resettled and that peasants become productive again as rapidly as possible. These plans were usually rationalized through typical Confucian morality. Simply, it was the duty of elites to promote the people's livelihood following a period of social and economic upheaval. However, reconstruction also made good economic sense because both tax and rent collection depended upon a fully flourishing agrarian sector. Such elite efforts were not unique to Wuxi; Mary Rankin has detailed the general parameters of agrarian rehabilitation of Taiping-devastated counties in northern Zhejiang and has demonstrated that these conditions allowed increased elite management activity throughout Jiangnan in the latter decades of the nineteenth century.^[29]

An important aspect of post-Taiping rehabilitation in Wuxi was the introduction of sericulture. Although Wuxi bordered the traditional silk handicraft areas of southern Jiangsu and northern Zhejiang, there had been no substantial mulberry cultivation prior to the Taiping period.^[30] This situation changed rapidly after 1870, with the development of the new Shanghai-based

filature industry. Under these circumstances, gentry members, behaving as traditional elites promoting agricultural improvement, introduced a substantially new form of mulberry cultivation. In contrast to the older sericulture districts south of Wuxi, where mulberry trees were grown on scattered embankments between rice fields, gentry members in Wuxi now encouraged cultivation in relatively large tracts.^[31] Moreover, these gentry-sponsored innovations linked the peasants who cultivated these trees to a new chain of production relationships, dependent ultimately on the Shanghai trading sphere and the international market for Chinese filature silk.

Memoirs of two elite members of Wuxi society discuss how they promoted new-style sericulture. The first case is Yan Ziqing, a Wuxi native, who had served as a county magistrate in Shaanxi province. Yan returned to his native town of Zhaimen in the northeastern Wuxi subdistrict of Huaishang in 1871. A posthumous collection of his writings records that one of his first projects upon returning home was to import three thousand mulberry saplings from Jiaxing county in Zhejiang, an area long famous for sericulture and the production of handicraft silk. Yan had these saplings planted on a thirty-mu plot and erected a large residence next to it. He proclaimed his hopes for the success of this endeavor with a stone engraving above the door of the residence, calling it the "Cottage for the Study of Mulberry Cultivation."^[32] The second case is Hua Guanyi, designated simply as a "large landowner." In a letter written to his daughter-in-law in 1873, Hua recounts the planting of three thousand mulberry saplings on fifteen mu of land. He also discusses the importance of establishing the ideal ratio of trees to land, in the case of his land, about 230 to 240 trees per mu . At this ratio, Hua knew that peasants could harvest about twenty jin of leaves per tree (one jin= one-half kilogram) and that each jin could be sold for 1.30 to 1.40 Chinese yuan (dollars).^[33]

Due to the efforts of men like Yan and Hua, sericulture spread rapidly in Wuxi during the next several decades. In the 1881 revised version of an important Wuxi gazetteer, a new entry for silk added to the category of local products stated:

In the past, silk was produced only in the subdistrict of Kaihua. From the early years of the Tongzhi reign period [1862-74], conditions have been in flux with much vacant land and many people [coming to resettle it]. This has created an ideal situation in which to begin the planting of mulberries and the rearing of silkworms. Sericulture has begun to flourish, and has spread to every subdistrict.^[34]

By the 1920s, one-third of all arable land in Wuxi was devoted to mulberry, and virtually all peasant families were involved in cocoon rearing as a subsidiary cash-earning occupation.^[35] Despite the positive gentry role in promoting sericulture, elite efforts alone did not cause this dramatic shift in

agrarian production patterns. Factors related to the structure of peasant farming in Wuxi were also at work, helping to rapidly integrate sericulture into the overall peasant family work effort.

Sericulture and Small-Peasant-Family Farming . The overriding consideration for Wuxi peasants when they opted for sericulture was the promise of substantial cash earnings.^[36] In turn, this search for new income led them into a new form of dependence on the marketplace and the elites who controlled it. Detailed data from three villages in the early twentieth-century countryside in Wuxi demonstrate that peasant families relied on sericulture for at least 50 percent of their families' cash incomes.^[37] In the 1930, such income was essential to maintain basic peasant livelihood in Wuxi because rice and wheat farming alone was insufficient to feed the existing rural population. Nearly all peasant families purchased grain to supplement their own production of rice and wheat and they needed cash income from cocoon sales and other subsidiary occupations to do so.^[38]

Behind these circumstances were factors traceable to the situation in the post-Taiping Wuxi countryside. Population statistics reveal that Wuxi lost as many as one-half to two-thirds of its people during the Taiping period.^[39] However, repopulation was rapid in the immediate post-Taiping years, given Wuxi's reputation as a rich agricultural region; this proved a powerful pull for in-migration from surrounding counties. Especially likely to immigrate were peasants from north of the Yangzi, an area much less fertile and productive than counties south of the river.^[40] Moreover, peasant farming families were also drawn to Wuxi by the attractive opportunity to earn cash through cocoon rearing. Via such subsidiary work, they could use the chronically underemployed labor power of peasant women to boost living standards and long-term prospects for economic well-being.^[41] By the 1920s, Wuxi's population had been restored to pre-Taiping levels, resulting in the second highest person per land ratio in all Jiangsu, with 1.29 people per mu of available farming land.^[42]

Ironically, the rapid repopulation of Wuxi under conditions of sericulture expansion undermined peasant expectations for new levels of prosperity. In three villages for which we have extensive household data, the average size of farming units was only 2.54 mu in 1940, approximately one-half acre.^[43] Because landholdings were small, peasants depended on cash income from cocoon sales to purchase additional grain. Although they were able to augment their grain supplies in this way, these peasants, on average, still fell only within the lower range of grain consumption patterns observed throughout China as a whole.^[44] Moreover, there was no guarantee that potentially profitable sericulture would always provide even this rather basic level of security. Climatic conditions in Wuxi constantly worked against the peasantry's cocoon rearing efforts.^[45] Frequent rain and high humidity during spring and

summer often wreaked havoc with silkworm crops, causing bacterial infections to grow at a rapid rate and wiping out the entire season's efforts. Facing these conditions, peasants in Wuxi, who had enjoyed favorable land-tenure conditions during the labor shortage of the post-Taiping years, gradually fell into patterns of subordination typical of poor peasant life—often mortgaging their land and seeking loans from wealthier members of their communities.^[46]

Thus by the early decades of the twentieth century, Wuxi peasants had replaced cotton cloth production with sericulture. In their hopes for prosperity, they actually committed themselves to continuing sericulture to ensure their basic subsistence. This situation created a new form of dependence between peasant farming households and the Shanghai-based filature industry. Commercial dependency in this new form was a step beyond what peasants had experienced before. Not only was sericulture a relatively risky enterprise ecologically, but it also put them in contact with the international luxury market for raw silk. They were therefore subjected to boom-bust cycles of demand over which neither they nor Chinese filature owners had much control. At the local level, this relationship between peasants and filatures was mediated by a new cocoon-marketing system, run by elites who did business in Shanghai but who, in many cases, came originally from Wuxi.

Development of the Cocoon-Marketing System . Cocoon hang were merchant firms that purchased cocoons from peasants on a contractual basis for early Shanghai filatures. Cocoon firm facilities consisted of a building or set of buildings in which cocoons were gathered twice a year, in spring and fall, and then dried before shipment to their final filature destination. The firms operated in market towns, buying cocoons from peasants living in villages that fell within the town's marketing scope.^[47] In Wuxi, cocoon firms were first organized in the 1880s, making use of Wuxi's position at the center of an extensive canal and river system, which proved ideal for cocoon transport. After 1904, with the completion of the Shanghai-Nanjing railway as far inland as Wuxi, the firms developed at an even more rapid rate. By 1909, Wuxi had seventy to eighty cocoon firms. British, American, Italian, and Japanese firms all sent comprador agents to Wuxi, with the Japanese firm of Mitsui, as well as the British firm of Jardine, Matheson, and Company, among the earliest promoters of cocoon-marketing activity.^[48]

The careers of four early cocoon firm agents demonstrate how Wuxi men with important ties to new urban-based commercial networks developed this enterprise. These men were Zhou Shunqing, a newly rising man of comprador backgound; Xue Nanming, son of the nationally prominent upper-degree holder, Xue Fucheng; and, finally, Rong Zongjing and Rong Desheng, brothers who started their careers in commerical positions established by their father and who became two of China's most important new industrialists.

Zhou Shunqing made his initial fortune as a comprador in the Chinese iron industry. Working first as manager of the British-backed Shengchang iron brokerage firm, he later became sole owner of the firm and opened branch operations in several cities. Zhou then began to follow more traditional routes to elite status. First, he purchased official rank through a personal contact he cultivated with an uncle of the Guangxu emperor. Through this connection, Zhou solicited deposits from members of the imperial court to found one of China's first modern banks, the Xincheng, with branches in Beijing, Tianjin, Nanjing, and Wuxi. When the dynasty fell, the bank closed, and depositors lost 4 million yuan ; yet Zhou survived the crisis, salvaging private reserves of 1.2 million yuan . Second, with his profits in comprador and banking activity, Zhou entered more traditional forms of elite activity, becoming a large-scale landowner in his home county. In the name of his lineage organization, he bought more than three thousand mu of land at very low prices from peasants forced to sell in the drought-plagued northeastern section of Wuxi. He also purchased land for himself in the vicinity of his native village of Dongze in the southern part of the county, an important new sericulture district. He then single-handedly turned Dongze into a bustling market town by building facilities for shops, various local marketing firms, and pawnshops. He called his new creation Zhouxin zhen —"Zhou's new market"—and built Wuxi's first silk filature there in 1904 Among his many early business ventures in Wuxi, Zhou operated seventeen cocoon firms.^[49]

While Zhou's career began with comprador activity and led into more traditional, locally based elite activity, Xue Nanming's career took the reverse route. Born into a bureaucratic, landholding family, only later in life did Xue invest in new commercial and industrial ventures. Xue's father was Xue Fucheng, a holder of the highest level jinshi degree, who had achieved national prestige and position in late Qing times, serving first as adviser to Zeng Guofan and later as ambassador to England, France, Italy, and Belgium. Xue Fucheng, a substantial landholder, also owned a large private granary. Xue Nanming began his own prestigious career in 1888, when Li Hongzhang appointed him to serve in Tianjin as a judge in court cases between foreigners and Chinese citizens. However, upon the death of his father in 1894, he returned to Wuxi and subsequently declined to resume office in Tianjin. Beginning in the late 1890s, Xue Nanming concentrated his efforts on the family's landholdings and new investments in silk production. In 1896, he collaborated with Zhou Shunqing in a brief filature venture in Shanghai. During this period, he also established ten-odd cocoon firms on behalf of Italian filatures.^[50]

The Rong brothers had more modest claims to traditional forms of elite prestige. However, they, too, chose Wuxi as a primary base for new-style investment activity after sojourning careers in commerce. Their father, owning only about ten mu of land, furthered family fortunes by traveling to

Guangdong in 1883, where a relative working in Zhang Zhidong's administration helped him become a collector in two different locations for new lijin , transit taxes on internal trade. He sent his elder son, Zongjing, to Shanghai at age seven to study in a firm manufacturing iron anchors. Soon after he arrived, Zongjing fell ill and returned home, but at age fifteen he went again to Shanghai to study old-style banking. He subsequently was a collection agent in Shanghai until the company for which he worked suffered financial difficulties at the time of the Sino-Japanese war. The younger son, Desheng, was apprenticed to a Shanghai old-style bank, and later worked under his father in the Guangdong Sanshui lijin bureau. Father and son returned to Wuxi in 1895, but in 1899, Rong Desheng briefly returned to Guangdong as an accountant in the provincial tax bureau. In 1900, he returned once again to Shanghai. At that point, he worked in an old-style bank that he and Zongjing had founded in 1896. Rong Desheng also set up a branch of this bank in Wuxi. Together, the brothers established several cocoon firms with funds borrowed from the Wuxi branch bank, on which they paid no interest as owners. They favored their native market town of Rongxiang zhen as the location for their first cocoon firm ventures.^[51]

As the first cocoon firm agents in Wuxi, Zhou, Xue, and the Rong brothers set the stage for the pursuit of business practices that would become increasingly bourgeois. Within a matter of years, each became an industrialist, although not necessarily a filaturist. Zhou and Xue took this route, but the Rongs organized flour processing plants and cotton mills. For our evolving analysis of silk industry elites, most important at this stage is that as comprador agents for the cocoon trade they first combined old-style commercial management with new content: filature-solicited cocoon marketing. As they did so, they set an important precedent for other elites who followed. By expanding their economic activities at home to take advantage of changing times, they established the cocoon-marketing system, an important new commercial network with which all Wuxi peasants would eventually reckon.

Industry, Cocoon Firms, and Gentry-Merchant Roles

Comprador agents with ties to Shanghai initiated cocoon marketing, but without the cooperation of lower-ranking members of the elite social hierarchy in Wuxi cocoon firms could not have been established. These firms needed managers familiar with local marketing conditions who would actually supervise the purchase of cocoons from peasants and oversee the first stages of cocoon processing. Through contractual arrangements with such men, comprador agents could assure the effective organization of cocoon marketing. As the years passed and filatures were built in both Wuxi and Shanghai, compradors became less important, and local cocoon firm owners became predominant in the Wuxi cocoon-marketing system.

The activities of Wuxi cocoon firm owners are significant for exploring the expanding role of local elites in the course of silk industry development. These men not only ran cocoon-marketing establishments but also formed the core of a new political organization, the cocoon merchants' gongsuo , or guild, which played an important role in both local and provincial politics.^[52] In these capacities, members of the Wuxi local elite became progressively more bourgeois, while also using traditional styles of gentry management to build the cocoon-marketing network.

It is difficult to determine the precise social origins of cocoon firm owners in Wuxi, but a few general descriptions of their backgrounds do survive. Japanese investigators visited cocoon-marketing sites in Jiangsu in 1916 and described the owners of cocoon firms as "local wine-makers or wealthy men."^[53] While these designations are somewhat vague, they suggest the sorts of local gentry-merchants previously engaged in an expanding array of commercial and managerial activities within their home locales. These men were ideal candidates for cocoon firm investment—close enough to rural marketing structures to be able to tap immediately into the system, but worth enough financially to afford the initial venture. The Japanese researchers further observed that to build the largest and most up-to-date cocoon firm facilities in the late teens took an investment of "several tens of thousands of yuan ."^[54] Although this figure was probably inflated, it clearly implied that only men of some substance at the local level could either have invested money for such a venture or have been in a position to borrow funds to support it.^[55]

Local men of economic substance were solicited by comprador agents via a complex contracting system to build and operate cocoon-marketing facilities.^[56] A surviving cocoon firm contract from 1887 demonstrates the dimensions of such a relationship between comprador and local cocoon firm owner. This contract—between two individuals, Gu Mianfu, a comprador agent dispatched from Shanghai, and Sun Boyu, a local Luoshe landlord— served as the legal basis of operation of the Renchang cocoon firm with three sites in Luoshe zhen , a major market town in the northwestern part of Wuxi. The signing of the contract was witnessed by a county official so that it could also serve as the official license, establishing local government recognition of the business and sanctioning its operation.^[57]

Ideally, we would like to know how much land Sun owned and what role cocoon marketing played in his overall economic situation in order to place him more precisely in the local social hierarchy. But profiles for nondegreeholding local notables like Sun are more difficult to develop than for men of higher social standing. The identification of Sun as a landlord at the market-town level came via independent discussions with two informants who had known Sun's son in the early 1960s. The informants, men with extensive experience of their own in the Jiangnan silk industry, insisted that men like

Sun Boyu, a landholder who also had commercial interests centered in his market-town environment, were the main builders of cocoon-marketing facilities.^[58] An article published later by one of these informants stated that cocoon firm owners were men of "local gentry standing."^[59]

As the economic power of cocoon firm owners grew, so, too, did their need for political organization to protect and develop their evolving interests. They responded by organizing the cocoon merchant guild, an organization that performed the all-important task of "self-regulation," imposing basic rules governing the conduct of the cocoon trade in Wuxi and acting as a self-collection agency for new tax revenues. The second form of guild activity was "self-protection," an area that could be further subdivided into three main protections: from other merchants who bought cocoons in the region; from the consequences of adverse government policies; and from the wrath of unhappy peasants when prices or other aspects of the cocoon-marketing process did not meet with their satisfaction. Given the breadth of these activities, I will concentrate here on only two aspects: the social origins and economic activities of guild managers, and their resulting relationships with the peasantry.^[60]

There was at least one precursor of the countywide cocoon merchant guild—the wenshe , or "culture association," of Kaihua subdistrict. Although this name is often translated as "literary society," I believe the translation "culture association" comes closer to its organizers' original intent to promote and enhance the general culture, broadly defined. Clearly, the name also had symbolic value, drawing upon what Pierre Bourdieu would call the "symbolic capital" of gentry culture.^[61] By using this name, gentry managers marked themselves as men of refined social status who also were interested in serving society at large via such organizations. The designation "culture association" was commonly chosen by gentry managers for their locally run community-service organizations, but in the early twentieth century its symbolic significance could also be used to enhance the legitimacy of entirely new activities.^[62] For example, a core group of Republican army officers who began the 1911 Revolution in Wuhan also called themselves a "culture association." Likewise, the first seven members of the Kaihua group were still firmly within the gentry-literati social world—all held the lowest examination degree of shengyuan —but the association's main tasks revolved around newly developing cocoon-marketing activities within the subdistrict.

The economic activities of the Kaihua Culture Association began informally in the early 1880s, with the seven sheng-yuan acting as troubleshooters of a sort, paying visits to the parties involved in disputes and suggesting ways of resolving problems that had arisen. One such issue involved a dispute between a local cocoon merchant and peasants over prices; a second involved setting the licensing fee for cocoon firms so that it would be on a par with the fee already existing for handicraft silk-buying establishments in the area.

