PART FOUR
VILLAGE ELITES AND REVOLUTION

Nine
Corporate Property and ocal Leadership in the Pearl River Delta, 1898-1941

Rubie S. Watson

Scholars of late imperial China have long debated the relative openness of Chinese society. In the 1950s and 1960s these debates focused on the nature and extent of elite continuity and tended to emphasize questions of personnel rather than structure. Few would disagree with the view that later imperial society was divided into status groups and classes. At issue has been the movement between these groups. For many scholars, China provides an example of an agrarian society with an open, highly fluid system in which families entered and left the elite with considerable regularity. According to this view the division between China's late imperial elite and the ordinary population was guarded not by a closed door but by a revolving one.

In recent years, studies of elite mobility rates have been combined with a concern for the institutions, structures, and mechanisms that affect those rates. The advocates of openness maintain that China's competitive examination system and the custom of equal inheritance among sons made it difficult for a family to retain its status for more than two or three generations.^[1] Because access to the bureaucracy and the economic benefits of office holding were determined by examinations, scholars have argued that families could not easily pass their status on to their sons and grandsons.

These arguments for an open elite have been criticized by other scholars who maintain that the mobility advocates have predetermined their results by their definitions and methodology.^[2] Early mobility studies, they contend, focused on the office-holding elite and measured mobility by the number of higher-degree holders whose fathers, paternal grandfathers, and paternal great-grandfathers all lacked examination degrees.^[3] Critics have argued that imperial degrees and offices are too narrow criteria on which to build a model of social stratification. As long as elite status is defined solely in terms of office and degree holding, and as long as the measurement of mobility is restricted

to three generations of direct descent in the patriline, it is statistically quite easy to demonstrate high rates of mobility, the critics contend. Robert Hymes and Robert Hartwell, it should be noted, have found that mobility rates are significantly lower if the measurement of mobility includes collaterals and affines in the universe of potential office-holding relatives.^[4]

Elite status, many of the critics argue, is not simply a matter of degree holding; more attention, they urge, should be given to the interlinkages of wealth, lineage organization, local leadership, and scholarly achievement. Recent studies, which have tended to focus on one particular locale, have found considerable continuity of the local elite, with the patrilineal descent group providing a powerful bulwark against downward mobility.^[5] Hilary Beattie's work on Tongcheng county, Anhui, during Ming and Qing times is an excellent example of this approach. In her book, Land and Lineage in China , Beattie argues that private landowning, cohesive lineage organization, and joint property constituted a powerful set of safeguards that protected families from the four-generation rags-to-riches-to-rags scenario so prevalent in the Chinese folk tradition and the writings of mobility scholars.^[6] In a study of status in the Song, Linda Walton summarizes the critics' position:

A new picture has begun to emerge in which the elite of late imperial China can be seen as a large group of lineages who prepared candidates for the examinations and provided office holders for the state, but who achieved, protected, and enhanced their status locally through a variety of social and economic means, such as marriage alliances with other elite families, and the establishment of the institutions of joint property.^[7]

Scholars who argue for high rates of mobility point to the examination system as the institution that kept the elite open to new blood, while partible inheritance produced unavoidable pressures for downward mobility of older elites. There is, of course, much merit in the view that partible inheritance directly effected rates of social mobility in late imperial China,^[8] and I have no particular quarrel with this argument. However, there is a danger that explanations of this sort may make us think that we understand more than we do. In fact, the asserted link between partible inheritance and high rates of mobility remains largely unexplored in China. In Europe, for example, equal inheritance did not inevitably lead to a highly fluid pattern of stratification. Late marriage, nonmarriage, status endogamy, cousin (or "close" kin) marriage, and high rates of infant mortality contributed to retaining wealth under regimes of equal inheritance.^[9] Perhaps China specialists have too easily accepted the dogma that equal inheritance leads to an open social system.

The landlord-merchant elite described in this chapter was embedded within a large, highly-segmented lineage. Unlike some members of China's late imperial and Republican elite, the group discussed here enjoyed considerable staying power. For more than 250 years a single patriline (the

descendants of a mid-eighteenth-century ancestor) dominated political and economic affairs in the village of Ha Tsuen and its hinterland. I explain how it was possible for a single line of agnates to maintain a position of wealth and power in a society known for its downward mobility.

In this essay control over both the lineage's organizational machinery and corporate property are seen as mainstays in the arsenal of formal and informal supports that made elite domination possible. I argue that a high ratio of corporate to private property is likely to affect local political organization, the ways in which elite and nonelite are linked, and the continuity of local elites. Half or more of the land belonging to the residents of Ha Tsuen was not subject to partible inheritance but rather was tied up in indivisible corporate estates. The role that corporate lineage estates played in diminishing the effects of equal inheritance and in creating political capital for the people who controlled those estates is examined in detail. In Europe, late marriage and high rates of celibacy appear to have reduced the pressures of downward mobility in at least some areas where partible inheritance was the norm. Is it not possible that corporate estates may have a similar effect on elite continuity in China?

In this article I enlarge on themes developed in previous work and situate the Deng lineage of Ha Tsuen in a larger regional context.^[10] By placing the Ha Tsuen material in a comparative perspective it is possible to clarify at least some of the links between corporate property and class structure. Whether the patterns described here fit the Tongcheng model so ably described by Beattie remains to be seen.^[11] There were, to be sure, some significant differences between the lineages of Guangdong and those Beattie describes for Anhui. The question of regional variation will be taken up in a later section of this chapter.

Setting: Ha Tsuen and Its Elite

This article covers the period from the British takeover of Hong Kong's New Territories in 1898 to the Japanese occupation in 1941. Changes in village politics in the postwar period are detailed elsewhere.^[12] This discussion is based on data collected in the village of Ha Tsuen, located in the northwestern sector of the New Territories. Prior to 1898 this area was part of Xin'an county in Guangdong province. At the time of my fieldwork in 1977-78, the village had a population of twenty-five hundred; all were Cantonese speakers. The males of Ha Tsuen share the surname Deng and trace their descent from Deng Fuxie, an official who settled in the Hong Kong region during the twelfth century. The Deng controlled a hinterland consisting of an administrative subdistrict, or xiang ,^[13] containing fourteen satellite villages inhabited by dependent (non-Deng) tenants. Until the early 1960s, most villagers were engaged in double-crop rice production. By local standards the village was a

powerful and wealthy community; it was dominated by a small group of landlords and merchants who until recently lived off the rents they collected from their agnatic kinsmen and unrelated tenants. These landlords lived in Ha Tsuen but had extensive interests in shops, businesses, and factories outside their home village.

The landlord-merchants described here rank well below the national elite of Qing or Republican times. The China historian is unlikely to have seen much of this local elite in the archival record. Ha Tsuen never boasted a jinshi or juren degree holder, although their agnatic kinsmen in Kam Tin (six miles from Ha Tsuen) did produce an occasional scholar-official. The Ha Tsuen Deng seem never to have been very concerned with education. A few men sought the trappings of literati status, but their efforts were never crowned with any real success.

These people did not travel in the power circles of Guangzhou or Beijing. However, their agnatic ties linked them to wealthy and powerful men from other Deng communities in Xin'an and neighboring Dongguan counties. In 1709, five Deng settlements joined to establish a lineage hall (Duqing tang ) in Dongguan City. Members of the Duqing tang were bound by their common descent from their twelfth-century founder, Deng Fuxie. The guiding force behind this hall was a Kam Tin Deng named Deng Paosheng, a jinshi , but we know little about Ha Tsuen's involvement in this higher-order-lineage except that they were members of the Duqing tang and did, according to my Ha Tsuen informants, take their turn at organizing the ritual sacrifices commemorating the founding ancestors.^[14] From interviews and genealogies I know that the Ha Tsuen and Kam Tin Deng shared a particularly close relationship. Until the Second World War they had mutual responsibility for the care of graves belonging to a number of shared ancestors. Ha Tsuen was in fact an offshoot of Karo Tin; Ha Tsuen's founders had moved from Kam Tin in the ninth generation (in the mid-fourteenth century). The Ha Tsuen Deng are proud of their close links to Kam Tin whose elite have for generations been considered the cream of local society.

Beyond their connections to agnates in the region, the Ha Tsuen elite's extensive economic interests also took them out of their village into the xiang , the standard market town of Yuen Long, and the intermediate marketing networks of the region. Some of these contacts involved kin; others did not. In the eighteenth and nineteenth century Ha Tsuen's elite owned and operated cargo boats that serviced parts of the Pearl River delta, Western District on Hong Kong Island, and the coastal area as far north as Swatow. In the latter half of the nineteenth century two brothers, members of Ha Tsuen's wealthiest line, established a pawn shop and match factory in Yuen Long Old Market, the standard market town that serves Ha Tsuen. Members of this line were also the original shareholders in Yuen Long's new market (established in 1916), which eventually eclipsed the older market. Beginning in the mid-nineteenth century men of this patriline built or expanded a

sugar-processing factory in Ha Tsuen and in the 1880s added a peanut oil factory.

That the marriage networks of this elite should mirror their expansionist economic outlook is not surprising. According to a survey to twenty elite women covering the period from about 1900 to 1978, all wives who married into this line came from beyond the immediate xiang , some from as far away as urban Hong Kong and the county administrative center at Nantou.^[15] For Ha Tsuen's landlord-merchants affinal ties were an important resource. Their mercantile activities and their role as political brokers depended on an extensive interpersonal network. The members of Ha Tsuen's local elite were never simply lineage or village personalities; their influence depended on extravillage ties as well as a solid economic and political foundation at home. As I have argued elsewhere,^[16] there is little doubt that affinity and ties to maternal kin helped to create and sustain the external relations that were so important to maintaining elite status.

No single factor can account for the privileged position of the local elite described here. Their ownership of important resources was certainly crucial, but their roles as merchants, political intermediaries, and managers of lineage institutions were also significant. Chinese rural elites were sustained by various institutions, many of which had little to do with everyday politics. In this chapter I examine in detail the relationship between control over lineage (corporate) resources and elite power.

Corporate Property

The list of corporate property types in China is extensive. There was, for example, temple land, temple association land, voluntary association estates, lineage land, and lineage segment land.^[17] Under these last two categories there were many subtypes: for example, charitable estates (yizhuang ), scholar estates, service land (yitian ), ritual land (jitian ), and ancestral estates (zu, tang ).^[18] Estates were formed by pooling resources or reserving the private property of a deceased ancestor (turning a private holding into a corporate holding). Here I focus on ancestral estates. The rents from these estates provide offerings for the estate's focal ancestor; leftover income is shared on a per capita basis among the estate's membership (i.e., the descendants of the focal ancestor or the descendants of those who formed the estate).

Clan or lineage estates in China are often seen as a form of charity in which benefits are provided for needy kin. The famous Fan estate founded by Fan Zhongyan is an example of a charitable estate, or yizhuang .^[19] That there is a benevolent, charitable aspect to many lineage estates is undeniable, but in the area I studied ancestral estates (zu ) are the primary form of corporate property, and they have little to do with charity. They were religious, economic, and political institutions. Here I stress their political role.

In 1905 the first land survey of the New Territories shows that about half

of all Deng-owned land belonged to corporate ancestral estates. Corporate landholdings were extensive and central to the local economy. When one asks villagers why they tie up so much of their valuable property in corporate estates, they do not answer by pointing to their charitable natures or the needs of their kin. Rather they say that they establish estates in order "to honor their ancestors" (a filial rather than a charitable act). No doubt their ancestors are indeed honored by these estates; zu rents guarantee that proper offerings and proper rites are carried out, giving their ancestors a kind of immortality. However, the formation of ancestral estates, especially large ones, also had the very practical effect of keeping land out of the cycle of family division and fragmentation that was so common in rural China. Estate land is inalienable, and those who share in it (the descendants of the ancestor) remain a group in both an economic and a political sense. As the number of descendants grows over the generations, the economic utility of the estate may decline, but its political significance may in fact increase.

Thus, in a society that practices equal inheritance among brothers, the formation of corporate ancestral estates is one strategy for avoiding or minimizing the disintegration of family property. I do not mean to imply that ancestral estates solve the problems of population pressure on the land; obviously this is not the case. Given a limited technology, the more people who have rights to the income from a piece of land, the less income each shareholder receives. However, in Ha Tsuen I found that members of well-endowed estates rarely shared equally in estate property. Some estate members controlled a disproportionate share of estate income and assets; some also monopolized the estate's political benefits.

Parallels with the use of "entail" among the English elite are striking. (See also Rowe in this volume.) From the mid-seventeenth to the nineteenth century entail was commonly employed as a strategy to preserve property for future generations.^[20] There are of course differences between entail and ancestral estate property: the zu is established in perpetuity, the entail for only a few generations (usually three). However, both were part of the strategic repertoire that protected property from sale, division, and fragmentation. Comparisons of entail and zu are particularly useful because they take the latter out of the lofty realm of the ancestral cult and place it on the more prosaic terrain of property holding, inheritance, and elite continuity.

Extent of Corporate Property in Ha Tsuen

According to the 1905 land records, 377.64 acres of Deng-owned land were tied up in corporate estates. Of this total, 369.69 acres were owned by ancestral estates (zu, tang ) and the remaining 7.95 acres belonged to hamlet or religious associations. According to my estimates, about 50 percent of land located in Ha Tsuen xiang and owned by the Ha Tsuen Deng was incorporated into ancestral estates. This figure is comparable to other estimates of

corporate property ownership in the area. For example, a survey of the standard marketing region of which Ha Tsuen is a part found that 44 percent of the land was "lineage owned in 1905."^[21] Two dominant lineage villages in the New Territories have a similar proportion of corporate to private property. Baker gives a figure of 52 percent corporate ownership for the Liao lineage of Sheung Shui,^[22] and J. Watson estimates a rate of 65 percent for the Man lineage of San Tin.^[23] All these figures, however, are much lower than Potter's reports for Ha Tsuen's near neighbor, Ping Shah. According to Potter, one of Ping Shan's eight hamlets (Hang Mei) registered a staggering corporate rate of 93 percent in 1960.^[24] Perhaps one reason Potter's figure is so much higher than those reported for neighboring lineages is that he based his calculation on the wealthiest hamlet in Ping Shah, the one that contains nearly all the community's large ancestral halls and many of its richest residents. (In the land records ancestral estates are listed under the hamlet where their managers live.) In surveys made of Guangdong province in the 1930s, Chen Han-seng calculates that what he calls "clan land" (presumably referring to lineage estates) made up about 35 percent of the total cultivated land.^[25] If only the Pearl River delta is considered, Chen's estimate climbs to one-half. Compared to these figures, Ha Tsuen is not unique in its percentage of estate land.

There is great variation in the size of Ha Tsuen's ancestral estates;^[26] the largest estate (Gouyue zu ) owns 40.96 acres, and the smallest has a miniscule 0.02 acres of poor hill land (table 9.1). Potter found a similar but more glaring pattern of difference in Ping Shah where 37 percent of the estates he surveyed (N30) owned less than 1 acre, and 50 percent owned from 1.1 to 15 acres. The largest owned 178.4 acres of land, an enormous holding for the area.^[27] To keep landholding in perspective, two facts are helpful: the largest private landholding in Ha Tsuen in 1905 was 71 acres, and about 1.5 acres of good paddy land supported a household in the period under discussion here.^[28]

Because a significant proportion of property in lineage-dominated areas was corporately owned, ancestral estates were important sources of rental land. According to my estimates, in Ha Tsuen in 1905 the landless rate was about 55 percent, and 97 percent of Deng households rented some of the land they farmed. Of those households that did own land, 84 percent had holdings of less than one acre.^[29] No doubt a considerable amount of valuable land was tied up in ancestral estates, which were often cast in the role of landlord. In Ha Tsuen some villagers said that their rents were reduced if they were tenants of their own zu , while others claimed that they paid the same rent as nonmembers. For the Pearl River delta village of Nanching, C. K. Yang reports that there was "no practical difference in the type of tenancy and the amount of rent charged" between private and ancestral land.^[30] Although there were no general restrictions on renting estate land in Ha Tsuen, some villagers claimed that members tended to have an ad-

vantage over nonmembers in competing for tenancy rights. There appear to be, however, no significant distinctions in terms of rental arrangement between lineage and private holdings.

It is clear that some estates produced large rents. However, few Deng could afford to live off the proceeds of ancestral estates without some other source of income. In 1905 the largest landowning estate in Ha Tsuen, Gouyue zu , had nearly 41 acres (see table 9.1) and in 1978 a membership of more than eight hundred men. During recent generations, rent from this estate produced a surplus after expenses that was sufficient to pay only a token cash dividend to members. In 1905, of course, a few estates produced large amounts of rent shared among a small membership. These estates were all recent in origin; two of the largest estates, with a combined holding of 51.34 acres, were formed in the 1880s and had only a few members. Until recently the yearly dividends from these estates were considerable, but it is unlikely that members lived solely on estate proceeds, for in 1905 these men. were among Ha Tsuen's wealthiest landlord-merchants.

The Benefits of Corporate Property: Estate Managers

In two recent articles concerning charitable estates in nineteenth-century Wuxi, Dennerline discusses the role that yizhuang played in political life.^[31] Although Dennerline is not primarily concerned to detail the exact processes

that connect yizhuang to political leadership, his discussion does make clear that local power and its exercise were inextricably bound to the control of charitable estates. Dennerline is no doubt correct in stressing the communitywide benefits of yizhuang , but he also states: "Once an estate was established, to control the management of it was a major political objective." He concludes that "charitable estates were ... tools by which the most powerful men in rural society exercised both their own influence within the lineage and the lineage's influence within the community."^[32]

The estates that Dennerline describes are not the same as those I found in Hong Kong's New Territories.^[33] Wuxi's charitable estates had an important welfare role (certified by the state), an insignificant aspect in the case of Ha Tsuen and, to my knowledge, other estates in the Pearl River delta. I place special emphasis on the unequal way in which corporate benefits were apportioned and the advantages estates gave to those who controlled and managed them. I do not deny that corporate estates offered important benefits to poor agnates; I do argue, however, that the wealthy benefited far more than did Ha Tsuen's smallholder-tenants.

I have already noted that ancestral estates make it possible to maintain a sense of group unity and purpose often endangered by the normal cycle of family division. Figure 9.1 represents the pattern of lineage segmentation among the Ha Tsuen Deng in 1905, and Table 9.1 provides a list of all Deng estates with more than ten acres of first- and second-class land. Ha Tsuen's landlord-merchants, it should be noted, were descended from ancestor Jingwu (generation 14). Taken together, Figure 9.1 and Table 9.1 make it clear that nearly all the largest ancestral estates in Ha Tsuen were clustered in the Jingwu line. Not only were Ha Tsuen's landlord-merchants members of the lineage's wealthiest estates, but they also managed these estates.

Ordinary villagers may have recevied a few dollars each year from their share of corporate holdings, but they were nearly always alienated from active control over the large estates. The real benefits were firmly in the hands of the wealthy, who served as estate managers (sili ). Managers handled, and continue to handle, the business affairs of the estates. They also act as the estates' legal representatives; in the words of a 1910 ordinance, the manager acts "as if he were the sole owner ... [of the estate land], subject to the consent of the Land Officer."^[34] In Ha Tsuen, once selected, managers usually retained the position for life. Often managers were the sons of previous managers, a pattern that is still repeated. Even today there is no formal procedure for choosing a manager, although the colonial administration has established guidelines for performing managerial duties. Nearly all the men who became managers, especially of the large estates, were members of the local elite. Potter notes a similar pattern in Ping Shan, "The last surviving gentry member of Hang Mei village, who died in 1920, was said to have been the manager of all important ancestral land estates in the village."^[35] In 1977-78

Source: Rubie S. Watson, Inequality Among Brothers, 46.

Figure 9.1
Outline of Major Segments of the Deng Lineage

managers of Ha Tsuen's largest estates were political leaders or wealthy men; often they were both. Among the Deng, managers have always been members of the estates they control.

The usual requirements for becoming an estate manager, especially of a large estate, were literacy, knowledge of the world outside the village, and wealth. Personal wealth, it was believed, ensured a minimum of honesty because the manager's private property could be confiscated if he strayed too far from acceptable behavior.^[36] The importance of wealth in choosing estate managers also ensured that a few men, mostly landlords and merchants, monopolized this source of local power.

Once a wealthy man became a manager he increased his opportunities for making money and acquiring influence.^[37] In Ha Tsuen, managers were especially powerful because they served life terms.^[38] Potter notes that in Ping Shana manager could take advantage of his position by loaning cash from the estate fund. At the end of the year when the accounts were due, he collected and put the principal back in the estate coffers and kept the interest for himself.^[39] Ha Tsuen residents report that managers often received "gifts" of money from prospective tenants or lessees of estate property who wanted to secure the tenancy of badly needed agricultural land. According to villagers, estate managers often acted toward tenants much like private landlords. At harvest time managers measured and determined the rent just as private landlords did, and, to the disgust of the villagers, they too used nonstandardized baskets (douzhong ) to measure the grain rents, keeping the extra for themselves.

It is no secret that on occasion the community of Ha Tsuen, and some of its ancestral estates in particular, have been badly served by their managerial elite. Many Deng believe that Ha Tsuen lost valuable property and suffered a general political decline through the corruption or incompetence of past leaders and estate managers. In particular, villagers point to the eclipse of their local market, darkly suggesting that at best their leaders were outmaneuvered by a group of outsiders and government officials or at worst their acquiescence was for sale.^[40]

Nearly seventy-five years after the event villagers are still upset by an area of marshland in front of the village that was usurped and leased by the colonial government to a consortium of overseas Chinese developers. The marsh-land reclamation, which created about five hundred acres of land, began in 1915 and set in motion a series of ecological changes that are still causing the Deng serious problems. Most notably and immediately affected was the narrow channel through which cargo boats serviced Ha Tsuen Market. Soon after the reclamation began, this channel became hopelessly silted and had to be abandoned, thus hastening, according to the Deng, the decline of their market. A number of villagers believe that the manager of the Yougong tang (the tang with responsibility for administering general village affairs, see be-

low) either refused to make a case for Ha Tsuen's rights over this marshland and/or accepted a payoff from the developers to let their arrangement with the government proceed unchallenged.

Ha Tsuen villagers also continue to feel bitter over the perceived betrayal by the manager of the Duqing tang , the estate of the Deng higher-order lineage. In 1892, in anticipation of commercial development in the area, the Duqing tang obtained rights (in the form of a reclamation certificate) from the Qing government to about 117 acres of coastal land near the present-day city of Kowloon. In 1894, before improvements were made, a Kam Tin Deng, a juren and manager of the Duqing tang , claimed personal possession of the certificate by taking it on lease (in perpetuity) apparently without the knowledge of other Duqing tang branch members. Palmer reports in his study of New Territories land tenure that after serving as both lessor and lessee in this transaction, the manager then proceeded to inform the five member branches, including Ha Tsuen and Karo Tin, of the new arrangement. He confessed that he had mortgaged the certificate to Fuk Tin (Futian) Company of Hong Kong, a commercial land enterprise owned by Li Sheng, a native of Xinhui county and a member of Hong Kong's trading and manufacturing elite.^[41]

The Deng eventually brought this case before the new Colonial Land Court which, to their great anger, decided in favor of Fuk Tin Company's rights to the disputed land. The court took the view that the Duqing tang had made no improvements in the land while it was in its possession and therefore had no right to the reclamation certificate. Villagers in Ha Tsuen still smart from this loss in which, once again, the local elite was hoodwinked out of their rights by what they consider the duplicity of a corrupt manager and the greed of a new class of commercial land developers. It is certainly worth noting that both these cases involve the victory of urban entrepreneurs over a traditional elite, two groups that for much of Hong Kong's history have remained highly suspicious of each other.^[42]

The 1905 land records include the names of the managers of each of Ha Tsuen's ancestral estates. An examination of these records shows that one man, landlord-merchant Deng Zhengming (a pseudonym), managed eight estates; during his lifetime he controlled, as either private owner or estate manager, a total of 193.75 acres in Ha Tsuen xiang . Zhengming's personal wealth included 68 acres of xiang land, a partnership with his brother involving at least three local factories, a money lending business, and a cargo boat company. Zhengming was the second largest landowner in Ha Tsuen; his brother was the largest, with a private holding of 71 acres.

Of the twelve ancestral estates in Ha Tsuen with more than 10 acres, Zhengming managed six, and of the nineteen estates with more than 5 acres he managed eight; thus, in his role as manager Zhengming controlled 52 percent of all ancestral land. According to my own calculations, the Deng

owned approximately 750 acres in Ha Tsuen subdistrict,^[43] which means that Zhengming either privately owned or managed more than 27 percent of Deng land in the area. Zhengming's philanthropy, his abilities as judge, and of course his wealth are still remembered; he remains a very popular figure in Ha Tsuen although he has been dead for nearly eighty years. Figures like Zhengming illustrate the complexity of managerial dominance. Resting in part on control of important resources, such dominance also relied upon the prestige that came from able management and the symbolic and political capital that prominent managers accumulated through their public and charitable activities.

Political Organization in Ha Tsuen

Chinese rural political organization in the nineteenth and early twentieth centuries must be understood in the context of a waning Chinese state and, in the case of Ha Tsuen, a colonial regime noted for its methods of indirect rule. As Ha Tsuen and other New Territories villages moved from Chinese to colonial control, changes were inevitable. There is little doubt that local commerce and manufacturing were directly affected by colonialization, but the British, like their Qing predecessors, continued to allow a considerable amount of local autonomy in the conduct of village political affairs. In the New Territories this remained true until the Japanese occupation.^[44]

The nature of local autonomy is well documented for this period. Hsiao Kung-ch'uan and Philip Kuhn have provided detailed studies of the mechanisms of local control in late Qing society.^[45] Although Mary Rankin focuses her analysis on Zhejiang, her characterization of rural political organization applies to areas well beyond the Jiangnan region. In Zhejiang she writes that the locally prominent dominated key public institutions, creating "the familiar end-of-dynasty pattern of autonomous control over peasants by socially conservative local oligarchies."^[46] Cole refers to lineages in Shaoxing as "mini-states" filling as they did a power vacuum at the local level, and Mann has utilized Max Weber's concept of liturgical governance to define and analyze these extrastate institutions and structures.^[47] There can be little doubt that liturgical forms of governance were widespread, and in this respect Ha Tsuen and other lineage-dominated areas were in no way unique.

There were no formal political offices and no formal arrangements for selecting community leaders in Ha Tsuen until the 1950s, when the colonial authorities introduced the Rural Committee system.^[48] Prior to this time there were, however, men who wielded great power and influence among the Ha Tsuen Deng. Elsewhere I have argued that lineage elders as a category did not play a decisive role in community affairs.^[49] Leadership fell not to the aged but to those with experience and economic influence.

Until the 1950s, community affairs and lineage affairs were not clearly

differentiated; politics was embedded in the structure of the lineage. Lineage institutions like Ha Tsuen's Yougong tang (Hall of Friendship and Reverence) served various functions. Yougong tang , founded in 1751, was one of two institutions that cut across hamlet and segment affiliations (the other was the village guard, which I will discuss below). Ha Tsuen's main hall (or da zu tang ) was at once religious center, meeting hall, administrative office, and symbol of Deng success and unity. In the absence of formal political organizations, it provided the framework for local decision making and administration. The term tang has a dual meaning in Chinese, referring to both the physical hall in which ritual offerings to the ancestors are made and the corporate estate, or "trust," of lineage property. In Yougong tang we see a structure of both religious and economic dimensions, but it functioned above all in the realm of politics.