They also had the capacity to solicit local government support for their efforts and summoned official county militia to assist cocoon firm owners in dealing with unruly peasants. Once these sorts of state-backed military actions materialized at the behest of Culture Association members, peasants realized, much to their initial surprise, that local gentry planned to use traditional means of coercion to promote a commercial activity as new and strange as cocoon marketing. Local peasants sometimes referred to these men as yangshang , "foreign merchants," a reflection of their growing awareness of the complex intertwining of old-style gentry managerial functions, elite-peasant marketing relationships, and the new international market for Chinese filature silk.^[63]

In 1886, the Kaihua Culture Association received county government recognition as the principal organization within the subdistrict involved in the regulation and protection of the cocoon trade. The original seven shengyuan formed its core managerial group, retaining their troubleshooting function and becoming the official tax collectors for lijin charges on the cocoon trade. There were now more than twenty members, including a clerical staff of an accountant, a manager, and a secretary. The core group of seven, headed by shengyuan Zhuang Hao, continued to make major decisions affecting cocoon marketing.^[64]

The Kaihua Culture Association set important precedents for elite organizing to develop and defend local cocoon-marketing networks. In the 1890s, this kind of activity spread throughout the county, as Shanghai and Wuxi merchants involved in cocoon firms began to meet during the spring cocoon-marketing period. At Shijinshan temple, located in a northwest suburb of Wuxi city, they decided matters related to cocoon marketing. Following a decade or so of such meetings, a countywide cocoon merchant guild was formed in 1902, with a permanent office at the temple.^[65] With this act, although they abandoned the literati-linked symbolism of the culture association, the members of the successor cocoon merchant guild substituted an obvious symbol of spiritual authority of still greater significance to the community at large: their temple-based guild organization thus underscored the legitimacy of their new commercial roles. Once again, in Pierre Bourdieu's terms, this move was a skillful use of "symbolic capital"—a means to convey their own moral consensus that cocoon marketing was good for the community and, crucially, that they were the ones controlling it.

In subsequent years, the seasonal regulation of cocoon-marketing operations became the task of guild managers, a hired staff who actually ran the organization. As in the case of the Kaihua Culture Association, the newly founded guild chose its managers from the ranks of gentry society. The first permanent manager of the countywide guild was Sun Xunchu, who soon thereafter left Wuxi to begin a tour of duty as an official in Hubei province. The guild then selected Zhang Dingan as its new manager. Zhang was son-

in-law of Zhou Shunqing, the comprador-turned-degree holder from Kaihua subdistrict already noted as a comprador agent involved in early cocoon firm development. Zhang Dingan managed various commercial endeavors in Zhouxin zhen , the location of Zhou Shunqing's new silk filature, cocoon firms, and pawn shops. Zhang was also the headmaster of a special school organized by Zhou to teach business skills. Upon Zhang's death, shengyuan Hua Zhisan became guild manager. Thus, among the first three managers of the Wuxi cocoon merchant guild, two were lower gentry and one had family ties to Zhou Shunqing, a rapidly rising figure in the upper tiers of elite society.^[66]

Gentry managers of the cocoon merchant guild in Wuxi also developed their links to county government, carrying out basic licensing and tax collection activities on its behalf.^[67] With state support, guild managers organized ever more elaborate systems of protection for the cocoon firm network. To this end, a primary duty of guild managers was to prepare formal petitions each spring to the county and provincial governments, asking for naval police to protect boats used for cocoon and silver cash transport.^[68] Although such protection was essential to guard against random banditry, we have seen that armed force could have a significant psychological effect on the peasantry. As in the case of the Kaihua Culture Association, when local gentry summoned county militia to defend cocoon firm interests, the cocoon merchant guild continued to rely upon the coercive force of government to extend the power and influence of the cocoon-marketing system.

Although few accounts of the actual marketing process survive, those extant confirm the growing power of cocoon firms backed by their guild organization. The following description of a cocoon-marketing situation in Wuxi is taken from an English translation of a Chinese journalist's account:

At the time when cocoon collecting was at its height, the cocoon collectors deliberately spread rumours of political unrest, impending civil war, slump in cocoon prices and even suspension of cocoon collection. These rumours were quite sufficient to trap the peasants, but there were also many other ways by which the collectors could get the better of the peasants. The peasants often had to bring their cocoons a long distance to the door of the collectors and, in spite of the crowd, the collectors would delay weighing for many hours. During the weighing the collectors would sham depression and poor business, thus deliberately lowering collection prices. Finally the peasants, exhausted by fatigue and hunger from early dawn, were forced to beg in pitiful tones for a little better price, which when granted only meant about ten or twenty cents extra. In addition, the Chinese system of "big" and "small" money [that is, silver and copper cash], gave the collectors a further opportunity to cheat the peasantry [through manipulation of the exchange rate]. At the end of the day there would still be peasants who had not sold their cocoons. They often made a great noise cursing the collectors, calling for fire from heaven, without realizing that the

collectors' property was insured and that such a burning would be of actual benefit to the collectors.^[69]

Other journalists' accounts confirm that factors such as fear of political turmoil or uncertain weather conditions frequently caused cocoon firm owners to lower prices offered for cocoons.^[70] Because peasants depended upon cash income from cocoon sales to purchase grain for family consumption, they sometimes took drastic measures, forming villagewide groups to protest the arbitrary lowering of cocoon prices.^[71] Sporadic, random violence also occurred over price issues, and local officials were called upon to arbitrate these disputes.^[72]

Despite defensive actions on the part of the peasantry, we have seen that cocoon firm owners also asserted themselves, using multiple methods of gentry authority to establish the growing power of the cocoon-marketing system. On the one hand, these practices and tensions were familiar in the highly commercialized rural society of Jiangnan. On the other hand, gentry interests were now channeled into an economic system related to new technology and an international marketing network for raw silk. These developments set the stage for a powerful new form of bourgeois practice. As it emerged in its full form in the 1920s and 1930s, we shall see that bourgeois practice in Wuxi was built not on a thoroughgoing transformation of productive relations but rather on a skillful merger of past and present methods of social and economic dominance.

Xue Shouxuan and the Modern Substance of Elite Control

By the 1920s, fifty filatures and more than two hundred cocoon firms operated in Wuxi.^[73] Several events in China and abroad contributed to this increase. The fall of the Qing dynasty and the establishment of a new Republic caused political uncertainty, but Chinese entrepreneurs on the whole seemed to prosper, gaining increased social prestige within the new system. Careers as businessmen became more acceptable in their own right, and new governments recognized that modern development was both necessary and desirable. World War I also changed the economic environment, hampering the foreign competitors of Chinese industry in both domestic and international markets. Economic recovery after the war provided an especially favorable situation for such products as Chinese silk because Western customers once again were able to purchase luxury consumer goods. From the mid-twenties onward, falling silver prices in world markets also favored China, which elected to remain on a silver standard while the rest of the world used gold. The relatively cheaper Chinese products made possible a more favorable balance of payments and accumulating supplies of silver. Finally, the estab-

lishment of the Nationalist government in Nanjing in 1927 brought commitment to a series of national policies to foster and promote industrial growth.^[74]

In this milieu, a single entrepreneur emerged as the most powerful force shaping Wuxi silk production. Xue Shouxuan, son of Xue Nanming, was one of Wuxi's earliest investors in the silk industry. His career is worth exploring in some detail, for it illustrates how bourgeois practice matured by linking traditional styles of economic and political maneuvering at the local level with new industrial management.

As the grandson of Xue Fucheng and the son of Xue Nanming, Xue Shouxuan began his silk industry career as a well-placed member of the Wuxi elite. Xue Fucheng had been a jinshi -degree holder and a high-ranking official, putting the family within the ranks of the late Qing upper gentry. Xue Nanming expanded family fortunes through land management at the local level and also by building five local silk filatures.^[75] Xue Shouxuan continued all aspects of the family tradition, retaining family landholdings, further developing the family's silk filatures, and also engaging in political activity at the highest levels possible by becoming an important participant in new Nationalist government committees for silk industry development.^[76] He solidified his position via marriage ties as well, a common practice among elite families, by marrying the daughter of Rong Zongjing, who had become one of China's most prominent cotton mill owners.^[77]

In the mid-1920s, Xue Shouxuan took over the family's silk filatures in Wuxi and began reshaping the management and investment patterns that had prevailed under his father's direction. Xue was educated at the University of Wisconsin, an experience that gave him insights into the organization of American business and also made him a representative of an emerging western-trained portion of China's twentieth-century elite.^[78] Upon his return, Xue's first important innovation was to end a system of organization in the family filatures called "split ownership/management" (shi/ying ye ), under which the majority of Wuxi and Shanghai filatures operated. This system created major problems because filature organizers did not operate the filatures themselves. Instead, they rented them annually to a separate stratum of filature operators, and neither owners nor operators took responsibility for filature maintenance or renovation. As a result, the output of Chinese filatures fell behind the Japanese in both quantity and quality, compromising Chinese competition in the world silk market.^[79] Xue aimed to change this situation by instituting a system of personal management under his direct control. To this end, he hired several men as his assistants in managing the family filatures; with their help Xue Shouxuan set about orchestrating a new phase of silk industry growth.^[80]

Although Xue first concentrated on introducing new filature equipment and instituting training programs for both male and female workers, ^[81] he

soon extended his work into sericulture as well. In 1930, Xue organized a sericulture affairs office under the direction of the Yongtai filature and built two egg breederies in Wuxi and Zhenjiang. The Yongtai office also began to organize sericulture cooperatives, promoting extension work in the countryside and the sale of Yongtai egg cards to local peasants. Under this system, the breederies prearranged contracts with peasants to purchase cocoons raised from Yongtai's improved egg cards. Xue's goal was not only to improve the quality of cocoons available to Yongtai but also to raise the quantity of cocoon output, thereby eliminating chronic problems the filature industry faced in maintaining an adequate cocoon supply.^[82]

With these ambitious efforts underway, Xue capitalized on his growing stature within the Wuxi silk industry even during the darkest days of the depression, when virtually all filatures in both Shanghai and Wuxi were forced to close because of the collapse of the world silk market.^[83] Not surprisingly, Xue's position as a well-connected member of the local elite gave him his initial maneuvering edge. As a first step, Xue turned to his powerful businessman father-in-law, Rong Zongjing. With Rong's personal introduction to the general manager of the Bank of China, Xue arranged loans to keep his filatures operating.^[84] Xue also became one of a select group of new business advisers to the Nationalist government in Nanjing. Their first project was a government bond program in 1931-32 to make loans to filature owners to assist industry recovery.^[85] This program set an important new precedent by involving the most prominent filaturists within central China in directly supervising and managing government programs to assist the silk industry. Among this group, Xue gradually emerged as the most prominent new leader, earning the title, China's "silk industry magnate."^[86] This title reflected not only Xue's political influence but also his growing economic position within the central China silk industry. By the mid-1930s, Xue either owned or leased approximately forty of Wuxi's fifty filatures.^[87]

To achieve this new stature, Xue continued to use his strategy of alliance with Nationalist government leaders. During 1934-35, the Nationalists established an elaborate system of governmental commissions designed to supervise both agricultural and industrial development, especially in central China. A Commission for the Administration of Sericulture Improvement was formed in Jiangsu to devise new regulations for cocoon firm operation and silkworm egg breeding, to be implemented in county-level model silk districts and sericulture development districts.^[88] Wuxi was made one of two model silk districts, and county magistrate Yan Shenyu was appointed its first head. There was one other government-appointed bureaucratic manager of this organization, Vice-head Cai Jingde, but the other fifteen members of the model silk district committee in Wuxi were all businessmen from within the Wuxi silk industry itself. This group was headed by three prominent filaturists: Qian Fenggao, Cheng Bingruo, and Xue Shouxuan.^[89] Xue

Shouxuan clearly wielded the most influence, serving also on the sericulture commissions of the provincial and national governments.^[90]

At this point, Xue's rapidly growing political influence, built from within the county and outward, affected the way his activities as an industrialist impinged upon rural society in Wuxi. Xue had not divorced himself from the traditions of old gentry society, in which locally based, multiple strategies for dominance were employed with the best connected, wealthiest, and most talented individuals emerging to serve at higher levels in the bureaucracy. By the 1920s and 1930s, the old mark of social pedigree, degree-holding status, had been eliminated, but the tradition of local power brokering within elite circles had not yet faded. Xue built on this tradition and changed it by using it to pursue modern industrial development. His actions had ramifications vis-à-vis both cocoon firm owners and the peasantry in Wuxi, as seen through initiatives that were part of the model silk district program.

New measures for cocoon firm control were the first aspects of model silk district policy in Wuxi that Xue Shouxuan used to his advantage. Regulations stipulated that all new cocoon firms must install in their drying rooms Japanese-style hot air blowers, the most modern equipment available.^[91] Small cocoon firms of four drying rooms or less, likely to be least technically advanced in their equipment and methods, were no longer allowed to operate.^[92] Both these measures were designed to guarantee higher quality cocoons by drying them under optimal conditions, thus insuring a more complete and uniform drying process. For Xue Shouxuan, however, who already had the most advanced drying machines, these measures meant that he would now have an advantage over his competitors.^[93] By maneuvering these restrictions he developed a plan to control cocoon marketing throughout the Lower Yangzi region.

Evidence of Xue's ambitions can be seen in neighboring Jintan county where Xue rapidly expanded his influence over cocoon marketing under the model district program. It is somewhat surprising that Jintan became the second model silk district; it was not a major producer of cocoons when the program was established.^[94] However, when we realize that Xue Shouxuan had important personal connections in Jintan, the rationale for granting model district status becomes clear. Xue cultivated close friendships with the head of the Jintan police bureau and also with the county magistrate who served simultaneously as vice-head of the model district. Assisted by these two county-level officials, Xue's agents were able to rent many cocoon firm facilities in Jintan. Once these facilities were rented, under model district guidelines, Xue renovated the existing drying rooms, installing new drying equipment. Although this required a substantial capital outlay, when the facilities were rented and the machines in place, Xue quickly established the capacity to buy and process a large proportion of Jintan cocoons. At the time he started this system in Jintan, Xue also began to cultivate friendships with

the magistrates of Yixing and Liyang counties, having similar plans for developing their cocoon trades.^[95]

Xue also used the model silk district restrictions on the selling of silkworm egg cards to peasants to his advantage. These provisions halted selling egg cards by private breederies—a new industry rife with speculation and abuse, that had arisen in Wuxi in the late 1920s.^[96] Now private breederies could continue to operate only under one condition: if a breedery was willing to take the time and effort to establish egg-rearing cooperatives with peasants, instructing them in new techniques and lowering the risk of cocoon crop failure, then it could retain the right to sell its improved egg cards.^[97] Under this provision, Xue Shouxuan's Yongtai and Huaxin filatures, which already had such a system of cooperatives, continued to run breederies that sold egg cards directly to peasants. In 1934, the Huaxin filature prepared female extension workers to organize new contractual cooperatives, where filature staff sold egg cards and required that the cocoons raised from those cards be sold back to filature-controlled cocoon firms.^[98] In 1935, the model district extension program sent out one hundred workers, but Yongtai and Huaxin also sent out another one hundred in Wuxi and surrounding counties to set up contractual co-ops.^[99]

The ultimate benefit reaped by Xue from this cooperative program was the ability to monitor all aspects of cocoon rearing and marketing in the Wuxi countryside, including the important matter of prices. Although the cocoon merchant guild had long been responsible for setting a cocoon price at the beginning of each marketing season, prices ebbed and flowed daily, sometimes even hourly, at the actual cocoon-marketing sites.^[100] Thus, it was to Xue's advantage to control more closely the price paid to peasant producers for their cocoons, and the cooperative system helped accomplish this end. An account of the resulting situation appeared in the Chinese press, relating that cooperative members had to carry cocoons from their households to the filatures, no matter how far the distance. Peasants were given certificates of the weights received but without a stated price. At least a month later the filatures announced their cash payments for cocoons divided into three grades. Prices for the relatively small portion of cocoons in the first grade were set slightly above the market, but Xue filatures paid well below the market for the second and third grades. If the market price for the first grade was thirty yuan per dan (one dan =fifty kilograms), the filatures would pay thirty-one or thirty-two yuan , but the second grade invariably brought one or two yuan less than the market price, and the payments for third grade cocoons were not much more than twenty yuan per dan . Only a small proportion of the cocoons were classified as first grade, so Xue's filatures, on average, paid substantially below market price by using this cooperative system.^[101]

Other press reports confirm that cocoon prices remained low in Wuxi throughout the mid-1930s—a situation caused by both the lingering effects of

the depression and Xue's power to influence prices. In 1936, the Nationalist government's Commission for the National Economy, on which Xue served as a silk industry adviser, announced that it would decide each year on a uniform nationwide price for cocoons based on raw silk prices on the international market and average production costs for Chinese filatures. For the spring season of 1936, the commission decided that the price offered to peasants for their cocoons would be thirty yuan per dan .^[102] Although raw silk prices had recovered to 74 percent of their 1929 level by that time, this suggested price for cocoons was only 42 percent of predepression highs.^[103] When the Wuxi cocoon industry association met on May 31, it merely reaffirmed the thirty yuan price. Hua Shaochun, one of Xue Shouxuan's personal associates, headed this association, accounting for the rapidity with which the organization agreed to the new system.^[104] Whether Hua had been installed by Xue or cultivated by him after assuming his position, the effect was the same—a decision that clearly favored Xue and his filatures by assuring that cocoons could be bought at lower prices.

In the final analysis, Xue Shouxuan used multiple methods to build a more fully integrated silk industry. He oversaw a partial reshaping of the cocoon-marketing system and in the process raised quality standards for Chinese silk production. However, the price to be paid for these advances was a silk industry increasingly dominated by Xue's personal networks. In addition, peasants found themselves with a diminishing capacity for price bargaining within the cocoon-marketing system. Xue thus combined old elite strategies with bourgeois leadership of modern industry and government programs designed to support it. At the level of local society, this combination comprised a powerful new system of social dominance linked to industrial modernization.