During the period under discussion the manager of Yougong tang was expected to play a leading role in local affairs. Although formally charged with the duty of administering the hall's corporate property, he was far more than a mere accountant. The position of manager carried no inherent political powers, but in the hands of a forceful individual it placed the incumbent in a powerful position. Ostensibly the manager of the Yougong tang was selected by lineage elders. However, because it was essential that the manager be literate and possess business skills and independent wealth, there were few possible candidates. The selection was further restricted by the fact that the manager usually held the position for life. Considering all these limitations, the freedom of the elders to choose a manager was more apparent than real; they were, however, entrusted with the right to legitimize the manager's authority. This was done when the elders gathered with village notables for a special banquet that marked the installation of a new manager.^[50]

In 1905 Deng Zhengming managed the Yougong tang . Zhengming was a descendant of Jingwu and Zuotai and therefore a member of Ha Tsuen's elite patriline (Figure 9.1, generation 18); at the turn of the century he was a leading figure among this group. Wealth, philanthropy, political acumen, and knowledge of the world are the qualities that the Deng themselves mention when speaking of Zhengming. Personal confidence and wealth allowed Zhengming to enter the political arena, but it is important to stress that his ability to dominate local decision making did not rest on his charisma or money alone. Zhengming's role in local politics and the role of other members of the elite (all drawn from the Zuotai patriline) were supported by their wealth and their institutional control over Yougong tang and the village guard. These two institutions were closely linked; the manager of Yougong tang was in fact entrusted with overseeing the local security corps. Members of the local elite had no direct role in the village guard organization, but nevertheless they had considerable influence through the oversight function of the hall manager.

As manager of the Yougong tang , Zhengming was also in a position to dominate many economic activities. The manager administered Ha Tsuen's market and therefore had an important voice in determining market fees and licenses. He also oversaw the maintenance of the channel that linked the cargo boat traffic of Deep Bay to Ha Tsuen's market pier and had overall responsibility for collecting the fees charged for the use of that pier. Community projects such as the construction of public paths and public wells were largely in the manager's domain. Finally, and perhaps most significant, the Yougong tang manager (together with the elders) organized the bidding process by which the leader of Ha Tsuen's village guard was chosen.

Managers, Landlord-Merchants, and the Village Guard

The village guard was a local security force; it was neither responsible to nor directly linked with the formal apparatus of the state. In the case of Ha Tsuen, the guard was firmly controlled by the local elite during the half-century discussed here. Because of their coastal location, the Deng had two guard organizations, a "water guard" (shuixun ) and a land-based patrol group (xunding ). James Hayes, in a discussion of the Hong Kong region from 1850 to 1911, notes that dominant lineages "got a sizeable revenue from leasing fishing stations and beaches to villagers and boat people."^[51] Among the Ha Tsuen Deng, this sizeable revenue was protected by their special water guard.

The water guard was responsible for watching over the Deng's coastal fishing stations and oyster beds whereas the land guard protected xiang households. The organization of Ha Tsuen's two guards was very similar. Both had their own written constitutions that set out the responsibilities of guardsmen and specified the procedures for selecting members. Each year a new leader for the land-based security force was chosen by an auction held at Yougong Hall. The leader of the water guard was selected every three years. As organized by the hall's manager, the bids for both offices were sealed, but, according to villagers, this did not protect the process from favoritism or manipulation. The manager who organized the auction retained considerable control over the selection. The bids were often substantial, amounting to as much as one thousand Hong Kong dollars in the 1930s. Villagers report that in the past landlord-merchants often staked potential guard leaders to the money they would need if their offer was accepted.

The highest bidder became the head of the guard, and the cash he bid went into the coffers of Yougong tang . Each new leader was then formally confirmed by the manager and the lineage elders. After putting their seals to a paper signifying their public acceptance of the new leader, the elders were treated to a banquet hosted by the newly installed guard head. As long as each of Ha Tsuen's eleven hamlets was represented by at least one member,

the leader had the right to choose the guardsmen who served under him. The land guard maintained a force of up to fifteen men, and the water guard had twelve members. Guardsmen were usually poor tenant farmers. All were of the Deng lineage; no outsider has ever served in either of Ha Tsuen's security forces.

Although the guardsmen did not receive salaries, they shared in the fees they collected from all households located in Ha Tsuen xiang . The leader of the guard received the largest share and derived his profit from the fees he collected in excess of his original bid. The income for the water guard came from their right to take 18 percent of the value of all oysters harvested along the Lau Fau Shan coast. The water guard had a more clearly defined commercial role than that of the village guard. Water guardsmen kept track of the ownership of individual oyster beds and restricted collection to only those who had the rights to use Lau Fau Shan's lucrative fishing stations.

Ha Tsuen's village guard, like that of other communities, must be seen in the context of the endemic violence along the coast during the nineteenth and early twentieth century.^[52] Village guard organizations were important and useful. They provided farmers with a primitive form of insurance; if guardsmen could not retrieve stolen or lost goods, they were required to replace them. The guard also defended the community against bandits, pirates, and encroaching neighbors. In the New Territories each dominant lineage had its own village guard organization that presented a united front against the security forces of neighboring xiang . The Ping Shan Deng and Ha Tsuen Deng were often at odds over land, water, public thruways, and control over satellite villages. On occasion their antipathies spilled over into open confrontation and violence. In 1978, Ha Tsuen villagers recalled with pride their once bellicose reputation when their guard brooked no challenges.^[53]

Obviously Ha Tsuen's guardsmen served the entire xiang ; there can be no doubt that their regular patrols did deter criminals and unfriendly neighbors.^[54] This was no small matter in an area where local people could count only on the support they themselves could muster. It is clear, however, that the Deng had more to gain from the guards' actions than did their dependents. Because the guardsmen themselves received direct rewards from their activities, they had a vested interest in maintaining the hegemony of the Deng. In effect, their fee-taking privileges depended on their ability to maintain the local status quo. Prior to the 1950s, migration out of Ha Tsuen was the only escape from landlord-merchant dominated politics. Because bids for the guard monopoly were large relative to an average householder's income and members of the local elite often staked guard leaders to the funds they needed to make their bid, it is not surprising that the heads of Ha Tsuen's guard organizations were usually closely associated with local landlord-merchant families. As late as 1978, during my fieldwork, the leader of the

water guard was in fact a close political ally of one of the village's elite families.

Ordinary villagers did receive some advantages from the guard and its activities, but there can be little doubt that those who gained most from security activities were those who had the most to lose. Ha Tsuen's water and land guards provided the force that stood behind the economic and political power of the local elite, and as such they were important factors in the maintenance of the landlord-merchant elite. This force, it should be remembered, was guided by the manager of Yougong tang and the members of his family and line.

Comparative Perspectives

As noted in the introduction to this essay many scholars have viewed high rates of social mobility as characteristic of late imperial and Republican China. In comparison to the caste hierarchy of India or the rigid status boundaries of Tokugawa Japan, China appears to have been a society in which mobility was not only tolerated but also expected. Two main reasons are usually given for this fluidity: China's system of bureaucratic examinations, and the practice of equal inheritance among brothers. Together these institutions are seen as central to creating an open society in which family background did not predetermine status or the opportunity for advancement.

Discussion of the influence of these two institutions has focused on two populations at opposite ends of the social hierarchy: the small bureaucratic elite at the top, and the peasantry at the bottom. We have already noted some shortcomings of the mobility literature focusing on the degree-holding elite; now we must turn our attention to China's villages, for many scholars have found significant evidence for mobility there as well. Mark Elvin, for example, has argued that "Chinese rural society in the nineteenth and twentieth century was ... one of the most fluid in the world."^[55]

The most sophisticated work on rural social mobility has focused on the peasantry of North China. For the authors of these studies, inheritance practices play an important role in rural mobility patterns. Philip Huang, writing about rural Hebei and Shandong, argues that "few rich households maintained their status for more than a generation or two." He concludes, "The main reason for downward movement was the division of family property among sons"; according to Huang a single partition could drive a rich household down to middle- or poor-peasant status.^[56]

Clearly the North China countryside described by Huang and others (including Duara in this volume) differed markedly from the New Territories village discussed in this article. Although in Ha Tsuen a small group of landlord-merchants maintained dominance over many generations, on the

North China plain, owner-cultivators predominated, landlords were often absent from the local scene, and villages exhibited low levels of internal stratification and considerable mobility.^[57] In contrast to the aggressively self-protective, single-lineage village I studied, Ramon Myers has noted the lack of a strong sense of village identity in North China.^[58] He argues that "the influence of clan [lineage] management in village affairs and farming was very small" and concludes by pointing out that the composition of the village elite "constantly changed from one generation to another and was not based upon hereditary succession." These changes "very much approximated," Myers argues, "the rise and fall of households in the village."^[59]

Myers's discussion of local political organization is supported by Sidney Gamble's survey of northern villages, conducted in the 1920s and including communities in Shanxi and Honan as well as Hebei and Shandong. In striking contrast to the oligarchic forms of leadership characteristic of villages like Ha Tsuen, Gamble found that most surveyed villages operated a system of regularly rotated leadership positions.^[60] One should not, however, conclude that northern villages were little democracies led by meritorious community servants. Property qualifications for leadership positions were not uncommon, and Gamble mentions cases in which leadership passed from father to son or circulated among a small number of "local worthies."^[61] In general, however, the leadership patterns described by Myers and Gamble are a far cry from the gentry-led politics of Zhejiang or the landlord-merchant-dominated regimes of villages and towns in Guangdong's Pearl River delta.^[62]

Studies based on 1920s and 1930s land surveys of North China show that the majority (sometimes 75 percent or more) of peasant households owned the land they farmed and few cultivators were landless (20 percent or fewer by some counts).^[63] In a survey of land distribution figures for pre-1949 China,^[64] Esherick cites a rate of landlessness for China as a whole (32.1 percent of peasant households) that, while slightly higher than some estimates,^[65] is far lower than those reported for many communities in the southeast. Not surprisingly tenancy rates as well as landlessness rates tended to be low in the North. In Hebei and Shandong, for example, rates vary from the Guomindang Land Commission's estimates that 15 percent of farm households were tenants or part-tenants to the National Agricultural Bureau figure of 28 percent tenant and part-tenant for Hebei and 26 percent for Shandong.^[66] Many southeastern villages do not fit this owner-cultivator model.

Rates of landlessness and tenancy are extremely high in southeastern China.^[67] Unfortunately there are not a great deal of data on landlessness in the Pearl River delta region during the first two decades of the twentieth century, but the available data from a slightly later period suggest that Ha Tsuen's landless rate of 55 percent and its tenancy rate of 97 percent are not improbably high. Based on a survey of a number of Pearl River delta villages,

Chen Han-seng found tenant rates in the 1930s that ranged from 70 to 90 percent;^[68] he also discovered that "nearly half of the peasant families [in Guangdong] are entirely landless."^[69] Reporting on the Pearl River delta village of Nanching in the 1940s Yang notes that "we did not encounter a single peasant who did not rent some land from others."^[70] In a study conducted thirty years later, Potter found that of the forty-two Ping Shah farmers he surveyed 83.3 percent rented all the land they farmed and only 11.9 percent owned part of their farms.^[71] Potter does not give any estimate of landlessness for Ping Shan as a whole, but his extremely high rate of tenancy suggests that the vast majority of Ping Shan farmers were without land.

One way to account for high landlessness and tenancy rates in Pearl River delta villages like Ha Tsuen is to remember that 50 percent of agricultural land was held by corporate estates,^[72] and of the remaining 50 percent available for private ownership 36 percent was owned by six Ha Tsuen landlords. In total, about 86 percent of Deng-owned land in Ha Tsuen xiang was held by estates or a few large private landlords. It is not surprising, considering these statistics, to find that many Ha Tsuen farmers were without land. Although it is easy to appreciate that extensive corporate holdings have an effect on landownership patterns, it is far more difficult to know what political and economic significance to attach to these patterns.

The contrast with the Lower Yangzi region is interesting in this regard. Beattie's work on Tongcheng and Dennerline's work on Wuxi support the view that lineage organization and corporate property were important factors in establishing elite leadership patterns and in preserving elite status.^[73] On the surface the Deng lineage of Ha Tsuen and those described by Beattie and Dennerline appear similar in form and content, and it is tempting to conclude from these similarities that the relationship between lineage and elite was the same in each case. There is, however, considerable evidence to suggest significant differences in the ways in which lineage organization and landownership patterns combined in these two areas.

In this paper I emphasize the importance of corporate property and management more than Beattie does, but the differences between the Ha Tsuen and the Tongcheng data exceed matters of emphasis. Two factors are especially important in distinguishing these two areas: (1) the ratio of corporate lineage property to private property, and (2) differences in the types of landed estates established. Both Beattie and Dennerline report low rates of lineage property (as does Rowe, in this volume, for the Middle Yangzi). Beattie does not provide exact figures but notes that corporate holdings were not large in Tongcheng.^[74] Dennerline estimates that in 1881 only three percent of Wuxi land was owned by lineage corporations.^[75] In 1931 this figure had risen to 7.81 percent,^[76] but even this higher proportion is a far cry from the 50 percent corporate property rate reported for Ha Tsuen.

According to my reading of Beattie's and Dennerline's work, the charit-

able estate (yizhuang ) was the primary form of incorporating lineage property in their two regions. As noted above, charitable estates, unlike the ancestral estates (zu ) described for Ha Tsuen, functioned primarily as welfare institutions. In Ha Tsuen, income from zu resources were available only to the descendants of the founding ancestor (however designated). After proceeds paid for sacrifices at the founder's grave, the remaining funds, which were often considerable, were divided among the membership. Some families in the Pearl River delta received substantial personal incomes from estate property.^[77] In contrast, yizhuang profits were generally reserved for lineage widows, orphans, and the poor. In Ha Tsuen there were immediate and direct economic benefits from estate membership; in Tongcheng and Wuxi benefits to ordinary and even wealthy individuals and families were far less tangible.

In both areas, however, lineage estates played an important organizational role, and the wealthy men who managed them were assured a voice in local decision making.^[78] In Ha Tsuen economic benefits and political advantage tended to be closely linked, but in the cases described by Beattie and Dennerline the links seem less firmly established. In fact, from Dennerline's evidence there appears little overlap with political advantage accruing to the managerial elite and economic benefits going to the weak and poor.^[79] Before making a final judgment on the extent to which yizhuang financially benefited local elites, more information on managing these charitable estates is required. For example, did the overseers of such estates obtain direct economic advantages from their managerial role? If charitable estates did strengthen elite political control and economic dominance, what mechanisms made it possible to achieve these ends? And, finally, how important was lineage property in preserving landholdings from the fragmenting tendencies of partible inheritance? At this point the evidence suggests that yizhuang played a different economic role than the zu and tang estates of the Pearl River delta.

There were high rates of tenancy in both the Lower Yangzi region and the Pearl River delta.^[80] According to Perkins, provincewide tenancy rates for the early 1930s (tenant farm families as a percentage of total farm families) ranged from 46 percent to 33 percent for Jiangsu and 65 percent to 44 percent for Anhui. The same surveys found rates of 49 percent to 52 percent for Guangdong.^[81] Based on a 1930s survey of Wuxi, Buck calculates a 6.6 percent landless tenant rate and a combined rate of 52.5 percent for part-owners and tenants. For Tongcheng, Buck finds that 43.4 percent of county farmers were landless tenants and 86.8 percent were part-owners and tenants.^[82] Based on these figures it is clear that half or more of the farmers in each of these areas depended on rented land for their livelihood. In contrast to Ha Tsuen, which had both a higher tenancy rate (97 percent by my estimate) and a substantially larger portion of rented land in the form of corporate

estates, lineage property in Tongcheng and Wuxi appears to have made up only a small portion of the available rental land.

In Ha Tsuen—where half of Deng-owned land was tied up in corporate property, where more than a third of the remaining land belonged to a small number of locally resident landlords, and where more than half of the better endowed estates were managed by one of these landlords—tenant farmers had few choices but to deal with Ha Tsuen's elite. In Tongcheng and Wuxi, farm families also depended on rented land, but, because corporate property did not make up a significant portion of this land, farmers were not forced to rely on lineage estates and estate managers as much as their counterparts in Ha Tsuen. Control over corporate property in Wuxi and Tongcheng does not seem to have had the economic significance that it did in Ha Tsuen. That is, tenant farmers in these areas were far more dependent on private landholders, including presumably both agnates and nonagnates, than on corporate ones. In this sense Tongcheng and Wuxi managers appear not to have wielded the same kind of economic power that was the stock-in-trade of Ha Tsuen's managerial elite.

In Ha Tsuen the peasant farmer depended on the village guard for his security, on intermediaries for dealing with the outside world, and on the small elite of landlord and estate managers for access to land. Among the Deng the roles of guard overseer, intermediary, and landlord-manager were monopolized by members of a small, tightly-knit group whose ancestors had dominated local society for more than two hundred and fifty years. The ties that bound Deng peasant to Deng landlord were complex and compelling, allowing great durability and cohesion to lineage communities like Ha Tsuen and giving them organizational advantages over villages where centralization was more difficult to achieve.

There are interesting similarities between the lineages of Guangdong and the Lower Yangzi region: in each case the political role of the lineage was a factor in establishing and maintaining elite control. There were, however, significant differences in the realm of economic control. One suspects that the contrasting ratios of corporate to private property, along with the differing forms of corporate estates (yizhuang versus zu/tang ), may have created different patterns of elite domination and continuity in these two areas. Whether the discontinuity between the lineage's weak economic power and strong political power in Tongcheng and Wuxi established a less cohesive and more permeable elite formation remains to be seen. Differences in the nature of elite control were, of course, not simply due to corporate property rates or the differences between yizhuang and zu/tang . Factors like land concentration, elite residence patterns, and the proportion of office holders also affected elite control. There can be little doubt that the extent of commercialization and industrialization in these two areas also played a key role in the political and

economic life of local elites. In Ha Tsuen, the lives of ordinary villagers were encompassed by the lineage; in Tongcheng and especially in Wuxi this appears not to have been the case.

As noted in the introduction, one case study does not allow us to draw general conclusions about rural elites in China. But the Ha Tsuen material is fully consistent with data collected from the New Territories and other Pearl River delta sites. It certainly leads us to ask whether the presence of well-endowed lineage estates created a more rigid, more hierarchical, and less permeable class structure than existed in villages where corporate property was unimportant. Furthermore, the existence of major estates may well have allowed for more political centralization in lineage-based communities. The Ha Tsuen data support the view of those who claim that the corporate landlordism of southeastern China produces a social formation that differs from the owner-cultivator regimes of the north and, I submit, the private landholding complexes of the Lower Yangzi region.^[83]

Ten
Elites and the Structures of Authority in the Villages of North China, 1900-1949

Prasenjit Duara

It has long been observed that rural society in North China was very different from that of South China, reflecting as it did a very different ecology. The prosperous rice-growing economy of the southern regions supported a highly stratified society where rural elites exercised domination through individual and corporate control of material and symbolic resources. In the north, rural elites at the village level were much less sharply differentiated from the bulk of the peasantry. This was certainly the case by the twentieth century, and it is unlikely that these elites had ever, since the early modern period, been able to attain the levels of wealth and power of their southern counterparts.^[1] In many ways the more sharply stratified rural society of South China conforms to the typical image of agrarian societies: relationships of dependency between landlord and tenant exist at the heart of all social arrangements. In the north, where dependency relationships were not quite so dense, the elite's grip over the structures of local authority was correspondingly weaker.

This chapter investigates the special relationship between the status of the rural elite and the structures of authority in North China through a study of customary law in four villages in Hebei during the Republican period. Each village was studied by Japanese investigators in the 1940s, and the four revealed a range of village elite types. Hou Lineage Camp (Houjiaying) is located in northeastern Hebei, near Manchuria, where many of its villagers worked as sojourners. Sand Well (Shajing), considered by its investigators to be an average North China village, is located just north of Beijing. North Brushwood (Sibeichai) in the cotton-growing district of south-central Hebei represents a village where commercialization led to an impoverished village elite and an unusually large role for absentee landlords and moneylenders. Finally Wu's Shop (Wudian), south of Beijing, had been frequently ravaged

Map 10.1.
Locations of the Four Villages in Hebei Province
Source: Prasenjit Duara, Culture, Power, and the State (Stanford, 1988).

by warlord armies, whose depredations led to the flight of the village elite (map 10.1).^[2]

Despite their marked differences, all these villages on the North China plain shared many characteristics of that entire region. Its ecology produced what Philip Huang has characterized as "a conjunction of low-yield, disaster-prone dry farming with high population density that laid the basis for severe scarcity."^[3] Not only did these conditions make large-scale land-lordism an unprofitable venture (at least until extensive commercialization of agriculture in some areas in the twentieth century), but the absence of significant amounts of corporate property also made it difficult for elite families to avoid diminishing their wealth at each successive generation through the practice of partible inheritance.^[4]

Historically, the area had been the theater for wars of succession to the imperial throne in Beijing. During the Republic, it was once again devastated by the battles between warlords and then by the Japanese invaders. Moreover, in spite of the political confusion, the peasants and elites of the region were subjected to the pressures of somewhat less than successful, but nonetheless intrusive, attempts at statebuilding by the successive regimes of the twentieth century. These efforts are too complex to evaluate in detail here. They include the state's attempts, for instance, to appropriate community religious properties for purposes of "modernization" and to force village leaders to pursue unpopular policies such as implementing land investigation or supervising the deed-tax payment that the rural folk had evaded for as long as they could remember. Most of all, state penetration involved heavy and unpredictable taxes on the village community to finance the expanding administration and its many programs at all levels—county, province, and center. Such unpopular state policies often led elites to evacuate leadership positions in the villages.^[5]

Under these conditions, the opportunities and incentives for local elites to exercise authority in the villages began to recede still further. In this essay I explore how, under these circumstances, an elite that did not directly control the means of production—land, labor, credit, and markets—managed to establish and exercise its authority in rural society, and further study the implications for a community in which the elite was unable to control the structures of authority.

Elites and the Nature of Authority in Rural China

I follow Max Weber in defining authority as legitimate domination,^[6] and by structures of authority I signify the means used to stabilize a social structure by guaranteeing the distribution of its material and symbolic resources. In advanced industrial societies, authority is formally located in the state and its legal system. Elites do not necessarily dominate, or otherwise participate in,

these structures of authority; they find other means to reproduce or extend their privileges.^[7] In small-scale societies such as the rural society of traditional China, however, local elites often dominate these structures, for reasons arising from the needs of both the elites and the community. It is easy enough to understand why these elites sought to monopolize the structures of authority. In a society where the state's reach was limited and the means available to maintain and expand resources were constrained by the absence of primogeniture, elites needed to ensure peasant compliance, or at least acquiescence, if they were to obtain their goals and maintain their position in society.

More interesting, however, are the ways in which communities might have needed elite participation in structures of authority. In China, various social roles—for example, sponsor and manager of religious societies and festivals, lineage leader, mediator with the outside world—typically conferred authority. However, in China's relatively commercialized peasant society, community tasks often required wealth, leisure, and influence for their effective performance. Structures of authority in this society gained credibility through elites participating to "make these structures work." To secure benefits for the lineage, for instance, village leaders in South China had to maintain informal networks among officials and county elites at a scale beyond the horizon of peasant capacities. Thus, authority derived from the performance of roles that were seen to maintain or bring collective benefits; but the effective performance of such roles often required elite resources.

Although participation in structures of authority was important to elite status, elites need not be synonymous with "persons in authority." I will, therefore, define elites only in terms of the first two attributes of the usual triumvirate of wealth, power, and prestige; the third might be acquired by nonelites as a result of the authority they derived from assuming community roles. Genealogically senior lineage members, ritual specialists, and others who did not necessarily have wealth and power might thus command prestige at the expense of richer and stronger elites. In practice, elites were likely to monopolize genuinely authoritative village roles precisely because of their wealth and power. Nonetheless, we shall see that prestige from community activity could under certain circumstances confer authority upon people who would not otherwise be considered elite and, therefore, might follow a dynamic other than wealth and power.

We see this most clearly in studies from the Canton delta region of South China. Although genealogically senior elders in Hong Kong's New Territories played a ritually important role in the lineage, the powerful lineage members—the elite of the community—dominated both lineage and community life.^[8] Rubie Watson's study (in this volume) of a New Territories village shows that a small number of wealthy landlord-merchant families of the Deng lineage managed to dominate the bulk of tenant-smallholders from the same lineage for centuries. The elite dominance in landownership was

reinforced by their control of the market and credit and their marriage alliances to other wealthy and powerful families.^[9]

The many-stranded relationships between landlords and tenants in the New Territories required elite domination of the structures of authority. For an economic elite that rented out land and lent money to ordinary peasants, authoritative roles ensured, for instance, that rents and interest were paid. However, even if the elite's aim was simply to protect its own interests, this elite could gain authority only by promoting, or appearing to promote, the general interests of the community. It was not enough for village landlords to exercise great power over their tenants and debtors; to gain legitimacy they also had to be responsible kinsmen and social leaders. Watson argues that the ideal of brotherhood within the lineage justified elite domination. But she makes it equally clear that lineages did not automatically confer authority to these elites; they needed to build up their symbolic capital within the lineage by engaging in philanthropy and other community tasks.^[10]

Unlike their southern counterparts, village elites in the north had neither the ability nor the compelling need to control local authority structures. Because these elites dominated neither landholding nor moneylending, their efforts to control the structures of authority were not as effective as those of the southern elites; nor was their desire to dominate so directly motivated by the need to preserve their economic interests. At the same time, rural North China was not a simple, self-sufficient society; this complex commercialized society required relatively strong leadership and authority figures—often stronger than available—to fulfill the peasantry's demands and coordinate community tasks. Genealogically senior lineage elders and ritual specialists were not especially powerful here and were hardly able to address its complex community needs. Who, then, would perform these roles? Who would be able to sustain credible authority structures, especially during the Republic, at a time of great stress in rural society?

Traditionally, authority in these villages derived from both effectively managing such community tasks as crop watching and religious activities and protecting or securing the villagers' interests through such roles as middleman in the innumerable contracts into which villagers entered. Such functions legitimated elite domination—that is, according to our definition, they gave elites authority. But this does not mean that there was no coercive basis for elite power. Occasionally, such coercive powers were built into community activities; for instance, the crop-watching associations were among the most popular and important village-level associations in Hebei during the Republic. Outsiders often assumed that these associations were meant to protect village residents' crops from outside bandits and robbers. In fact, the crop guards would hardly have been able to take on a band of robbers; rather, the crop-watching system was designed to guard the crops from the local poor, including the poor members of the village itself. Nonetheless,

many poorer landowners also benefited from this protection. The militia in some localities may have worked in a similar way. However, in the villages studied here, the militia were organized by higher-level offices, often by the ward (qu ) of state administration.^[11] Whatever the local variation and despite our focus on legitimacy and acceptance of domination, we can hardly overlook the fact that, in peasant society, this domination was inevitably tied up with some threat of coercion.

The remainder of this essay focuses on the authority that accrued to village middlemen operating within the realm of customary law and convention. Peasants relied on these middlemen to arrange contracts for many purposes, from securing loans to partitioning family property. The middlemen often indicated village authority more clearly than did managers of community activity. Not only were middleman functions vital to peasants' livelihood, but other community roles traditionally performed by village leaders, such as managing village religious ceremonies, were diminishing in importance in the twentieth century.^[12]

Elites and Middlemen in Customary Law

Ramon Myers and Fu-mei Chen have demonstrated the crucial role of customary law in facilitating the informal market for land, labor, and credit in rural North China.^[13] Here I focus on three kinds of contractual arrangements: taking out a loan, leasing land, and selling land. All three types of contracts were critical to the life of the commercialized agricultural society of North China, and all required a middleman (zhongren, zhongbaoren ).