Bourgeois Practice in a Chinese Context

Traditional form, modern substance: these were the social dyad that shaped bourgeois practice in the context of the Chinese Republic. In the late nineteenth century, members of the Wuxi gentry elite expanded their commercial initiative, bringing the peasant economy into contact with a new international system of mechanized silk production. The trajectory of social change already underway in Jiangnan created an environment suitable for these developments. Patron-client landlordism had waned during Qing times, and gentry members developed new strategies to build or maintain their local social dominance. As commercial activity became an important source of income, new economic relationships between elites and peasants resulted.

The story of Xue Shouxuan demonstrates the culmination of these trends. I have outlined the parameters of his bourgeois practice, emphasizing the

social power resulting from this new hybrid form behavior. Xue built his modern silk industry by using strategies of the earlier gentry. To maintain his symbolic capital, he carefully cultivated relationships with selected local officials and elites, as well as high-placed acquaintances in Nanjing, and provided essential markets for peasant producers even while depressing the prices they received for their cocoons. When one adds the eminence of his family background to the aura derived from his business success, Xue still appears as a prestigious patron-broker as well as an aggressively successful capitalist. His status in higher arenas reinforced his local dominance over dependent peasants, and his local prominence contributed to his stature in national circles.

However, Xue's political style was derived from the tradition of gentry-merchants before him, and his industrial activities occurred within a rural society that, though evolving, was still substantially within the late imperial framework. In the early twentieth-century context, Xue's activities neither simply recreated nor totally transformed the previous relationships between gentry-merchants and peasants. He modified elite dominance by introducing new technology and factory organization, but he did not try to displace or transform peasant society by divorcing his suppliers of cocoons from their rural households. The resulting pattern of development differed substantially from those we think of as constituting classical capitalist modes. Peasant family farming was not eliminated as part of this process, but rather it became an integral link in a chain of new production relationships serving modern industry.

Further studies of early industry in China might well begin here—with an examination of how modernizing elites linked industry with the existing rural order. Until we have more studies that examine evolving relationships such as these, we will be unable to argue conclusively about either the nature of the Chinese bourgeoisie in the early decades of the twentieth century or their roles in promoting early industrial growth.

Five
Power, Legitimacy, and Symbol: Local Elites and the Jute Creek Embankment Case

R. Keith Schoppa

During the last two decades analyses of Chinese local elites have concentrated on identifying functional elites; delineating their social, political, and economic groupings; and describing their functions. The focus on identification was in part a necessary reaction against the long-held view that local society was dominated simply by the traditional gentry (i.e., civil service degree holders). Important studies have spelled out a range of more apposite functional elite categories (in the late Qing, for example, scholar-gentry, gentry-merchant, merchant, gentry-manager; and in the Republic, a much wider array, including militarist, capitalist, educator, landlord, and bandit chief). The study of elite groupings (lineages, associations, parties) has provided a fuller sense of key collective actors in local society. Examining spatial contexts for elite actions has suggested variations in patterns of elite organizations and political culture. In treating elite functions, scholars have detailed changes over time and space in their nature or mode.

Because of the paucity of detailed sources, however, relatively few studies concentrate on the processes and discourse of local decision making and of changing local power relationships, studies that might offer insights into elite stratification and patterns of dominance vis-à-vis lesser elites and nonelites. One generally well-documented case that provides such evidence is the struggle during the last years of the Qing dynasty and the early Republic over the Jute Creek Embankment (Maqi ba ) in Tianyue subdistrict (xiang ), Shanyin county, Zhejiang province in the Lower Yangzi macroregion.^[1] In a series of crises relating to local dominance, the episode reveals elites in a core subdistrict struggling to continue domination over elites in a backwater sub-district and the latter's paramount leader maneuvering to maintain his control over the subdistrict's lesser elites and nonelites. Furthermore, an analysis of the affair's context and course provides considerable insight into the wide

range of resources for local-elite legitimacy and power and the strategies for their maintenance.

The Setting

Tianyue Subdistrict . Contiguous to Xiaoshan county and to the flourishing Puyang River port of Linpu, Tianyue subdistrict was located in the southwest corner of Shanyin county, which was merged with Guiji to form Shaoxing county in 1912. The area's political agenda was dominated by flood control issues. Situated along the narrow-channeled and precipitously sloping Puyang River, both Linpu and Tianyue frequently had to fight rampaging floodwaters after storms in the mountains to the south. Before the mid-fifteenth century, the Puyang turned to the east immediately before it reached Linpu to join the Little West River in Tianyue (map 5-1).

With mountains on three sides and the Puyang to the west, the subdistrict had traditionally been divided into three sections: upper, middle, and lower—designating placement along the Puyang River and delimiting its varying topographical orientation (map 5.2).^[2] Upper Tianyue, south of Dayan Mountain, sloped to the west with its streams flowing directly to the Puyang. Lower Tianyue, north of Qinghua Mountain, also sloped to the west with its streams flowing to the Little West River. Between the two mountains was Middle Tianyue, sloping to the northwest and drained by Jute Creek, a stream which joined the Little West River east of Linpu. The land at the foot of the mountains in Middle and Lower Tianyue was very low, in the words of the subdistrict's gazetteer, "sunken like the bottom of a pan." Floods were a continual problem; the area reportedly experienced natural disaster nine out of every ten years, becoming commonly known as "the barren subdistrict."^[3] The plight of Lower Tianyue and sites down the Little West River in northern Xiaoshan and Shanyin counties was especially tragic during floods; the river did not empty into Hangzhou Bay until it reached the Three River Outlet, northeast of the prefectural capital of Shaoxing. Large areas along its channel were frequently flooded.

To deal with these problems, in the reign of the Chenghua emperor (1465-1488), the prefectural government sponsored the diversion of the Puyang River through the excavation of a new channel northwest of Linpu and the construction of the Maoshan sluicegate where the Puyang River had previously turned east. From that time, the main force of the Puyang River flowed west along the southern edge of Linpu on its way to join the Qiantang River.^[4] For Lower Tianyue and the northern part of Shanyin county, the river's diversion meant freedom from the floodwaters of the Puyang.

In its fervor to end the flooding problem, the prefectural government also constructed an earthen embankment across Jute Creek to check the runoff from the mountains to the east of Middle Tianyue. Although this construction

Map 5.1.
Major Rivers and Streams in the Three-River Drainage System, Xiaoshan and Shanyin Counties, 1911.

Map 5.2.
Waterways and Water Facilities in the Tianyue-Linpu Area

made water problems farther downstream even less likely and theoretically provided Middle Tianyue with a reliable irrigation source, in reality it consigned much lowland in that part of the subdistrict to swamp. Not only did storms in the mountains to the east send cascades down to submerge the crops; but also, when dikes broke along the Puyang River to the west, floodwaters pouring into the area could not be drained off. Forty-eight villages were, to use the gazetteer's language, "abandoned like entrenchments on the frontier" to face the ravages of flood unprotected, "outside the embankment."^[5] Whether the prefects and their advisers had foreseen the unfortunate results, they did nothing to correct the situation. Once the embankment was in place, the rest of the county declared its efficacy and refused to countenance any change. There appeared little recourse for redress.

The political impotence of the forty-eight villages is shown by the very phrase "within the embankment," which came to stand generally for the lower reaches of the Little West River. By any objective evaluation, however, Middle Tianyue was "within the embankment," a fact that had turned it into the flooded realm of "fish and turtles." The language of description was infused with the perspective of the dominant. Productive farm land was soon covered by two lakes, Upper Ying and Lower Ying, which in turn had to be diked to prevent even further inundation of farm land. The 1588 reconstruction of the embankment included an opening to be used for irrigation "within the embankment" in case of drought, but it was narrow and low, and moreover the Middle Tianyue elites did not hold the prerogative to open the watergate.^[6] As time passed, the forty-eight villages came to be marked by both bleak poverty and outbursts of banditry and unrest. Although generally surrounded by the flourishing inner core of the late imperial Lower Yangzi region and bounded by the flourishing economic center of Linpu, Middle Tianyue remained, in fact, peripheral—poor, troubled by social unrest, and embittered at being held hostage to nature.^[7]

The World of Linpu Town . The people of Middle Tianyue were, in part, unable to change such a disastrous situation because of the subdistrict's dichotomous orientation. Administratively it was under the jurisdiction of Shanyin county even though it was separated by a mountain watershed from the rest of the county and oriented naturally to Xiaoshan county and the town of Linpu. Shanyin county treated Upper and Middle Tianyue like unwanted stepchildren. Linpu created yet another problem of identity because its administration was divided between Xiaoshan and Shanyin counties. Its main market area, to which Tianyue sent rice and paper products, was located not in Shanyin, but in the Xiaoshan section of the town.^[8] The sub-district's natural and administrative status cast it into a political no-man's-land where local inhabitants were often forced to handle problems (like water control projects along the Puyang River) that in other areas the state might

help to solve. This sense of isolation only exacerbated the problems of unrest spawned by the area's poverty and banditry.

The presence of the important entrepot of Linpu so near to the peripheral pocket perpetuated Tianyue's poverty and stimulated continual awareness of its powerlessness. A flourishing market already in the Song,^[9] the town had apparently been seen by administrative decision makers as potentially too wealthy, and therefore powerful, to function as a single administrative unit. Its administrative duality, however, was a source of continual political and economic problems, making it harder, not easier, to control.^[10] One of Zhejiang's six great rice markets, in 1912 Linpu had sixty rice firms (hang ), most of which participated in the town's powerful rice merchants' guild (miye gonghui ). The price of rice in the three-county area (Xiaoshan, Shanyin, and Guiji) depended on the Linpu market, which operated under the direction of rice and shipping firms, brokers (yahang ), and large numbers of outside merchants.^[11] Because the three-county area had to import large quantities of rice annually, many merchants from Jiangxi and Hunan represented so-called "long route" (changlu ) rice firms in those provinces; present also were large numbers of shippers from Ningbo and Shaoxing prefectures who delivered rice along the region's canals. The presence of large numbers of sojourners contributed to the wide open atmosphere of Linpu, where opium dens and brothels flourished.^[12]

When industries (rice-husking mills, oil presses, breweries) were established in the late nineteenth century, rice merchants played major roles in their operation, thereby expanding their share of the local wealth.^[13] In the Republic, the Linpu merchant "establishment" compiled an infamous record of hoarding and corruption.^[14] Officials sometimes plotted with commercial elites: in 1920 the Linpu police force became involved in a bribery and kickback scheme that became one of the county's most notorious twentieth-century scandals.^[15] Perhaps not surprising in such a situation, frequent social outbursts resulting in injury and death against the merchant-official establishment are recorded: rice riots (May 1911; March and April 1925); a salt riot (April 1917); and a riot against paper taxes and charges (August 1921).^[16] For these reasons, merchant militia units were active in Linpu; and in September 1914 the town became one of only five sites in Zhejiang for the establishment of a subsidiary county yamen.^[17]

The unrestrained commercialism of Linpu and the poverty of Tianyue seemed two different worlds. The subdistrict's geography and administrative location virtually invited exploitation. Linpu rice merchants moved aggressively into Tianyue and attempted to make killings on rice purchases by taking advantage of desperate farmers in those reputedly exceptional years of decent harvests.^[18] The sense of being exploited led to social explosions like the 1921 riot against paper stamp surtaxes, when an estimated five thousand Tianyue people involved in the manufacture of paper attacked and destroyed

the Linpu paper firms and demanded the abolition of the new tax.^[19] Although this collective violence may have had a class dynamic, the rioters' locale-determined identity pointed again to the perceived subordinate status of the subdistrict.

Tang Shouchong and Subdistrict Elites: Sources of Local Dominance

The leading figure in water control projects undertaken in Tianyue and at Linpu in the first five years of the Republic was Tang Shouchong, a resident of Datangwu, a village located at the approximate center of Middle Tianyue subdistrict and the Tang lineage home since the twelfth century. The village was about 4.5 kilometers southeast of Linpu and 3.0 kilometers from the Jute Creek Embankment in an area of low ricelands and bamboo- and tea-producing mountains.^[20] In addition to the Tang's agricultural base,^[21] the lineage also had strong financial interests in Linpu, where Shouchong likely had a subsidiary residence. A list of Linpu contributors to the 1922 reconstruction of lake dikes in Tianyue shows that at least eight of the town's twenty-one individual contributors were from the Tang lineage.^[22]

Tang Shouchong himself clearly straddled the worlds of agriculture (Tianyue) and commerce (Linpu). In his role as the Tianyue subdistrict manager (xiangdong ) in the late Qing self-government structure, he was active in various local issues in addition to water control.^[23] In Linpu, where his management of waterworks repair was predicated on a channel's significance as a merchant thoroughfare, he was codirector on the Fire God Dike project with Lu Zumei, head of the town's chamber of commerce branch.^[24] He was personally involved in the ownership and/or management of the Yuantai old-style bank (qianzhuang ) in Linpu.^[25] In constructing public service institutions apart from water works, Tang focused not on the subdistrict but on Linpu. There in 1905 he established a school with three other elites from nearby Ruluo subdistrict in Xiaoshan county; in contrast, the earliest recorded school establishment in Datangwu was 1931.^[26] In 1910 Tang also established an orphanage in Linpu at substantial personal expense.^[27]

Sources do not disclose whether the subdistrict's population resented Tang's public works focus on Linpu; whether they distrusted Tang's close relationship with Linpu commercial elites who often took advantage of the subdistrict; or even whether Tang himself may have been involved in underwriting the financial activities of rice merchant exploiters of Tianyue. We also cannot tell whether in the case of a conflict, his interests in town affairs perhaps weakened his will to defend subdistrict interests. But we do know that some people in the area considerably resented Tang.^[28] It is certainly safe to say that, as in any dominant-dependent relationship, subdistrict residents experienced ambivalent feelings—resentment and antagonism as well as appreciation and admiration—for Tang.

The Tang lineage also straddled the world of native place and the world beyond. While Tang Shouchong took Linpu and neighboring Tianyue sub-district as his stage, his elder brother Tang Shouqian strode into the provincial and national arenas.^[29] Shouqian, a jinshi or metropolitan graduate of 1892, headed the Zhejiang Railway Company in the first decade of the century, taking a popular nationalistic stance that catapulted him to the governorship of Zhejiang in the aftermath of the 1911 revolution. Although he was appointed minister of communications in the provisional Nanjing government, once Yuan Shikai assumed power, Tang left for Southeast Asia to visit overseas Chinese, never again taking an official position.^[30]

Sources indicate that the nationalistic fervor and activity of Shouqian had a great impact on Shouchong and his Linpu allies. They purchased more than five thousand yuan of railroad stock from proceeds of the land set up to finance the school that Shouchong had established with others in 1905.^[31] Because the brothers' relationship was reportedly very close, their conversations on Shouqian's visits home doubtless dealt with the elder brother's activity in the railroad affair, in the Society to Prepare for Constitutional Government, and in other nationalistic and reformist efforts.^[32] Some later developments in the Jute Creek episode reflect Shouchong's own absorption of the spirit of nationalism and political change. In any event, the prestige accruing to the lineage and its local leaders, particularly Shouchong and his younger brother Shouming, from the accomplishments of Shouqian was undoubtedly great, enhancing their other bases of local power.^[33]

Among lineages in Tianyue, the Tang had cast a dense marriage net.^[34] In Middle Tianyue, circumscribed by mountain and river, the lineage had marriage connections with at least thirteen of the twenty-one major lineages.^[35] A frequent Tang liaison was with the Ge lineage of Shantoubu, less than 2.5 kilometers east of the Tang lineage home; the mother of Shouchong and Shouqian was a Ge. From the end of the Qianlong reign, the Tang lineage produced only one upper-degree holder—Shouqian. Ties to the Ges, however, linked the Tangs to the most prolific degree-attaining lineage in Middle Tianyue from 1796 to 1912.^[36] In the closed defensive society of the sub-district, the Tangs were able to utilize marriage ties to solidify their leadership position and link themselves to the subdistrict's other most powerful lineage.^[37]

Joining Shouchong in the Jute Creek Embankment case and other water-control work were Ge Bilun, Kong Zhaomian, and Lu Luosheng. Ge of the Shantoubu lineage, resident in Tianyue since the Yuan dynasty, was a scholar who chronicled the events of the early Republic for the county gazetteer.^[38] Kong, whose lineage home was established in the fifteenth century about two kilometers southeast of Linpu, was the only one of the four with an upper-level degree (a juren or provincial graduate of 1894).^[39] Though he was his lineage's only upper-degree holder after 1777, his lineage played important roles in the locality into the Republic. A brother, for example, served as dike

administrator on the first section of the West River Dike near Linpu in the late 1920s.^[40] Lu, a scholar who also chronicled events of the times, represented a lineage that was not particularly distinguished, but its village was the closest (less than half a kilometer) to the Jute Creek Embankment.^[41] Like Tang, Ge and Kong served in local self-government positions in the late Qing and early Republic: Ge, in the Shaoxing county assembly from 1912 to 1914, and Kong, as the subdistrict council chairman (yizhang ).^[42] Heading the new self-government institutions apparently added little substantial power to their roles, but the positions officially confirmed their preexisting local power.^[43]

An analysis of Tang's local social and political dominance by 1912 shows his success emanating from an ability to operate in several dichotomous but complementary spheres: Tianyue (agriculture, periphery) and Linpu (commerce, core); local (leader of a lineage with strong socioeconomic base) and beyond (ties to provincial and national affairs through Shouqian); private (voluntary managerial roles) and public (self-government position).^[44] In addition to operating effectively in these spheres, he had obvious expertise and energy in the field of local water control. He directed all the major repair work in the area from 1911 to 1916. His appointment by the county assembly in September 1912 as honorary inspector of the West River Dike noted his public-spirited enthusiasm and the seriousness with which he undertook water conservancy responsibilities.^[45]

It should be noted that Tang's power in the locality derived in the main from resources legitimated by traditional cultural values.^[46] His social status as gentry-manager, his reliance on lineage-marriage ties, and his fraternal connection to higher places were social resources based on cultural patterns that lent legitimacy to his leadership position.^[47] His self-government leadership invested him with officially recognized legitimacy. His expertise brought what one writer has called "competent authority."^[48] With the latent political power of Tang's wealth undergirding his social, political, and personal legitimacy, his local leadership position appeared secure. Yet, power has been defined as "legitimacy put to the test of social action."^[49] The case of the Jute Creek Embankment not only tested Tang's power but also began to erode his legitimacy.

The Jute Creek Embankment Case: Suoqian Elites Versus Tianyue Subdistrict

When proposals to change the embankment into a bridge reached the government in 1911, it was not the first time. As early as 1643, the subdistrict's cause was championed by local native Liu Zongzhou (1578-1645), philosopher, Donglin adherent, and teacher of the philosopher, Huang Zongxi.^[50] Liu, whose pen name (hao ) was Jishan, advocated demolishing the embank-

ment, pointing out that the Maoshan sluicegate provided the security from floodwaters that Lower Tianyue and points farther downstream needed.^[51] Such a move was strongly opposed by the spokesman for downriver sites in Xiaoshan county, Ren Sanzhai, who credited the Jute Creek Embankment with preventing serious floods in the preceding two centuries—a curiously myopic view that discounted the Maoshan sluicegate, which had actually prevented floods.^[52] In the views of Ren and others, the issue of the embankment constituted a zero-sum game: "They speak of [Middle] Tianyue not having half its land [because of floods]. The only question is who is going to have the difficulty."^[53] Though nothing came of Liu's effort, the people of Middle Tianyue remembered him with respect and appreciation while their situation festered. In 1881 Tang Shouchong's father, Tang Pei'en, resuscitated Liu's suggestion by setting down principles of a Ji [shan] Society to remember Liu's spirit.^[54] Although there were no immediate tangible results, a decade later the Shaoxing prefect broached the idea of changing the embankment to a bridge. Despite the fact that neither he nor his successors followed up the suggestion, Middle Tianyue villagers invoked the proposal in subsequent years.^[55]

In fall 1911 the subdistrict's bitterness expressed itself in the Tianyue self-government council's petition, drafted by Tang and Kong, to the provincial assembly in Hangzhou.^[56] Although the petition came during a barren harvest period in the aftermath of two serious floods, this calamity alone doubtless did not catalyze local elite action because countless others had not. More likely, elites were activated by the establishment of local self-government districts and institutions. In that year, because of its elites' arguments, Lower Tianyue had been separated to form an independent sub-district named after its main village of Suoqian, the first administrative recognition of the separation wrought centuries earlier by the embankment's construction.^[57] Whether this decisive action motivated the petition, it seems apparent that, though self-government titles added little to the actual credentials of local leaders, the idea and institution of self-government began to transform local political processes. Local leaders quickly grasped the utility of this new institution for handling local matters.

The petition had two goals: demolishing the embankment (to relieve at last the flood-stricken Middle Tianyue) and attaining provincial assistance in repairing river dikes in Middle and Upper Tianyue (now called simply Tianyue). In their presentation, Tang and Kong set forth the plight of their native place: through the existence of the embankment, "evil gentry" in Lower Tianyue (now Suoqian), motivated by their own private advantage (sili ), continued to prey on the subdistrict's people.^[58] The embankment was a "question of life or death."

By pointing to the area's economic poverty, the petitioners had in mind not only the recurring pattern of cropless years but also the recurring use of

scarce private funds for repairing and maintaining dikes—amounts traditionally contributed by households whose fields were thereby protected. In part, lack of any outside support stemmed from the county's general inattention to this area and to what the petition pointed out as the political and cultural poverty of the subdistrict: very few people were literate; and the subdistrict produced no officials, no muyou (the private secretaries and legal advisers that made Shaoxing both renowned and reviled in the empire),^[59] and few gentry. There was some exaggeration here: Tang Shouqian was an important outside "connection." But the main thrust was true enough. The thirteen subdistricts of Shanyin county in the Qing dynasty produced 2,032 upper-degree holders, an average of 156 per subdistrict. Although Middle Tianyue contained 59 per cent of Tianyue subdistrict's population, its villages produced only 18 (0.8 percent) of the county's total degree holders.^[60] In sum, in Tianyue there was an obvious paucity of local leaders legitimized by upper degrees and possessing outside connections who might be able to effect desired changes. In contrast, Suoqian subdistrict had no chronic flooding problem (thereby escaping the financial burden of continual dike expenses), and, the petition noted, the subdistrict had many "connections" to officialdom.^[61] From the subordinated's point of view, Suoqian's main instrument of continuing dominance was its connection to officialdom, a tool Middle Tianyue felt deprived of by its political poverty. Although not specifically mentioned in the petition, we may surmise that Suoqian's newly independent administrative identity further enhanced its political visibility and influence.

The provincial assembly reacted to the petition by asking the governor to investigate the situation, but the coming of the 1911 Revolution frustrated this effort. Before the revolution Tang and Kong had served as Tianyue (i.e., former Middle and Upper Tianyue) self-government leaders. After the revolution, as evidence of renewed determination to solve the problem, the forty-eight villages in former Middle Tianyue elected Tang, Kong, Ge Bilun, and Lu Luosheng as representatives (daibiao ) to bring the issue to a conclusion. Probably the impetus for the election of "representatives" and for the designation itself came from Tang and his allies, an evidence of the constitutionalist ethos common at the time among core elites and perhaps specifically of the oppositionist mobilization of the railroad movement.^[62] Its use at the moment of elections for a national parliament was symbolic of Tang and the others' absorption of the new ethos, which, in turn, might impart some measure of legitimacy to their effort.

By early 1913 Tang, the acknowledged subdistrict leader, had assumed two social roles vis-à-vis the farmers of Tianyue in the embankment case. In his position as long-time gentry-manager in the subdistrict, he acted as patron to subdistrict residents. In exchange for his expenditure of political, instrumental, and economic resources in the community, he expected support and loyalty from the farmers for whom he acted with paternalistic

authority.^[63] A second role was as representative of the subdistrict's forty-eight villages, a role in which Tang's authority at least theoretically came from subdistrict residents. Although Tang's relationship to the state was defined by his self-government role,^[64] his relationship to the villagers was defined by both patron and representative. The dynamics of the latter roles, however, were sharply different. We know neither how the villagers perceived Tang's role nor the depth of their understanding of representative forms, but it is likely that in their view he remained primarily "patron."

In November 1912, the four representatives called for the embankment's end in letters to the Shaoxing county assembly, military governor Zhu Rui, and the provincial assembly.^[65] The arguments in these letters emphasized the increased possibility of action following the recent revolutionary change to a republican polity. The spirit of the time, they alleged, was one of change and innovation.

The counterattack by Suoqian subdistrict and the rest of Shaoxing county nonofficial elites was multifaceted. Their past dominance over Tianyue had been based simply on political power and the legacy of the past. With their control now challenged, the elites quickly devised a host of methods in an effort to maintain their pattern of outright domination over the elites and masses of Tianyue. The county assembly chairman, Ren Yuanbing, who had close connections to Suoqian elites, stonewalled, not answering the Tianyue letter, not introducing it for discussion in the county assembly, not investigating or soliciting other subdistrict opinions. At the same time, he sent telegrams to powerful Shaoxing figures outside the county, claiming that the "scoundrel (gunpi ) Tang lineage" was trying to destroy the crucial dike.^[66] To the cabinet in Beijing, Ren specified the villain to be none other than Tang Shouqian.^[67] Suoqian subdistrict leaders sent the provincial assembly a specious map of the area to make it appear that the danger to Suoqian would be acute if the embankment were destroyed.^[68]

Stonewalling, deception, ad hominem attacks were the defensive reactions of elites "within the embankment." In addition, the moral discourse of those upholding the status quo was similar to that used by those placed on the defensive in other local crises of the rapidly changing late Qing and Republican periods.^[69] It should be obvious that in elite culture there may be an inherent contradiction between the particularistic sectional interests of elites and the universalistic moral values they are seen to represent.^[70] When sectional interests are being promoted or protected, discourse on the basis of general moral values both masks the struggle for dominance and enhances the legitimacy of the cause. Thus, in contrast to Tang's invocation of innovation in the political symbolism of the revolution, Suoqian elites retreated to the legitimacy of the traditional social symbols—the morality of the group and its harmony—to combat threatened change. Suoqian and seven other

Shaoxing subdistricts claimed that the issue was one of private (si ) advantage for Tianyue subdistrict: if the embankment were destroyed, the other subdistricts could never again be defended against serious flooding; the morality of the collective opposed the immorality of private interests.^[71] Jin Tanghou appealed to the graves of ancestors (the group's past) and warned of the likelihood of outright struggle between Suoqian and Tianyue subdistricts (the value of harmony).^[72] The chairman of the Suoqian subdistrict council, Lou Kehui, even threatened staunch resistance to the government if it decided to destroy the embankment.^[73]

Whatever the merits of the case, the arguments were couched in categories of traditional moral symbols that purposely lent legitimacy to the cause. Nowhere was there discussion of the political, economic, or ecological facts of the situation; nor was the argument placed in current spheres of discourse (such as the growing debate over the meaning of public and private, or the trend in water-control controversies to investigation by outside conservancy organizations).

Winter 1912-13 brought several provincial and county investigators to the scene.^[74] In December two deputies of the provincial government reported that removing the embankment would have no harmful effect for those downstream. While the magistrates of Shaoxing and Xiaoshan delayed a formal recommendation until spring, Civil Governor Qu Yingguang was dispatched in January to meet with the magistrates and representatives of both sides (for Tianyue, Tang, Kong, Ge, and Qiu Shaoyao, leader of another Tianyue lineage; and for the downstream sections of the Little West River, two representatives of the Suoqian subdistrict council, the subdistrict manager from Ruluo [north of Linpu], and the vice-chairman of the Xiaoshan county assembly).^[75] Unfortunately, there is no record of these discussions. From what we know of later events, it is unlikely that the Suoqian elites yielded at all in their resistance to change. The burden of Tang and other Tianyue spokesmen was to demonstrate to officials and elites that the situation was not (as Ren Sanzhai in the sixteenth century had depicted it) a zero-sum game. For the discussions to have continued at all, compromise must have been pursued. Whether Tang was willing to make concessions or whether Governor Qu, serving as broker, was able to effect the compromise, we do not know. But, at the conclusion of the discussions, the governor announced his support for changing the embankment in some fashion.

When the Ministry of Agriculture and Forestry authorized proceeding, Military Governor Zhu Rui dispatched another deputy to discuss the methods. Pressed by elites from the two sides, the provincial government compromised: the embankment would not be demolished, but the opening installed in the late sixteenth century would either be widened and heightened or another would be cut in the embankment to allow better and faster

drainage. Projected expenses would determine the method.^[76] In March Zhu Rui announced the decision to widen the existing opening.

Unable to accept any change in the status quo, Suoqian elites and their county allies, acting through the Shaoxing assembly, wired their opposition to the Ministry of Agriculture and Forestry. In almost immediate response, the ministry and its agricultural affairs office ordered Zhu Rui to delay any action.^[77] The source of such a rapid, favorable response to the assembly was almost certainly long-held connections, a reminder that, despite the Tang lineage, Tianyue still suffered from poverty of appropriate connections. Noteworthy also are the connections of county and Suoqian elites to Beijing bureaucrats revoking decisions by the provincial officials. The decision seemed to perpetuate the dominance of the downstream areas over Tianyue. Government ultimately responded to the most powerful and best connected, not to those stricken by economic, political, and social poverty, however just their cause.

The Jute Creek Embankment and After:Tang Shouchong and Tianyue Subdistrict

For the Tianyue subdistrict, the imposed indefinite delay on any action meant that Tang Shouchong, despite his considerable resources for local power, simply could not parlay his local legitimacy into influence beyond the Tianyue-Linpu area. Even more significant, however, because of his unsuccessful efforts to persuade Suoqian and county elites to assent to changes in the embankment, Tang lost the support of those subdistrict farmers whom he represented. It is unclear whether the expectations Tang raised among the farmers caused their discontent with his leadership; but the sources do make clear that, once the decision to delay action became known, Tianyue farmers applied unremitting and antagonistic pressure on Tang and the other three representatives to insure a favorable solution. The four representatives, caught between immovable official delay and growing pressure from the sub-district, reportedly had no alternative but to resign from their positions in the spring. Tang and Kong also resigned their self-government posts.^[78]

Perhaps the elites meant the resignations as a ploy: by resigning Tang and Kong, able leaders in an area without many educated or outward-looking elites, may have been trying to signal the government about the urgency of positive action. It did not work: Zhu Rui responded that according to self-government regulations,^[79] they could not resign. Whether or not their resignations were legal, there is no record that from this point they actively fulfilled self-government functions. The resignations of the four "representatives" of the villagers may also have been a ploy directed against the farmers of Tianyue subdistrict. Threatened or actual resignations, they may have

thought, would lead the farmers to see that their able, prestigious representatives (and patrons) would be unable to continue in the face of such unremitting pressure from below. This approach implied that the four men believed the subdistrict depended on their ability for future negotiations. If that was their other strategy, it also did not work: within a short time, the farmers had formed their own Forty-eight Village Association (hereafter called the Association) under the leadership of Qiu Shaoyao.^[80]

Although Tang's recently established relationship to institutions of state as a self-government deputy brought a certain mechanistic impersonality in social exchanges, his mode of domination over Tianyue farmers remained largely one "in which relations [were] made, unmade, remade by interactions between persons."^[81] The necessary "continuous creation" of dominance as patron required more than Tang's economic preeminence in the area and his claims to traditional cultural legitimacy. It may have included personal characteristics—demeanor, voice, habits, dress—which enhanced his dominance in any situation. But it definitely entailed his continuing success and expertise in the day-to-day "practice" of water control. Undoubtedly word had reached the villagers that their initial demand for demolition of the embankment had been compromised. There may thus have been some suspicion of the role of the patron/representative even before the word to delay had come from Beijing: had the case not been argued forcefully enough? What role did Tang's long-time friendship with He Bingzao, negotiator on the opposing side, have in the deliberations?^[82] These doubts likely contributed to the subdistrict farmers' willingness to choose another spokesman when Tang was unsuccessful, for Tang's role as patron had been badly damaged when defeat in the larger political arena drastically depleted his symbolic capital.^[83]

Little is known about the Qiu lineage or the leader of the Association, Qiu Shaoyao. With a lineage base since the Southern Song in two villages within two kilometers of the Ge lineage home, the Qiu produced no upper degree-holders during the Qing.^[84] The Qiu were also not included in the 1937 county gazetteer's list of nine important subdistrict lineages. Circumstantial evidence suggests that the Qiu, however, may have been one of the most significant Tianyue lineages whose arena of action was limited to subdistrict boundaries. Although there is no evidence that any Qiu served beyond the subdistrict in any capacity, in the 1922 subdistrict repair of the Ying Lake dikes no fewer than seven of the nineteen managers were from the Qiu lineage.^[85] These repairs only benefited area farmers, in contrast to the broader commercial implications of water projects in which Tang was involved.

Although Shaoyao had no recorded degree, he was obviously literate. The Association met three times from its inception in the spring until July 1913. Following those meetings, Qiu petitioned the authorities requesting rapid

action on the embankment case and an end to the officially imposed delay.^[86] During that three-month period, spring and summer rains had once again broken the Puyang River dikes, flooding the land in Tianyue. With the embankment opening insufficient for rapid drainage of the area, crops drowned or were felled in the mud by the force of the floodwaters. The Association wrote to Zhu Rui saying that "unless the outlet is widened, there will be no end in sight to the flood-caused famines"; the letter was signed "the men and women refugees of the forty-eight villages."^[87]

In the face of continued government silence, on the night of July 17, 1913, the families—men and women, old and young—of the forty-eight villages, armed with shovels and hoes, demolished the embankment. When police and troops arrived from Shaoxing, no one was at the site. Jute Creek flowed unobstructed into the Little West River for the first time since the mid-fifteenth century. As late as November 1913 Shaoxing elites requested restoration of the embankment, but heavy rains in late summer showed clearly that the river system could handle the additional water without a problem.^[88] There was general agreement that the river should be dredged and its banks should be made higher as a precaution. At the site of the former embankment, construction of a bridge began in December 1913 and was completed in July 1914.^[89]

Farmers, acting on their own authority, had ended the crisis. Qiu had clearly been their leader in communications with the government and had probably been involved in planning the demolition—though the sources list no leader of that episode. The government's choice neither to investigate the act nor to prosecute anyone involved indicates its probable satisfaction that the social and political tension had been relieved. Why did the Tianyue farmers act only at this time after over four centuries of bitter acquiescence to Suoqian elites? I would suggest that the rhetoric of change accompanying the 1911 Revolution and the active leadership of Tang and his allies raised hopes as never before that change would come at last. When these hopes were unexpectedly dashed, the long pent-up frustration exploded. Perhaps the widespread local political and military mobilizations in the aftermath of the revolution provided models for the farmers' response.