Loans were perhaps the most frequent type of contract in North China, and peasants acquired credit by mortgaging property or providing a guarantor. For loans based on mortgaged property, the middleman mainly acted as a witness to ascertain the contents of the contract—in particular to investigate any competing claims on the pledged or mortgaged property—and to mediate disputes. Then, if the loan was not repaid, or more commonly, if interest payments were not forthcoming, it was his duty to either arrange for an extension or dun the debtor. The middleman for a guaranteed loan, in addition to performing these tasks, also undertook to guarantee the repayment of the loan.

In the villages I have studied, guarantors were not usually required for rental contracts, probably because tenures were short or unspecified in many villages.^[14] Under these circumstances, nonpayment of rent meant certain eviction and difficulty in finding another tenure. The importance of the middleman lay in his ability to locate a tenure, to gain favorable terms for the tenant, and intercede when rent payment was contested. In a land sale transaction, the middleman had several functions. He was usually approached by the seller to seek out a buyer and negotiate a favorable price.

He was responsible for ascertaining the quality and the size of the plot. More important, he had to determine, by inquiring of neighbors and the lineage, whether there were any competing claims on the land. Finally, he ensured that payments were properly made and mediated all disputes.

All contracts required middlemen, but not all contracts required that the rich and influential elite play this role. Probably only loan contracts requiring guarantors needed propertied middlemen; in practice, it was distinctly advantageous for villagers to employ a man of wealth and influence as the middleman. Wealthy and influential middlemen often had extensive connections, and they could contact people interested in seeking credit or purchasing land from a much wider radius than could an ordinary villager. Even more important was the credibility such men brought to their authority as middlemen.

What authoritative roles was the middleman expected to play? In case of a dispute over a contract or a violation of its terms, the authority and responsibility to mediate fell upon the middleman. Moreover, a contract was almost always embedded in an unequal power relationship, and the middleman's authority was crucial not only to prevent the violation of the contract terms by a stronger party but also to negotiate and renegotiate the terms on behalf of a weak party. This was not an uncommon scenario and took place, for instance, when a villager entered into a contract with an absentee landlord or moneylender from the city.

Whether to prevent violations, mediate disputes, or negotiate the terms of · a contract, the effectiveness of the middleman's authority was considerably shaped by his "face." For example, in one village, Hou Lineage Camp, a price dispute broke out just as a land transaction was about to be completed. After many attempts to find a mutually agreeable price, the buyer finally resigned himself to the seller's price, saying, "I agreed because it is difficult to do the job of the middleman. If I decided not to buy it, it would affect the middleman's face."^[15] The mere presence of a third party known to two contracting strangers was thus itself a means of facilitating negotiations because it added a personal element to the contract. Similar considerations worked to discourage contract violations. The ties of personal obligation to a middleman weighed much more heavily upon a would-be violator if the middleman had considerable "face." Moreover, the higher the status or "face" of the middleman, the more successful dispute mediation was likely to be.

When the middleman introduced a villager seeking a contract to a powerful figure, it was believed that the more "face" the middleman had, the greater was his ability to negotiate terms favorable to the villager. If the middleman was of sufficiently high standing, there was said to be no need for a written contract. A middleman with "face" could fetch a good price for property, get a loan for a longer period, arrange for its extension, negotiate the interest rate, or get the lender to forgive part of the interest.^[16] Elites, particularly

declining ones like Zhang Yueqing of North Brushwood, took pride in what their face could obtain. Zhang, a village headman for many years, remarked, "If a seller wants to return a deposit [having changed his mind about the transaction] and the buyer refuses to accept it, there is no hope for the seller if his middleman is a poor man. On the other hand, if I was the middleman, I would use my 'face' and ask him accept the deposit."^[17]

Thus, what seemed to count in the exercise of authority, in customary law at least, was what the Chinese call face (lian, mianzi ), a subject that deserves more attention than it has been accorded. In his study of Taitou, a Shandong peninsula village near Qingdao, Martin Yang observed that everybody had face, although one person may have more or less than another. His notion of face brings it close to ideas of honor and shame. In his discussion of village conflict, Yang reveals that gaining, preserving, or causing the loss of face was the object of much social interaction and strategizing. The gain or loss of face was shaped by a highly complex matrix of social variables, including not only the statuses of the people involved but also their relationships to other people. Yang gives us an instance of this social complexity:

Another kind of circumstance in which face is involved is when a youngster offends a senior member in the village. When the offended man is about to punish the boy, other villagers may pacify him by saying: "For his parents' face, you may forgive him." Then the senior member may say: "All right you are old neighbors. For your face and for his parents' face, I forgive him this time."^[18]

Yang could have added here that, had no one witnessed the event, the older man might have let the matter slip, and the whole question of losing face might never have arisen. But had the event been witnessed by a different set of villagers, with less status and thus less ability to plead the boy's case, the boy might have indeed been punished, and his parents might have lost face. Thus the social context of an event was also important.

Yang's analysis of face as an object of social contestation helps us understand the concept as a desired cultural goal that people both aspired to attain and valued in others. It enables us to understand, for instance, why an elite that did not control the means of production in a peasant economy was still motivated to undertake various seemingly thankless public tasks with remuneration only in prestige. However, Yang's analysis only illuminates one dimension of a manifold phenomenon. To acquire face was also to acquire authority, and nowhere was this connection clearer than in the realm of customary law. In this sense, face was not just an attribute everybody possessed; rather, some people possessed more face than others, and, moreover, possessed a kind of face that was efficacious: it could get things done without the overt use of wealth or power.

Although face did not always give people of wealth and influence author-

ity, often the face of the elite was authoritative. Consider, for instance, the matter of trust. Several peasant informants equated effective face with trust. Trust was not something particularly associated with elites, but in economic matters the trustworthiness of an individual was not easily separable from his wealth and standing in the community. Indeed, wealth and trust were the two sides of a face whose authority was judged by its efficacy in obtaining practical results. In a loan contract, for instance, the size and duration of the loan and even its rate of interest could be determined by the face of the middleman. A moneylender who made a loan to a villager whose middleman was an ordinary person, not particularly known for his trustworthiness or standing in the community, had to consider the higher risk involved. He might, therefore, loan out a smaller sum for a shorter period and charge a higher rate of interest than if the middleman's ability and reputation ensured repayment.

Of course, face did not always work according to strict economic logic. Probably the face of an elite middleman would be more successful in pressuring a powerful figure than the face of a common peasant, if for no other reason than because the elite was likely to be part of a more elaborate network of reciprocal obligations with other elites than the peasant. Thus, when villagers sought middlemen with face, these middlemen were often—though by no means always—elites with wealth and power in rural society. For a member of the elite, gaining face represented a mechanism for converting wealth, power, and influence into prestige and authority. By gaining face, he was accumulating what Bourdieu has called "symbolic capital," which might be re-deployed at some later time in the service of politics or honor.

One qualification is in order. Face was only relevant within the middleman's arena, defined both territorially and organizationally. Nonelites could exercise influence within the village or with outsiders when they had some special connection. Village elites might have been able to drive a bargain not only with other clients in the village but also with clients and business partners in lower-level markets; however, they may have been completely powerless at higher-level marketing centers where some absentee landlords resided—those of a class with few ties to the village elite. In the following analysis, we see how the arena in which villagers conducted their business varied from village to village, and the ability of the elite to control the structures of authority also depended in some measure on the arena in which they operated.

Elites and Middlemen in Four Villages

Hou Lineage Camp . This village, located in Changli county in the northeastern corner of Hebei, had become increasingly prosperous through the first third of the twentieth century, mainly given increased incomes of so-

journing workers and traders who went to Manchuria. Of the approximately 110 households in the village in 1940, about 11 percent (or 12 households) owned between 60 and 150 mu , while about 70 percent owned less than 30 mu . This 70 percent had to supplement their income by leasing part of the land belonging to the 12 or 13 households with surplus lands. The presence of a small, visible elite, which controlled part of the land cultivated by other villagers, makes Hou Lineage Camp more similar to the southern villages of the Canton delta than any other village studied here. However, differences in wealth between the ordinary peasant and this elite did not even begin to approach the sharp differences between elite and peasant in the Canton delta.^[19]

This village elite undertook most middleman roles for loans and land transactions;^[20] with its large circle of friends and acquaintances it could negotiate a good deal. The middlemen were never paid for their services, but, if they performed special favors, their client would bring over a gift for their children as a token of this gratitude.^[21] The most common type of loan taken out by villagers requiring guarantors was often made by merchants from outside the village. Such loans placed heavy responsibilities on the guarantor-middleman, and five of the most prestigious and powerful village leaders were said frequently to play this role. Four of them were wealthy villagers with face, unquestionably part of the village elite.^[22] Let us look more closely at these elite middlemen, in particular at the careers of Xiao Huisheng and Liu Zixing.

Hou Lineage Camp had a tradition of elite participation in community activities. Three plaques awarded by villagers to their leaders attest to this tradition. One recipient of such a plaque worked in the county tax office. The plaque, dated 1870, noted that he often advanced money to villagers unable to pay taxes or requested the magistrate to grant extensions. Another plaque, from around the turn of the century, had been granted to the first modern headman of the village and hung on the outside wall of his son's home. It represented a token of the villagers' gratitude for his various services, such as mediating disputes, giving aid to the poor, and assisting people entangled in legal suits.^[23]

The third plaque, inscribed "ardent in public service" (rexin gongyi ), was granted to Xiao Huisheng, undoubtedly the most influential leader in the village. Although the leaders of Hou Lineage Camp had promoted the idea of a plaque, it was presented to him in 1937 by the residents of thirty-eight villages—all within the fourth ward of Changli county. The villagers were grateful for his mediating disputes that might have become ruinous legal suits, especially when he was director of the county telephone bureau from 1934 until 1937. Later he was awarded yet another plaque, this time by his fellow villagers only, to acknowledge raising money for the village school.

Underscoring the practical value of face, plaques were apparently granted to benefactors during their lifetime to enhance their face.^[24]

Despite some similarity between Hou Lineage Camp and Canton delta villages, a leader like Xiao could probably never have emerged in the lineage-dominated society of the south. Not only did Xiao have no lineage in this village, but also he had not even been born there. Xiao's father had settled in Hou Lineage Camp, the native village of his business partner, after the two men's business had failed in Manchuria. Xiao grew up in the village, attended school in the county seat, and graduated from a university in Manchuria. He returned to the county and headed up the telephone bureau of the county government. He was said to be well versed in the law and able to arbitrate disputes before they became legal suits; but it was also said that once a dispute became a legal case he never tried to influence the decision. He had many friends and extensive contacts at the county level, which he appeared to use generously for the villagers. For instance, he could get loans for villagers without collateral, and he often acted as a middleman in contracts. He had sixty mu of land when he returned to the village in 1937 and served as the assistant village headman and a member on the school's board of trustees.^[25]

Although Xiao managed well without a lineage, Hou Lineage Camp was not without lineage politics. Lesser-elite leaders relied on combinations of kinship bonds and patronage to build bases of support. Liu Zixing (sometimes known as Liu Zixin), the leader of the Liu lineage, had been village headman twice, once in the early 1920s and again in the late 1930s. Among the richest men in the village with 170 mu of land, he himself lent money at interest. In the late 1930s, he began to accumulate more and more land by foreclosing mortgages. At the same time, however, he often guaranteed contracts requiring no collateral, thereby undertaking the risk to pay for a peasant should he become bankrupt.^[26]

As the leader of the Lius, Liu Zixing had been deeply involved in lineage feuding. In 1921, when he was village headman, the village undertook a survey of all village lands for tax purposes, and Liu was found to be one of those with concealed land. His Hou lineage rivals on the council threatened to take him to court. The matter was ultimately dropped because Liu, like most people, had concealed only a few mu of land, and nobody in the village really cared. But the Hous had achieved the purpose behind their threat to go to court: Liu was made to resign from office and taught at a county school until 1929. Thereafter he served on the county education board but returned to live in the village.^[27]

When Hou Dasheng became the village headman in the early 1930s, Liu got his opportunity to exact revenge. Hou Dasheng ignored village councillors in decision making and misused public monies, following an increasingly

common pattern in the late Republic when the pressures of warfare, heavy taxation, and increasing government intrusion into village affairs made it extremely difficult for responsible leaders to stay in office. Finally, in the mid-1930s, a group of villagers reported the matter to the county magistrate, and Hou Dasheng was forced to resign.^[28] The principal leaders of the opposition included Liu Zixing, who probably took the matter to the magistrate himself. As a member of the board of education, he had access to the county authorities. In 1939, several years after the incident, he became the headman of the village and remained in the post until 1941.^[29]

From these plaques and biographies, we see how face and authority came to those with wealth and influence only if they utilized those resources for the benefit of the community. Needless to say, this would hardly prevent these elites from also using their prestige to pursue their own personal goals—as we glimpsed in the career of Liu Zixing. Nonetheless, the fact that village authority rested with this kind of elite is underscored by the official posts in the village that most of them held at one time or another in their careers.

The fifth middleman, Kong Ziming, did not fit this pattern. Kong was not rich and owned less than twenty mu of land, but he was educated and enterprising. We do not know how much education he really had, but he was highly regarded in the village because he could quote the Confucian classics and talk smoothly (a skill he learned as a shop assistant in Manchuria). This, of course, made him an adept mediator and an invaluable middleman during negotiations. He mediated the conflict between the Lius and the Hous in the village. Kong also initiated the move to grant the plaque to Xiao Huisheng. In 1940 he became the assistant village headman, and in the same year he leased fifty mu of rice land from a Japanese company and farmed it with two hired laborers. Kong's successful career reveals that one did not need to be wealthy to command authority in the village; clearly education and resourcefulness went a long way in rural North China. It would probably have been rare for this kind of person to command authority in the Canton delta where generations of domination by a few elite families made it extremely difficult for one who was not wealthy, and especially one who, like Kong or Xiao, was not from the dominant lineage, to break into this monopoly.^[30]

Sand Well . Located in Shunyi county, approximately thirty kilometers north of Beijing, this was considered by Japanese interviewers to be an average North China village. In the late 1930s, there were seventy households in the village. There were no large landlords employing many tenants, but several managerial landlords owned more than seventy mu and employed wage labor. Distribution of land was unequal: 60 percent of the households owned 14 percent of the land, and 15 percent owned 52 percent.^[31]

In Sand Well, the wealthy had actively participated in middleman roles, locating partners for land and credit contracts and serving as guarantors and

mediators. Here grain loans requiring guarantors were made by the store in the nearby market town or by speculators on the grain market. The well-respected village headman, Yang Yuan, had extensive contacts in the market town and often acted as the guarantor.^[32] Yang Yuan was descended from a line of wealthy village councillors, and two of his brothers were also councillors. He owned a handicraft store in the county seat; for villagers this was their most important opening to the commercial world of the city. Here Yang introduced his fellow villagers to prospective moneylenders, to landlords looking for tenants, and to those seeking to buy or sell land.^[33]

By our standards, the Yangs might seem corrupt. He and his brothers concealed more land than most others in the village, and he doubtless used his position as village headman to sell off some infertile land to the village government. However, the villagers did not seem to care a great deal; they seemed reasonably tolerant of people who pursued personal gain from public office as long as these people contributed to the welfare of the community. As an indication of his standing, he was an important mediator in disputes both between villagers and between villages. Although in 1940, after eight years of service, Yang was keen to give up the post of headman, which had become an especially high pressure job under the Japanese occupation, he was clearly regarded as a local leader. He was, for instance, appointed principal of a school established jointly by four villages around this time.^[34] Yang Yuan was a typical elite leader whose authority was built upon his role as the gate-keeper of his small community, protecting its interests from, and securing its needs in, the outside world.

In Hou Lineage Camp we saw, in Xiao Huisheng, the rise of an elite patron and the growth of his authority; in Sand Well, we see the decline of elite patrons. Two elite members who had frequently served as middlemen actually lost some property as guarantors. Zhao Tingkui came from a line of wealthy and prestigious councillors. His father had owned 130 mu of land, but he himself had inherited only 70 mu . By 1940, his property had dwindled to 19 mu , in part because he had to pay for a defaulter; the person he had stood for as guarantor had died, and the family was simply unable to repay the loan.^[35] Du Xiang, himself a councillor, was also descended from a family of councillors; his uncle was said to have owned 700 mu of land.^[36] In the early years of the Republic, Du Xiang owned 54 mu of land. During these years he was an active middleman arranging loans and rental tenures for the villagers. Once, he acted as a guarantor for an affine who was unable to return the loan, and Du lost 10 mu of his own land. For various reasons he lost more and more land. By the 1930s, although he still acted as a middleman, Du Xiang did not have the same face that he had held earlier.

Yang Yuan and Du Xiang were still the most commonly used middlemen in Sand Well, especially in loan contracts with outsiders.^[37] Villagers also turned to friends and relatives, often in neighboring villages, to serve as

middlemen in land transactions and land tenure contracts.^[38] By the 1930s, however, another very different type of middleman had appeared in the village. An impoverished peddler, Fu Ju, settled in Sand Well in the early 1930s. His job took him all over the area and secured him a wide range of contacts, which he put to good use. In the late 1930s, he became a tenant of a powerful landlord living in the county seat and frequently used this connection and his other contacts to act as a middleman. In particular, he arranged for villagers to mortgage their land to landlords in the county seat. Unlike elite middlemen in the village, Fu charged a commission for his services, although he denied it. By the late 1930s, Fu Ju served as the middleman for most land tenure contracts and many land transactions in Sand Well.^[39]

North Brushwood . The circumstances of the two remaining villages, North Brushwood in Luancheng county and Wu's Shop village in Liangxiang county reveal a picture of elite activity quite distinct from the other two villages. By the late 1920s and 1930s, these villages had become so thoroughly poverty-stricken that one could scarcely speak of a village elite any more. At the same time, absentee landlords from the county seat became significantly involved in the market for credit and land; thus, the arena in which peasants were involved to secure their needs was widened considerably. But these new participants in land and credit markets came from social strata so distant from the world of the village that what remained of the village elite could scarcely exercise influence on them. As a result, this elite became increasingly irrelevant to the life of the village.

The peasants in North Brushwood principally grew cotton. In this region of south-central Hebei, cotton began to outstrip millet cultivation sometime between 1910 and 1930. The county gazetteer of 1871 registers the land owned by the villagers of North Brushwood between 2,400 and 2,500 mu .^[40] A number of bad harvests during the early 1920s and early 1930s, together with the depressed prices for cotton, had rapidly impoverished the village.^[41] By the early 1930s, over 1,600 mu of peasant-owned land had shifted to the hands of big absentee landlords in the county seat. However, 627 mu of this land had been mortgaged and could still be redeemed. With the rise in agricultural commodity prices in the late 1930s, many peasants might have been able to redeem their land, but they still had to clear rent arrears before they could do so.^[42]

The poverty of North Brushwood is indicated by the fact that the average family owned less than ten mu and cultivated a total of less than fifteen mu in an economy in which a family of five needed twenty-five mu to make ends meet.^[43] Only about ten families in the village were able to manage without taking out loans in a normal year. Stratification was not very sharp within the village, and only three households owned as much as thirty to forty mu . The one large landowner, Zhang Yueqing, had mortgaged forty-eight of his eighty mu to landlords in the county seat.^[44]

In terms of property relationships, North Brushwood seemed less like a northern village populated by peasant proprietors than a poorer version of a Lower Yangzi village, with the bulk of the peasants cultivating land owned by large absentee landlords. Three landlords in the Luancheng county seat controlled 723 mu of the total of 1,372 mu of land leased by the North Brushwood villagers.^[45] Much of this land had been mortgaged by the villagers to landlords who kept them on as tenant-cultivators on their own mortgaged land. Relations between the mortgagor-tenants and the moneylending landlords were not harmonious, and they worsened after the Japanese invasion. Even before 1937, tenants delayed rent payments and landlords frequently had them arrested. Tenants retaliated by burning the homes of the landlords. The turmoil accompanying the Japanese invasion made it difficult for tenants to harvest their crops and pay their rents, as bandits and soldiers pillaged their fields. In 1940, when the Japanese army gained fuller control of the region, the landlords began to demand these rent arrears in current prices, which were much higher as a result of the wartime inflation. Apparently backed by the county government, they refused to allow tenants to redeem their mortgaged lands unless these arrears were paid.^[46]

Some evidence suggests that in earlier periods an elite in the village included individuals who owned two hundred to three hundred mu of land.^[47] However, by the late 1930s, scarcely a person in North Brushwood had wealth and connections comparable to those of the elites of Sand Well or Hou Lineage Camp. The single partial exception was Zhang Yueqing, a former village headman of fourteen years' standing who was descended from wealthy village councillors. But Zhang's wealth had also diminished, and he too had mortgaged forty-eight of his eighty mu . Part of Zhang's financial problems arose because, as village headman, he had advanced money to pay the heavy irregular taxes demanded by the Nationalist government in the late 1920s, and the wretchedly poor villagers had been unable to pay him back. He resigned from the village headmanship soon after but returned to the position again in 1934. Following the Japanese invasion in 1937, he was kidnapped by bandits who demanded a ransom for his release. After that experience, he gave up his post permanently.^[48]

Despite his reduced circumstances, Zhang strove to retain his status as a community leader. He still worked as a traditional doctor in the village, dispensing free services, and ran a traditional-style school in his home. He claimed that he was still the most effective middleman in negotiating loans because of his "face"—as he put it—and he was an expert at measuring land in the village.^[49] But Zhang was fighting a losing battle. As the urban money-lenders' power over the village increased, his influence began to dwindle, and villagers turned to other middlemen. Zhang himself had mortgaged more than half his property to the biggest moneylender, Wang Zanzhou; because Zhang too had incurred rent arrears in 1937, Wang refused to let him redeem his lands. Not unexpectedly, Zhang was North Brushwood's

bitterest critic of these moneylenders and a vocal champion of the rights of the mortgagor-tenants: he was losing more to these outsiders than just his land.^[50]

It is easy to see how the absentee landlords were increasing their power over the villagers. Mortgaging property to these landlords had become the principal means of acquiring credit in North Brushwood. Villagers also depended on these landlords to allow them to continue cultivating their lands as tenants. They were beholden to them for additional loans based on the increased value of their mortgaged land, for extensions on rent payments, and the like. As the hold of these landlords tightened, there was little room for an independent patron to provide access to credit, land, and labor markets on terms favorable to the villager. Instead, to communicate their needs to the absentee landlords villagers became entirely dependent on a different kind of person—one who did not necessarily represent their interests.

Two such individuals, Zhao Laoyou and Hao Laozhen, lived in the village. Like Fu Ju in Sand Well, neither owned more than a few mu of land; nor did they have prestige or authority in the lineage. Nonetheless, they were the middlemen in all negotiations with the three big landlords. Villagers approached these men when they wanted to sell or mortgage land, and these middlemen checked the productive capacity of the prospective mortgagor. Before they leased out the land, the money-lending landlords wanted these middlemen both to see that the mortgagor and prospective tenant had adequate labor, implements, and farm animals to work the farm and to determine that the tenant's own property was not so great that it would absorb all of his productive capacity. Landlords also used these middlemen to dun tenants for arrears and act as witnesses when legal action was brought against the tenant.^[51]

It would be difficult for such middlemen to acquire face in village society. Indeed, it was claimed that they were agents of the landlords, receiving money secretly for each transaction and keeping an eye on each mortgagor-tenant to see that he did not secretly sell off the land. They themselves received interest-free loans from the landlords and were allowed to lease 80 mu and 60 mu under the best conditions of tenure. Villagers often wined and dined Zhao and Hao in hopes of obtaining favorable terms from the landlords, but the evidence suggests that the two always protected and promoted the business of their patrons.^[52] Zhang Yueqing captured the difference between this type of middleman and the community-oriented elite-patron:

If the mortgagee feels that he has already lent out too much money and does not wish to raise the value of the loan, he tells that to the middleman who simply communicates that to the mortgagor. But as long as I am the middleman I would never simply communicate [but rather negotiate] such a decision because I would incur a loss of face.^[53]

These new middlemen were a far cry from the village elite-patrons who sought to enhance their status by employing their face to secure favorable terms for the villagers. In fact, villagers often did not know these new men personally and had to approach them through a relative or friend—in effect using a middleman to secure a middleman. Men like Zhao and Hao performed their brokerage roles as clients of the big landlords or as specialists who charged a fee for their services. The decline of previous village elites seems to have been spurred by the increased tax demands of the expanding state and by the disorder and warfare of the twentieth century. But the accelerated commercialization of the rural economy in the cotton-growing regions of Hebei also had a clear impact on North Brushwood. As falling prices pushed more peasants into poverty, absentee landlords secured a grip on land and credit markets. The face of the older village elites was inadequate to deal with such men, and a new group of specialists—tied to landlords as clients—monopolized middleman functions.

Wu's Shop Village . Located just south of Beijing, the village was squarely in the path of armies battling their way to the capital. Like North Brushwood, the villagers here depended on absentee landlords, but here the cause of poverty had more to do with the depredations of warfare and weather than the price of crops. In the late Qing period, Wu's Shop residents cultivated some two thousand mu of land; by 1941, the residents cultivated only eleven hundred mu , of which six hundred mu were owned by landowners living outside of the village. Although some absentee landowners were Wudian residents who had fled to the county capital in the 1920S and 1930s, most of the decline represented villagers' sale of land to landlords living in the county seat, some of whom were officers in the county government.^[54]

Poor harvests, the frequent disorder, and excessive taxation beginning in the early 1920s explain the impoverishment of this village.^[55] In the late 1930s, 77 percent of the fifty-seven families in the village owned less than twenty mu in an economy where twenty-five mu was the minimal standard of living.^[56] Although these peasants supplemented their incomes by working as tenants and wage laborers, most were still compelled to borrow at the end of the year. Thus they were caught in a cycle of debt, the distress sale of land, and further debt.^[57]

In late Qing times, the village elite had managed temple festivals and other public activities such as the crop-watching association. But the demands of the Republican era warlord-armies for taxes and requisitions of food and draft animals were so intolerable that many elite fled to the city. Warlord battles began to ravage the area as early as 1919: first, the battles between Wu Peifu and Duan Qirui, and then three successive battles between Zhang Zuolin and Wu Peifu in the early 1920s. With the Japanese invasion in 1937, the last two village families of some substance also left the

village.^[58] From that time on, village office, especially that of headman, was occupied mainly by reckless villagers who felt that they could squeeze some personal gain from an otherwise unrewarding role.

In Wu's Shop, as in North Brushwood, the need for a guarantor for loans often posed a problem because, by the 1930s at least, it was difficult to find villagers with sufficient property to play the part. Moreover, the precarious financial condition of the villagers made the guarantor's role truly perilous. But life had to go on, and ordinary relatives and friends acted as middlemen in all contracts and even guaranteed small loans for one another. Needless to say, these guarantors were often unable to pay when called on to do so, and cases often ended up in the county courts. According to a former headman of Wu's Shop, the state did not enforce the guarantor provision very strictly, and sometimes let both the borrower and the guarantor off the hook.^[59] Such occasional acts of state paternalism hardly addressed the basic problems of this hapless village. Unlike North Brushwood, where the decline of old elites was at least accompanied by the rise of specialists or professional agents, however slight their value to the peasants, most residents of Wu's Shop had no place to turn but to their own diminishing resources.

A Typology of Middlemen

We have identified three types of middlemen employed in rural contracts: the elite patron, the professional-agent, and relatives and friends. The elite patron with face was propertied, well-connected, and involved in various community tasks. As a middleman, he could often secure a contract in terms favorable to the villager, and for this he did not expect a material reward. The elite middlemen in Hou Lineage Camp and Yang Yuan in Sand Well exemplify this type. It is likely that elite middlemen existed in all four villages until as recently as the 1911 Revolution, and Zhang Yueqing of North Brushwood presents a poignant example of a declining elite patron.

It is not easy to distinguish the particular causes for the decline of elite middlemen in the individual villages. In Sand Well, Du Xiang and Zhao Tingkui probably represented downwardly mobile families in a society where, historically, elite families circulated fairly rapidly. Not many families in the villages of North China could prevent the diminution in their fortunes through the equal partition of family property over a few generations. But in North Brushwood and Wu's Shop the causes of elite decline seemed part of a more secular trend of social change resulting from warfare, commercialization, and state building. We see here the disappearance of elites per se, not the replacement of individual elites by other village elites.

What does this tell us about elite domination over the structures of authority? In Hou Lineage Camp, apparently the elite continued to guarantee the basic needs of the community, and thus maintained the patronage

relationships upon which their authority rested. In Sand Well, although Yang Yuan, the headman, continued to perform important middleman roles, the few other elite men in the village stayed away from these roles, possibly mindful of the fate of the two who had lost money as guarantors in this village. By the late 1930s, the peripatetic peddler, Fu Ju, was handling most land tenure and mortgage contracts. Given that elite involvement in middleman roles in this village was languishing by the 1930, we may conjecture that the community-centered authority structures of this village were weakening.

In North Brushwood, elite participation in brokerage roles had declined by the early 1930s, and the type of middleman had clearly changed. The domination of the village by powerful absentee landlords living in the county seat led to the emergence of contracts secured by professional agents. The professional-agent middleman may have appeared as a response to the accelerated monetization of the region's cotton-based economy in the twentieth century. Indeed, in this respect, this type of middleman resembled the professional middlemen of the Lower Yangzi valley. A study of the middlemen for absentee landlords in Suzhou indicates that they were full-fledged professionals whose job it was to seek out tenants and sellers of land from a catchment area often covering several counties.^[60] Middleman services there had clearly developed into a depersonalized, contractual business, with little room to exercise or develop one's personal authority. Perhaps, people like Fu Ju and the North Brushwood professional-agents, who had little authority in the old world of paternalistic elites, might have come to establish new contractual relationships appropriate for an economy in which depersonalized contracts were replacing face-to-face relationships. Meanwhile, however, peasants continued to view them as agents of powerful outsiders and insensitive to villagers' needs. This made for a very ambivalent relationship.

In Wu's Shop, heavy tax demands and warlord battles left the village with hardly any prestigious or substantial figures to play the roles of elite middlemen. There is not even much evidence of professional agents of the absentee landlords, although they possibly did exist. In the absence of alternatives, friends and relatives performed the roles of middlemen, especially in loan contracts, as best as they could. We can find examples of friends and relatives serving as middlemen in the other villages, especially North Brushwood, where old village elites were also weakening. However, the impartiality of such a middleman who was particularly close to one party in an agreement was always suspect, and he was unlikely to be as successful in his efforts as the elite patron. Aside from the suspicion of partiality, poor friends and relatives did not have the kind of face that elite middlemen could bring to the bargaining table; and their thinly scattered individual connections could not compare with the special access to the market that the professional agent could boast.

Conclusion

This essay has sought to illuminate two themes: the manner in which an elite with no direct control over the means of production establishes its dominance over the structures of authority; and the implications for a community where the elite's exercise of authority is weak or nonexistent. Both phenomena were found in the villages of North China. The peculiar ability of the elite in parts of South China to dominate authority structures came from their direct control over the ordinary peasants' means of livelihood. While northern elites had no such direct control, they were still in a position to significantly affect the livelihood of the peasantry by, among other means, performing the role of middleman in contracts. Peasants' gratitude and obligation for these brokering elites conferred on them a certain authority, exemplified, for instance, in the cultural ideal of face or in the awarding of plaques. In a word, these elite figures were patrons.

S. N. Eisenstadt has analyzed the kinds of societies dominated by patron-client relationships, and the rural society of North China fits his typology.^[61] Patron-client relationships are marginal in developed capitalist societies where the market, backed by the legal code, governs access to resources. They are also marginal in societies where closed corporate groups completely control access to these resources. Patron-client relations typically occur in societies where the market, a major agency regulating the flow of resources, is imperfect and poorly integrated. Perhaps one of the greatest imperfections of the market is the absence of an effective civil or commercial code. This lack reflects the inability or disinclination of the state to guarantee the rights of partners in an economic exchange, thereby allowing their regulation by the authority structures of local systems.

As a consequence of the imperfect market and legal system, the individual peasant or village household often depends upon a powerful local figure, a patron, to ensure the fulfillment of a contract, to provide access to the market on terms not impossibly weighted against him, and to protect him from predatory local government functionaries. In return, the patron receives expressions of gratitude and loyalty upon which he builds a stock of political capital. Thus the relationship between patron and client is marked by reciprocity; though as Alvin Gouldner has reminded us, reciprocity should not be mistaken for equality.^[62] Indeed, in many societies the patron exercises considerable domination over his client, and the attitude of the client toward his patron is correspondingly ambivalent.

In many ways, the principal characteristics of this type of society—an imperfect market and a weakly developed system of formal legal guarantees—were found in North China into the twentieth century. We see this most clearly in Hou Lineage Camp where elite figures played the role of patrons and no doubt used the support of their clientele in contests for power

and office in the village. We see it to a lesser and diminishing extent in Sand Well, but by the 1920s and 1930s we hardly see this phenomenon at all in North Brushwood and Wu's Shop. The reasons for the impoverishment and evacuation of the elite from these villages can be found in complex inter-related factors involving economic decline, dependency, warfare, and state building. What happened to these villages when they were deprived of patrons who could secure their needs, especially in the world outside the village? What was the fate of the structure of authority, embedded, for instance, in the middleman's role, in places lacking elite patrons to lend their weight?

People, of course, continued to use customary law and enter into contractual arrangements to satisfy their needs. They were able to do so because of the nature of customary law in China, which did not absolutely require the regulating authority of a powerful local figure for its operation. The very act of contract violation would make it nearly impossible for the violator to obtain the services of a middleman in the future. Thus we can see whole communities managing to get by without influential elite patrons, but many peasant families probably did not get by very well or very long. Given a society where peasants often faced powerful forces in the markets and political centers, the appeal to elite paternalism could make a crucial difference in their ability to obtain tax remissions or delay rent payments.

In a world with few other subsistence guarantees for peasants, the disappearance of the village elite or the renunciation of its responsibilities could ultimately lead—as it often did—to rebellion or banditry. If so, what might this tell us about elites and the peasant revolution sweeping across the North China countryside in the 1940s? To the extent that powerful elites outside the village were at least partially responsible for the weak position of the peasantry, class interests were an important element in creating revolutionary conditions; but to the extent that communities were without elites who might have been able to shore up the authority of the old system, the absence, rather than the presence, of class differences within the village made for revolutionary conditions. Any explanation of China's peasant revolution will have to consider this shifting nexus of community and class in rural society.

Eleven
Local Elites and Communist Revolution in the Jiangxi Hill Country

Stephen C. Averill

In May 1930 Mao Zedong accompanied the Red Army on one of its sweeps through the hill country of southern Jiangxi. While the army carried out mass mobilization activity in the countryside, Mao remained around the capital of Xunwu (formerly Changning) county for nearly two weeks gathering information on conditions there through conversations with revolutionary cadres, local merchants, and peasants. Mao's long survey report of what he considered a typical part of the Jiangxi-Fujian-Guangdong border region constitutes by far the richest and most detailed contemporary description available of the economy and society of this area during the early stages of the Chinese revolution.^[1]

By this time Mao had already begun to familiarize himself with the life of the hill-county peasantry, and he now wished to learn more about the elites and market towns that mediated peasant access to the outside world. His "Xunwu Investigation" therefore devoted only one section to the conditions of the local peasantry, and barely mentioned the bandits, sworn brotherhoods, and feuds that were an important part of the hill-country social scene. Instead, it provided a rare and painstakingly thorough delineation of the county's trade routes and marketing conditions, and of its mercantile and landlord elite community.

By combining Mao's account with information from other sources, it is possible to obtain a revealing picture of conditions in a peripheral region quite different from the prosperous and sophisticated Yangzi valley areas that have figured so prominently in scholarship on modern Chinese society. In the following pages I examine the local elites of the region during the early stages of the Chinese revolution. More specifically, I trace the transmission of the electrifying impulses of radical political change from their urban ori-

gins to their ultimate point of discharge in the Jiangxi hill-country hinterland and explicate how these impulses affected and were affected by the complex and highly stratified local-elite society through which they passed.

Xunwu County and the Jiangxi Hill Country

Xunwu is located deep in the Nanling mountain ranges of southernmost Jiangxi, astride the administrative and economic divide separating the Gan Yangzi macroregion to the north from the Lingnan macroregion to the south. The county, in fact, has a foot in each region: its northern parts drain into the Gan River, while streams in its southern half feed into Guangdong's East River system and flow into the South China Sea. At some point, merchants (probably outsiders) had taken advantage of these divergent waterways to establish a trade route through Xunwu linking the major regional city of Ganzhou in Jiangxi with cities of the Guangdong coast. In the late nineteenth century the county's modest mercantile community still depended heavily on this route.^[2]

Most of Xunwu's inhabitants were Hakka (guest people) peasants whose ancestors had settled in the region several centuries earlier, and their customs, dialect, and life-styles were similar to Hakka elsewhere in the Jiangxi hill country. Most supported themselves through terraced paddy rice cultivation along valleys and in mountain basins or by the manufacture or growth of a variety of handicrafts and mountain products such as timber, tea oil, and grass cloth. The little agricultural land available in this mountain region was widely but inequitably held: most peasants had at least a little land, but few had enough to support themselves fully, and the bulk of the land was owned by a few individuals, lineages, and associations. Probably most of these wealthy people and groups were from the major market towns of the county that were strung along the river valley trade route. Certainly this was the common pattern elsewhere in southern Jiangxi, where long-established valley communities used their wealth, locational centrality, and historical headstart to dominate more peripheral and recently settled mountain hinterlands.

Lineages were as common in the Jiangxi hill country as they were in nearby Fujian and Guangdong, and they appear to have had the same organization and activities that scholars (including Rubie Watson in this volume) have found in the coastal provinces. Massive, multisegmented lineages with thousands of members and extensive corporate property were common throughout the hill country, especially in established local (bendi ) communities and among Hakka living in the triprovince border region (including Xunwu) known as the Hakka heartland. Coexisting uneasily and often servilely alongside these behemoths were many smaller and weaker lineages.

Deep in the mountains, a few Hakkas and non-Han minority peoples lived scattered in small multisurname hamlets with little or no apparent lineage organization.^[3]

The assorted pressures of living in this difficult environment, the legacy of conflicts engendered by earlier migrations, and the area's general inaccessibility had long combined to make Xunwu and the rest of southern Jiangxi an extremely violent and anarchic region. Bandit gangs, Triad sworn brotherhoods, gambling societies, and sectarian groups were endemic to the area, and massive lineage feuds frequently mobilized thousands of people against each other for years at a time. Traditions of resistance to governmental authority were so deeply ingrained that yamen runners, feared as oppressive "tigers" in many parts of China, were frequently beaten and killed in this area. Even county magistrates dared not venture into the countryside.^[4]

Into this hard, wild country had begun to creep by the early twentieth century, a variety of institutions, products, styles, habits, and inclinations—largely Western in origin or inspiration—that, for lack of a better term, we may call "modern" and that in various ways were altering the lives and attitudes of the region's people. We will discuss a few of these changes, such as a new school system, but most we must pass over, though Mao took obvious delight in discussing the changes at great length, including such minutiae as the introduction of new shoe styles and haircut fashions.^[5] Both trivial and significant changes notwithstanding, by the early 1920s the main structures and processes of Jiangxi hill-country society—particularly rural society—remained essentially intact, and life for most people went on much as it had for centuries.

Stratification of Hill-Country Elites

Presiding over this conflict-ridden society was a remarkably multilayered, many-faceted, and socially volatile group of local elites, whose diversity and rapacity reflected their environment.^[6] These elites were themselves stratified over a rather broad continuum, with members varying substantially in power, prestige, and sophistication. Before examining elite involvement in arenas of particular relevance to the growth of the Jiangxi revolutionary movement, let us first subdivide the elite continuum into several broad strata, each crudely, but for our purposes sufficiently, defined by the locus and range of activity of its members.^[7]

At the upper level of the local elite in a given county stood a small group of "great households" (wanhu ) whose influence or reputation extended throughout the area and sometimes beyond. At least in very rural areas such as Xunwu, these households almost by definition were large landowners, defined by Mao as those having land producing more than five hundred shi (about 45,000 kilograms) of grain per year,^[8] but they also engaged in such a

range of other activities that perhaps they are better viewed as small-scale conglomerates rather than simply landlords. In Xunwu almost all heads of the twenty great households resided in the county, though it is difficult to determine precisely where. Presumably most lived in or near the county capital or largest market towns, though they also typically maintained a strong presence (complete with fortified dwelling) on their respective rural "turfs."^[9]

The Xunwu family of Pan Mingzheng indicates the contours of this group and illustrates the range and extent of its members' influence. Pan himself was by far the largest landlord in the county; his yearly ten thousand shi (900,000 kg.) of grain production ranked him well above the other great households of the county. In addition to his paddy landholdings, Pan also controlled mountain and forest lands, houses and several stores in the nearby market town of Jitan (the county's commercial center). All told, his assets were said to have totaled nearly three hundred thousand yuan (dollars). Moreover, Pan's individual holdings were only part of those controlled by the important lineage to which he belonged: at least six other households in the lineage had individual holdings greater than one hundred mu (one mu equals about one-sixth of an acre), and numerous others had lesser amounts. The lineage had formed an association known as the Rongyang Hall (tang ) to handle its affairs, and Pan's family almost certainly played a major role in administering its corporate property.

Beyond controlling these immediate economic resources, Pan and his relatives wielded influence in other aspects of local life. Though an elderly and stubbornly conservative man himself, Pan had personally run one of the several middle schools that functioned briefly in Xunwu after 1911. One of his sons ran another school, and several of his grandchildren went outside the county, or even the country, to study "the new learning." Two of his sons (one of them a Qing lower-degree holder) also at different times headed the office of financial administration within the county government, and one of them later successively headed the county militia force, joined the Guomin-dang (GMD) county committee, and in 1930 acted as county magistrate. Various members of the family were also connected through marriage to other important figures within the county's upper elite.^[10]

In the multiplicity of his interests and the range of his contacts, Pan was typical of the upper layer of the local elite. This stratum of great households included the largest landlords and merchants and leaders of the most powerful and prosperous lineages. Even when, as in Pan's case, their initial power probably stemmed from landholding and lineage leadership on a rural "turf," almost all the upper layer of the local elite also established a presence in county capitals or strategically located market towns where they developed extensive and varied networks of influence. From these bases they moved to build up ties with elites from other parts of the county and some-

times extended their influence to still higher levels of the central-place hierarchy.^[11] Often, of course, the main attributes of elite status suggested by these examples—wealth, family pedigree, official position, educational accomplishment, and military strength—were interlocking, overlapping, and mutually reinforcing. Wealthy landlords might simultaneously be important lineage leaders and merchants, their sons prominent educators or local administrators, and their daughters wives of other important elite figures.

Often connected in various ways with the "great families" of this upper elite, but considerably more limited in their resources and connections, was a much larger stratum of middling elite individuals and their families. This stratum included medium-sized landlords, defined by Mao as those with lands producing two hundred to five hundred shi (18,000 to 45,000 kg.) of grain per year, merchants, principals, and teachers of Western-style upper-elementary or middle schools, many members of the county bureaucracy, and some militia and police leaders. Although members of this stratum of the local elite had much less impressive reputations than leaders of the "great families," they were often well-known and important individuals within more limited arenas.

One such Xunwu elite figure about whom we have some information is He Zizhen. He's father had begun in humble circumstances, as a clerk in the county yamen's punishment office and later as a geomancer. Then he reaped a minor bonanza by somehow (probably through his old yamen contacts) being awarded local tax-farming rights for the cattle tax. The income from this license was sufficient for He's father to buy paddy land that produced "several tens" of shi in rent.

These resources were sufficient to send He Zizhen to middle school in nearby Pingyuan, Guangdong province, and then to a mining school in Henan. Following his education, He taught school in Xunwu for eight years, during which time he also organized a night school for adults and an association for Xunwu students who had studied in Guangdong. Later he served for a time as county police chief and used the proceeds of his work to buy land of his own. His wealth grew further after he followed his father's lead and in 1927 took up tax farming, in this case obtaining a license to control the sale of beans in the county seat. By this time he had also become a member of the GMD's county committee. Thus, He's family had worked its way to a position of considerable influence.^[12]

In other cases we can see clusters of related middle-elite families that mark the fission products of the breakup of great households. Such, for example, was the case with many relatives of the Chinese Communist Party (CCP) leader Gu Bo, from Tangbei village in the southwestern part of Xunwu. Gu's great-grandfather was apparently already wealthy, though sources provide no details. At any rate, his grandfather was wealthy enough to give land pro-

ducing one hundred shi (9,000 kg.) of grain to help provide an educational endowment for the county and to marry into a large landlord family. An amazing total of twelve men in Gu's father's generation attained the rank of shengyuan (lowest degree) before the imperial examinations were abolished; in 1930 eleven of them were still alive (out of a total of about four hundred shengyuan in the entire county), giving this village of approximately a hundred families by far the highest concentration of such degree recipients in Xunwu.^[13]

Whether degree holders or not, in the late 1920s several direct descendants of Gu's great-grandfather were well-established members of the middle elite. For example, Gu Lesan and his unnamed older brother together owned land producing three hundred shi (27,000 kg.). The brother was a shengyuan who had twice served as a member of the provincial assembly, while Gu Lesan had formerly served with a warlord officer of Jiangxi origin and had been a local official in Guangdong. Among their relatives, Gu Guangrui was an upper-elementary school graduate with land producing more than two hundred shi (18,000 kg.) and a son who had graduated from middle school in nearby Meixian county. Other family members were small landlords or, like Gu Bo's own father, local school teachers. Although this cluster of families (actually probably a branch of a larger Gu lineage) could clearly not match the much more extensive and powerful networks of a great household like Pan Mingzheng's, they were nonetheless a comfortable, respectable community with considerable aggregate resources that doubtless exerted substantial local influence.^[14]

As these examples indicate, the middle stratum of the local elite engaged in diverse occupations and activities but on a somewhat more limited scale than the upper elite. The examples of He Zizhen and Ge Lesan's brother show that members of the middle elite could be well traveled and well educated and could tap into extensive and significant networks of power and influence at the county level or beyond. The other Gus, however, are perhaps more typical of this group in their predominantly rural power base and localistic orientation. Although they were often connected through clientelistic ties with upper-elite local power holders, the primary locus of activity of mid-elite figures was the local market town and its hinterland, and their power was based primarily on local land ownership, kinship ties, and government service.

Coexisting with the middle stratum of the local elite in the countryside, and often even more confined in their sphere of activities, was yet another, much larger, group of local elites. The members of this lowest layer of the elite resided largely in villages and small market towns in close contact with the peasantry. In this rural setting they held various leadership roles: heads of small or branch lineages, militia captains, traditional local school teachers, or village elders. Economically this stratum was quite diverse but generally

much less secure than their counterparts in the upper-elite strata. Many of this group were small landlords, though as was true of the middle-elite figures cited above, such small landlord households could have quite different origins. Mao took pains to divide Xunwu small landlords into two major groups. The first was the "declining households" (poluo hu ), descendants of large landlord families now reduced in wealth and status after dividing family estates. This group (nearly one-third of all landlords) Mao further subdivided into three segments: families that still had a surplus beyond their needs; families that were occasionally forced to sell or Mortgage property to survive; and families that were en route to real destitution. The second landlord group was the "newly emerging households" (xinfa hu ), aggressive and hard-nosed families just risen from the ranks of the peasantry or small merchants. Many members of this group, too, were not well off; indeed, some Communist cadre preferred not to consider them true landlords at all but rather a special subgroup of rich peasants they awkwardly termed "rich peasants of semilandlord character" (ban dizhu xing de funong ). Whatever their precise character, Mao estimated that the newly emerging households comprised almost half the entire landlord stratum.^[15]

Whether small landlords or rich peasants, almost all members of this elite stratum necessarily relied upon activities other than landowning to provide at least part of their livelihood. Around 10 percent of the small landlords in Xunwu, for example, also managed small businesses in local market towns, and the "newly emerging" households were notorious for the rapacity of their money-lending operations. Other lower elites either made money by participating in governmental or quasi-governmental activities such as dispute mediation or proxy tax-remittance (baolan ) or served as collection agents for larger landlords. Some also colluded with bandits and sworn brotherhoods, set up gambling houses, grew opium, or engaged in other illegalities.^[16]

Although the resources of the "great households" were far greater and their reputations far broader than those of lower-elite households, the latter arguably had the greatest cumulative impact on the mass of the population. Certainly in peripheral counties like Xunwu, where great households were fewer and poorer than their counterparts in core regions, lower elites were in much closer contact with the peasantry, and their actions and attitudes were more influential in shaping popular perceptions of elite culture.

The very nature of the lower elites' position on the ambiguous and shifting boundary between the elite and the mass of the population, as well as the limited extent of their power and prestige, has meant that little detailed information has survived about their lives and careers. As one brief example from the upper echelons of this stratum, however, one might cite Yan Guoxing, a small landlord-cum-merchant in southwestern Xunwu who had land producing about one hundred shi (9,000 kg.) of grain, ran a small drygoods

store in the market town of Gongping, and acted as a minor paper merchant. Yan obviously had considerable influence in the local merchant community, for he persuaded them to loan him several thousand yuan for an abortive electoral campaign for the provincial assembly; in 1930 he was still mired in debt.^[17] Another example might be Gu Bo's father, Gu Guangming, a "bankrupt small landlord" who taught at the local village school (sishu ) and had to rely on wealthier relatives to finance his son's middle-school education.^[18] Many other members of this lower-elite stratum, lacking the business or familial connections of Yan or Gu, were forced to rely much more upon their own meager resources.

How large was the local elite in a county such as Xunwu? The question is complicated by gaps and inconsistencies in Mao's data and the lack of population estimates for the county, but by extrapolating from the available information, I have roughly estimated the elite at about 3.5 to 4.0 percent of the population. This figure is probably at the low end of the scale for southern Jiangxi counties.^[19]

In sum, local elites in Xunwu (and by extension, the entire Jiangxi hill country) were a small, diverse, three-tiered collection of families combining considerable access to the outside world with an emphatically localistic orientation, strong emphasis on landowning with substantial commercial activity, and commanding collective dominance with remarkably insecure individual tenure. If in the 1920s some elites sought election to the provincial assembly, sent their children out of the county or the country for "modern" educations, and (as we shall see) joined national political movements, plenty of others were still ignorant of the outside world—men disparaged as "old mountain rats" (shan laoshu ) by impatient Xunwu youths.^[20] Although local elites had hopelessly blurred the traditional distinction between landlord and merchant, they were still in far closer contact with the countryside and its peasant inhabitants than many of the Jiangnan elites on whose lifestyles so many stereotypes about Chinese elites are based. And if the elite as a whole controlled a grossly disproportionate amount of land and other resources in the hill country, that was little consolation to the nearly one-third of all landlords who made up the "declining households," as they plunged through the lower layers of elite society toward reentry into the peasantry.

Elite Power and Conflict

As the foregoing description of elite stratification indirectly indicates, local elites in the Jiangxi hill country maintained social dominance through intertwined webs of power whose important strands included educational achievement; wealth earned through land, commerce, and usury; access to state authority and resources; and private control of the means of coercion. Elite dominance was likewise typically exercised through a variety of institutions,

many of which had a long history in the region. All these various sources and expressions of elite dominance constituted limited resources to which elites had differential access, and keen competition over them often led to violent conflict. It is impossible to discuss here all the ramifications of elite power. Nevertheless it is important to highlight certain aspects of the structures and uses of elite power and of the strife they generated, for it was amid these entrenched institutions and nagging insecurities that the revolutionary movement eventually emerged.

The Maintenance and Exercise of Elite Power . Lineage ties constantly entered into social calculations, and lineages are one of the first arenas to consider when examining how elites maintained dominance. Lineages in Jiangxi, like those elsewhere in southern China, often served as the institutional focus for many activities other than defining and perpetuating kinship relations. Thus lineages commonly owned and managed collective property, ran local schools, helped finance higher education for worthy members, established and enforced codes of social behavior, mobilized armed forces, and sometimes helped the state collect taxes, spread its ideology, and maintain local order.^[21]

Although these functions were carried out in the name of the collective good, elites within the lineage frequently gained disproportionately from them. Most lineage land and other collective property, for example, was managed by elites, who, habitually used their positions to profit financially in various, mostly illicit, ways. Elites likewise appear to have received disproportionate benefit from lineage-run schools and scholarship funds because their children were more likely than peasant youths both to attend school and to seek advanced education outside the community. Lineage codes of conduct generally expressed sentiments congenial to, and were interpreted and enforced by, lineage elders who were usually also local elites; they also mobilized and directed (often for their own purposes) armed lineage forces and served as intermediaries between lineage members and the state. Controlling the means of coercion and opportunities for mediation further enhanced the power and prestige of local elites.^[22]

Whether in lineages or other arenas, elites differed among themselves in how and why they exercised power. Money lending provides a clear example of these differences. Hill-country elites of all strata made loans, but their ways and purposes differed. The few upper and numerous middle elites who made loans wanted either safe investments or (in the event of default) the chance to acquire property used as collateral. Therefore, they loaned relatively large amounts at moderate interest rates, mostly to lower-level elites rather than directly to peasants. Lower-level elites, particularly the aggressively entrepreneurial "newly emerging households," sometimes sought such loans as capital for small-scale commercial enterprises. More often, however, they sought interest income and so reloaned the money at moderate interest

rates to other lower elites or in small amounts at high rates to peasants. Lower elites who controlled small lineage trusts or other collective property also often loaned part of the resources they managed to peasants, expropriating part or all of the proceeds. The variable amounts and interest rates in this elite loan "industry" were doubtless based primarily on relative investment security, but they probably also helped strengthen patron-client links among elites of different strata; they certainly helped perpetuate overall elite economic domination of the peasantry.^[23]

Similarly differentiated patterns of elite activity almost certainly existed in other institutional contexts—such as management of different public bodies (gonghui or gongtang ) or relationships to the state—though it is difficult to determine their precise extent. Merchants and landlord/merchant hybrids, for example, often dominated the management of temple associations and other organizations in market towns. Members of the middle and upper elite probably managed most large lineage and social welfare associations, and small branch-lineage associations appear frequently to have been dominated by lower-level elites.^[24] Likewise, upper-stratum or urban-oriented elites doubtless had much better access to benefits provided by the state (and may therefore have reciprocated with greater support) than did the more rural lower-stratum elites.

How does this pictures of elite structure and power relate to recent scholarly emphasis on the long-term stability and continuity of the local elite?^[25] According to this scholarship, elite power and status were more frequently and reliably attained and maintained via landholding, lineage development, and other localistic methods than through the rewarding but unpredictable route of examination success and high bureaucratic office. Lending support to this view we see in Xunwu (and doubtless other southern Jiangxi counties as well) that some surname groups persisted for centuries in the elite-centered historical record despite the county's poor overall record in the examinations and despite the underrepresentation in the degree-holder lists of some of the county's most powerful and well-known surnames.^[26]

However accurate the general argument for continuity may be, the frequent references to "former great households" now fallen in wealth and status that fleck Mao's accounts of the Xunwu middle elite, and the large number of "declining households" he found in the county's lower elite, suggest that we must be wary of viewing the complex question of elite continuity too simplistically. Most analyses of elite mobility have studied imperial rather than twentieth-century elites and dealt with lengthy time spans. They have also focused on prosperous core regions, and their authors have been forced by the nature of their sources to give more exclusive coverage than does this essay to the great-household stratum of the local elite. Mao's comments remind us that in more peaceful times and more prosperous areas, some rise and fall of elite households must have occurred, even within the

most dominant and enduring lineages or patrilines. In troubled times and out-of-the-way places, and among the volatile lower elites whose lives are so much less well-documented, both the actual occurrence and (equally important as a factor in elite behavior) the subjective fear of downward mobility must have been a constant fact of life.