The main losers in the episode were Tang Shouchong and his Tianyue-Linpu allies.^[90] They had failed to prevail against outside opposition to the relatively minor change of widening the outlet, only to be rejected by embittered local farmers who then illegally solved the problem by destroying the dike. By spring 1914, Tang's local leadership credentials had been further diminished and discredited. The self-government system that had officially legitimized Shouchong's local leadership position was abolished in February. In April his brother Shouqian, former nationalist symbol and garnerer of prestige for the lineage, was also discredited, becoming in effect an added liability to Shouchong. Shouqian had at last assented to nationalizing the

Shanghai-Hangzhou-Ningbo Railway; he had vigorously opposed that action during the last decade of the Qing, and it now aroused considerable hostility. Shouqian's decision followed a 20,000 yuan bonus bestowed by the government: many interpreted that as a bribe, and Tang used the bonus to help make up a large cash deficit in a Hangzhou native bank of which he was the largest stockholder.^[91] Two weeks after Shouqian's decision, a telegram reached Shanghai that Shouchong had been assassinated at the Linpu native bank with which he was associated. Although the report was false, the reporter attributed a spate of such rumors to those who were "nursing grievances" (xiechou ) against Tang.^[92]

Restoring Legitimacy: the Ji Society

In this context of the rapid decline of his personal and lineage prestige, Tang Shouchong espoused a powerful local symbol from the past to reestablish his legitimacy and strengthen his damaged clientelist relations with Tianyue villagers in water control matters.^[93] He resurrected his father's idea of a Ji [shan] Society to revere and do obeisance to the spirit of Liu Zongzhou, the first advocate of changing the embankment into a bridge and relying solely on the Maoshan sluicegate.

The timing of the founding of the Ji Society—and its leadership, goals, and roles—indicate that it was primarily a vehicle to recoup the local standing of Tang and his allies and their authority in water-control matters. As for timing, Tang did not establish the society in the crisis of 1912-13 when the legacy of Liu might have been used as a rallying point for the subdistrict. Nor did he establish it during fall 1913, when the county and even the civil governor toyed with the idea of erecting another embankment. He organized it instead in spring or summer 1914,^[94] when the embankment was already destroyed, the bridge almost completed, and the area already totally dependent on the sluicegate. Such timing suggests more than simply remembering Liu for his particular water-control proposal.

Tang clearly dominated the society. Of eleven original "disciples" (houxue ) of Liu who joined in establishing the society, there were five Tangs (including the three brothers, Shouqian, Shouchong, and Shouming); Linpu Chamber of Commerce head Lu; Ge Bilun and Lu Luosheng, earlier representatives of the forty-eight villages; Tong Zhaolin, a close friend of Shouchong; and two representatives of the Xu and Hua lineages.^[95] At least nine of the eleven were close personal allies or relatives of Shouchong. Notably not involved was Qiu Shaoyao, who had led the Association in summer 1913. That Tang did not choose to organize a modern association or club (then fashionable in the core zones) is significant. In a time when Tianyue farmers had taken matters into their own hands—an obviously disconcerting episode for elites not in control—and in a place where Tianyue and Linpu lawlessness had bred out-

right unrest and rumors of assassination, Tang's claim to legitimacy and his opportunity for social and political dominance lay in reasserting not ideas of change but values of traditional cultural morality.

Specifically his goal was unifying the subdistrict community around a cult of Liu, a symbol of its past unity (albeit one produced by shared poverty and weakness) to heal the breach between key subdistrict elites and villagers.^[96] The old temple to Liu at the site of the Maoshan sluicegate was repaired; a new room was constructed to house a newly inscribed tablet and a portrait of Liu. Rites were held on the second day of the fourth month to observe the day of Liu's birth; essays were written to commemorate the founding of the society and celebrate Liu's significance.

If rituals "express in a condensed dramatic way the meaning of the shared values that bind a community together and, in the very act of expressing these shared values, deepen the community members' commitment to them,"^[97] Tang clearly hoped to deepen the subdistrict commitment to values he and his allies felt especially significant. Delineated on the temple tablet and in Ge Bilun's record of the society's establishment, the central values were reverence and gratitude for Liu's past contribution.^[98] Recalling the common heritage embodied in Liu's efforts was further designed to foster the subdistrict's sense of community, especially important at a time when the changed environment (the removal of the dike) would physically open the community to greater contact with areas downstream and likely provide challenges to old social arrangements and situations. A third value, the indispensability of social harmony, focused on Liu's renowned sense of morality as a salutary antidote to the threat of lawlessness in the Linpu-Tianyue area in general and the spirit of illegality that had given rise to the dike demolition in particular. Ge began his account with an implied comparison of the decline in morality and the threat of disorder during Liu's life at the end of the Ming with the situation of the early Republic. But, just as Suoqian elites had earlier utilized traditional symbols to maintain control over the embankment, Tang and his allies now used symbols of morality and harmony to restore control over their former clients.

Finally, but very important, in an emphasis that reveals Tang not only as keeping the traditional symbols but also as manipulating those symbols for change, the founders of the society stressed Liu's love of country, a value from Liu's own life that had never been central in the subdistrict's remembrance of his importance. In the past Liu was revered for his concern for the people of the subdistrict. In his commemorative essay, Ge stressed emulating the spirit of Liu's patriotism—he had starved himself to death out of loyalty to the Ming in the wake of the Manchu victory. This manipulation of the past had clear meaning for the present.

The perceived arena of the Ji Society was the world of the subdistrict and the world of the nation beyond—the worlds of the elites who founded it; it

was not solely the world of Tianyue led by the Qiu lineage. The society and the Liu cult were efforts to coopt the leadership of Tianyue water issues from Qiu, restore fully the leadership of Linpu and Tianyue to the Tang and their allies, and help redirect the vision of Tianyue farmers to the outside world. This nationalistic reemphasis on the meaning of Liu suggests that much subtle but important change could be effected amidst the continuity of old forms. As keepers of the Liu cult's symbols, Tang and his allies dominated not only the old forms but also the meaning of the past and the movement to new thought patterns and attitudes.^[99]

With the repair of the temple came renovation of the temple grounds. A large plum orchard was planted where there had once been a fish pond; to one side of the temple a new lotus pond was formed. These symbols of the new productivity of the subdistrict followed the dike's demolition. Whether the first year of the remodeled temple brought vandalism, we do not know. But in late 1915, Ji Society members petitioned the Shaoxing magistrate to give official protection to the temple and its grounds. Their stated concern was that ignorant rustics would cut down newly planted trees, harvest lotuses in the pond, and harm the temple structure itself. The Shaoxing magistrate ordered the desired protection, thus underwriting the concerns and dominance of the elite organization. One elite petitioner, it should be noted, was Qiu Shaoyao; he was now brought into the society, his leadership of the villages coopted.

Tang Shouchong dominated the reconstruction of water projects after the establishment of the Ji Society. Having tapped the subdistrict's reservoir of historical respect for Liu and having officially subsumed Liu's remembrance within the society, Tang could use the society to mask his own leadership and circumvent possible opposition. In 1914, for example, the Ji Society announced the establishment of a Forty-eight Village All Elder Society to discuss rebuilding the New Sluicegate Bridge near the Maoshan Sluicegate.^[100] Nine men were named directors of the project; only one, Tong Zhaolin, was a close associate of Tang. Yet the sources make very clear that, even though one Hua Xuchu was chosen head of the project, the superintendence and financing of the affair were completely in the hands of Tang Shouchong. He and Tong Zhaolin reportedly worked day and night, totally controlling the project. Given the unfriendly climate for Tang and his allies during the first half of 1914, the mechanism of the society allowed him to reassert his legitimacy and control.

In 1915, Tang's reputation had apparently been revived sufficiently to permit him to act on his own without reference to the Ji Society in calling together elders from the forty-eight villages to discuss the nature and cost of repairs on the Xiashao dike, southeast of Linpu.^[101] Once again, Tang managed the whole project, reportedly at great personal sacrifice. Although this effort was a personal success, Tang used the authority of the society to

announce regulations on dike maintenance from the stage of a dramatic performance at the April rites for Liu. With the power of ritual and symbol supporting the regulations, Tang and the Ji Society leaders undoubtedly hoped to impress their import more deeply into the minds of the people.

The Ji Society did not play any specific role in water projects after 1915. Tang managed major work on Linpu's Fire God Dike in 1915-16 and 1922 without reference to the Ji Society, which had always applied more directly to Tianyue subdistrict.^[102] In the repairs of the Ying Lake dikes in 1922, however, the society also was not a factor; in this effort, the Qiu lineage once again emerged as the major actor. As its last known act, the Ji Society published the documents concerning the embankment from the late Ming until its demolition in 1913. It thereby imprinted its domination of both the distant and recent past onto the future by selecting and arranging the materials that would constitute the historical record.^[103]

Conclusion

Two struggles for dominance in the Jute Creek case occurred within Tianyue subdistrict: the first between elites in former Lower Tianyue and elites and nonelites in Middle Tianyue; the second between upper-ranking elites (with bases in Linpu and Tianyue) and lower-ranking elites and nonelites within Middle Tianyue itself. In the two episodes, the nature of the elites and their patterns of dominance—specifically, the nature, mode, and application of dominance—differed.

The historically developed social ecology of the subdistrict is crucial in understanding the dynamics of the first episode of contention. While Lower Tianyue (Suoqian) had continued from the Ming period to thrive in a core region with close ties to Shaoxing via the Little West River, a flourishing artery of trade, Middle Tianyue had been relegated by political decision to a waterlogged periphery, isolated from direct water ties to the rest of the Three-River system. In addition to Suoqian's productive agricultural lands, the intermediate market town and salt entrepot of Suoqian provided lineages there with an opportunity for economic diversity through trade.^[104] The wealth of Suoqian was in sharp contrast to the poverty of Middle Tianyue, which, although oriented to the central market town of Linpu, most often found itself the town's exploited target rather than advantaged satellite.

More important were the social and political results of the economic situation in both areas. In the Suoqian case there was reportedly a greater diversity of elites, including wealthy farmers, commercial magnates, and larger numbers of degree holders. Many looked beyond the locality, and many had connections to elites in Shaoxing and beyond. In the Middle Tianyue periphery, there was a greater homogeneity of elites. Most were farmers, a few were involved in Linpu trade, and a small number were degree holders.

Connections to elites beyond the subdistrict were few, though Tang Shouqian was, at least at times in his roller-coaster career, a formidable "connection." In the end, the openly confrontational and antagonistic domination of Suoqian elites was both guaranteed by social ties that could tap political power at the capital for their purposes and continued (until the dike's demolition) through direct political orders to provincial officials to desist in their actions.

The dynamics of the second episode—the effort of Tang Shouchong to reestablish his domination over other Tianyue elites and nonelites through water control—were sharply different. Here the issues were patron-client ties and their sociocultural base. Tang exploited the culturally symbolic Liu cult and Ji Society to restore his symbolic capital and reassume the subdistrict's paternalistic leadership. In the culturally and socially homogeneous subdistrict, Tang reapplied his domination over the populace subtly and indirectly, using personal and value-based methods. If, following Bourdieu, we consider any application of domination in society to be "violence," then it can be said that Tang used the Ji Society to "euphemize" his "symbolic violence" against the people of the subdistrict.^[105]

Why did the people of Middle Tianyue accede to Tang's domination? The answer lies in the basically pragmatic, self-interested reactions of nonelites within their general social acculturation of submission to elite dominance. Throughout the episode, the farmers pursued their own interest as they saw it. Despite several centuries of inaction in the face of the calamity wrought by the embankment, their quiescence was not necessarily permanent. Specific circumstances finally motivated them to act: heightened anticipation that the problem would at last be solved led them to pressure Tang to reach a favorable conclusion. When Tang resigned, they formed the Association; when it too failed, they demolished the embankment. But just as it would be misleading to see the farmers as an always quiescent mass, it is equally incorrect to see them as a continuing interest group aggressively intent on forcefully directing their affairs on their own initiative. In subsequent water work, the villagers were again at the beck and call of Tang.

Their willingness to follow Tang over Qiu or other Tianyue elites undoubtedly came from his superiority in material and moral resources and their perception of that superiority.^[106] A simple comparison of Tang's and Qiu's resources makes the point obvious. In any sort of synthetic status ranking of their resources,^[107] Qiu, whose credentials were those of a literate lineage leader respected by Tianyue villagers, comes up short. In comparison, Tang was a wealthy and literate lineage head with a creditable record in public works, business involvement in Linpu, and connections to Linpu elites and his famous brother. Qiu could not transmute his degree of local legitimacy to permanent power equal to that of Tang any more than Tang could transmute his local legitimacy and dominance to power equal to that of even

better credentialed elites at a higher level.^[108] Pragmatic township residents could easily see Tang's advantages and the greater likelihood of his long-term success in water conservancy leadership. In this sense, Tang could fully reestablish his dominance because the villagers were willing to be dominated.

The general categories of elite dominance—economic wealth and diversity, political status and position, social status and connections—provide the bases for understanding in general the relationship between contending elites or between elites and masses. But in every time and place, the specific "practice" of dominance differed according to the issues and people involved. In the Jute Creek case, the ultimately decisive tools of domination were either based on or attained legitimacy through cultural values. I have argued that Tang regained his leadership and attempted to maintain local dominance through cultural symbols, ritual, and personal suasion. It should also be noted that the social connections bringing Suoqian elites temporary dominance were, in their origins, a cultural phenomenon (sharing a native place where respective ancestors were interred), objectified into a potent sociopolitical mechanism. In addition, in both episodes, the discourse of dominance was often symbolic, tied for import to cultural ideals. These episodes indeed illustrate that "domination is as much a matter of cultural and psychological processes as of material and political ones."^[109]

Finally, these episodes point to the fluid means of dominance in a society where, even in "objectified" state structures, social relationships (including the dominant-dominated nexus) formed a continually evolving dynamic core. In Tang's dealings with the township, the continuous maintenance, if not creation, of relationships was the means by which his legitimacy was established and dominance maintained. Suoqian's long-term dominance, upheld by government decree, was overthrown overnight by rebellious Tianyue farmers grown exceedingly weary and upset by the unyielding outside domination and the failure of their own patronlike leaders to bring satisfactory results.^[110] In the study of local history, it is finally the matters of human relationships—their continuities and discontinuities, their timing and expectations—amid the full array of contexts, contingencies, and processes that complicate the analysis of patterns of elite dominance. Undergirded by the very particularity of social and cultural norms amid which all Chinese acted, these relationships pose analytical challenges for the social scientist and the humanist alike.

Six
Local Military Power and Elite Formation: the Liu Family of Xingyi County, Guizhou

Edward A. McCord

In his path-breaking study of militarization in nineteenth-century China, Philip Kuhn identified the control of local military forces, specifically militia, as a key component of local-elite power in the late Qing and early Republican periods.^[1] There is, however, a paucity of supporting studies showing how militia leadership functioned in practice for specific elite families in the establishment, maintenance, or expansion of local power.^[2] In this article I attempt to meet this deficiency with a case study of the role of local military power in the rise of one family, the Lius of Xingyi county in Guizhou province, to a position of local, and eventually provincial, dominance.

The most renowned member of Xingyi county's Liu family was Liu Xianshi, military governor of Guizhou province from 1913 to 1920 and again from 1923 to 1925. As Guizhou's preeminent warlord, Liu Xianshi played a leading role in the political struggles and civil wars that ravaged southwest China in the early Republican period. Contemporary and historical accounts of Liu Xianshi's rise unfailingly note one distinguishing feature of his background: the militia base of his family's power. Both Liu Xianshi's grandfather and father, as well as three of his uncles, were actively involved in militia organization in the mid-nineteenth century; through militia leadership the Liu family originally established itself in a position of importance in Guizhou society. Heir to this family tradition, Liu Xianshi also led militia in Xingyi county in the last decade of the Qing and used this local force as the military base for his climb to provincial power.

One advantage to Liu Xianshi's notoriety as a warlord is that it has increased available historical materials on his family. With these materials it is possible to trace the family's use of local military power for a period extending from the mid-nineteenth century through the early twentieth. This

history shows not only that military power played an essential role in determining and advancing the social and political power of the Liu family but also that the control of local military force was most effective as a medium for expanding elite power within specific contexts. The successful upward mobility of the Liu family depended, in the long run, on a family strategy that adjusted to both military and civil opportunities in the changing social and political milieu of nineteenth- and twentieth-century China.

The Setting: Xingyi county and the Liu Family

Xingyi county is located in the southwest corner of Guizhou province precisely at the point where the province's border meets Yunnan and Guangxi provinces (map 6.1). To a certain extent, the county resembles the rugged, underdeveloped frontier areas often found along the borders of Chinese provinces. In terms of the core-periphery paradigm applied to China by G. William Skinner, Xingyi falls within the periphery of the Yun-Gui macroregion. Because the Yun-Gui macroregion was one of the most backward and least integrated in China, many of Xingyi's frontier features actually characterize the entire region.^[3]

Although most of Xingyi, like most of Guizhou, is covered by mountains, the terrain here is particularly rugged even by Guizhou standards.^[4] The rivers of southwestern Guizhou flow southeast into the West River system, which enters the sea at Canton, but they are not navigable until well into Guangxi province. Therefore, until the twentieth century, all transport to and from Xingyi went by human bearers or pack animals along well-worn paths. As for much of the Yunnan-Guizhou region, Xingyi county has only a small amount of level, arable land. Nonetheless, rich soil and plentiful rainfall make this land extremely fertile. Xingyi's county seat is located on a rich, irrigated plateau, one of the larger stretches of level ground in the province.^[5]

Despite its location, Xingyi county was not as isolated as many peripheral areas. Xingyi city was established at the site of a preexisting market town, Huangcaoba, that marked the juncture of two important interprovincial trade routes: one, with Huangcaoba at its center, linked Guizhou's provincial capital, Guiyang, with Kunming, the capital of Yunnan; the second went southeast from Huangcaoba into Guangxi where it tied into river transport systems ending in Canton. In the eighteenth century, Huangcaoba was already a well-established marketing center for local products from the forests and mines of southwestern Guizhou and eastern Yunnan. Raw domestic cotton imported into Xingyi in exchange for these products helped supply a local weaving industry. During the nineteenth century, the content of this interprovincial trade underwent some important changes. Opium was introduced to southwestern Guizhou at the end of the eighteenth century and

Map 6.1.
Xingyi County in Relation to the Provinces of Yunnan, Guizhou, and Guangxi.

eventually became the main product exported out of Huangcaoba. Penetration of foreign goods also increased in this period, and by the end of the nineteenth century foreign yarn from Canton steadily replaced domestic cotton as a main import into the Xingyi area.^[6]

Despite the large volume of trade carried along these interprovincial routes, the level of commercial organization in Xingyi county remained relatively low. In the 1890s, a visiting British consul described the bustling market at Huangcaoba as a rather wild and unregulated "free mart." According to his report, Huangcaoba had no banks, and very little money actually changed hands in the market. Most commercial transactions were conducted by barter, with opium serving as the most frequent medium of exchange.^[7] This low-level monetization reveals the relatively underdeveloped and peripheral character of Xingyi's economy.