Intra-elite Factional Conflict . As one might expect given the complex classification of hill-country elites and the intense pressures to obtain scarce resources and avoid the disastrous consequences of family fragmentation, elite relationships were often characterized by competition and strife. Although information on it is limited, this intra-elite conflict is important to our understanding of political behavior in the hill country and demands at least a few remarks.

Let me again begin with some examples. In Chongyi county, in southwestern Jiangxi, two major elite cliques struggled with one another during the 1920s. Adherents of one clique lived in the vicinity of the county capital and the market town of Yangmei. According to the Communist reminiscence that is our main source on the subject, this group had long "colluded" with county magistrates and had a firm grip on all "public property." The other clique was based in a group of market towns on branches of a river system separated from the county capital by a range of hills. In the mid-1920s this clique, considered more progressive than its rival, supported the local GMD organization.^[27]

In Yudu county north of Xunwu there were also two cliques, the Changcun and the Yushui factions, named after the middle schools that served as their headquarters. According to another CCP leader, the Changcun faction was based in parts of the county—largely in the east and north—where lineages were large arid strong, and its adherents were the large landlords and "local despots" (tuhao ) who controlled these lineages. The Yushui faction had its strength in the county capital and in the western and southern sections of the county, where lineages were relatively weak. This faction included many middle and small landlords who were also engaged in commerce. Around these cores, each faction gathered various "poor intellectuals" and students, presumably recruited from their respective middle schools. Both cliques competed for influence within the county government and access to profits from proxy remittance and pettifogging.^[28]

In Xingguo county just north of Yudu, factional conflict also revolved around different middle schools. Here a coalition of elites from around the county capital controlled the county's public middle school, while rural elites, organized into a body called the Federated Township Self-Government Assembly (Lian xiang zizhi huiyi) controlled a nearby private middle school. The city elites had better access to successive county magistrates and with their support obtained a larger share of local spoils and

forced the rural elites to pay a disproportionate share of local government costs. As in Yudu, students in the local schools became involved in the factional struggle started by their eiders.^[29]

Several points emerge from these examples. First, elite conflict focused on the county capital and government, most obviously because the magistrate's yamen and the center for such government functions as public security, tax collection, and education were in the capital. Elites interested in obtaining government office, profiting from tax collection or litigation, or protecting their private activities from government functionaries or other local elites had good reason to compete for influence with the county bureaucrats. For their part, officials also inevitably became more involved in local elite politics. Because Qing magistrates could not be regularly appointed to serve in their own provinces, magistrates serving in Jiangxi were unfamiliar with local conditions and perforce relied upon local elites to help them govern. After the fall of the Qing, people often served in their own provinces, but magistrates in Jiangxi were rarely assigned to their home counties and still required local advice. Magistrates lived and worked in the yamen, so they were naturally most likely to come in contact with members of the elite who lived nearby, a fact that gave such elites definite advantages in local factional struggles.^[30]

Second, economic and social centrality of the county capitals and their immediate surroundings also made them the focus of factional strife. Although some county seats in Jiangxi, including Xunwu, were not the largest commercial centers in their counties, the majority were. Virtually all were well located along transportation routes, surrounded by fertile and densely populated lowland paddy lands. By contrast, much of the remaining hill-county territory consisted of rugged mountain country, with difficult access, low productivity, and sparse population. Moreover, the county capital was generally the site of the region's largest temples and lineage halls and its most prestigious schools. The vicinity of the county capital was thus likely to be both a base of important elite interest groups and a prize to be competed for. In many Jiangxi hill counties, elite conflicts also replicated and intensified general and long-standing divisions between urban and rural or between established communities and later-arriving "guest people" (Hakka) immigrants.

The Yudu county example also indicates another type of intra-elite competition, namely interlineage factionalism and feuding. That such feuds were often really disputes between the respective lineage leaderships rather than the mass of lineage members is underlined by a conflict in Ruijin county, where in 1923-24 two upper elites quarreled over whether to establish new local transit-tax stations and eventually mobilized more than a thousand men from each lineage to settle the issue through armed conflict.^[31] At least as common as open fighting between leaders of large lineages, however, were

situations similar to the Yudu case: one or more large lineages and their leaders seeking to dominate surrounding areas inhabited by many small lineages.^[32]

The Yudu case is particularly intriguing because social and economic attitudes seem linked to elite type, with large lineage leaders representing older, more "feudal," landlord attitudes and the small landlords-cum-merchants opposing them embodying more modern "bourgeois" sensibilities. Agriculture and handicrafts in the Jiangxi highlands had been highly commercialized well before the twentieth century, and elites at all levels were substantially involved in trade; thus, we must be careful not to overdraw such distinctions. Nevertheless, as we shall see shortly, elite attitudes toward revolution were often affected by their economic and social circumstances, and it is plausible that some such factors influenced earlier elite factionalism as well.

Networks of Influence for Elite Revolutionaries

Although elite factionalism was surely influenced by socioeconomic distinctions between urban and rural elites, or landed versus commercial wealth, it is notable that education was a key arena for political conflict in both Yudu and Xingguo. This politicization of conflict in education was quite natural. Education and scholarship had long been linked to elite status and political activity in China. Moreover, in the modern era, most early CCP members were intellectuals from elite families, politicized during the student activism of the May Fourth era. Back in the hill country, these revolutionaries exploited connections within the factionalized elite "educational circles." But they also came to rely upon quite different networks of influence—hill-country bandit bands and sworn brotherhoods. With these connections in mind, let us first investigate the role of elites in hill-country education and then discuss elite ties to bandits and brotherhoods.

Hill-country Education . Although Jiangxi hill counties produced few successful candidates in the imperial examinations, traditional education there had many of the same social effects as in other parts of China. Mastery of the Confucian classics, measured by success in the imperial examinations, was for centuries a major criterion of elite status, and a large, ramified, and conservative institutional apparatus had grown up at all levels of society to support the quest for examination achievement. Moreover, education in imperial China, as in most other societies, provided young people with opportunities to expand their intellectual and social horizons. Aspiring students had to travel periodically to county, prefectural, and provincial capitals for examinations, and most spent some time in local or regional academies (shuyuan ). Both gathering for examinations and attending academies normally involved boarding away from home in company with other wealthy and/or

bright young men from different localities, and the experience frequently led to lifelong friendships and student-teacher ties.

Abolishing the examinations in 1905 and developing a new Western-style educational system significantly affected existing elites. With a few strokes of the bureaucratic brush, the Qing leadership abandoned the entire formal process by which many of its own members had been selected and threatened to negate the educational efforts of millions of younger Chinese currently involved in the system. Within a few years, Western-style schools became disseminators of new ideas and incubators of political agitation instead of guardians of intellectual orthodoxy.^[33]

The important new impact of the Western-style schools, however, should not blind us to the many institutional and attitudinal links they maintained with the past. Many of the first new schools were simply renamed academies, located on the grounds and inheriting the endowments of their predecessors. Like the academies, the most successful and prestigious of the new schools remained clustered in the largest administrative centers. After the imperial examinations ended, teaching in the new schools was one of the few remaining respectable job opportunities for many Qing degree-holders, and some of them remained influential in local educational establishments well into the 1920s.^[34]

Moreover, although the avowed aim of education changed from preparing students for the examinations to teaching them the skills that had made the West strong, the main concern for most people continued to be education's role as a route to individual and family—not national—wealth, power, and prestige. Possession of a modern school degree still brought special community respect; and modern schooling was either an important passport out of the hill country or a route to acquiring important jobs in one's own county bureaucracy or educational establishment. Moreover, a school degree, unlike land, was a portable resource that could be used in different places and arenas of elite interaction. It might well be that in the twentieth century school degrees became more easily renewable from generation to generation than were landholdings, enhancing the effectiveness of education as an elite strategy to help ameliorate the effects of partible inheritance.

Management and certification of education had, however, changed. No longer did the government set tight quotas on the number of students per province who could receive, say, a middle-school degree or determine the subject matter to be studied. Although there was still some governmental regulation of education (particularly of provincially run middle schools) and many schools still had entrance examinations, now school leaders themselves much more firmly controlled education.^[35] Greater local control over access to a still-important educational process appears to have been at least as significant as the prestige of a scholarly vocation in explaining the continuing power of educators in local-elite affairs during the 1920s.

Control over educational access was aided by well-developed networks

based on school experience and student-teacher ties. Ties of this sort had been important under the imperial examination system, but their influence was perhaps even greater during the Republican period. Local educators who had graduated from prestigious urban middle schools facilitated their brightest protegés' entrance into the same schools and then frequently hired them for county and subcounty schools after they graduated. Even when students pursued different occupations they often maintained contact with their old teachers. Through these networks of contacts both within and without the educational system, local educators could exert influence on, or call for aid from, people in a wide variety of places and occupations.^[36]

The value of education as a route for career advancement, the position of local educators as brothers controlling access to influential networks of personal connections, and the important role played by schools and scholars in the political discourse of the late nineteenth and early twentieth century together made educational circles a prime arena for the widespread elite contention and factional struggle previously noted.^[37]

By the 1920s, then, local schools throughout Jiangxi were ambiguously modern institutions in which new concern for Western culture and the abhorrence of Western imperialism coexisted with older conceptions of the function of education in the social order and the role of educators in local politics. In the southern Jiangxi hill country these diverse currents were accentuated by problems of peripherality: the region's distance from better-developed centers of culture and communication enhanced the importance of local schools as one of the few sources of information about new ideas and trends and one of the few routes out of the hill country for ambitious young people seeking rewarding personal careers and new resources to maintain elite status.

Ironically, the deficiencies of the hill-country schools themselves enhanced their importance as links to the outside world. The overall poverty of the hill country meant that it was difficult to find funds to build and maintain schools, especially higher-level, Western-style schools. The density and quality of such schools in a given area was therefore generally low, and students were forced either to use them as stepping-stones to further training elsewhere or to bypass them entirely. Most children still began their education in old-style village or lineage schools (sishu ), often taught by aging Qing lower-degree holders; education beyond this level was mostly confined to children of local-elite families. Only a few counties in the Jiangxi hill country were able to maintain local middle schools during the 1910s. In Xunwu, for example, four middle schools had opened before 1930, but all had closed quickly (three within a year of opening), and most students seeking a middle-school education either went to Ganzhou or crossed the border to Meixian or Pingyuan in Guangdong.^[38]

According to Mao's research, elite attitudes toward the new schools

varied somewhat depending on age, geographical location, and position on the continuum of elite wealth and status. Most patriarchs in upper- and middle-elite families were willing to manage new schools, serve as county educational officials, and send their children away to middle schools, but they were motivated more by desire of personal power and profit than by enthusiasm for the new education. Those upper elites who lived close to major market towns or near the riverine trade route connecting the county with Guangdong, Mao found less conservative—as were the children and grandchildren of conservative stalwarts like Pan Mingzheng. Given the substantial resources of most upper- and middle-elite families and the length of time they could afford to wait before receiving benefits from their children's education, it is not suprising that three-fourths of the county's university graduates and two-thirds of those who studied abroad had this sort of background.^[39]

Mao also found that attitudes toward education varied within the lower stratum of the elite, depending on whether a familiy was "declining" or "newly emerging." Declining lower-elite households embraced the new education almost desperately. Sometimes they were following family traditions: the bulk of the county's surviving lower-degree holders were apparently small landlords and/or village schoolteachers. Often, however, creating an educated "man of talent" was viewed as a last chance to revive sagging family fortunes. A high percentage of children from such families therefore attended upper-elementary and middle schools. If the declining families looked fearfully to the future, newly emerging elite households anxiously recalled their recent past; they tended to retain the aggressive, single-minded pursuit of short-term profit that had just enabled them to scrabble a bit above their former fellows, and they saw more immediate and certain return on their hard-earned capital from today's usury than from tomorrow's middle-school graduate. Children from this group were thus educationally underrepresented in the county.^[40]

As we shall see shortly, these differing elite attitudes toward education paralleled their differing stances toward revolution. Before we can properly discuss the interaction between education and revolution, however, we must consider briefly a very different arena of elite concern, equally important for the eventual growth of a revolutionary movement: the world of "the brothers of the greenwood," the bandit gangs and sworn brotherhoods endemic in the hill country.

Bandits and Brotherhoods . For all that the schools of the Jiangxi highlands suffered by comparison to their counterparts in higher-level urban centers, they were still indisputably part of the respectable mainstream of hill-country elite life. The bandit and sworn brotherhood gangs that swarmed through the hill country, on the other hand, seemed clearly and unambiguously to belong to the world of the peasants, and to its most disreputable

segment at that. After all, the gangs were primarily composed of peasants. To the desperately poor, banditry offered a last chance to make a new life or prolong a miserable existence. To the able and ambitious, banditry constituted one of the few routes for social mobility open to those starting without wealth and family connections. And to peasants exasperated and discontented with the arbitrary injustice so prevalent in their lives, banditry provided an opportunity, however limited, to strike back at their oppressors. In each function, banditry remained apart from, and potentially antagonistic to, the elite political system.

Yet there were connections as well. Hill-country bandits and brotherhoods by no means totally divorced themselves from local society, including its elites. Economically, hill-country bandits supplied their everyday needs, fenced their plunder, and spent their profits in local market places. Socially, they remained in close contact with family and friends in their home communities, even while they (or at least their leaders) acquired the wealth and power that set them apart from the populace. And politically, the gangs organized armed forces that intervened in factional disputes and occasionally vied for control of local government.^[41]

In all these areas the gangs interacted with the elite power structure. Indeed, southern Jiangxi elites sometimes had such close and symbiotic ties with the region's bandit and brotherhood organizations that officials found the two groups difficult to distinguish. Some elites established sworn-brotherhood or patron-client relationships with bandit leaders, covered for sworn-brotherhood gambling operations, sold goods to and fenced booty from gangs, enlisted gangs to browbeat the local populace or intimidate rival elites, or recommended gang leaders for bureaucratic appointment.^[42]

As with other aspects of elite behavior, relationships among elites and gangs varied according to the situation and the self-interests of the elite members involved. Upper and middle local elites—wealthier, further removed from the countryside, and more sympathetic to official concerns than the lower elites—tended to be more antagonistic toward "heterodox" groups. Their persons and their property were, after all, prime bandit targets. Lower elites, on the other hand, were often much closer to bandit leaders in sympathies, activities, and origins. Both lower elites and bandit leaders were often recently, and only barely, removed from the peasantry; both were, to use Eric Hobsbawm's phrase, "men who made themselves respected."^[43] Lower elites themselves sometimes became bandit leaders, while particularly effective bandit leaders might be rewarded with militia commands and other perquisites of elite life.^[44]

Nevertheless, despite the attempts by local groups to draw gangs into their orbit and under their control, to coopt gang leaders and profit from their activities, bandit and brotherhood activity in the Jiangxi hills remained at least partially distinct from the elite-dominated political and economic

system. It is precisely this ambiguous quality of gang activity, simultaneously associated with and yet separate from elite life, that compels our attention here. Just as educational circles addressing both long-standing needs and new concerns served an important transitional function that facilitated a nascent revolutionary movement's entry into local elite society, so the hill-country "bandit world" later played a corresponding role in facilitating the revolution's movement out of elite society and into the wider peasant world. As powerful peasant-based armed bodies outside the mainstream of elite life, yet tantalizingly susceptible to elite manipulation, bandits and brotherhoods were logical objects of attention by Communist cadres of elite background seeking to enter and transform rural society.

Elites and the Origins of Revolution

In 1921 the young teacher, He Zizhen, formed a Guangdong Schools Alumni Association (Liu Yue xueyou hui) in Xunwu to organize the many local youths who, like himself, had studied there.^[45] Gu Bo, once He's student in elementary school, joined the group later while attending middle school in Guangdong, where he also became a Communist. In the mid-1920s Gu and other young radicals split with the alumni association to form a separate organization that eventually emerged as the core of the county's revolutionary movement. This group, rather misleadingly named the Xunwu Common People's Cooperative Society (Xunwu pingmin hezuo she), recruited more than one hundred followers in Xunwu's educational circles, founded the Zhongshan (i.e., Sun Yat-sen) School, and began mass movement activity. In response, He Zizhen and several conservative associates formed a rival body known as the Young Revolutionary Comrades Association (Qingnian geming tongzhi hui) and founded the Xinxun (New Xunwu) School to compete with the Zhongshan School. People referred to these groups as the Cooperative Society Clique (sometimes called the Zhongshan clique) and the New Xunwu Clique.

These groups were formed during the 1923-27 Guomindang-Communist Party United Front that culminated in the Northern Expedition and the nominal unification of China under the Nationalist government of Chiang Kai-shek. Events in the hill country unfolded against the backdrop of this national revolutionary upsurge and the accompanying conflict within the United Front that led in 1927 to the GMD's break with the CCP and the suppression of radical elements.

Competition between the two groups heated up as the alliance between the CCP and GMD gradually deteriorated. By mid-1927 the conservative New Xunwu Clique appeared to have the edge, aided by the spreading "white terror" that had violently suppressed mass movements and purged radical leaders across the province. Nevertheless, the Cooperative Society

Clique retained control of the Zhongshan School, and its members for a time operated relatively openly as they sought to make the transition to a more rural stage of the revolution.

As part of the transition, the radicals established several rural branches of the Zhongshan school, which they used to propagandize among the peasantry and contact a local Triad sworn brotherhood known as the Three Dots Society (Sandian hui). After Gu Bo and other CCP leaders themselves entered the brotherhood, its lodges around the country joined the schools as centers of secret preparation for an armed uprising that occurred in March 1928; students and brotherhood forces coordinated attacks on the Xunwu School and the government yamen in the county capital, and other Communist-led forces attacked elites in the southern part of the county where Gu Bo's family lived. Although initially successful, the uprising was soon crushed by government troops and forces of a local bandit working in collusion with threatened local elites.

Despite its failure, the attack clearly marked the transition of the struggle between conservative and revolutionary elites in the county from a conflict waged largely by factional groups based in the mainstream elite educational institutions to a violent civil war with widespread peasant participation. Following their defeat, Communist forces retreated, reorganized, and built a base in the southern part of the county. By the time Mao visited in 1930, most of the county was in Communist hands, land redistribution and other radical reforms were well underway, and the stage was set for the eventual incorporation of the region into the emerging central soviet.

Even this brief account of events in Xunwu sufficiently reveals the close but complex relationships between the emerging revolutionary movement and several enduring institutions and processes—the educational system, bandit and sworn brotherhood gangs, factional strife—important to local elites. To explore these relationships more fully, however, it is necessary to move for a time beyond Xunwu to discuss these and other aspects of elite society in the context of the revolutionary movement in the hill country as a whole.

A distinct Communist-led revolutionary movement in Jiangxi first appeared during the mid-1920s in the province's elite-dominated school system, following and building upon the earlier use of local educational circles for disseminating radical ideas and forming new political organizations. After the May Fourth incident of 1919, students throughout Jiangxi had formed study groups and associations to further both their understanding of the "new culture" and their ability to exert local political influence. Later, as vague radical sensibilities were channeled into formal party affiliations, many early CCP leaders returned as teachers to organize party branches in the local schools from which they had graduated or in new schools established in small towns and villages. Evening classes at some of these same

schools provided opportunities for early efforts to organize the peasantry. Eventually after the collapse of the GMD-CCP alliance in 1927, schools also functioned as way stations for Communist cadres seeking to reenter hill-country society, centers for propaganda and recruitment, and headquarters for armed uprisings.^[46]

If the elite-dominated school system thus nurtured the revolutionary movement, then the pervasive factionalism afflicting hill-country elite society also simultaneously shaped it. In the 1920s, factionalism in education generally took the form of school-centered struggles between young and liberal "new cliques" (xinpai ) and "old cliques" (jiupai ) dominated by more conservative and elderly educators. During these struggles, smaller factional groups similar to the Cooperation Society often formed within the larger ill-defined cliques, coalescing on the basis of hometown, kinship, or school ties, common ideological interests, or patron-client connections to a particular leader. Over time such groups became both the scaffolding within which even smaller party branches could be organized and the institutional vehicles for some early assaults on the established power structure.^[47]

At this point, communist attacks on elite powerholders were still well within accepted parameters of elite activity and still largely indistinguishable from the background clutter of local-elite factionalism. But factional tactics were not merely calculated ploys used by the Communists to disguise their intentions and cover their tracks. They were also expressions of a deeper and less conscious style of elite political behavior absorbed from the environment in which the young CCP leaders had grown up, and as such continued to manifest themselves from time to time throughout the early history of the revolutionary movement. The CCP leaders' temporary alliances with armed local powerholders after 1927, the continued influence of distinctions like urban/rural or large lineage/small lineage on the development of base areas, and the numerous internal disputes that plagued the communist leadership, all reflected the revolutionary movement's persistent tendency to both expand and fracture according to patterns of political interaction typical of hill-country elites.^[48]

Other institutions important to the conduct of elite politics—bandit gangs, sworn brotherhoods, and lineages—also became involved in similarly complex ways in the emerging revolutionary movement. In its initial stages gangs and brotherhoods that had habitually colluded with established elite power holders frequently continued to support their erstwhile patrons by helping to attack Communist party branches and mass movement organs. Especially after 1927, however, the dynamics of the situation changed: conservative elites relied more on government troops and personally raised militias, and communist leaders found alliances with gangs among the few available alternatives to the now-disbanded peasant associations.^[49]

Communists recognized that in addition to military striking power the

gangs also provided useful and familiar routes for contacting the peasantry and ready-made organizational nuclei around which peasants could be mobilized. Many CCP cadres, therefore, went beyond simply negotiating alliances with gang leaders—a method that produced quick but often impermanent results—and sought to win over gang members from within. They became initiated into gangs and brotherhoods themselves and assigned CCP political agents to the gangs to help "reorganize" them. In this way cadres both established close ties with potential peasant recruits and helped mitigate the dangers of intrigues and mutinies by gang leaders. Eventually most gangs were fully incorporated into the rapidly expanding Red Army, with their original leaders either firmly committed to the Communist cause or (more frequently) replaced by people who were.^[50] CCP leaders also tried persistently, though with only moderate success, to curb the power of lineages. Generally larger, more cohesive, and more permanent than gangs, lineages were also solidly rooted and widely ramified institutions of elite dominance. Any revolutionary restructuring of Jiangxi rural society required destroying, or at least neutralizing, these centers of elite power. That Communist cadres made some progress toward this difficult goal is clear, but the scarce available sources do not clearly reveal their methods.

It is reasonably certain, however, that CCP cadres initially sought whenever possible to accommodate and take advantage of lineage ties rather than force any immediate confrontation. Thus cadres sometimes commanded attention, obtained protection, or appealed for support on the basis of their prominent surnames. They might also use their status as scions of elite families with major lineage branches, or, conversely, arouse relatives to redress wrongs done to their own poor lineage by some nearby giant oppressor. This latter practice was apparently particularly effective. Cadres united numerous small lineages to oppose, under the banner of the revolution, the largest and most dominating lineages and their elite leaderships.^[51]

Because hill-country elites played such a prominent role in the early stages of the revolution, we must ask where, within the highly stratified local elite, support for the revolution was strongest and where resistance to it was most determined. The answer to these complicated and different questions appears essentially the same: the lower stratum of the hill-country elite.

We have seen that members of different elite strata differed markedly in their access to public bodies and bureaucratic officials, their connections with bandit gangs and brotherhoods, and their attitudes toward education. We might, therefore, assume that many lower elites would also support at least some measure of political challenge to an upper-elite power structure from which they derived relatively little profit. Mao supports and elaborates on this hypothesis in his "Xunwu Investigation," asserting that elite attitudes toward the pace of political change in hill-country society paralleled their attitudes toward the new education. Those with substantial wealth and considerable security were basically conservative, though most would accept or

even promote some change if it worked to their own advantage. Lower-level elites just emerging from the peasantry had little time for politics but clung like bulldogs to every shred of economic and social advantage they had accrued within the existing system. And declining lower-elite families were most anxious to see changes, political or otherwise, that would arrest their downward social slide.^[52]

Mao's assertions receive empirical support from both the evidence he provides on the differing memberships of the New Xunwu and the Cooperative Society cliques in Xunwu and biographical information collected on other Jiangxi revolutionaries. Mao confirms that members of the Cooperative Society were predominantly scions of declining families from the local elite's lower stratum, leavened with a few progressive middle or upper elites.^[53] Less detailed biographical data on other Jiangxi revolutionaries generally supports this picture of a revolutionary leadership drawn mainly from the lower-elite stratum, especially its "declining households."^[54]

If the declining households provided much early elite support for the revolution, the newly emerging households of the lower elite appear later to have become its most stubborn and effective opponents. The revolutionary movement originated in the intellectual ferment and factional strife of the Jiangxi educational system, an arena that lay, both by choice and circumstance, largely beyond the purview of the newly emergent elite households. This situation changed, however, once the revolutionary movement percolated into the countryside. With their hard-won and tenuously held positions now directly threatened, and lacking the resources that made flight a viable option for higher-level elites, the newly emergent elites had little choice but to fight, and their resistance was widespread and tenacious.^[55]

In sum, the lowest of all the elite strata was simultaneously the most open and most resistant to change: most open because its many declining households were willing to tamper with a status quo that offered them little prospect but further decline; most resistant because its newly emerging households were unwilling to jeopardize the smallest morsel of their hard-won gains. The dual nature of the lower elite—both facilitators of and obstacles to change—made dealing with this group one of the most delicate and exasperating problems facing the Jiangxi revolutionary leaders.

Conclusion

Most studies of the Chinese revolution have given little systematic attention to the role of local elites, dismissing as a minor irony the fact that numerous Communists came from prosperous families and treating elites almost exclusively as targets of the revolution.^[56] As we have seen, however, elites in southern Jiangxi instigated as well as obstructed social change: revolution was disseminated via elite-dominated schools, structured in elite-run organizations, shaped by elite-centered patterns of factional politics, and aided by

elite-influenced bandits and brotherhoods. Of course, few elites fully supported the revolution throughout its course; some resisted from the start, and many others joined the opposition as the movement's challenge to the established order deepened. Nonetheless, the growth of the revolution in rural Jiangxi—and elsewhere in China—depended far more and far longer than commonly realized on a support structure provided by local-elite society.

In contrast to many China experts, scholars of comparative revolution have long recognized the importance of elite action, but they have generally discussed elites only in macrosocietal terms, as abstract, largely undifferentiated groups acting on national or even international stages.^[57] This essay, however, has looked at revolution in a local context and emphasized elite diversity rather than uniformity. Elites in southern Jiangxi were far from the unitary social category still often implied in discussions of "the local elite." They were, in fact, a diverse, multitiered collection of people who varied considerably from one another in resources, attitudes, and roles in the established power structure. Recognizing these differentiated layers allows us to appreciate more clearly the nature of the interlocking, highly articulated elasticity of local-elite society that for so long absorbed the shocks and transferred the energy of contacts between the peasantry and the outside world.