Until the late eighteenth century most of southwestern Guizhou's inhabitants were members of non-Chinese, primarily Miao, ethnic groups. As in many parts of Guizhou, these ethnic groups came under increasing pressure from Han Chinese immigration in the eighteenth century. Because of its important trade routes, the Xingyi area was a particular target for Han settlement.^[8] The ethnic balance in the Xingyi area began to swing in favor of the Han Chinese following the harsh suppression of a major Miao rebellion in the late 1790s. As Miao and other ethnic groups involved in the rebellion in southwestern Guizhou were pushed into more remote areas, a new wave of Han immigration flowed into the Xingyi area. By the mid-nineteenth century, Han Chinese guest people (kemin ) were estimated to make up 70 to 80 percent of Xingyi county's forty thousand inhabitants.^[9]

One characteristic Xingyi shared with much of the Yun-Gui region was a low level of social and political integration. In fact, only after the defeat of the Miao rebellion in 1797 was Xingyi county officially established as an administrative unit as part of a larger reorganization designed to strengthen bureaucratic control over Guizhou's southwestern border.^[10] Nonetheless, the county retained a reputation for independence as a place "where the room [distance] of officials has been preferred to their efforts at government."^[11] The Xingyi area also had relatively few of the higher-degree holders or retired officials that formed the upper social elite in more settled areas of the country. One mid-nineteenth-century source noted that because of Xingyi county's position on the "Miao frontier," it was only beginning to produce degree graduates qualified for assignment to official posts.^[12] For the entire period from 1796 until 1854, Xingyi prefecture as a whole produced only fourteen provincial degree holders (juren ) and one metropolitan degree holder (jinshi ); of these, only one provincial degree holder came from Xingyi county.^[13] Although no exact correlation can be made, the comparative weakness of the state and the thinness of traditional gentry elites may have contri-

buted to the extraordinary rise of the Liu family to a position of local power in Xingyi county.

Although the arrival of the Liu family in Xingyi county cannot be precisely dated, it appears that they were part of the wave of Chinese immigration entering Xingyi at the end of the eighteenth century. Liu Xianshi's great-great-grandfather, Liu Taiyuan, a traveling peddler of literary supplies from Hunan, made the move to Xingyi county (figure 6.1). Liu Taiyuan settled in Nitang, a remote village on Xingyi's southern border, where he gave up his previous occupation to establish a tung oil extraction business. Thus, like many other immigrants into southwestern Guizhou, Liu Taiyuan had been attracted not by land but by commercial opportunities. Under Liu Taiyuan's son, Liu Wenxiu, the family oil business expanded to the point of hiring nonfamily labor. Real prosperity, however, was only achieved when Liu Wenxiu's son, Liu Yanshan, moved the family business from remote Nitang to Xiawutun, a village in a fertile irrigated valley only a few miles from Xingyi city. With greater access to the Huangcaoba market and the assistance of four able sons, Liu Yanshan's business flourished. In a short time, the Lius had become one of Xiawutun's richest families.^[14]

The Liu family's prosperity was accompanied by a shift in their economic interests. After moving to Xiawutun, Liu Yanshan began to use his commercial profits to acquire agricultural land. In the end, he and his sons abandoned the oil trade entirely to concentrate on managing their landholdings. By the mid-nineteenth century, the Lius had gone from being a leading merchant family to one of richest landlord families in Xingyi county.^[15] Given the success of the Lius' oil business, it seems unlikely that this shift to agriculture was simply economic. Although the sources are silent on this matter, one can assume that the family may have been responding to traditional values that emphasized the prestige and security of landholding over commercial activity.

Within traditional Chinese society the next logical step for the Lius might have been to seek advancement into the ranks of the scholar-gentry by competing for degrees in the state-sponsored examination system. In some regards, the Lius were already poised for such a step. According to one account, the Liu family established a private school at their original home in Nitang.^[16] Liu Yanshan's four sons, and most likely Liu Yanshan himself, were literate and well versed in the classics on which the examination system was based.^[17] Whether or not the Lius intended to pursue this route to educational advancement, they were soon distracted by more pressing concerns. Beginning in the 1850s, a number of increasingly serious rebellions began to erupt in different parts of the province. When the Lius stepped forward to organize local defenses against these rebellions, they embarked on a path that would bring them greater rewards than they could have hoped to gain through the examination system.

Figure 6.1.
The Lius of Xingyi

The Military Base of Elite Power

Although coinciding somewhat with the outbreak of the Taiping Rebellion in neighboring Guangxi in 1851, Guizhou's rebellions had indigenous roots in increasing ethnic tensions and economic inequities. From 1850 to 1875 there was always at least one, and usually more than one, rebellion active in some part of the province. Before the end of this twenty-five-year period over 90 percent of Guizhou's administrative cities would fall at some point to rebel forces, and at the height of the rebellions more than 60 percent of the province would be under rebel control.^[18]

For the first half of this period, Xingyi county, apparently escaping much disorder of these rebellions, may have even prospered. According to one account, the market at Huangcaoba flourished during the rebellions, "keeping the peace in its own streets and selling opium, arms, and foreign goods to all who could pay for them."^[19] War finally came to Xingyi with an ethnic Hui (Muslim) rebellion that arose along the Yunnan-Guizhou border in 1858 and sparked other local ethnic and secret society uprisings. In 1860 the county seat of Pu'an directly north of Xingyi county became the first in a chain of administrative cities to fall to the rebels. By 1862 the rebels had advanced to take Xingyifu, the administrative seat of the Xingyi prefecture, less than sixty miles from Xingyi city. As the rebellion spread, local officials, following precedents established elsewhere in Guizhou and other provinces, encouraged local communities to organize militia for their own self-defense.^[20]

The Lius were one of many families in southwestern Guizhou who stepped forward to manage local self-defense when they saw their own localities threatened. In 1860 Liu Yanshan and his four sons led local residents in constructing a stone fortress and then began to train local residents as militiamen. The Lius' defensive efforts at Xiawutun came none too soon. Six months after the fall of Xingyifu, Xingyi city also fell to rebel forces. The fortress at Xiawutun, which became a place of sanctuary for the local refugees, was soon besieged. After seven desperate months, the Lius were finally forced to negotiate a token surrender to the rebels.^[21] Despite this questionable tactic, by successfully guiding Xiawutun through this particularly difficult period, Liu Yanshan and his sons established their reputations as local military and community leaders.

The Liu family's militia leadership provides a case in point for Philip Kuhn's observation on the inadequacy of defining the local elite solely in terms of degree-holding gentry; not only local degree holders but also many wealthy or influential commoners assumed militia leadership, clearly an elite function.^[22] In a study of Sichuan province in the 1850s and 1860s, Keith Schoppa has also shown that the majority of the men who performed conventional elite functions in local defense, education, and philanthropy were not degree holders.^[23] Thus attention to elite activities instead of status categories

both provides a-more accurate picture of local elites and particularly identifies nondegree-holding elites such as the Lius of Xingyi county.

Once degree holding is no longer the sole criterion for elite status, an examination of elite activities presents a more dynamic picture of change within the structure of local-elite power. The Lius, emerging as a wealthy landlord family, must at least be considered members of the village elite of Xiawutun, but prior to the 1860s the scope of their influence was limited. Before this time, for example, there is no sign that the Lius had assumed any community leadership roles. Likewise there is no indication that the Lius functioned as power holders in the tradition of "local strongmen" often seen in frontier areas.^[24] The Lius' landholdings, after all, were located in one of the county's more settled areas and did not require private armed retainers for their protection. Only in their organization of local defenses in the 1860s did the Lius begin to act as community leaders and become local strongmen; also as militia leaders the Lius extended their influence beyond the borders of their own village. Militia leadership therefore provided the Lius with a means for upward mobility unavailable to them in more peaceful times. Thus, militia organization served as an arena for not only elite activity but also elite formation.

In the years following the siege of Xiawutun, the Liu family became involved in local defense and militia activities on an ever wider scale. While Liu Yanshan guarded the family's base at Xiawutun, his sons led militia in campaigns against various rebel strongholds in both Xingyi and neighboring counties. There are no exact records on the size or organization of Liu family forces, but at different times estimates of the Liu brothers' command ranged from several thousand to more than ten thousand men. Three times (1864, 1866, 1868) militia led by the Lius helped retake Xingyi city from rebel forces. Subsequently, the Lius also participated in recovering other important cities in southwestern Guizhou, including Xingyifu and Xincheng. The Lius' militia activities were not pursued without some family losses. Liu Yanshan's second son, Liu Guanlin, was killed in an assault on Xingyi city in 1863, and his first son, Liu Guanzhen, was killed in a feud with neighboring militia leaders in 1865. Liu Yanshan himself died, apparently of natural causes, in 1867. After this, Liu Yanshan's third son, Liu Guanli, assumed primary leadership of the family's military forces, assisted by his younger brother, Liu Guande.^[25]

The activities of the Liu family in this period, however, were not limited to commanding militia forces; to support their forces, the Lius were also active in collecting funds and provisions. No doubt benefiting from Xingyi's wealth as a commercial center, the Lius not only supplied their own forces but also supported regular army units campaigning in southwestern Guizhou and provided funds to obtain the services of militia from neighboring areas to assist in their own campaigns. Liu family leadership and fund-raising skills

were also applied to community construction projects, primarily defensive. Besides the stone fortress they built at Xiawutun, the Lius aided in constructing ten other forts in neighboring areas. After the third recovery of Xingyi city, Liu Guanli and Liu Guande led efforts to rebuild and strengthen the walls of the county seat. Finally, after the death of Liu Guanzhen, the Lius directed the construction of a loyalty temple (zhongyi ci ) to commemorate him and other local "martyrs." When this temple burned in 1875, the Lius again contributed to its reconstruction.^[26] The Lius' attention to this temple no doubt reflected its symbolic importance to the family, a concrete reminder of their community leadership.

The position of local leadership attained by the Liu family in suppressing rebel forces of the 1860s and early 1870s was enhanced by the receipt of official titles and brevet offices. Liu Guanzhen was awarded an army rank of major (youji ) for his role in the first recovery of Xingyi city. After his death, this title was passed on to his younger brother, Liu Guande. Liu Guanli received the honorary rank of subprefect (tongzhi ) for his role in the second recovery of Xingyi city. His military abilities were also recognized by appointing him with official authority over all militia in Xingyi prefecture. After the recovery of Xingyifu, Liu Guanli was further rewarded with a prefect's rank (zhifu ) and the right to wear a peacock feather.^[27] In a society where the state had always claimed the right to determine the criteria for social status, such titles and awards were not considered empty honors. Indeed, the paucity of regular upper-degree holders in Guizhou made these awards even more valued, something officials sponsoring militia realized and used to their advantage.^[28] The titles received by the Lius were important, then, because they both legitimated their positions as local leaders and considerably advanced their community status. In the eyes of the state and society, the Lius' official titles removed them from the category of commoners and placed them within the ranks of the gentry.

The Lius also found ways to translate their newly acquired local power as militia leaders into more material benefits. First, the family no doubt gained from the plunder of territories recovered from rebel hands.^[29] Second, the Lius were not above using their new authority to consolidate their family landholdings by intimidating other landowners into selling or ceding to them prime lands around Xiawutun.^[30] Finally, the Lius benefited from their access to funds and provisions collected for support of their military forces. Sources sympathetic to the Lius attributed their effectiveness in obtaining contributions to the high esteem in which they were held by the public.^[31] But charges brought against the family for illegal exactions and embezzlement suggest that these contributions were often more coerced than voluntary and that the Lius were not averse to pocketing part of the proceeds for their own gain.^[32]

Although the original purpose in raising militia had been to aid in sup-

pressing rebellion, the Lius also used their military power to better the family's position in struggles with other "loyal" forces. Contrary to views of elite solidarity in the face of rebellion, conflicts over manpower, funds, or the command of military operations often led to feuds and even open warfare among various militia leaders. The death of Liu Guanzhen in 1865 was the result of one such feud between the Lius and another group of Xingyi militia leaders.^[33] Such conflicts with militia rivals intensified after 1866 when Liu Guanli attempted to assert his authority as general militia head for Xingyi prefecture. Sometimes these power struggles even brought the Lius into conflict with local officials and regular army units. One such case occurred in 1866 when Liu Guanli initiated an attack on the forces of a neighboring militia leader who refused to accept his authority. Liu's adversary, however, received the support of Xingyi's prefect and a local brigade-general who sent official troops to repel Liu's assault. Only after a new prefect, a previous military ally of the Liu family, was appointed, was Liu Guanli able to defeat this opponent.^[34] In another instance, the Lius reportedly attacked Xingyi city to force the flight of a magistrate who had sought to curtail their growing local power. After gaining access to the city, the Lius proceeded to massacre a number of gentry families who had supported the fallen magistrate.^[35] Although eagerly accepting official recognition for their role in the suppression of rebel forces, the Lius could and did use their military power to challenge official authority when their own interests were at stake.

In these local power struggles expedient alliances with rebel forces could also blur the distinction between orthodox and heterodox forces. For example, one weak militia leader caught between powerful rebel forces on one side and aggrandizing militia leaders, including the Lius, on the other, changed sides no less than five times within a three-year period.^[36] Despite their own questionable surrender to rebel forces during the first siege of Xiawutun, Lius were not reluctant to use charges of rebel collaboration, true or not, against their enemies; thus Liu Guanli finally avenged his brother's murder. Labeling these enemies rebel collaborators, Liu attacked their forces, pillaged their villages, and massacred their families.^[37] Through such tactics, the Lius eliminated possible militia competitors while expanding their own military power by absorbing rival forces into their own.

As can be seen from these cases, the military conflicts of this period were hardly limited to the war between elite forces of order and armies of rebellion. The organization of militia, although meant to combat rebellion, also introduced military force as a new element in determining elite power, and elite power structures had to be reconstituted to incorporate this new element. To some extent, then, the conflicts between militia leaders, and between militia leaders and officials, were the violent manifestations of readjustments in the hierarchy of local-elite power.

Abuses of power and disregard for authority by militia leaders were natu-

rally matters of official concern.^[38] While rebellions still flourished, however, the government was often willing to overlook the transgressions of important militia leaders such as the Lius in exchange for their support. Nonetheless, in the early 1870S the Qing court finally felt compelled to respond to the accumulating charges and complaints against the Lius and ordered Guizhou's governor to investigate officially the family's activities. Because military resistance to a court order would have placed the family irrevocably in rebellion, this challenge demanded a political rather than a military solution. Liu Guanli immediately departed for the capital of Yunnan to seek higher-level protection from the governor general. Liu Guande, meanwhile, remained in Guizhou to oversee the bribery of local officials. In the end, the governor general interceded on the Lius' behalf, and the investigation was dropped. Furthermore, on the recommendation of his new benefactor, Liu Guanli was even appointed expectant intendant (dao ) and was retained for a time on the governor general's staff.^[39] The generous application of wealth and the careful cultivation of official patronage enabled the Lius to emerge from this affair not only unpunished but also even more secure in their position.

Although the Lius of Xingyi may have originally assumed militia leadership as a matter of self-defense, militia power helped the family do more than simply survive. Before the rebellion the Lius were already a prominent and wealthy landlord family, but by commanding local militia the Lius gained new status and influence as community leaders and local strongmen. By applying military force, the Lius increased their family wealth and eliminated local rivals. As a result of their military achievements, the Lius gained official honors that gave them status equivalent to the degree-holding gentry and established valuable official contacts that extended as far as the governor general's office in Kunming. In the course of their military campaigns their influence grew beyond the confines of their own village; by the end of the rebellions they were considered the most powerful family in Xingyi county and one of the most important families in all southwestern Guizhou. For the Lius, then, militia leadership served as a medium for a remarkable leap in local status and power. However, the family's escape from official prosecution also attests to limits on the uses of military power. Ultimately, maintaining the Liu family's elite position would depend on the extent to which they could be more than military strongmen.

Giving Up the Gun

Conditions in southwest Guizhou had changed considerably when Liu Guanli returned from Yunnan in 1875. The previous year, the last of the Hui rebel strongholds had been taken by government forces, and the Muslim rebellion had ended. Although scattered bands of ethnic or secret society rebels continued to trouble parts of Guizhou until 1880, by and large the

major rebellions had by this time been defeated. Liu Guanli responded to the restoration of peace by announcing that he would retire from government service to devote himself to the care of his ailing mother. After fifteen years of almost continual fighting, Liu then proceeded to disband all his family's military forces.^[40]

Given the importance of militia leadership to the expansion of Liu family power, Liu Guanli's willingness to dissolve his family's military forces appears at first glance almost quixotic. Nonetheless, Liu's response to restoring peace was not all that unusual. There is considerable evidence that most elite-led militias raised throughout China to combat midcentury rebellions were disbanded in the more peaceful postrebellion period.^[41] The only question that requires further examination is why militia leaders such as the Lius abandoned so willingly the control of local military power that had been for a time so essential to their elite positions.

One of the most important reasons for Liu Guanli's decision to disband his family's militia was certainly his recognition of the risks in attempting to maintain military power after the rebellions ended. As the government's military strength recovered, officials were better positioned to curb the excesses of militia leaders, especially those whose activities had been too blatantly illegal or whose loyalty had been too inconsistent. The Lius had participated in enough campaigns against "rebellious" militia commanders to know how easily this label might be applied to themselves.^[42] There was also a lesson to be gained from the family's costly struggle to defend itself from prosecution in the early 1870s. Their strong-armed self-aggrandizement had antagonized local officials and many members of the local elite. Continuing to behave as local military strongmen would allow these antagonisms to fester and might make the government more attentive to their enemies' complaints. With the restoration of peace, the best hope for maintaining the family's position lay in abandoning, not preserving, their militia power.

Liu Guanli was perhaps also willing to give up militia leadership with few hesitations because to do so by this time detracted little from his family's local power. Through militia leadership the family had gained new power "resources" that, once acquired, did not depend on military force to be maintained. One resource was Liu Guanli's expectant intendant rank, which gave him both community status and access to high officials. Another hidden strength was Liu's former position as general militia head for Xingyi prefecture. Through this position Liu could claim a superior-subordinate relationship to most militia commanders of southwestern Guizhou who, like the Lius, emerged from the period as important local community leaders.^[43] Marriage ties also helped solidify these relations. Liu Guanli married his daughter to the son of Wang Peixian, who, after the Lius, was perhaps the most important militia leader in Xingyi county and one of the Lius' closest allies.^[44] The Lius' influence with such men helped balance the enemies they

had made among other members of the local elite. Finally, the Liu family had emerged with more wealth and property than they had had before the rebellions. The political advantages of wealth were readily evident in the bribery that helped the family avoid prosecution in the early 1870s. In the years to come, the Lius continued to use their wealth to cultivate official goodwill by generous "gifts" to local and provincial officials and their staffs.^[45]

The dissolution of the Lius' military forces thus had little effect on the Lius' local power. Indeed, in the postrebellion period they continued to be recognized as the dominant family in Xingyi county, and to a certain extent in southwestern Guizhou as a whole.^[46] The Xingyi magistrate reportedly would not act on any issue without first consulting with Liu Guanli, and Liu's opinion generally settled any local affair, big or small. As a result of his far-reaching power following the rebellion, Liu Guanli was pictured as heading a "small court" in Xingyi with influence over the entire Pan River watershed of southwestern Guizhou.^[47]

From the case of the Liu family alone it is impossible to determine the extent to which state power may generally have devolved into hands of local elites in the postrebellion period. In the 1890s the British consul Frederick Bourne reported that the Xingyi county magistrate had candidly acknowledged that he had little authority in the county.^[48] But Bourne saw this as merely continuing government weakness inherent in the region. Nonetheless, the relative scarcity of upper gentry in southwestern Guizhou through the mid-nineteenth century suggests that the rise of the Lius to a position of such influence was an elaboration of elite power that, at the very least, changed the structure of Xingyi society. Although the state's position may not have been any weaker than it had been before the rebellion, the power of the Liu family in Xingyi county may have helped thwart any attempt to strengthen government authority.

The Liu family's continued domination of Xingyi local society was not, however, simply the result of its achievements from the period of militia leadership. Viewed over time, elite power in local society appears not as a static construct but as a set of relationships in a constant process of formation and reformation. The Lius' activities in the postrebellion period seem to show an awareness that the maintenance, let alone enhancement, of their position and influence required that they seek out new means of renewing their power as local elites. The particular strength of the Lius as elites was their flexible response to changing circumstances and opportunities. Just as they adapted to possibilities for elite power found in militia leadership during the period of rebellion, the Lius turned their attention to civil potentialities for elite power in the peaceful period that followed.

Traditionally, education served as the main channel for status enhancement in Chinese society; and in the last quarter of the nineteenth century Liu Guanli focused his public efforts in this field. Although the titles Liu had

received as a militia leader placed him officially at a level equal to the upper-degree holding gentry, ranks attained through military achievement were viewed in practice as inferior to those acquired through education. The power the Liu family wielded in local affairs would not necessarily translate directly into social acceptance within the ranks of the scholarly elite. Thus, Liu's attention to education at this time not only revived family interests that had been interrupted by their military activities but also initiated a strategy to transform his family's social status for more peaceful times. As a militia leader Liu had made many enemies among the local elite by his bullying, but in his new guise as educational patron he presented a more cultivated image by being "courteous to the wise and condescending toward the scholarly."^[49] Personal behavior and public activity thus combined to reflect the Liu family's gentrification along more socially acceptable lines.

Liu Guanli's educational pursuits involved many traditionally approved activities. He established an academy in Xingyi county and filled it with the books needed for a classical education. He also provided financial assistance for the education of promising local students. Finally, to raise the quality of local education he provided generous stipends to bring prominent scholars from the provincial capital to Xingyi either to supervise the county's educational programs or to lecture in its schools.^[50] Largely through Liu's efforts, Xingyi county became, despite its peripheral location, something of a regional educational center.

Besides the reputation Liu Guanli gained as a patron of local education, his activities had even more concrete benefits. His assistance to young local scholars established patron-client relationships with the men most likely to emerge as the next generation's social elite. Likewise, Liu established social contacts with the provincial elite through the scholars he brought to Xingyi to assist in his education projects. This same end was achieved by generous gifts sent to promising scholars in Guiyang.^[51] By assuming the role of a patron of men of talent, Liu solidified his social influence within the Xingyi area and extended it to the provincial capital.

Liu Guanli's goals in promoting local education were not, however, simply limited to enhancing his own reputation. Liu was also providing for the educational advancement of his own sons and nephews. While he and his brothers had their educations interrupted, Liu Guanli could hope that the younger generation might contribute to the family fortunes by succeeding in the traditional examination system. In the mid-1880s the family achieved some success when Liu Guanli's elder son, the future warlord Liu Xianshi, and his nephew, Liu Xianqian (the son of Liu Guanzhen), both earned the lowest level degree (shengyuan ).^[52]

The year 1875 therefore was a turning point in the history of the Liu family whereupon they shed the military power that had made them southwestern Guizhou's preeminent strongman family in favor of pursuing civil

status through educational promotion and advancement. For the Lius, military and civil activities were not exclusive categories of community leadership; rather, the sharp contrast between the family's activities before and after 1875 appears determined primarily by the different contexts of the two periods. Indeed, when changing conditions at the turn of the century offered new opportunities for community leadership in both military and civil fields the Lius responded with alacrity in both areas.

The Revival of Family Military Power

After a quarter of a century of relative peace, Guizhou in the first decade of the twentieth century was again beset by periodic local disturbances and ethnic rebellions. By one account, a total of seventy-three separate uprisings occurred in the province in this period.^[53] With official approval, Guizhou's local elites again responded by raising militia for local self-defense.

For Xingyi, the main threat came not so much from local uprisings as from a series of secret society rebellions that began in neighboring Guangxi in 1898 and produced roving rebel and bandit bands that harassed Guizhou's southern border.^[54] In the face of this threat, Liu Xianshi gave up his studies to devote himself to militia work, accepting a position as manager of Xingyi county's General Militia Bureau (tuanfang zongju ).^[55] The greatest danger to Xingyi came in 1902 when a large rebel army crossed the river that formed the county's border with Guangxi. At this point Liu Guanli also stepped forward in response to official appeals to help organize the county's defenses. Because the county's own garrison troops had been called away to quell rebel forces in Yunnan, the Lius first attempted to use their militia to defend Xingyi city. They soon found themselves outnumbered and abandoned the city to defend their own fortress at Xiawutun. For six days Xiawutun itself was then besieged by the rebel army. Before they could be forced to surrender, though, government troops fought their way back into the county, and the Lius reemerged from their fortress to join the offensive. The Lius then helped government armies recapture Xingyi city and push the rebels back into Guangxi.^[56]

As a result of this action, the Lius revived their military reputation and gained new official rewards for their military services. First, in the course of the campaign the militia units led by the Lius were officially given an irregular army designation as the Border Pacification Battalions (Jingbian ying). Liu Guanli was appointed general commander (tongling ), and both his son, Liu Xianshi, and his nephew, Liu Xianqian, were given battalion commander (guandai ) posts.^[57] According to one source, Liu Xianshi was also rewarded with the honorary rank of a county magistrate.^[58] Later, in recognition of his military abilities, Liu Xianqian became the commander of the Guangxi governor's personal guard.^[59] Thus, militia leadership again rein-

forced the Liu family's standing by helping its members acquire new official titles and posts.

Although the original rebel problem had been handled, the unsettled conditions in Guizhou as a whole argued against the complete disbandment of the Liu's militia. With Liu Xianqian's departure for Guangxi, the Border Pacification Battalions were reduced from two to one. But, Guizhou officials continued to see the Lius' forces as an important element in southwestern Guizhou's defenses and encouraged their maintenance. Thus, Liu's battalion was not included in a 1906 reduction of old-style irregular military forces carried out by Guizhou's governor in order to provide funds for a modern new army. Instead his unit was one of twenty battalions of these old-style forces retained for incorporation into a new standardized patrol and defense force system (xunfangying ). Liu Xianshi's battalion was renamed the second battalion of the West Route Patrol and Defense Forces, and Liu Xianshi retained his position as its commander.^[60] It should be noted, however, that as one of the poorest provinces in China, Guizhou had no funds for the extensive reorganization and retraining required by national orders establishing the patrol and defense force system. Thus, to some extent recognizing Liu's force as a patrol and defense battalion was merely an attempt to create the appearance of compliance with military reform requirements at minimal cost to the provincial treasury. For all practical purposes, Liu's battalion remained a militia unit under Liu's personal control. In this instance, the patrol and defense force reorganization simply provided further legitimation for the Liu's local military forces.

Prior to assuming militia leadership Liu Xianshi appeared prepared to seek advancement through the degree system, but, as a result of his role in suppressing the Guangxi rebels, he obtained a military office and established a military reputation. Because of his later emergence as a warlord, Liu Xianshi's military activities are often seen as the most salient feature in his background. But, in fact, despite his retention of this military post, much of Liu Xianshi's attention after suppressing the rebel army was taken up with civil not military matters. Proceeding naturally from Liu Guanli's previous educational work, the Liu family became actively involved in the new educational and reform programs that were among the most important features of elite activity in the years immediately preceding the 1911 Revolution. These civil activities in the end proved as important as the Lius' renewed military power in Liu Xianshi's eventual emergence as a warlord.

Late Qing Reforms and Expanding Elite Power

The Liu family's connections to late Qing educational reforms had early beginnings. In 1897, Yan Xiu, an eminent scholar with ties to Kang Youwei's reform party, was appointed as Guizhou's educational commissioner.

Arriving in Guiyang, Yan established the Statecraft School (Jingshi xuetang), the first Guizhou educational institution to introduce Western elements into its curriculum.^[61] Only a select group of forty talented students was enrolled in the new school; among them was Liu Xianzhi, Liu Xianshi's younger brother. In the following years, Liu Guanli's program of inviting prominent scholars to lecture in Xingyi schools also brought leading reformers into contact with the Liu family.^[62] Despite Xingyi's peripheral location, these educational contacts kept the Lius well informed of the reform programs emerging in the provincial capital. The Lius quickly realized that these programs were creating new arenas for expanding elite power, and they moved to confirm and enhance their own local leadership by introducing these reforms into the Xingyi area.

The Lius' own participation in reform activities began with locally implementing the modern educational system mandated by the court to replace the civil service examination system in 1905. Liu Guanli actively promoted educational reform, but due to his ill health (he died in 1908) his son, Liu Xianshi, acted as the effective head of the family in these matters. Thus, although Liu Guanli supported the establishment of a Xingyi educational promotion office (quanxuesuo ) to oversee the county's educational revitalization, Liu Xianshi actually took the lead by serving as its general manager. Many of Liu Xianshi's efforts in this post were directed toward setting up modern schools in accordance with new educational directives. The private academy originally founded by the Lius was converted into an upper-level primary school (gaodeng xiaoxue ). At the same time, Liu assisted in establishing about twenty lower-level primary schools within Xingyi and others in surrounding counties. Finally, Liu Xianshi was also involved in founding other educational institutions, such as a girls' school, a teacher training institute, a military school, and a public reading room.^[63]

The Lius also continued to foster their patron-client relations with other Xingyi families by encouraging and supporting the education of their sons. The Lius paid special attention to promising students, helping them attend the higher-level professional or technical schools that were beginning to be established in Guiyang, and the family even provided stipends to help more than twenty students seek advanced education in Japan.^[64] Among these students was He Yingqin, the scion of a prominent Xingyi family who would later gain fame as a Guomindang general and Nationalist China's minister of war. He's military education in Japan was supported by Liu Xianshi, and his tics to the Liu family were later enhanced by marriage to Liu Xianshi's niece.^[65] Naturally, as in their previous efforts, the Lius did not fail to insure that the younger generation of their own family would also benefit from these new educational opportunities. For example, Liu Xianshi's younger brother Liu Xianzhi, his eldest son Liu Gangwu, and a nephew

Wang Boqun (the grandson of Wang Peixian) were among a number of family members sent to pursue higher education in Japan.^[66]

Although concentrating on education, Liu Guanli and Liu Xianshi also responded quickly to other types of elite reform. In the economic field, Liu Guanli supported a proposal for constructing a Yunnan-Guizhou-Sichuan railroad and encouraged Liu Xianshi to set up forestry companies. In politics, Liu Xianshi headed a Xingyi association to prepare for local self-government. In support of social reforms, Liu Xianshi even helped establish a society for the elimination of footbinding.^[67] In regard to the content of their activities, the Lius remained imitators rather than innovators. All the reforms the Lius introduced into Xingyi followed precedents established in the provincial capital. Still, their role in diffusing reform programs belies the usual perception that late Qing reforms were the preserve of urban elites in more advanced core areas.

One important result of the Lius' participation in reform activities was the enhancement of their reputation with and connections to members of the emerging provincial, and even national, reformist elite. The Lius' efforts to bring eminent scholars from Guiyang to Xingyi continued, in this period, to help the Lius establish personal relations with members of the provincial elite. Liu Guanli's reputation in educational affairs had already been established, and Liu Xianshi's assumption of leadership in reform activities helped him gain the respect of Guizhou's educational and constitutionalist leaders in his own right. Equally important were the contacts made by the younger members of the Liu family while attending new schools in Guiyang or higher institutions abroad. For example, while attending the Statecraft School in Guiyang, Liu Xianzhi's teachers were Guizhou's most prominent educators, and many of his classmates were the children of Guiyang's eminent families. After going to Japan, Liu Xianzhi entered the circle of reformers that gathered around Liang Qichao, and he even served for a time as Liang's aide. Upon completing his education in Japan, Liu was invited to take up a position on the Yunnan governor-general's staff.^[68] By achieving this post Liu Xianzhi illustrates the opportunities for elite advancement that many found in the new educational system. At the same time, the contacts he made while pursuing his education also contributed to his family's broader social connections at provincial and national levels.

The most telling sign of the social standing the Liu family had achieved in the first decade of the twentieth century was the marriage of Liu Xianshi's daughter to the son of Tang Eryong, a provincial degree holder (juren ) and the scion of one of Guizhou's eminent families.^[69] As the founder and head of Guizhou's provincial educational association, Tang Eryong was the single most powerful man in Guizhou educational circles and a leading member of the provincial reformist elite. The matchmaker who arranged the Tang-

Liu marriage had some reason to boast that it marked "the reach of Tang Eryong's power out into the province and the advance of Liu Xianshi's power toward the center."^[70]

Even though reform programs became the main focus of the Lius' public activities in the last years of the Qing, they did not abandon their local military concerns. Indeed one of the Lius' educational projects, their military school at Xingyi, helped bolster the family's military power by training a cadre of officers for Liu Xianshi's local military force. No matter how pressing his other activities, Liu Xianshi also continued to inspect his troops regularly and insure their loyalty by personal attention to their needs.^[71] Even after the death of his father in 1908 Liu Xianshi did not give up his military post but rather moved his battalion headquarters into the family residence so he could continue to oversee its affairs while in mourning.^[72] In terms of the Lius' total influence, however, their educational and other reform activities, not their military power, projected the family into a wider social and political arena. Although Xingyi county remained the Lius' main power base, this base stood at the center of an emerging network of elite relations that prepared the family for a new outward extension of power during the 1911 Revolution.