Understanding local-elite differentiation in Jiangxi also invites closer attention to the pivotal role played by Chinese elites occupying the liminal social terrain near the imprecise boundary separating elite from peasant. In the 1920s and 1930s, Communists looking at Chinese society from the bottom up frequently spoke of "rich peasants of semilandlord character"; officials viewing the same scene from the top down talked of "local bullies and rotten gentry" (tuhao lieshen ); this essay has described both "declining" and "newly emergent" elites. Although these various terms are imprecise and by no means synonymous, there is certainly much overlap in the groups they represent; their zone of convergence largely encompasses those members of the lowest, largest, and most volatile stratum of the local elite, whose attitudes and actions were crucially important to all who sought to change the contours of Chinese rural society.

Mao Zedong was just such an individual; and more than the simple curiosity of an admittedly avid observer and analyst of Chinese life, the pressing, practical worries about revolutionary policy doubtless motivated his inquiries in Xunwu in May 1930. Nonetheless, his concern then was, on one level, much the same as ours is now: comprehending the rumbustious society of the Jiangxi hill country, and most particularly the jostling, arguing local elites who wrapped the region in such strong but anxious embrace. The "Xunwu Investigation" is testament to the significant progress he made in this effort, and it remains today a vivid and valuable record of a vanished way of life.

Concluding Remarks

Mary Backus Rankin and Joseph W. Esherick

The eleven chapters in this volume provide a picture of diverse and changing Chinese local elites. Clearly we cannot capture the essence of these elites in such simple static definitions as "gentry," "scholar-officials," or "landlords." But if we conceive of elites as the people exercising dominance in local arenas, we can describe a coherent range of different patterns of dominance as a variety of elites adapted their strategies to the available resources. That provides a broader basis from which to consider the questions raised in the introduction. How much mobility was there in traditional elite society? What were the natures and resources of late imperial Chinese elites? What strategies did elites use to dominate other groups in local society? We can then move to questions of how these elites changed over time and to the still critical issue of elites and the state in China.

Although the local focus calls attention to diversity, the articles in this volume do not inspire images of endless fragmentation and disorder. Instead they suggest that social patterns in China were shaped less by the political center and woven more in local society than the usual picture of a uniform gentry-elite suggests. Local patterns varied markedly, but the elites in this volume, following their own routes and using the available resources, created institutions, expressed cultural values, used symbols, and interacted in ways generally accepted in Chinese society. The emphasis thus shifts from the question of how government controlled local leaders and unified society to issues of how local elites, acting within Chinese historical and cultural contexts, dominated local arenas and interacted with elites in other arenas in ways common to the larger society and culture. Most institutions under local leadership—lineages, poetry clubs, guilds, or militia—were not neatly linked to hierarchies ending up in the capital but were shaped at home and joined to other localities and to government in looser and messier ways. By examining

elites in their local contexts, we can work from the bottom up to identify the resources and strategies they employed to maintain their dominance.

One fundamental factor shaping the strategies of Chinese elites was their relative insecurity. Elites in China had few reliable long-term guarantees of their status compared to old-regime elites in Europe or to elites in caste societies like India. The European nobility, especially on the continent, was a hereditary status group with legal guarantees, rights to landholding, and tax privileges that far exceeded those available to the Chinese gentry.^[1] Although recent research suggests that the long-accepted contrast between primogeniture among European elites and partible inheritance in China has been overdrawn,^[2] there is little doubt that rather strict Chinese adherence to the practice of dividing property equally among sons made it more difficult to maintain the family patrimony than in societies with such customs as entail and "strict settlement."^[3]

Acquisition of at least the lowest examination degree, with its attendant legal privileges, was the surest reasonably achievable route to elite status in late imperial China. But this status was marginal in many areas—a shengyuan degree per se provided little prestige in prosperous Jiangnan, with its abundance of higher-degree holders—and it could not be passed on to heirs. Furthermore, and again unlike caste or estate societies, status in China was not an effective substitute for wealth because in the long run it depended on wealth—to educate heirs for the examinations, cultivate connections, and maintain an appropriate life-style. The cultivated gentry life-style set elites apart from commoners. Although the examination system and gentry networks were important in spreading the social practices of the literati throughout China, these life-styles could mark a family as elite through long periods without any examination success.^[4] The styles were remarkably uniform throughout the country, spread by both the examinations and officials and the sojourning merchants and scholars. By the end of the nineteenth century even the frontier elites of Taiwan, Sichuan, and Guizhou were adopting such ways of life.^[5]

To determine the success of Chinese local elites in devising strategies to maintain their dominance we must distinguish between the continuity of elite personnel and the social continuity of elites. Continuity of personnel means that specific families maintain their elite status over long periods of time even though the nature of the elite they belong to may change. Social continuity refers to the persistence of a particular elite type, although the families belonging to that elite may change over time.^[6] Previous studies of Chinese gentry have emphasized its social continuity. We later argue that the late imperial elite was changing in important ways, but the persistence of characteristic, elite-identifying life-styles through the nineteenth and even

into the twentieth century indicates that in comparative terms there was remarkable social continuity of the Chinese elite. Influential studies have also suggested that the Chinese gentry elite maintained its dominance in part because the examination system kept it open to the talented and ambitious, allowing it to absorb potential rivals. Thus social continuity was linked to discontinuity of personnel.^[7]

Recent research, however, has seriously questioned the earlier work on elite mobility, suggesting substantial continuity of elite personnel as well.^[8] The work of Beattie on Anhui, Meskill on Taiwan, and the articles by Rowe, Brook, and Watson in this volume strengthen the argument for elite continuity at the local level.^[9] Local dominance could be maintained for long periods of time without relying on degrees and office, and it is time to recognize that there was both mobility in the scholar-official elite and continuity in the local elite.

How can we explain the combination of social continuity and continuity of local elite personnel in China? We might expect that landed wealth was the essential foundation for Chinese elite status, but the landholdings of the Chinese elite were comparatively very small. Twentieth-century surveys indicate that landlords, comprising 3 to 4 percent of the population, owned only about 39 percent of all cultivated land in China as a whole,^[10] and given the enormous number of petty landlords, local-elite families probably owned less than half that amount. In late nineteenth-century England, by contrast, thirty-five hundred to forty-five hundred great gentry families owned between 70 and 75 percent of the land.^[11] In Russia, nobles and the crown essentially monopolized landholding. France was more similar to China, but ecclesiastical, noble, and bourgeois landholders still owned roughly 60 percent of all land on the eve of the French Revolution.^[12]

The Chinese elite could not maintain its status simply through officeholding and links to the state, for that entailed unbroken success in the examinations, which no family could guarantee. Nor were the comparatively limited landholdings of the elite adequate to explain its persistent dominance. Rather it was the flexible application of a broad repertoire of strategies relying on multiple resources that enabled local elites to preserve their positions. Chinese local elites became masters at dealing with ambiguous mixtures of security and insecurity; their multiple resources allowed options in dealing with new situations. This flexibility served them well. When marked and rapid social discontinuities occurred in the twentieth century, a significant number of elite families maintained their social standing in the face of changes that had little precedent in the Chinese historical experience. To understand this flexible elite repertoire we must look more closely at the nature of the late imperial elite and the resources and strategies it employed.

Status, Class, and Stratification of the Late Imperial Elite

Insecure status in late imperial China and the interdependence of status and wealth produced a different interaction between the principles of status and class than one finds in other parts of the world. In seventeenth- and eighteenth-century continental Europe, separate status and economic stratifications developed in which rich bourgeoisie outranked less wealthy merchants but could not move in the same circles as the nobility at the top of old estate hierarchies.^[13] In China, merchants were not kept apart by rigid status boundaries, and commercial wealth interacted with status throughout the Qing period. The Chinese social structures contrast differently with India and Japan: in India, wealth and caste stratifications fitted increasingly poorly during the nineteenth and twentieth centuries, but local caste lines remained intact; in Japan, despite blurring of barriers toward the end of the shogunate, relatively firm status and legal lines led to a merchant-urban culture distinct from that of the samurai elite.

The flexibility and interchangeability of status and class in China by no means indicates that distinctions were absent. First, a clear consciousness of class and status existed both among elites and between elites and masses. This situation appears similar to E. P. Thompson's description of eighteenth-century England.^[14] There was strong consciousness of social distinctions and a certain degree of tension, but no strong class cohesiveness based on common relationships to the means of production or other economic factors. Superiority was demonstrated through life-styles, honor, and cultural display, all of which required both wealth and a cultural mastery that could not simply be bought. This consciousness permeated relationships in which social differences were small (as between a long-established great family and a recent arrival in elite society) as well as the greater gap between elites and masses.

It is also important that, despite the strong consciousness of social distinctions, Qing society was on the whole distinguished by weak personal dependency between elites and nonelites. Legally dependent, serflike bonded tenants began to disappear during the late Ming in the face of commercialization and peasant uprisings. The Manchu invasion and consolidation of Qing rule further disrupted the remnants of manorial society. By the eighteenth century, agricultural bondage was rare, although it occasionally persisted in places like Huizhou, Anhui, where tenants continued to till corporate property, living in a house provided by the master and bound to the land; even they, however, enjoyed security in return for their small obligations and were allowed to sublet their land.^[15] Such holdovers were unusual, and so-called feudal relations between landlords and tenants rested on informal, possibly oppressive practices, underlining peasant inferiority

but without legal standing. Bell's article gives examples of how gentry landlords could get official support to force peasants to pay rent or meet other obligations, but elites could not always rely on the backing of a government that also wanted to prevent lower-class uprisings.

If neither legal privilege nor the relation to the factors of production, nor personal social ascendancy seems adequate by itself to account for elite dominance, then were relations with the state the key? The state was the source of the degrees, titles, and honors held by the gentry. Stratifications of Chinese elites in terms of degree holding have been made familiar through the works of Chang Chung-li, Ho Ping-ti, and Ch'ü T'ung-tsu so often cited in this volume. Lines are drawn between upper-gentry jinshi, juren , and gongsheng who could hold office, lower-gentry shengyuan and jianshen who had some legal privileges but did not qualify for office, and commoners who had no degrees and no legal protections against official power. Philip Kuhn has recently and powerfully argued that Chinese views of hierarchy embodied in Confucian theories of social relations were strongly reinforced by the hierarchy of degrees and state office and that the distinction between rulers and ruled overwhelmed social divisions, dividing the state sector from all others.^[16] State-certified status was certainly of great importance, but this still seems too much a view from the center. Examination and office were not the only sources of the hierarchical perceptions that seem ingrained in Chinese culture. Such perceptions were also nurtured within family and local social systems not constantly impinged upon by the state.

The other main approach to stratification, described in Bell's paper, has focused on economic relationships between landlords and tenants and the effects of the Ming-Qing commercial expansion in Central and South China. It suggests growing class distinctions between wealthy landowning gentry living in towns and cities, small landlords and wealthy cultivators in the villages, and poor peasants and wage laborers at the bottom of the social hierarchy. Averill's article in this volume introduces Mao Zedong's division of the local elite into three layers based essentially on the size of their landholding.^[17]

These approaches through state-conferred status and economic position have not been systematically related to each other, and neither in itself seems to offer a comprehensive model. Scholars have, therefore, begun to look for broader alternatives. David Johnson's nine-part model plotting education/ literacy (with its associated legal privileges) against dominance widens the degree-based criteria into a more general one that can be applied from the top to the bottom of society.^[18] Others like Min Tu-ki have elaborated upon earlier work on distinctions within the gentry.^[19] Jerry Dennerline, in particular, has examined the interactions of such factors as kinship, marriage, lineage organization, local philanthropy, and network alliances in establishing social stratification.^[20]

The articles in this volume provide much material to support a more complex view of stratification, and they suggest that, rather than focus on degrees or land, it is better to start by looking at the arenas dominated by local elites. The possibilities of the arena approach are particularly illustrated by Averill. Starting from Mao Zedong's economic classification of the elites in Xunwu, he draws in the many other factors that defined family position in the Jiangxi hill country. The result is a map of power and status in the county. At the top were a few great households, based in the district seat or biggest market centers, who had the highest reputations, broadest connections, greatest influence, and the most resources. Below them were middle-elite families, often in smaller towns, who commanded smaller quantities or a smaller range of the same resources and whose influence did not extend as far. At the bottom of the elite were insecure families whose resources were barely adequate to warrant their inclusion in the upper reaches of society. The categories of classification are necessarily vague because what ultimately counted was the reach of a family's power. Yet the identities of the dominant families were common local knowledge, and local people would have had no difficulty ranking elite families within county or subcounty arenas.^[21] Still, no single factor defined elite status, which leads us to consider the diversity of elite resources and their uses in dominating local arenas.

Local Elite Resources

We may think of local elites in China as competing for a limited store of resources that were generally effective in Chinese society. The relative importance of each resource varied with the context of local arenas and geographical regions. They also served different functions in acquiring or maintaining elite status. The most basic of these resources are to be found in the familiar list of education and office, commercial wealth, military power and land.

Education was surely the most prized resource—as a means to gain examination degrees during the late imperial period or to acquire school or university degrees and professional expertise during the Republic. A high examination degree under the old system not only allowed entrance to bureaucratic office and a chance to enter national circles of power, but it also guaranteed access to the local magistrate simply on the presentation of one's calling card and insured that officials would treat the holder with appropriate courtesy and respect.

Education thus functioned as a stepping-stone to higher and broader arenas outside the locality. It put elites into official positions or gave them contacts they could use to act as patrons for kin and community. Within the local arena it increased community status, broadened social alliances and marriage prospects, and conferred the prestige that was such an important factor in local dominance.

Even though education was one of the most important values in elite soci-

ety, its importance as a local resource, and especially the importance of degree holding, varied markedly. Except on the most unruly frontiers, elites sought to educate their children; but the fruits of this education in the form of upper degrees were unevenly distributed across the map of China.^[22] As a result, jinshi and juren degree holders were so numerous in the core zones of the Lower Yangzi that a lowly shengyuan degree conferred little status. Even in provinces that produced the most upper-degree holders during the Qing—Jiangsu and Zhejiang in the Lower Yangzi, and Hebei (Zhili) and Shandong in North China—there were enormous differences in the number of upper-degree holders from core and peripheral counties.^[23] The gongsheng degrees, conferred for protracted but unsuccessful efforts to pass the provincial examinations, and the lower shengyuan degrees could be effective local resources on peripheries or areas like the North China plain where higher-degree holders were rare, but they provided negligible access to higher arenas.^[24]

The diligent scholar from a poor family who succeeded in the metropolitan examinations was a persistent, but seldom realized, ideal in late imperial China. In reality, it was difficult for a family to leap into elite status through the examination system. Some wealth was needed to support a boy through years of study, so a degree, especially an upper degree, was likely to be a return on previous generations' economic resources that had already moved the family to at least the lower local elite. Degree holding was, however, effective in maintaining elite status. Once examination success and office holding had established a family in the local elite, the family was well-placed to garner resources to educate sons and win more degrees in future generations. Even if a family did not win more degrees, the prestige from one major success could, as Rowe points out, maintain elite standing for generations to come. Upper-degree holders usually came in the middle of a cycle of rise and decline. As Brook observes, they were likely to be preceded by several generations of more moderate success in the prefectural examinations and followed by another period when the family only acquired lower degrees but was still locally recognized as part of the elite. Given the competition in the imperial examinations, upper degrees were the most prestigious, but also the most uncertain, resources of late imperial elites.

Wealth from trade was perhaps even more important to the rise than the maintenance of elite families, although it was often vital to both. References to "raising one's family through trade" abound in the biographies of elites in local gazetteers. Commercial wealth was such an important resource because of the growing opportunities in the Ming and Qing periods and also because this wealth could be so easily converted into the resources of status—in the form of education, degrees, or life-styles—or into landholding. Multiple strategies of family maintenance based on commerce, education, and landholding became well established in the Lower Yangzi during the Ming.^[25]

Chapters in this volume repeatedly illustrate the impact of commercial activity on local-elite societies from the coastal cores of the Lower Yangzi and

the southeast to the peripheral hill country of Jiangxi and the southwest.^[26] Zelin shows that a local elite far removed from cultural centers might even be defined by mercantile resources, and Rowe suggests that commerce was an effective alternative to upper degrees in maintaining local lineages in the relatively developed Middle Yangzi county of Hanyang. Commercialization made mercantile wealth a widespread, basic resource for gentry and military men as well as traders during the Qing, and in a few places even before.

What, however, of military power in a society where soldiering was theoretically even more disparaged than trade? Military degrees, in contrast with civil-service degrees, brought little status in core areas, although they had more weight in peripheral zones with few civil-degree holders. Nonetheless, the articles in this volume add to the evidence that military power was, in certain places and times, an important resource. Coercive power was most important on frontiers or peripheries not firmly under governmental control. Elites of core zones were more likely to use their greater rapport and influence with officials to obtain protection from governmental forces. In certain other areas, like parts of Guangdong and Fujian, piracy, banditry, and lineage or village feuding were common over protracted periods, even though the areas had left their frontier origins behind.^[27] The village guard that Watson describes in Ha Tsuen originated in such circumstances, even though its relative importance as a resource seems to have declined in the twentieth century. In other places, like the North China plain, control of local forces became increasingly vital with spreading banditry and warlordism during the late nineteenth and twentieth century.

In general, military force was more useful in acquiring than in preserving elite status. The founders of several Hanyang lineages in Rowe's article rose as officers in the Ming armies during the Yuan-Ming transition, but they quickly used their military resources to acquire land and begin the pursuit of education and civil degrees. Military power did not have the legitimacy to sustain elites over long periods. As a more practical matter, although the government could not discipline unruly elites everywhere, even the declining Qing state proved in the northwest during the 1870S that it could smash local forces if its power was focused on a particular area. Thus even frontier strongmen needed and acquired more reputable resources to maintain their positions in the long run.^[28]

Military power tended to be a temporary resource that assumed importance in troubled times of rebellion or dynastic transition but diminished in importance once order was restored. Local societies might become militarized as Philip Kuhn showed for Hunan and neighboring provinces during the Taiping Rebellion, but it is more problematical whether that condition persisted.^[29] McCord's material on the Liu family suggests militarization was likely to be a temporary response to specific conditions that favored military power. The Liu's progress toward acquiring gentry educational resources

was interrupted by mid-nineteenth-century rebellions in Guizhou. As militia leaders they rose more rapidly than they could have done by more conventional routes. Once that crisis had passed, they again began building up their educational resources but returned to military pursuits when disorder increased at the beginning of the twentieth century.

Finally, land was the most commonly acquired resource of local elites and was the ideologically approved source of wealth in Confucian China. Land ownership demonstrated status and brought wealth through rents or direct management, more stable sources of income than trade. The image of the gentry landlord is particularly associated with the Lower Yangzi, but peripheral elites also bought land to establish their positions, and Rowe points out that several of his Middle Yangzi lineages were able to buy large tracts of land after wars and rebellions. Merchants in these lineages invested large amounts of their profits in land. They also reclaimed flooded land—a common way of acquiring land in the southeast and Yangzi valley.^[30]

Landholding was probably more useful for maintaining status than for entering the elite, though Rowe suggests that wealth from land alone was not enough to maintain elite families over long periods. It was possible for peasants gradually to improve their economic positions and build up their landholdings, and Averill's chapter indicates that such progress did occur. However, to buy land a family needed money, and it was difficult to accumulate such wealth from agriculture alone. The wealthy merchant acquired more land than the diligent peasant. Furthermore, land was a scarce resource in China, and growing population pressures made it particularly difficult to accumulate. Large tracts could be acquired on frontiers or in areas depopulated by rebellion. But in most times and places, land was accumulated only slowly, in small pieces and scattered plots. Even when large holdings, by Chinese standards, were acquired, they were not impressive in European terms. Nor was land always safe: records of ownership could be lost or destroyed in times of war or rebellion.

Unlike other elites in the world, Chinese elites were not defined by a few key resources—land, caste status, or inherited titles. In one sense they might be considered weaker than an elite with an unassailable claim over a basic social resource. However, they compensated for their lack of monopoly by a remarkable flexibility in using various resources selected according to the opportunities available in their local arenas. Mercantile wealth, particularly important in commercialized regions like the Yangzi valley and the southeast, could be significant in some peripheries and frontiers as well. Landed wealth was more common in the fertile south than the north; education was more useful in peaceful cores, and military power in troubled peripheries. But the resources of elites were not all internal to their local arenas. In fact, a crucial aspect of elite power was access to external and overlapping arenas.

The Span of Elite Activity: Locality and Overlapping Arenas

Frequent reliance on combined outside and local resources was basic to local-elite strategies. Their interests were not limited to their home turf, and we need to ask what forces pulled elites away from their home areas. Joan Vincent pointed out at the Banff conference that anthropologists working on other parts of the world have found that external resources are essential if local elites are to maintain their dominance for extended periods of time. In a closed local system, leveling processes will gradually narrow elite-mass distinctions.

Both outside resources and outwardly directed strategies appear particularly significant to understanding late imperial and Republican Chinese elites. Indeed, as Philip Kuhn provocatively asked at the Banff conference, were there any genuinely local elites in China at all? Were they not all culturally bound to a China-wide system defined by the Confucian state, which offered every well-educated boy a hypothetical chance to pass the civil service examinations? Did other factors also pull elites away from home? These questions raise important issues about the articulation of local and wider arenas.

From their local arenas in village, subdistrict, town, county, and city, elites ventured into what can be conceived as both wider (extending over larger geographical areas) and higher (focusing on a higher level in administrative or commercia1 urban hierarchies) arenas. Participation in these higher and wider arenas typically required more resources (higher degrees, greater wealth) and sometimes different resources (education rather than land) and brought greater opportunities for power. Some successful local elites permanently moved up and out of their original arenas, often migrating to other provinces, but during the late imperial period it was more typical for them at least to maintain contact with a home base. Many also brought outside resources into their home localities and returned home during and at the end of sojourning careers that had increased their families' resources and local power. For this reason it is often difficult to draw clear lines between local, provincial, and national elites in China because the same person might appear in each role as he moved back and forth between arenas at different points in his career. In such cases, patterns of local dominance typically had an outside dimension.

The outward pull of the examinations and the bureaucracy and the externally derived status from state-conferred office of degrees directed elite interests toward the center and caused the often-noted circulation between national and local elites. But this is only part of the picture. G. William Skinner introduced the concept of marketing systems with ascending orders of central places grouped into macroregional economic systems.^[31] The com-

pleteness of this economic integration varied markedly with the levels of development in different parts of the country, but it increased decidedly during the late imperial period. Through this system trade fostered outside elite interests as much as did bureaucratic office. The form of urbanization in China, which dispersed small towns throughout the countryside in an interlocking hierarchy of central places, influenced the nature of local elites' external strategies. With urban activities scattered over the levels of this hierarchy rather than concentrated in a central megapolis,^[32] official control over elite movements and interests was much more difficult. After long-distance trade increased, and major manufacturing centers developed; traders established regional and interregional networks and associations with others from their locality. The growing numbers of scholars in places like the Lower Yangzi also formed their own provincial or regional networks outside the bureaucracy.^[33]

Skinner has called attention to the importance of sojourning in late imperial China, suggesting that as a key family strategy it also served to bring resources into otherwise disadvantaged areas.^[34] In areas like Huizhou, Anhui, and certain Shanxi and Shaanxi counties, sojourning merchants brought back wealth to buttress their local position; and the clerks and private secretaries from Shaoxing county in Zhejiang did much the same thing, relying upon their bureaucratic connections.^[35] By the nineteenth century, sojourning had become a pervasive phenomenon. Scholars as well as merchants, the poor as well as the elites, people from wealthy areas as well as those from peripheries found employment away from home. Significant for our purposes are the elite families' links to wider commercial, bureaucratic, or academic arenas from which they derived resources to enhance dominance at home.

Articles in this volume provide numerous examples of such external links. The salt elites in Sichuan initially relied on venture capital provided by Shaanxi merchants; the elites in Rugao had contacts in both Shanghai and Nantong. Hanyang lineages extended trade and kinship networks, and Jiangxi hill-country elites also had outside commercial links. In Schoppa's chapter, superior contacts to higher bureaucratic arenas helped the elites of one locality to force unfavorable water-control arrangements on a neighboring subdistrict. The importance of contact with people outside one's home arena also helps explain why brokerage between arenas and patronage by elites with access to higher arenas were so essential to local elite activity. Above all, local elites' transactions, mediations, and gatekeeping roles between intersecting arenas required them to face two directions: in toward their local arenas and out toward nonlocal determinants of their local power; and they carefully cultivated connections and developed strategies in this double context.

Local Elite Strategies

All elites must plan and maneuver to some extent to preserve their status. Early modern European elites rigorously husbanded their economic resources, carefully arranged advantageous marriages, and practiced strict family discipline including birth control.^[36] In China, the insecurity of elite status and the comparative scarcity of local elite resources required them to plan carefully for the fixture. These strategies were frequently devised in a context of intense competition for scarce resources.

Such elite competition was exacerbated by divisions based on such factors as kinship, ethnicity, locality, networks, and differential access to markets, productive resources, water, or political power. We see competition to win examination degrees or dominate markets, divisions between old and new money or local people and outsiders, fights among militia leaders and different religious groups. This competition could become violent, particularly in the peripheries and frontiers but also sporadically in more settled regions.

If elite strategies often assumed social conflict with other elites as well as other classes, they also illustrate some ways in which conflict was tempered through expected patterns of behavior that fostered civility. Cultural ideals of harmony reinforced unities arising from community or kinship, resistance to outside threats, common education, experiences, and associations.^[37] Elite strategies were crucially divided between those that tended to enhance elite competition and conflict and those that tended to unite elites as a self-conscious dominant class.

A catalog of the many strategies pursued by local elites would be long indeed. They constructed marriage alliances, sent sons into different occupations to diversify resources, and carefully cultivated connections with their equals and superiors. Such strategies might be pursued by families seeking either to rise within the elite or to maintain their positions. We have already indicated that certain strategies were particularly useful in bringing new families into the elite. During the Ming and Qing trade became the preeminent strategy for upward mobility. Men could also rise to prominence by commanding military force, but strongmen had to broaden their strategies to retain power for long periods. Marriage served as an avenue both to enter the elite and to remain there, and it deserves further study. Education-based strategies to acquire upper degrees offered the possibility of moving beyond the local level into the highly prestigious national scholar-official elite. But because a family could not expect to remain in those circles for many generations, degrees were also symbols to demonstrate family status within local arenas.

The topic of this volume is local elites rather than social mobility, and we are moreover persuaded that—despite fluctuating family fortunes, regional variations, and particularly unstable lower boundaries between elites and

nonelites—significant social continuity in the late imperial period extended to a lesser extent into the Republic. Therefore, we will here stress certain strategies that the articles in this volume suggest were particularly important in maintaining elite status and dominance: husbanding assets and extending local political power through lineages; defining local elite circles through horizontal networks and associations; creating vertical networks of obligated clients through patronage; and enhancing their prestige and community standing through essential functions of mediation and brokerage. Most of these strategies assume elite status. Together they illustrate how elites used basic resources to create social organizations and relationships to protect their social positions and extend their dominance over others.

Shaping Lineage Organization . Chinese lineage organization, a particular focus of anthropological and historical study, plays a central role in a striking number of articles in this volume.^[38] We see lineage not just as a kinship organization but a socioeconomic institution growing out of elite strategies to maintain local power. Lineages were defined by locality as much as by genealogy. Rowe points out that lineages often date their founding from the patriline's relocation to the locality, and Watson shows how the coincidence of lineage and territorial community could result in locally powerful social organization. Lineages were also closely associated with elites, particularly in South and Central China—their founding typically followed a family's rise to elite status.

Lineages thus play an important role in the "localist strategy" Hymes sees emerging among Southern Song elites. Lineage formation tended to come in waves, and peaks tended to come in periods of devolving state power, such as the late nineteenth century, when elites were turning their attention more to maintaining themselves in their locality than to advancing through the national bureaucracy. This was not, however, the only factor governing the growth of lineage power. Beattie's study of Tongcheng, Anhui,^[39] and Rowe and Watson in this volume all note that the accumulation of lineage land took place in the wake of major social disruptions and depopulation: the Yuan-Ming transition in the Middle Yangzi, the 1660s relocation of coastal populations in Guangdong, and the rebellions of the Ming-Qing transition in southern Anhui. Some late-Qing growth of lineage landholding in Jiangnan may have resulted from the availability of land in the wake of the Taiping Rebellion, and it certainly reflected the increasing initiatives taken by Lower Yangzi elites in this period.^[40]

When newly wealthy merchant families formed lineages, they may have been claiming membership in the local elite, but lineage formation appears above all as a defensive strategy to protect resources, particularly in contexts of competition, uncertainty, and change. Official permission was needed to establish charitable and other lineage estates, which nearly made gentry con-

nections a necessity. However, the variety of lineage forms and uses makes it clear that lineages evolved in local settings and beyond any unifying governmental direction.

The resources accumulated in lineage organization benefited the whole lineage, and the charitable estates of the Yangzi valley, in particular, indicate genuine concern for the well-being of relatives. Watson clearly shows, however, that lineage structure also helped perpetuate the position of the elite managers. It is useful to distinguish between elites in a lineage and the more problematical concept of elite lineages—that is, those with unusual local power or many degree holders. Elite lineages included many nonelite members, but certain families or branches used the lineage as a resource to maintain their status. Then they might not only act as patrons for nonelites in the lineage, but they might also dominate nonelite lineage members, whose interests might not coincide with their own,^[41] Thus lineages both defended group interests and enhanced elite power within them.

This they did in various ways. Zelin's and Watson's chapters underline the nature of the lineage as a corporate body with institutions that could assume various economic roles. Zelin most dramatically illustrates how the lineage estate (tang ) could be adapted to function as a business corporation. In his comments at the Banff conference, Rowe further suggested that in the nineteenth century, lineage, welfare, or commercial tang all acted like "trusts": legal entities with designated uses and limited liability.^[42] Whatever the exact nature of these trusts, they clearly demonstrate the institutional flexibility of lineages. They combined a measure of sanctified inviolability, by virtue of their association with an ancestral legacy, with a capacity for corporate action that was easily adapted to business requirements. Traditional legitimacy and economic rationality seem effectively combined in the trusts of the salt-well merchants of Fu-Rong.

The lineages described by Brook and Rowe fall into the different pattern of the Middle and Lower: Yangzi valley, where lineages were also widespread and numerous but owned less land in their charitable estates (yizhuang ) and did not go into business.^[43] Even without huge assets they were corporate organizations for the primary lineage purpose of defending against the progressive fission of family wealth through the custom of partible inheritance. These lineages thus functioned much as entail did in Europe, producing an undivided and inalienable patrimony.^[44] Brook succinctly sums up lineage formation in Ningbo as a strategy for transfering elite family resources from generation to generation. Lineages limited the claims of collateral kinsmen on elite family wealth while providing a "somewhat broader pool of elite resources and junior agnates" for the difficult, recurrent task of succeeding in the examinations. Rowe further points out that even though his Middle Yangzi lineages did not have sizable corporate estates they still husbanded the "aggregate resources" of their members to protect their interests and

increase their opportunities. Jonathan Lipman's paper at the Banff conference showed that elite strategies to adapt kinship forms transcended religious and ethnic boundaries: The saintly patrilines (menhuan ) of Muslims in the northwest sought to confine sanctified charisma to a single patriline, thereby excluding others from claiming socioreligious power.

Scholars have found evidence of kinship solidarity, transmittal through patrilines, and even lineages, within North China,^[45] but these lineages were usually much smaller and weaker with little or no corporate property or organization. If any organization existed, it usually focused on the ritual honoring of ancestors. Elites on the North China plain simply lacked the resources of land, wealth, and education to support the kind of lineages seen in the south. In North China weaker kinship organization contributed to the rarity of the kind of elite continuity seen in the Yangzi valley and the southeast. Northern elite families tried to preserve their wealth through the patriline and rose and fell more rapidly.

The Construction of Networks . Lineages, in general, protected existing resources and excluded outsiders from access to them. At the same time, elites also had to look outside their own kin group to preserve their local positions, and for this purpose networks were important. Whereas lineages were constructed to exclude even kin who might drain resources, networks included selected useful outsiders; whereas lineages were especially useful in elite competition, networks enhanced elite cooperation and solidarity; whereas lineage protected existing resources, networks expanded resources; whereas lineage activity often focused on the local arena, networks reached beyond locality. Networks, pervading all societies, have been defined as "quasi-groups," unbounded social fields with no clear leaders and organization linking friends, neighbors, affinal relatives, and occupational associates.^[46] Both horizontal networks between approximate social equals and vertical networks linking inferiors and superiors were essential components of Chinese social relations.

Certain characteristics of late imperial society encouraged Chinese local elites to form horizontal networks: the linkages between local and larger political, commercial, and social structures; the connections and support systems originating in the academies and examinations and continuing through government service; the lack of firm status guarantees, which put a premium on support from personal allies; the weak legal protection for real and mercantile wealth, which made powerful connections always useful; and imperial prohibitions on formal associations with any political implications. Networks were an important vehicle for guanxi (connections, relationships), a pervasive element in Chinese social interaction. In the absence of the firm criteria and guarantees of elite status (estate, caste, class) found in some societies, the bonds of guanxi were constantly created and recreated through elite social practice and extended through ongoing networks.

Brook's article in this volume gives the fullest picture of an elite subcommunity defining itself within the local arena of Yin county, Ningbo, where an aristogenic upper gentry provided its own social certification, creating a long-lived upper class held together in horizontal networks. If not completely impermeable, this status-conscious social network maintained its identity against less prominent elites by a dense web of interconnections. Marriage alliances were a particularly powerful form of social connection, and the Chinese principle of mendang hudui (literally: "the gates are matched and the households paired") favored marriages between social equals. Affinal ties, poetry clubs, historically oriented scholarship with links back to dead heroes of the anti-Manchu resistance, and the organization of local relief were among the many activities through which this Ningbo group defined its exclusive membership and reaffirmed its social solidarity.

Seemingly inconsequential and often ephemeral literati associations like the Discarded Silk Society and the Mirror Lake Poetry Society were intimately tied up with the networks. The clubs had little organization beyond the personal connections of their members; but they provided members a reason to meet, thus solidifying and demonstrating elite cohesion. Because the Qing effectively banned any political organization until the end of the dynasty, such organizations might also provide a place to discuss local or national affairs, discussion that could easily take on political overtones. On a less exalted level, government-sanctioned, but locally run, community schools in Guangdong province during the eighteenth and nineteenth centuries also served as meeting places for elites to discuss local affairs.^[47] Bell shows that similar associations could be used for economic dominance. Through the Culture Association, established in late nineteenth-century Wuxi county, a group of lower gentry sought to control silk marketing and peasant producers.

In addition to forming networks to solidify their position within the local arena, local elites were also linked outward by both vertical and horizontal networks. Gentry, especially upper gentry, cultivated friendships with upper-degree holders and scholars in other arenas or with officials in the bureaucracy. Outward-reaching literati networks were particularly visible in the Lower Yangzi where dense commercial networks and a large community of upper-degree holders fostered broad contacts among local elites and with their friends in government. Barkan's chapter shows that ties formed through the metropolitan examinations and office in Beijing were instrumental in bringing Sha Yuanbing into reformist networks of prestigious scholars and wealthy merchants in Jiangsu province after he retired from office. Elsewhere, Kuhn has noted the networks of militia formed in Hunan during the Taiping Rebellion, networks that sponsored the rise of a powerful group of provincial officials in the late nineteenth century.^[48] Networks also linked so-

journing merchants to one another and to their native arena. Long-distance trade gave rise to highly visible associations of merchants from Huizhou, Shanxi, Shaanxi, Fuzhou, and Ningbo. Common place of origin was a principle about which networks were frequently established, and when enough people from the same locality sojourned within a given town or city, they gave institutional expression to such ties by forming native-place associations (huiguan ).^[49] Thus, the vehicles for networking varied with the resources available to the elites of a particular area at a particular time, but the propensity to form networks was widespread, encouraged both by culture and the overlapping arenas.

Vertical Networking and Patronage . Vertical networks linking men of unequal wealth and status were equally common. Many of these were not strictly authority structures. Thus the Liu family studied by McCord increased its power by making connections at higher administrative levels and becoming part of provincial reformist networks. Others incorporated unequal patron-client relationships that provided one side with loyal supporters and the other with access to influence, employment, or reflected glory. Late imperial society was replete with such patronage ties, linking civil service examiners and successful candidates, teachers and students, wealthy merchants and artists, elites at different levels of the administrative hierarchy, military commanders and subordinate officers, lineage elites and ordinary members, shopkeepers and apprentices. Such ties continued easily into the Republican period. Averill, for instance, notes in the Jiangxi hills the importance of school ties between local educators and the protegés they helped enter middle schools. Vertical networks might also link elites of that area downward to marginal figures like the bandit leaders who provided occasional military muscle in return for protection and markets.

Patronage was a pervasive phenomenon in China. As in other partly commercialized agrarian societies, village-level patronage was encouraged because peasants did not have the power, knowledge, or contacts to affect outside decisions impinging upon their lives.^[50] The role of patron ideally fell to a powerful member of the community—landlord, lineage head, or manager—who would be trustworthy in representing the community and effective in pressing the interests of its members. Thus Watson says the manager of the major lineage trust in Ha Tsuen was chosen because his wealth, knowledge of the world, and connections would make him an effective patron; such credentials implied that the patron was well-educated and wealthy, with status relative to others within the community. As Duara suggests, however, the community credentials of patrons varied with the village social structures of different regions. In North China villages—with few landlords, less resources, and relatively undeveloped kinship structures—patrons skillful at

arranging affairs might rise from quite humble circumstances and, in striking contrast to village elites in the southeast, from outside the dominant kin groups of a village.

Above the village level, patronage was equally important. If imperfect market development made patrons and middlemen necessary for economic transactions, the imperfect political integration of the late imperial polity made patrons necessary for bureaucratic access and official appointment. The law of avoidance banned men from holding office in their home province and made local officials always outsiders. The best access to such outsiders was through men who shared their culture and life-style—the gentry. Because the powerful upper gentry of core areas also had links to higher levels of officialdom, which could be utilized to overturn an unfavorable decision by a magistrate, gentry, when locally available, were particularly desirable patrons. A successful examination candidate was virtually expected to act as patron for his home area and community by providing introductions, encouragement, and sometimes jobs and by using his connections in government to promote and defend local or kinship interests. In his account of the dispute over the Jute Creek flood-control embankment, Schoppa shows that access to such outside spokesmen was crucial to the ability of elites in wealthier towns to advance their interests at the expense of poor villages in a disadvantaged geographical location upstream.^[51] At still higher levels, the great surplus of degree holders seeking office in the Qing meant that official appointment also depended on patronage networks, which inevitably worked to the advantage of core areas with prestigious and well-connected elite families. Nongentry elites also used patronage to enhance their status and further their own and community interests. Frontier strongmen might patronize young scholars, thus building clientage networks that could later be used to enhance their elite status. As commercialization fostered the practice of sojourning during the late imperial period, sojourning merchants also became patrons providing jobs for relatives or people from their home area.^[52]

Patronage implies paternalism and reciprocity, but patron-client relations were unquestionably asymmetrical, particularly when clients were from lower social strata. An act of patronage underlined the patron's superiority, conferred or confirmed authority, and offered a step in building a useful clientele. Aside from providing material advantages to both sides, patrons dealt in symbolic capital. Through patronage they enhanced their own reputations, made these reputations available to their clients, and often increased the legitimacy of their claims to superiority by softening the more direct dominance they exercised as landlords, usurers, or merchants.

Brokerage and Mediation . Brokerage was closely linked to patronage, for successful patrons often performed brokering roles. But patrons had particularistic ties of mutual obligation to their clients, ties brokers often lacked.

Like patrons, brokers were most effective if they had the wealth, status, and reputation that tended to make brokerage an elite function. Conversely, however, because brokering was an empowering role, in areas where elite brokers were weak or unwilling to serve, nonelite individuals could rise to perform brokerage roles and thereby gain power over local arenas.

Duara's article provides the fullest analysis of brokerage's impact on elite dominance in China. Brokers were needed because buyers and sellers had to be brought together in incompletely integrated market systems; also, business discussions through intermediaries, reflected in the customary legal practice of middlemen guaranteeing agreements, was culturally preferred. Brokers acted between two arenas, and precisely the constant intersection of overlapping arenas made their role so common. The ideal broker was a community patron with outside contacts and some wealth relative to the other villagers. Even if he lacked significant material resources—as many North China brokers did—this ideal broker possessed "face" that inspired trust on both sides. The prestige associated with face enhanced influence in both the local community and the external arena to which the broker gave access. In the resource-poor northern villages with weak kin structures, this symbolic capital might be enough to bring a man into village elite circles.

It is striking that although brokerage, like patronage, might indicate incomplete integration, both economic development and attempts at state building in the Qing and the Republic probably increased the need for brokers. Compradors rose as brokers between foreign traders and Chinese merchants or producers. Tax farmers became more visible in the nineteenth and twentieth centuries, inserting their own interests while brokering between a revenue-hungry state and reluctant taxpayers. Middlemen purchased handicraft products or raw materials from peasant households for factories in China or resale in urban or foreign markets. Third-party introductions continued to be necessary in many social and political situations, and, as Strand points out, the breakdown of order during the Republic created a need for patron-brokers who could defend their arenas against the incursion of coercive state power.

If brokerage connected two different arenas, mediation was a community function that required a person of some local stature. Traditionally, dispute settlement fell to those with demonstrated authority derived from status and wealth; but new state initiatives, such as the creation of modern police, bureaucratized some twentieth-century mediating roles. The cultural value placed on harmonious relationships, coupled with the hesitancy to submit to corrupt judicial processes, placed a premium on informal mediation of the very frequent disputes within Chinese society. One apparent distinction between the powers of the old nobility of England and the Chinese gentry is that the former had local judicial authority denied to Chinese elites by the state. This difference narrows, however, if one considers the social authority

possessed by a mediator: his powers were not legally secured; his aim was reconciliation, not judgment or punishment; and the sanctions at his disposal might include censure or fines but rarely any corporal punishment or confinement. Nonetheless, the social authority derived from mediation, like that from brokerage and patronage, enhanced the influence of elite leaders.

This discussion of networks, patronage, brokerage, and mediation indicates the extraordinary importance of interpersonal relations in defining and preserving elite status in China. The brokers in North China villages illustrate how people with very few material resources could claim elite status by performing certain structurally necessitated roles. The theoretically autocratic imperial state monopolized authority but possessed limited power. As a result, most local governance was worked out in the local arena through informal arrangements. To guarantee and maintain these arrangements required patrons and brokers who could reach agreements with local officials, negotiate tax rates and remissions, guarantee loans and land sales, and mediate disputes between competing interests. Over time, performing these roles came to be regarded as elite functions, and the behavior patterns associated with patronage, brokerage, and mediating roles—serious demeanor, cultivated bearing, proper regard for ceremony, broad community concern, and apparent impartiality tempered by a respect for human feelings—became integral parts of elite culture. To the extent that these roles were necessary and the behavior associated with their performance was accepted as proper, the elite could use them to maintain its cultural hegemony over the rest of the population.

Cultural Hegemony and Patterns of Dominance

We began our discussion with the notion that elites are defined by their dominance in local arenas, and we have explored the resources and strategies that individuals and families employed to attain and maintain elite status. We must now focus on the strategies employed by these elites to dominate nonelites in local arenas. The resources used for domination were much the same as those for defining social position, but here the issue is their use in exercising sociopolitical power over others.

We have noted above that ties of personal dependency—of bondservants and servile tenants—were rare after the Ming dynasty. In general, the Qing and Republican elites dominated free men and women. Coercive resources available to elites were designed for use against outsiders or in outside arenas, but militia could equally well be used against peasants at home; and the knowledge that elites commanded such force deterred insubordination. Such military resources were certainly important on peripheries, frontiers, and the southeast coast, but, except in times of rebellion, they were far less

important than the economic resources of elites in core areas. Landlords demanded half the harvest in rents, merchants controlled markets and manipulated prices, and moneylenders extracted usurious interest payments. Despite community norms that might restrain collections causing excessive hardship, and even though rather secure tenancy rights were won by peasants in more commercialized areas, such relations were clearly unequal and the threat of coercion was large. If elites and peasants were, as Watson observes for South China, often bound together by many-stranded ties, the patterns of coercion and symbolic violence embedded in these ties were also many-stranded and reinforcing.^[53]

Economic forms of domination by landlords and merchants were no doubt increasingly important as commercialization intensified and spread across China, a point well-illustrated in Bell's chapter. What place was then left for cultural hegemony as economic relations became increasingly prevalent?^[54] An answer is suggested by further exploring patron-client relationships. These were not undermined by the commercial economy. On the contrary, the roles of middlemen or guarantors required by commercial transactions often had to be filled by well-connected elite patrons. The resulting patron-client ties subordinated more than just tenants: freeholding peasants might especially require such services, for they (unlike tenants) had to deal with local functionaries who collected their land tax and validated their land deeds. The patron, whatever his actual shortcomings, could display wealth, education, and knowledge of the world to justify his dominance; acts of patronage demonstrated status and increased legitimacy. When a patron was successful as protector and broker, he added still more to his symbolic capital and created bonds of obligation and loyalty among his clients. Cultural symbolism could be turned into networks of power without usually resorting to open coercion. At the same time, the implied threat in these unequal relationships—that the patron could withhold benefits or turn his power against an ungrateful or insubordinate client—was greatly reinforced when patrons held other power over the resources of their clients. Mutually reinforcing ties, to some degree reciprocal, lent stability to unequal relationships within a community setting, but they did not eliminate the coercive underpinnings of dominance.

As patrons, elites maintained support by, as Strand points out, conveying a sense of reciprocity and mutuality in personal relations. The benefits from elite patronage, brokerage, and mediation were tangible justifications for elite claims to superiority, as were their other community functions like welfare, education, and contributions to maintain buildings and roads. At the Banff conference Robert Forster pointed out that the English gentry was more successful than the French nobility in retaining the deference of the lower classes because in England the upper classes retained administrative, legal, and judicial functions that enabled them to play a paternal role, where-

as in France the local elite had little to justify its claims to superiority after such functions were assumed by the developing state. Although the functions of Chinese local elites were different, they were clearly more like the English gentry in this respect—an example of why the term gentry was originally borrowed from English society. Elites, or some of them, were quite continuously engaged in performing the functions that justified their authority, and, as Duara shows, they might invest considerable time and effort in cultivating the relationships necessary to maintain hegemony.

The most essential relationships to cultivate were those with other elites in shared or higher arenas, for through those relationships local elites could make the deals requiring their patronage and brokering. Thus the very same strategies of resource accumulation and network building that brought men acceptance among their elite peers also earned them deference from the wider population. It follows that elites came to be much concerned with displaying their elite status, for that very display validated their status and indicated that they might successfully perform elite roles. The theater of symbolic display allowed them to broadcast and justify their claims to their social equals and superiors and to legitimize their power over inferiors.

For such an exercise of cultural hegemony to be effective required shared values that would induce the populace to accept the elites' claims to superiority. The state examination system played an important part by encouraging respect for education and the educated and by spreading hierarchic Neo-Confucian values in the learning appropriate to examination success and at lower levels of education as well. The examinations, monthly Confucian lectures under the xiangyue system, handbooks of family instructions and almanacs spread values like unity, harmony, hierarchy, and respect, which benefited both state and elites.^[55]

Another range of values—including reciprocity, magnanimity, and community—were less exclusively related to, although present in, orthodox Confucianism, and they were more closely intertwined with Buddhist concepts of charity and the communal orientations of agrarian society. These norms were ingrained more by social practices than by state policies. The diffusion of cultural norms thus flowed from several sources, and values were not shaped solely by one Confucian elite that controlled national institutions and the "media of indoctrination" and consciously integrated Chinese culture on the basis of a single ideology.^[56]

When we move from the values themselves to the ways in which elites used them to inspire respect and create power, culture appears even more an active process of people expressing values rather than a code of predetermined meanings.^[57] Late imperial elites gained legitimacy by demonstrating that they were bearers of Confucian gentry culture, but they did not follow any one script in translating values into life-styles that marked them as elites and into symbols that conveyed their superiority. Confucian learning was

almost always important, but so was local service and conspicuous consumption, especially in ritually important matters like weddings and funerals. Nondegree-holding lineages might use such cultural symbols even more effectively than families with many degree holders,^[58] and the masses were familiar with these symbols even if they did not interpret them exactly as elites did. Life-styles then played a critical role in converting the basic resources of status into the stuff of social domination. Cultural hegemony could under many circumstances replace more overt, forceable, and illegitimate dominance while ensuring that elites retained their privileged positions.

Probably the most famous picture of the life-styles of the highest national elites is in the novel, The Dream of the Red Chamber , which describes the refined pleasures and sorrows of the Jia family within the confines of its opulent mansions and lovely gardens. We sense the many dimensions of display when scholar-friends show off their knowledge of poetry in suggesting names for features in the new garden created for the visit of a daughter who was an imperial concubine; when thousands of townspeople line the streets to view an elaborate and costly funeral procession for a family member; and when a woman in the household is told to hire a new maid to keep up the number of servants at a time when parvenu families were adding to theirs.^[59] Few could match the Jias, but the historical record is full of accounts of cultural display in the life-styles of the rich and famous. Brook and Rowe in this volume argue that elites in local settings were defined by criteria of education, refinement, opulence, and pedigree—not simply examination degrees. Scholars communicated with each other through poetry more than through classical scholarship, and literati networks like those in Ningbo (or even the far more modest local leaders studied by Schoppa) kept alive local intellectual traditions that might not be entirely approved by the state. These cultured lifestyles spread from the economic cores to the frontiers, and Meskill's book is an excellent illustration of elite families on Taiwan slowly shedding their military strongmen's garb and donning the robes of the Confucian scholar.^[60]

Life-styles and high culture became part of the symbolic capital of elites operating within specific arenas, and the symbols of status required public display. Thus gentry wore scholars' robes and buttons on their caps to indicate their rank. They rode in sedan chairs and built elaborate ancestral halls. Weddings cementing alliances between elite families and funerals demonstrating filial piety were carried out with lavish and expensive ceremony. Such symbolic display was composed of a host of individual acts, which together formed a holistic, culturally infused image that elicited deference and respect. The expense involved in such display ensured that elite status would be available only to those of some wealth, but simpler behavioral patterns were equally important to the elite image. These are captured by Barkan's portrait of Sha Yuanbing: a stern but just man, intolerant of moral laxity and ignorance, but concerned over the affairs of district, province, and

nation; a learned scholar sitting in his library surrounded by books. This image, not just his prestigious jinshi degree, contributed to Sha's community prominence for two decades after the end of the imperial examination system.

Symbolic capital was created in another way through welfare activities benefiting the community. Such activity reflected the value placed on magnanimity in China as in India and other agrarian societies. Other forms of periodic largess, like paying for major religious feasts also reflected this value, and served to establish wealthy elites as a force in the community.^[61] Such charity was always offered in a highly public manner. Soup kitchens made good works visible to all, and the names of temple donors were carefully recorded on the walls or a specially carved stone stele. The eighteenth-century English gentry may have been masters of theatrical display,^[62] but Chinese elites more than matched them in this respect.

Such symbols and behavior created capital in the form of obligations of deference, respect, service, or favors through social intercourse. Cultural symbols thus opened the way to future material rewards through opportunities to gain wealth or amass a following of clients.^[63] What F. G. Bailey calls the "small politics of reputation" pervaded social relations.^[64]

Cultural symbols were used most obviously to maintain hegemony, although Duara shows that symbolic capital, in the guise of face, could in itself establish authority in the social structures of North China villages. During the Qing, cultural hegemony often maintained the unequal balance of elite-mass interests in uneasy and periodically punctured equilibrium. Late imperial elites and the masses shared the "field of force" that E. P. Thompson saw encompassing upper and lower classes in eighteenth-century England. Higher and lower poles were held together by certain shared values, similar views of the proper social scheme of things, and a sense of limits beyond which it was impractical or improper for power to go.^[65]

Elites in both city and countryside often, as Strand notes, lived side by side with the poor. This proximity made their cultured life-style more visible and, at the same time, inspired—or forced—them to assume responsibility for poor neighbors. They easily fell into the role of patrons to the potential clients next door—a role that expanded into continuous involvement in local philanthropy during the Qing. It was also easier for the poor to evaluate the rich and, as suggested by both Schoppa's account of the Jute Creek embankment dispute and Watson's article, to set forth terms for their continued deference.^[66] Because the status of Chinese elites was not entirely secure, and because expectations of magnanimity were as widely diffused as those values that directly reinforced hierarchy, the Chinese poor were periodically able to hold their superiors accountable. Rioting was one way to do so when circumstances had surpassed the poor's tolerance, and this is just what hap-

pened when local elites were unable or unwilling to get rid of the Jute Creek embankment that periodically flooded the fields in their home area.

The limited latitude for disorder in still basically ordered society also reduced social tensions by providing an outlet for the energies and grievances of the lower classes that stopped short of full social conflict.^[67] Chinese elites did not necessarily approve, but mainly tolerated, disorderly festivals, unorthodox sects, and minor riots that were more likely to target the state than themselves. Sometimes they also participated in unrest. A stock figure in Chinese history and literature is the disgruntled shengyuan who led or joined the masses in protest. Lower-class riots might even be useful to more locally dominant elites in their conflicts with other elites or officials. Schoppa shows that subdistrict leaders tolerated the community riot that destroyed the Jute Creek embankment and then worked to restore their social authority. The inhibitions on indiscriminate use of power created by the interaction of elite and lower classes were reinforced by the power of the Qing government, which had its own interest in social equilibrium.

The centrality of the patron image in legitimizing power and exercising dominance is revealed in situations where it was inoperative or broke down. Much of the social conflict in China, as in premodern agrarian Europe, occurred between different localities. Such conflict not only directed we-they consciousness away from inequalities at home but also reinforced the patron-client nexus by increasing the need for elite leaders capable of defending the community. A more serious threat to elites came from the portion of the masses with whom they could not form patron-client relationships. Sectarian bands and the geographically mobile underclass, including laborers along transport routes, peddlers, miners, seasonal workers, disbanded soldiers, beggars, and criminals—all of whom escaped the vertical networks of local dominance—presented a more fundamental problem. Traditional elites were not very successful in developing strategies—other than exclusion or periodic philanthropy like the "winter defense" distributions of food and clothing—to meet this kind of challenge as it escalated after the mid-nineteenth century.

Esherick's study of the Boxer uprising illustrates just how destabilizing the unrooted poor might be in an area where community ties were eroded and local elites lacked material and symbolic resources to maintain control.^[68] But the itinerant lower fringes of society did not often invade the structures through which elites maintained dominance. This kind of challenge only arose during the Republic, when left-wing and revolutionary elites began to organize groups in the old vertical structures of authority. Strand describes how, in the 1929 campaign to unionize shopworkers and clerks in Beijing, long-standing grievances surfaced that gave "the lie to the ideal of elite paternalism." Class solidarities and conflicts were revealed as Beijing shopowners co-operated to suppress the movement. The still more forceable reactions of

landlords to the first Communist attempt to organize peasants in Haifeng county, Guangdong, during the early 1920s suggests a rural version of the same class conflict.^[69] Elites did not want to pay higher wages, accept lower rents, or give up land. These economic interests alone do not seem to explain the strength of their resistance, however. Organizations like peasant unions threatened the very structure of patronage and denied the symbolic capital upon which so much elite authority rested. At that point conflict, not paternalism, dominated elite-mass social relations, and the stakes escalated to matters of life and death.