The Making of a Provincial Warlord

One of the most important factors influencing the course of the 1911 Revolution in Guizhou was an antagonistic split in the province's reformist elite over the control of Guizhou's public organizations. One faction, organized as the Self-Government Study Society (Zizhi xueshe) under the leadership of Zhang Bailin, successfully won a majority in the first Provincial Assembly elected in 1909. Their opponents, the Constitutional Preparation Association (Xianzheng yubei hui) led by Tang Eryong and Ren Kecheng, based their power on control of the province's educational institutions.^[73] At the beginning, the division between the two factions was primarily personal and political not ideological. Over the course of their struggle, however, Zhang Bailin's group slowly became more inclined toward a revolutionary republican program. When news of Hubei province's successful revolutionary uprising against the Qing dynasty reached Guizhou in mid-October 1911, Zhang's "revolutionary" faction, not their more conservative "constitutionalist" (i.e., constitutional monarchist) opponents, set about plotting a similar uprising in Guizhou.^[74]

The outbreak of the 1911 Revolution in Hubei found Guizhou's governor, Shen Yuqing, in a precarious position. Although new to his office (he had only arrived in Guizhou in April 1911), Shen was well aware of growing revolutionary sentiment in the province. He was particularly concerned, quite rightly as it turned out, that Guizhou's modern New Army, which

served as his capital's main garrison, might follow the lead of Hubei and other provinces and carry out a revolutionary coup. In the face of this danger, Shen turned for advice to Ren Kecheng, head of the constitutionalist faction. Besides suggesting that the leaders of the Self-Government Society be rounded up and executed, advice Shen declined to follow, Ren proposed that Liu Xianshi be called upon to lead a force of local troops to Guiyang to defend the capital against a possible New Army uprising. With his personal military experience and his family's long history of loyal military service, Liu Xianshi was a logical candidate for this assignment. At the same time, given his family's strong ties to the constitutionalist faction through Tang Eryong, political considerations no doubt also influenced Ren's recommendation. In any event, Shen accepted Ren's advice and sent an urgent message to Liu Xianshi asking him to lead five hundred men to Guiyang. As a "sweetener" to insure Liu's speedy compliance, Shen offered to supply Liu's men with new guns from the provincial arsenal. Upon receiving this message Liu showed no hesitation and immediately set out for Guiyang at the head of five hundred men.^[75]

Guizhou's revolutionary plotters, meanwhile, were emboldened by the news that a successful New Army uprising had taken place on October 30 in neighboring Yunnan. At the same time, they learned of Liu Xianshi's advance toward Guiyang and decided their own plans could no longer be delayed. On the evening of November 2, New Army soldiers and military students rose in support of the Revolution. With no reliable military forces at hand, the governor was forced to surrender to the provincial assembly, and on November 3 the assembly declared Guizhou's independence. While a Japanese-educated new army instructor, Yang Jincheng, was elected military governor, the real power in Guizhou's revolutionary regime was a policy-making cabinet (shumi yuan ) headed by Zhang Bailin.^[76]

Liu Xianshi and his army had only marched halfway to Guiyang when they received news of the uprising. Although Liu's chance to be a counterrevolutionary hero had been snatched from his grasp, he was unwilling to miss an opportunity to play some role in the events at Guiyang; therefore, instead of retreating to Xingyi, he sent a bold message to the new regime proclaiming, "Although I have approached the capital this time under the orders of Governor Shen, for a long time my heart has inclined toward the revolution you gentlemen have supported."^[77] Liu then asked that he be allowed to proceed to Guiyang to bolster revolutionary military power. Because his own contacts had been strongest with the constitutionalist faction, he sent this message in the hands of a young nephew, Wang Wenhua (Wang Boqun's brother), who had become close to several members of the Self-Government Study Society while studying in Guiyang and was known to have revolutionary sympathies.^[78]

Neither Liu's message nor his messenger was sufficient to allay the doubts

of some revolutionaries over his supposed change of heart. Zhang Bailin called a meeting of revolutionary leaders to discuss Liu's offer; some felt that Liu should be ordered to return to Xingyi with his men, and others argued that the only safe course would be to arrest and execute him. In the end, though, the decisive factor in this issue was Zhang Bailin's conviction that the strength of the new regime required healing the antagonism between Guizhou's two reformist factions. In pursuit of this goal, several prominent constitutionalists, including Ren Kecheng, had been offered posts in the revolutionary cabinet. These men supported Liu Xianshi's entry into Guiyang and argued that his military experience could be useful to the new regime. As a sign of conciliation, Zhang overrode the objections of his colleagues and welcomed Liu Xianshi into Guiyang.^[79]

Liu Xianshi's opportunistic jump from counterrevolutionary to revolutionary ranks paid off handsomely. Upon arriving in Guiyang, his unarmed band of five hundred men received the new guns originally promised by Governor Shen. Liu's men were then recognized as a regiment (biao ) in Guizhou's new revolutionary army, and Liu himself was allowed to retain his command. Finally, Liu was also given a place on the revolutionary cabinet and was appointed head of the new government's military affairs section.^[80] Thus Liu was allowed to combine actual military command with an active role in the regime's military administration and policy making.

Zhang Bailin's attempts to reconcile Guizhou's competing elite factions proved a failure. After his entry into Guiyang, Liu Xianshi became the constitutionalist faction's main military ally in an ongoing and increasingly bitter struggle with revolutionary leaders for control of the provincial government. The revolutionaries were weakened by the departure of Yang Jincheng, with most of Guizhou's best troops, for the revolutionary front in central China. On February 2, 1912, Liu and his allies incited a dissatisfied military commander to lead an attack on revolutionary leaders in Guiyang. Liu's main military opponent, the head of a revolutionary force organized from secret society followers, was killed, and Zhang Bailin was forced to flee the city. When Zhang attempted to gather a military following in west Guizhou to fight his way back to Guiyang, Liu ordered his cousin Liu Xianqian, who had returned to Xingyi after Guangxi joined the Revolution, to lead a militia unit from Xingyi to attack Zhang's rear. Zhang was defeated and forced to flee the province.^[81] Even at this point, however, revolutionary forces in Guiyang remained sufficiently powerful to prevent Liu from eliminating them. To achieve this end, Liu and his allies sought military assistance from the Yunnan army, an action in which Liu family connections again played an important role.

From the beginning of the 1911 Revolution, Liu Xianzhi, from his position on the governor-general's staff in Yunnan, had been alert to the Revolution's opportunities for extending his family's power. When news of the

Hubei uprising reached Yunnan, he had urged the governor-general to have Liu Xianshi lead a force from Xingyi to Kunming to help defend the city from revolutionaries. Before anything could come of this, however, Yunnan's new army had successfully seized Kunming in the name of the revolution. Despite his original opposition to the revolution, Liu Xianzhi was protected by Cai E, Yunnan's new military governor, because of Liu's strong ties to Cai's mentor, Liang Qichao. On the basis of this connection, Liu was even given a staff position in Yunnan's new government. Unaware that Liu Xianshi had already set out for Guiyang at the request of Guizhou's loyalist governor, Liu Xianzhi then proposed that a detachment of Yunnan troops be despatched to join up with Liu's troops in Xingyi in order to march on Guiyang in the name of the revolution. This plan was also abandoned once the news was received of the successful uprising of Guizhou's New Army.^[82] After Guizhou's revolutionary government was established, Liu Xianzhi, along with other Guizhou men connected to the Liu family on Cai's staff, became a secret channel of communication between Liu Xianshi and the Yunnan government.^[83] Following long negotiations, Cai was finally convinced to send a military force led by Tang Jiyao to seize control of the Guizhou government.

On February 29, 1912, Tang's forces arrived in Guiyang where they were welcomed by the revolutionary government, which had been deceived into believing they were simply passing through to relieve other revolutionary forces in central China. On March 2, Tang Jiyao joined with Lu Xianshi to carry out a military coup. Revolutionary military forces were surrounded and disbanded or in some cases disarmed and massacred. The military governor who had replaced Yang Jincheng was forced to flee. Other revolutionary leaders who still held military or civil posts were killed or forced to flee the province.^[84] Thus, with Tang's aid, Liu and his allies were finally able to defeat their revolutionary rivals.

The price for Tang's military assistance was his assuming Guizhou's military governorship. At the same time, Tang recognized that he required an internal base of support to rule in Guizhou. Thus, Liu Xianshi's civil allies were allowed to control the province's civil administration. Liu himself was rewarded with the post of provincial minister of war, placing him even more firmly in administrative control of Guizhou's military forces. Besides his own original unit, which he was now allowed to expand, Liu was also placed in command of a new Citizens' Army consisting of reorganized patrol and defense force units. Certainly, the Yunnan occupation to some extent limited Liu's military power. Nonetheless, with Tang's assistance, Liu completed the elimination of revolutionary or otherwise unreliable military forces, effectively removing all his Guizhou military rivals.^[85] Although Liu's army remained small in relation to Tang's army, Liu had become, by default, Guizhou's most important military leader.

The final reward from Liu Xianshi's Yunnan alliance came in October 1913 when Tang Jiyao returned to Yunnan to succeed Cai E as Yunnan's military governor. Having proven himself a useful ally, Liu was allowed to take over Guizhou's military governorship. Liu quickly undertook a large-scale expansion of his army to insure his military control of the province. As was often the case with Republican military governors, Liu also dominated Guizhou's civil administration, using the same civil elites who had been his political allies since 1911. Finally in 1916 he ended any pretense of a division between civil and military powers and directly assumed the post of civil governor. By establishing his control over Guizhou's military and civil administrations, Liu Xianshi joined the ranks of China's Republican warlords.

The political upheaval of the 1911 Revolution provided an ambitious Liu Xianshi with the opportunity to transform local into provincial power. There was no question that his rise to become Guizhou's preeminent warlord would have been impossible without his family's local military base and his experience as a local military commander. At the same time, this local military power alone does not explain Liu's success. The crucial points of Liu's advance—Governor Shen's request for Liu's military assistance, Liu's admission into Guizhou's revolutionary government, and the procurement of Yunnan's military assistance—all depended not so much on Liu's own military power as on the broader network of social and political influence established by Liu Xianshi and other family members in the decade before the 1911 Revolution. The civil as much as the military aspects of Liu family power helped Liu Xianshi exceed the limits of his family's original local power base.

The Liu Family in the Republican Period

After Liu Xianshi's 1911 march from Xingyi county to Guiyang he no longer functioned as a member of the local elite. From this point on, Liu Xianshi's attention was not focused on Xingyi's local affairs but on the broader struggle for provincial and national power. Only in 1925 did he finally retire to Xingyi after being forced from power by a military rival, and less than two years later he was dead. Although Liu's career as a warlord is beyond the scope of this study, it is perhaps fitting to look briefly at the effects of Liu's political success on his family and its Xingyi power base.

One of the most obvious effects of Liu Xianshi's rise was the opportunity it gave him to obtain official posts for other family members. The loss of effective central control over local and provincial appointments after the 1911 Revolution opened the way for a degree of nepotism by Republican period power holders that would have been unimaginable under the imperial system. Liu Xianzhi, on his brother's insistence, was "elected" as a Guizhou national assemblyman and served for many years as his brother's political

representative in Beijing.^[86] Liu Xianqian held a series of high Guizhou military posts as well as important civil positions including a circuit intendancy and even, for a time, Guizhou's civil governorship.^[87] Wang Wenhua, the nephew who acted as Liu's intermediary during the 1911 Revolution, rose under Liu's patronage to the position of division commander in the Guizhou army.^[88] Wang's brother-in-law, He Yingqin, also held a number of posts in the Guizhou army including brigade commander and chief of staff.^[89] Wang Wenhua's older brother, Wang Boqun, held a circuit intendancy in Guizhou after the 1911 Revolution and then served as Liu's official emissary on a number of political missions.^[90]

Although Liu Xianshi's patronage was essential for the initial posts received by his relatives, many of them used these positions as stepping stones for official careers independent of Liu's direct influence. Liu Xianshi's son, Liu Gangwu, made contacts within the Nationalist Party while serving as his father's emissary to Sun Yatsen, and, after his father's death, obtained a series of posts in the Nationalist government.^[91] After leading a Guizhou expeditionary force into Sichuan, Wang Wenhua emerged as an autonomous warlord in his own right. He Yingqin eventually left the Guizhou army to take a post as an instructor at the Whampoa Academy in Canton and then became a leading Nationalist general. As a result of extensive social and political contacts made early in his career, Wang Boqun obtained many posts from the Nationalist government in the 1920s, culminating in an appointment as minister of communications.^[92] It is highly unlikely, however, that any of these men would have achieved these positions without the original impetus of their relationship to Liu Xianshi.

In traditional Chinese society, elite families saw official posts as a means to replenish family wealth and status. Under the imperial system, however, the direct combination of official and local family power was prevented by the "law of avoidance" that kept an official from holding an office in his home province. This principle was abandoned after the 1911 Revolution, as exemplified in Liu's own rise to provincial power, and this change added an extra dimension to the benefits of officeholding in the Republican period. In the late Qing, the Lius had to apply much money and effort to cultivate relationships with provincial officials to protect their local position. Once Liu Xianshi assumed control over provincial administration, however, his family's local interests were not only freed from official interference but also placed under official protection. Indeed, to watch over the family's power base, Liu Xianqian often held west Guizhou military or civil posts, which gave him direct authority over the Xingyi area.^[93] While Liu Xianshi remained in power, his family's dominant position in Xingyi was unchallengeable.

The manner in which Liu Xianshi's official position could be used to protect local family interests, however, suggests a subtle change in the importance of the Lius' local base to their family power. In the late Qing, the Lius'

Xingyi base was the family's main source of both military and economic power. After 1911, more important sources of power came from outside Xingyi. For example, the family's military power originally came solely from its Xingyi-recruited forces. In the Republican period, family members in Xingyi still maintained their local military power, at times using this power to bolster the family's broader political interests; for instance, Liu Xianqian's 1912 attack on Zhang Bailin with Xingyi militia forces. Another example can be seen in 1923 when Liu Xianshi recovered Guizhou's governorship with Yunnan assistance after a two-year fall from power: a local force from Xingyi, led by one of Liu's nephews, joined the vanguard of Yunnan troops escorting Liu back to Guiyang.^[94] Nonetheless, while Liu Xianshi ruled as governor he had the military resources of the whole province at his disposal. In this period, this greater military power, not the family's local militia, contributed more to the family's overall position.

A similar change in the sources of Liu family wealth can also be seen in the Republic period. By 1911 the Liu family was already one of the wealthiest landlord families in southwestern Guizhou. In the Republican period, however, the family's total landholdings exceeded four thousand mu , making the Lius possibly the largest landlord family in all of Guizhou. Given the extravagant life-style of the Liu family in Xingyi, reported in the Republican period, it is unlikely that this leap in landed wealth came solely from the frugal reinvestment of the profits from previous holdings.^[95] Rather, it suggests new external sources of wealth that no doubt also derived largely from office holding. In the Republic, as under the empire, official wealth came not from salaries, which remained quite low, but from graft. For warlords like Liu Xianshi, authorities unto themselves, the opportunities for personal profit were in general limited only by the size of the public treasury. Although records on this type of graft were seldom kept or made public, some reports reveal the sums the Lius were able to extract from their public offices: During a 1916 political crisis, Liu Xianshi sent more than four hundred thousand yuan , reportedly taken from the Guizhou treasury, to Shanghai to serve as a personal emergency fund in case he lost power.^[96] In two separate cases, Liu Xianzhi and Liu Xianqian were each accused of embezzling two hundred thousand yuan in public funds.^[97] Even granting some bias in these reports, they suggest the amount of wealth the Lius were able to acquire from their positions. There is no question that much of his wealth eventually found it way back to Xingyi, where it was invested in the family's growing estates.

Liu Xianshi's fall from power in 1925 did have some impact on the Lius' local power base in Xingyi because it removed the political shield of invulnerability that his position had provided for family interests. There had been no such danger to the family's Xingyi base in his earlier loss of power from late 1920 to 1923 because in that case he had been ousted by troops loyal to his nephew, Wang Wenhua, in what was to some extent an intrafamily

conflict.^[98] In 1925, however, Liu was forced out by an unfriendly rival who decided to establish a military presence in Xingyi county under the guise of a bandit pacification campaign. The Liu family tried to resist this intrusion into their base with disastrous results. The Xiawutun fortress fell to the invading army, and the Lius' military forces were disarmed and disbanded. A large store of weapons the Lius had hoarded at Xiawutun (enough to arm two regiments) was also confiscated, along with a large quantity of the family's accumulated "loot." ^[99] Liu Xianshi's 1925 fall thus meant the decline of not only the family's provincial political influence but also its local base.

In the end, however, the Liu family landholdings were so extensive and its social status so high that even the 1925 disaster was only a temporary setback that could not undermine the family's dominant position in Xingyi county. Of course, some members of Liu family, such as Liu Gangwu (who never returned home after his father's death), had official careers that drew them more or less permanently away from the Xingyi area. But other members of the family remained in the county to husband the family's interests and continue the family's tradition of local leadership. For example, several of Liu Xianshi's nephews later held district head (quzhang ) posts in Xingyi and thus maintained the family's influence over local administration. In the 1930s and 1940s, Liu family leadership in Xingyi was primarily assumed by one of Liu Xianshi's nephews, Liu Gongliang, who had become Liu Xianshi's trusted secretary in his last years as governor and had followed his uncle back to Xingyi after his fall. There he upheld family traditions by his involvement in many local affairs, including a special emphasis on local defense. During the Red Army's passage through the area in 1934, Liu Gongliang helped organize an anticommunist defense committee, directed the construction of fortifications throughout the county, collected arms and funds for local defense, and trained local militiamen. Much like descriptions of the Liu family's "small court" in the late Qing, Liu Gongliang came to be seen as Xingyi county's "de facto magistrate."^[100] Only the coming of the Communist revolution in 1949 ended the Liu family's century of local power.

Conclusion

The history of the Liu family of Xingyi illustrates the manner in which local military force could function in Chinese society, forming and enhancing local-elite power. Through the organization of militia in the mid-nineteenth century the Lius rose from the ranks of village landlords to become a dominant elite family in southwestern Guizhou. Likewise, the family's local military power later provided the base for Liu Xianshi's emergence as provincial warlord. Nonetheless, the Lius as local elites were never simply military strongmen. Indeed, in the last quarter of the nineteenth century the Lius gave up their militia leadership with no appreciable loss in elite status. In-

stead the family concentrated on enhancing their social position by more traditionally approved forms of elite behavior, notably educational patronage and network building. Although they revived a local military base at the turn of the century, in the following decade they also 'sought to perpetuate and expand their local power by introducing reform programs into the Xingyi area. The long-term maintenance of the Lius' position as local elites ultimately depended on both military and civil activities.

The case of the Liu family suggests that the real key to understanding the role of military power in local elite formation is to identify the context in which it was applied. In the mid-nineteenth century, militia leadership served as the agency for the Liu family's remarkable rise to a position of local dominance, but the condition of widespread rebellion and social disorder ultimately made this rise possible. Likewise only the special conditions of the 1911 Revolution in Guizhou gave Liu Xianshi the chance to parlay his family's local military power into provincial military domination. For the Lius, then, extraordinary times had presented extraordinary opportunities to benefit from local military power; in peaceful times the Lius switched just as easily to civil means of enhancing their social position. The Lius' particular success lay in the ability to adjust their social strategies to new conditions and to employ their military and civil resources in ways that took best advantage of changing circumstances.