Changes in Elites Over Time

The resources and strategies discussed in the previous section may be considered the stock of an elite repertoire of practices. Actual practices varied as conditions changed, and Chinese elites changed as well during the entire period from Ming to Republic. The chapters by Brook and Rowe indicate that periodic political crises of rebellion, war, and dynastic change destroyed old families and created conditions favoring social mobility. They find that the Yuan-Ming transition in the fourteenth century was probably even more disruptive in this respect than the seventeenth-century transition from Ming to Qing.^[70] In the nineteenth century, the Taiping Rebellion again loosened elite social structures in large parts of China; this time when the character of social institutions was beginning to change. The more fundamental transformation was, of course, part of larger processes of economic, social, and political change: commercialization, increased foreign trade, militarization, functional specialization, growing and politicized voluntary associations, and the shifting political context of republican China. During most of the late imperial period these processes changed the elite without fundamentally altering the elastic social structures. Toward the end of the nineteenth century they began to transform social structures as well. Here we consider various dimensions of these changes, their impact on Chinese elites, and their effect on elite relations to the state above and the general populace below.

The Growing Importance of Commerce . Although some have traced China's commercial revolution to the Song,^[71] the really dramatic growth in local markets, handicraft production for the market (especially cotton textiles), interregional trade, and a vigorous money economy began during the Ming. Population growth, new crops (corn, sweet potatoes, tobacco), an extended reign of peace, and the influx of silver from the New World fueled this expansion. The new commercial activity affected Chinese elites in several ways. It provided an important new avenue to elite status. Some merchants like the Fan family of Shanxi and the salt merchants from Huizhou made enormous fortunes during the seventeenth and eighteenth century through government

monopolies in salt, copper, and other commodities.^[72] More fortunes were made in trade beyond strict government supervision: first in grain, then in handicraft products (especially cotton cloth and silk), and, in the southeastern coastal provinces, in foreign trade with Japan, Taiwan, and Southeast Asia.^[73]

In the sixteenth and seventeenth centuries, references to parvenu merchants are often colored by classical Confucian disdain for profit making—the attitude that formally relegated merchants to the lowest stratum of the social hierarchy. Some literati certainly felt compelled to protect their status against the challenge of commercial wealth,^[74] but literati disdain for mercantile activity eventually became more pro forma than real. Already in the Ming sumptuary laws were breaking down, merchants were purchasing degrees and interacting with gentry, and gentry families were rising from merchant backgrounds.^[75] By mid-Qing times gentry families commonly engaged in usury and trade, and merchant families supported Confucian academies and gentry publishing projects. The cultured sons of merchants were accepted into the gentry elite, and merchants eagerly assimilated the norms of the literati, bought land, and joined in philanthropy and other public works.^[76] The social/cultural fusion of merchant and gentry elites was largely accomplished in the commercialized zones by the end of the eighteenth century, setting the stage for broader political collaboration in the nineteenth.

In the nineteenth century, trade with the West opened up further commercial opportunities, which were effectively exploited by some established trading families. The merchants of Foshan, west of Canton, provide an interesting example of the elite transformation that resulted from this process. With backgrounds in handicraft production and trade with Southeast Asia, these merchants were among the first licensed by the Qing to trade with the British in Canton. When the treaty ports were established in 1842, following China's defeat in the Opium War, such men had skills and connections to work as compradors for Western firms.^[77] Ambitious men from Canton and other coastal entrepots like Ningbo followed opportunities for foreign trade in Shanghai. Although these treaty ports were marginal to Chinese society as a whole, many successful Chinese business-elites escaped marginality by acquiring official titles, purchasing land, and becoming patrons and philanthropists in their home towns; some advised officials on the new problems of intercourse with the West. Such men were strategically located to insert themselves into a changing nineteenth-century elite, which required men with particular specialized expertise as well as the classical cultivation that had traditionally qualified one to rule.^[78]

The gradual erosion of literati exclusiveness was further illustrated by the massive sale of examination degrees in the late nineteenth century. As a result of the Qing's desperate efforts to raise revenue, one-third of the gentry class had purchased their degrees, and two-thirds of the official establish-

ment had qualified by the "irregular" route of purchase.^[79] As merchants routinely purchased examination degrees and official titles, and as venality of office undermined Confucian assumptions about bureaucratic qualifications, merchants and gentry were increasingly drawn together in various local arenas where they were routinely identified by a newly popular term: shen-shang , or "gentry-merchants." Bell's article shows how, in the twentieth century, this new hybrid class became an integral part of an emerging business elite, engaging in bourgeois practice but also relying on well-established elite strategies of networking and local management.

This changing relationship of merchant and gentry elites in China is particularly important for comparative history· Robert Forster provocatively noted at the Banff conference that trade and commerce appeared more respectable in China than in Europe before 1900; this observation is an important corrective for those still believing in the efficacy of classical Confucianism's anticommercial bias. There is no question that the Chinese gentry was more open to mercantile wealth than the nobility of continental Europe, but the parallel to England—where the purchase of a country house and the assumption of a proper life-style could qualify wealthy merchants for gentry status—is striking indeed· Because this allegedly "open" English elite is often credited with both England's political stability and the country's economic modernization, it is fair to ask why a seemingly comparable Chinese elite had the former effect, but not the latter.^[80] We suggest three things that seem to distinguish the Chinese and English cases.

First, the type of trade accorded proper status differed in the two countries. The 1700 edition of Edward Chamberlayn's Angliae Notitia , the standard reference of the day, dropped all disparaging comments on "shopkeeping" and proclaimed that "in England as well as Italy to become a merchant of foreign commerce, without serving any apprenticeship, hath been allowed as no disparagement for a gentleman born, especially to a younger brother." But domestic wholesale or retail trade was clearly beyond the pale.^[81] By contrast, the Ming and Qing dynasties periodically restricted and prohibited foreign trade. The prohibitions were demonstrably ineffective and not seriously enforced for long periods by the Qing, but they still left those engaged in foreign trade open to disparagement as "criminal merchants and sly people ... [who] secretly trade with foreigners in prohibited goods."^[82] But because China was a single empire, long distance domestic trade was comparable to foreign trade between the countries of Europe—and was also capable of producing great wealth that brought high status. The Chinese in effect reversed the European evaluations of domestic and foreign trade, which certainly inhibited the sort of overseas trade that fueled so much of Europe's early modern expansion.

Second, the greater Chinese acceptance of domestic commerce may also he related to the structure of Chinese commercial enterprises, which main-

tained a reasonably firm separation of ownership and management. We see some of this among the salt merchants described by Zelin, but they seem perhaps more involved in day-to-day management than many Chinese merchants, especially those with gentry aspirations.^[83] Xue Shouxuan in Bell's article deliberately sought to end this separation as part of his new "bourgeois practice." Because earlier Chinese "merchants" left most direct commercial dealings to their managers, they were less tainted by money grubbing and freer to pursue the cultivated life-style of the gentry.

Finally, the economic geography of merchant-gentry relations was different in England and China. The landed gentry who ruled England through the mid-nineteenth century were always closely tied to London, where they maintained townhouses, rubbed shoulders with the great merchants, invested in banking and overseas ventures, and formed the connections which made England "a nation of aristocrats and squires ruling in the interest of bankers and overseas merchants."^[84] In effect the structure of elite power in England concentrated capital and influence in London, where it fueled the nation's economic modernization.

In China, a different structure prevailed. The most powerful commercial interests in China were not concentrated in the nation's capital, nor did they derive from the Jiangnan economic heartland. The typical Chinese merchant was a sojourner. Merchant groups from Huizhou (in the hills of southern Anhui), from Jiangxi, and from the interior northern provinces of Shanxi and Shaanxi dominated much trade in the early and mid-Qing. They were joined by Fujianese and Cantonese from the southeast. Later, Ningbo merchants rose through coastal and foreign trade and banking during the nineteenth century. Thus, the Chinese merchant was always something of an outsider, who met the gentry on the latter's ground and, almost inevitably, on the latter's terms. As an outsider, the merchant was also relatively dependent on bureaucratic favor, which left him vulnerable to sometimes extravagant demands for "contributions." The wide range of sojourning also meant that investments were not concentrated in one or a few centers.^[85] Sojourning merchants might sometimes trigger the rise of local entrepreneurs, but they also remitted a substantial portion of the profits of commerce back to widely separated, and in some cases peripheral, areas from which they had come.^[86] In fact, the differing merchant-gentry interaction in China deflected the London-style concentration of wealth by returning some capital to the hinterland.

In this context the sort of entrepreneurial elite described by Bell in Wuxi and epitomized by Zhang Jian in Nantong is so important. In the twentieth century, these local gentry-merchant elites began operating their own enterprises in the economic heartland. Not only were profits kept in areas of potential economic development, but money also began flowing from land to commerce; most critical, this all happened in a context where commercial

and industrial elites were not outsiders, but the dominant elements in the local political arena.

Militarization and the Rise of Military Elites . In the mid-nineteenth century, the threat of rebellions throughout China led to organized militia for defending the established order. Philip Kuhn has studied this process of militarization in the Middle Yangzi valley, and in this volume Edward McCord has documented the phenomenon in Guizhou. In provinces like Hunan, the gentry under Zeng Guofan seem to have maintained firm control of militia networks, but McCord shows that in Guizhou militia networks were an avenue for new men to enter the elite. The same was probably true in Anhui and in North China—where the orthodox gentry were weaker than Hunan, and the state had to rely on new elements to combat the threat of rebellion.^[87]

The militarization of local society was not a continuous process. McCord shows that the Liu family abandoned much of its original militia base in the late nineteenth century to concentrate its power on patronage networks and promote educational and other reforms. Elsewhere, he has demonstrated the discontinuity of local militarization between nineteenth- and twentieth-century Hunan.^[88] However, clearly by the early twentieth century the status of military men was rising significantly, as nationalistic ideology preached the need both to train a new core of military professionals and to spread military training and values among the general population.^[89] This late Qing rise of the military was the foundation for the rise of the men who would dominate China during the warlord era and (to a somewhat lesser extent) under Chiang Kai-shek's Guomindang.

How much did the militarization of national and provincial elites affect the nature of local elites? Here the record is surely mixed, with military power playing a fairly minor role among the Lower Yangzi elites. On the frontiers, where coercive resources were always important, they seem to have become even more so in the twentieth century. On the peripheries and in North China,^[90] where gentry and commercial elites were weak, local elites did become significantly militarized. In an era when modern rifles were readily available from foreign suppliers and provided a weapon to check or challenge the state or other rivals that was infinitely more threatening than the swords, spears, and flintlocks earlier dissidents could wield, local military elites became a serious force to be reckoned with. This greater power of military technology did not, however, normally increase the status of local militarists. The small-scale local commanders were too close to bandits and "local bullies"—and their use of coercive power was too destructive to the fabric of local societies—to be legitimate. When militarization occurred in once relatively settled societies it had a different social impact than did the military force integral to newly forming frontier societies or long incorporated into peripheral community structures. In the twentieth century, externally trig-

gered militarization was likely to erode communities rather than strengthen them. The character of elites, new or old, was also compromised as they substituted more direct forms of domination for cultural hegemony.

Functional Elites . The twentieth-century rise of business and military elites was the most visible example of a process wherein discrete functional elites arose alongside (or in place of) the old gentry elite. But the process was not simply elite functional specialization creating a pluralist society; occupations were not that separate. Averill refers to his great households as "small-scale conglomerates." Xue Shouxuan, described by Bell, was not only an industrialist but also a landlord, a scion of a prominent gentry family, and a member of the new Nationalist regime. Functional categories do not show how elites diversified to protect their interests. They underplay the importance of networks and associations linking elites together and may overstate the social distinction between merchants and gentry in the late Qing and Republic.

Nonetheless, increasing evidence of specialization clearly begins before the twentieth century. Benjamin Elman has argued that the eighteenth-century complex of academies, libraries, and printing houses in the Lower Yangzi supported a prestigious group of professional academics.^[91] The medical profession provided respectable employment for upwardly mobile men or sons of gentry families who would or could not follow the standard examination route—much as the younger sons of English gentry families became attorneys in the seventeenth and eighteenth centuries.^[92] Merchants specialized in particular trades or in banking. Local management, another respectable career, might involve special skills in water control, famine relief, or philanthropic activity. Law and administration required more and more specialists during the Qing, including respectable occupations such as private secretaries in the bureaucracy, less respectable jobs such as clerks, and clearly disreputable callings such as the pettifoggers (songgun ) who serviced the everyday lawsuits of ordinary citizens.^[93] Specialization was slowly redefining the acceptable range of elite occupations, but on the whole it elaborated established gentry and merchant roles. Skill in certain approved occupations might bring social approbation, but it did not markedly improve status unless combined with other factors—birth, education, social connections, personal rectitude—that were commonly accepted social markers.

There are, however, serious indications that elites were becoming more functionally specific in the twentieth century; certainly professional elites were becoming more prominent and more organized . Various "worlds" (jie ) of educators, merchants, industrialists, journalists, lawyers, and financiers emerged as publicly identified and frequently organized contenders for influence in the new China. The elite functional repertoire broadened, and the required level of expertise increased with greater specialized training. More important, functional competence was gradually becoming an independent

source of status. In the seventeenth and eighteenth centuries, elites may have been significantly involved in commerce, but such involvement did not per se contribute to elite status. Commerce provided the wealth a family might use to seek status through scholarly or philanthropic activity or education for the exams. By the twentieth century, with the formation of chambers of commerce and the official encouragement of trade, commercial or industrial prominence alone was likely to earn one a position among the local elite. As this shift took place during the late Qing New Policies and the early Republic, one finds hybrid cases in which upper-degree holders like Barkan's Sha Yuanbing, with local standing and presumably also business interests, headed chambers of commerce and other new organizations influential in local affairs; thereby they conferred status on these institutions while simultaneously receiving influence from them.

As the Republic progressed and elite functions proliferated, more and more elites were defined by their competence in functional roles rather than by their embodiment of gentry cultural ideals—that is, by their jobs and occupations rather than by their backgrounds and behavior. We see this at the village level in the new "professional" brokers described by Duara; in the villages, these new elites do seem to be new men—upwardly mobile hucksters. We see it at the county level in the silk industrialists of Wuxi and the succession of Shas of Rugao. At the county level it is striking, as Barkan, Bell, and Rowe all indicate, that new functional elites were drawn so often from old elite families. Now more occupations conferred elite status and more routes provided upward mobility, particularly through the army. New men appeared, but elite, or at least near-elite, families were often best positioned to acquire new skills. Long-established patterns of elite flexibility and adjustment to changing conditions meant that even when the nature and institutions of the elite changed, there was still continuity in personnel. Though the social continuity of the old elite was broken, the biological continuity of elite families was often preserved. Averill describes a particularly interesting variation of this process: young men who were able to challenge the old order due to the respectability and status they commanded from their family positions within it.

Finally, we should note that in many cases, this emergence of functional elites in the twentieth century is described as the rise of less legitimate, less prestigious elite types. This is certainly true, and the new men out to make their fortunes were likely to be more openly exploitative than established elites. Even so, legitimacy is a normative judgment, and not all change represented social erosion. Normative standards are always slow to change, but they do, in time, catch up with social realities. We should remember that it was late in the nineteenth century, one hundred years after the onset of the industrial revolution, before English industrial elites were regarded as respectable.

The Emergence of a Public Sphere of Local Elite Activity . The imperial state had always been extremely suspicious of independent elite associations, though in the late Ming, the Jiangnan elite had managed to transform scholarly and cultural networks into organizations for quite overt political activity. The Qing cracked down severely on any associational activity that looked at all political, but they did not prohibit all organized gentry activity on the local level. Philanthropic associations—including foundling homes, societies providing support for widows, and famine-relief bodies—were the most important types of local elite organization. Angela Leung has pointed to the seventeenth-century emergence of united welfare associations (tongshan hui ), as local gentry and merchants took over responsibility for moribund governmental institutions in Lower Yangzi cities and towns. Although the Qing reasserted state interest in welfare in the eighteenth century, the resulting united welfare agencies (tongshan tang ) were run by local elites under official supervision.^[94] Even during the height of Qing state power in the eighteenth century, local elites were gradually outstripping bureaucratic authorities in other arenas as well, such as the small, but important, community-oriented water control projects of the Yangzi valley and South China.^[95]

State policy did much to encourage this tendency toward extrabureaucratic elite initiative. Madeleine Zelin has shown how the eighteenth-century Qing government pulled back from giving county magistrates adequate funds to run local government and fully control local arenas.^[96] In the nineteenth century, the Qing further compromised its authority by delegating certain tax collection powers to local gentry, merchants, brokers, and guilds.^[97] Commercial taxes in particular were collected outside the bureaucracy, and, as they became increasingly important, the fiscal power of merchant and gentry groups expanded.

The mid-nineteenth century certainly represented a major watershed in the growth of autonomous local-elite power. Philip Kuhn has shown how elite autonomy was enhanced by militia building in this period. Though the militarization may not have been permanent, Susan Mann's work on taxation, William Rowe's on Hankou guilds, and Mary Rankin's on elite initiatives during the postrebellion reconstruction further demonstrate the emergence of increasingly autonomous structures of local-elite power in the last half century of the Qing.^[98] Rowe and Rankin have analyzed this process in terms of a growing, community-focused "public sphere" of organized elite activity between the official bureaucracy and the private sphere of families, kin groups, and business enterprises.

If the Qing actively discouraged any overt political role for local elites before the 1890s, the 1902-1911 New Policies of the Qing government explicitly permitted elite political mobilization in local assemblies and councils and also gave them larger powers in education, economic development, and public security. Elite mobilization of this period built upon the "public sphere"

activities and networks of the nineteenth century, but the new political context of the reform period fundamentally altered the import of elite behavior. China's defeat in the Sino-Japanese War of 1894-95, the threat of partition in the 1897-98 Scramble for Concessions, and foreign occupation of the capital and of Manchuria in the wake of the Boxer Uprising gave rise to an unprecedented wave of nationalist sentiment. The rise of the Chinese press fanned this sentiment; general politicization of elite activities followed—a politicization that soon promoted hostility to the Qing dynasty. Provincial assemblies, local chambers of commerce and educational associations, study societies, and political associations, really proto-parties, provided powerful institutional bases for political opposition and helped bring about the republican revolution of 1911.

Republican elites controlled even greater organizational resources in private associations and quasi-governmental professional associations (fatuan ). These resources could. as proposed by Strand in this volume, lead to the appearance of a "civil society," a more autonomous successor to the public sphere, in which independent groups might organize, explicitly pursuing the interests of their members. A new type of local politics emerged in this civil society, a constituency politics of associations as interest groups, which threatened to replace the old politics of networks and personal relations. This politics developed in the cities of core areas, but as Averill's article indicates even the peripheries were ultimately affected by the new politics, with revolutionary consequences.

At the same time a new mass politics emerged, capable of threatening elites from below but using the same organizational forms that elites devised to press their interests against the state. Whereas the Qing had quite successfully prohibited open political activity for some 250 years, politics dominated the twentieth century; the genie was out of the bottle, and the forms, structures, and discourse of elite politics changed irrevocably. Once the public pursuit of private interest was legitimized, once elites began to function as representatives as well as patrons, once appeal was made to progress and change as well as harmonizing accommodation to cultural norms, the polity began to be transformed in ways that presented both new opportunities and new challenges to local elite dominance.

The Fragmentation of the Elite . The late imperial gentry elite was arguably the most unified (though not uniform) elite in the world. Even while it included more than the degree holders who had passed the state-administered examination system, those degree holders set the behavioral norms and defined the life-styles that set the elite apart. The examination system unified the gentry nationally, guaranteeing a comparable number of lower degrees to each county and providing preferential quotas of upper degrees for isolated and underdeveloped provinces on the frontier.^[99] No clerical, aristocratic, or

military elite challenged the dominance of the gentry; and the commercial elite imitated, joined, and slowly transformed the gentry—but did not compete with the gentry for local power as a separate group.

The late-nineteenth- and twentieth-century transformation of the elite placed unprecedented strains on elite unity, however. The emergence of a public sphere of elite organization provided a growing number of local arenas for elite activity. Elites came increasingly to focus on their separate local arenas as much as on the national examination system. But the environments, the resources, and the interests of these local arenas differed markedly, encouraging differences in elite behavior as well. Both Rankin and Schoppa have noted a widening gap between economic cores and peripheral zones. The cores benefited more from the commercial expansion of the nineteenth and twentieth century and recovered more quickly from the rebellions. When new schools were established to replace the examinations abolished in 1905, official funds favored the provincial capital. Core elites demonstrated the will and the resources to finance schools in their own communities, and the peripheries fell further and further behind. The same was true of other associational activity. Core elites organized more quickly, more extensively, and more effectively and managed to press their agendas in higher political and economic arenas.^[100]

Although Rankin and Schoppa have stressed the gap between core and periphery, others have stressed the rupture of urban and rural elites. Esherick has emphasized the role of an "urban reformist elite" in the 1911 Revolution, and Kuhn has written of a modernizing "new urban elite that found it increasingly hard to identify itself with the problems of rural China."^[101] In fact, there is substantial overlap between these urban-rural and core-periphery cleavages. The major urban centers were all in regional cores, and the elites of these cities diverged most markedly from those in peripheries. The urban elites were more commercialized, more functionally differentiated, more likely to be trained in modern schools and affected by Western culture, and more committed to models of economic modernization that favored the cities and the cores. These characteristics often put them at odds with the elites of less developed rural hinterlands, who were more concerned with protecting their own security against a rising tide of bandit attacks and communist insurgency than they were with the core elite's fancy blueprints for economic development and political reform. To the hinterland, the urban elites were "false foreign devils"; to the cities, the rural elites were "local bullies and evil gentry."^[102]

Just as important as these broad social fissures in the elite were more particularistic splits that came with the functional differentiation of the elite and its politicization. Warlord struggles could make bloody affairs of such divisions in the elite; and the ideological polarization of the 1920s tended to color a wide variety of local factional quarrels. With different functional

groups represented by their own associational structures, and with political groups forming factions to compete for political advantage, it was inevitable that the old elite would lose some of its former cohesion. We should not overstate those divisions, however. Broad consensus still existed on who the elites were and what qualified them to rule; but the challenge to that consensus was growing, and the internal solidarity of the elite was not as strong as it had once been.

Local Elites and the State

All these changes in Chinese elites affected their relationship to the state—a relationship that has been a central concern of scholarly inquiry on China. The early work on Chinese elites stressed the state-conferred examination degrees as authentic cachets of elite status and the role of the local gentry in defending the imperial order. Although more recent work, including the chapters of this volume, has focused on the local sources of elite power, almost all scholars agree that only in relatively limited portions of the frontier can Chinese elites be considered totally apart from the state. The critical question from our perspective is the transformation of this elite-state relationship in the modern period.

First we must confront the problem—complicated by changes associated with the dynastic cycle—of a baseline from which to chart the process of change. In many respects, the late Qing looks like the late Ming, where historians have found expanding social organization, elite political initiatives, and weakened governmental efficacy. But the early Qing state reversed most of these trends, and it makes some sense to begin our account in the eighteenth century, which appears as the most effective era of Chinese bureaucratic governance. In the monographs of Peter Perdue and Pierre-Etienne Will and the survey by Susan Naquin and Evelyn Rawski, we see a state capable of impressive initiatives in land reclamation, resettlement, water control, and famine relief, but its powers are increasingly circumscribed by local interests as the century progresses.^[103] In Perdue's work in particular, the devolution of state power to local elites that Philip Kuhn associated with the mid-nineteenth-century rebellions can be seen as a process beginning in the eighteenth century.^[104]

It is also clear, however, that the process involved more than the devolution of state power. We have just reviewed a series of initiatives, dating back to the late sixteenth century, through which local elites gradually established a public sphere for their activities in philanthropy, education, local defense, water control, public works, fiscal affairs, and, in the twentieth century, in professional associations, journalism, political organization, economic development, and local self-government. Local elites were doing more than just stepping in to fill gaps left by a disintegrating state. They were developing

new strategies and new institutions to protect their positions and to guide political developments in accord with their interests and ideals.

By 1911, provincial and local elites perceived their interests and ideals to be incompatible with the Qing state. The symbiosis of imperial state and gentry elites had come to an end, and the latter joined the Republican revolution. But the revolution was not simply a triumph of local elites over the state; and the Republican period was more than just the political disintegration of the state.^[105] Most important, the modern competition between state and local elites was not a zero-sum game. At the same time that local elites were organizing in new ways and increasing their resources, the state was doing the same. Beginning with hesitant nineteenth-century reactions to foreign imperialism and internal rebellion, the state was enlarging and modernizing its armies, increasing its fiscal resources with new commercial taxes, and improving its communications with telegraphs, steamships, and (by the 1890s) railways. In the first decade of the twentieth century, the Qing state took on an impressive range of economic development activities through the new Ministry of Agriculture, Industry and Commerce and the Ministry of Post and Telecommunications, plus the new Bank of China and all their local appendages. The new army and the police greatly strengthened the coercive forces of the state; and the campaigns to eradicate opium demonstrated that the Qing was willing to use its new powers.

Under the Republic, economic development projects expanded to include the efforts at cooperative formation and sericulture improvement described by Bell. The social reforms broadened into widespread attacks on footbinding or popular "superstition," which could be dramatically intrusive even when sporadic. Tax increases and the conscription of corvée labor unquestionably brought the weight of the state much more directly to bear on the average peasant. Duara shows the impact that such changes could have on village society as old elites with status in the community were replaced by new and more exploitative parvenus connected to the bottom rungs of the new state apparatus.^[106] Even if it was not very effective, a process of state building was certainly going on, and local elites were involved in it and reacting to it.

One local-elite reaction was simply to distance itself from the state, increasingly separating local and national elites. In Schoppa's article the absence of the nationally prominent Tang Shouqian from the struggle over the Jute Creek embankment, so important to his home village, symbolizes this separation. Local elites found few positive qualities in the predatory national warlord governments of the late 1910s and 1920s. In her Banff conference paper, Helen Chauncey described how some retreated defensively to pursue independent reformist aspirations at home; others, with a military following, became local strongmen or regional warlords protecting their own turfs.

It is probably best to think of elite-state relations in the late Qing and the Republic as reflecting the conjuncture of a cyclical devolution of state power with secular trends of state building and elite self-mobilization. The competitive expansion of both state and elite power, in a context of increasing elite organization and politicization, resulted in a more profound conflict between elites and the state than was characteristic of the late imperial period. Authoritarian/bureaucratic/militarist governments failed to build a broad constituency for modernizing state building but kept elite associational politics from developing into a serious bourgeois democratic alternative. Expanding state resources were not effectively translated into control or legitimacy; and the most effective republican era government, the Guomindang's Nanjing regime, showed, in Strand's words, "a remarkable capacity to ... punish its natural allies" among the local elite. As the Guomindang shifted from a mobilization to a bureaucratic style of state building at the end of the 1920s,^[107] it found that it could neither replace local elites nor win their loyalties. Barkan's work on Rugao indicates that, far from successfully establishing a corporatist state,^[108] the Nationalist government vitiated locally run services without providing effective substitutes. For example, Western-educated Xue Shouxuan in Bell's article cooperated closely with the Nationalists, but he used the official sericulture agencies so profitably in Wuxi that one may ask whether the state brought him into the fold or whether he captured a bit of the state.

The changing elite relationship to the state, and what it meant for local dominance, can be illustrated by looking again at brokering roles. In areas like North China or regional peripheries, where elite material resources in land and commercial wealth were limited, the elite's privileged access to the magistrate and local bureaucracy was one of the most important foundations of its influence in the community. The literature on the late imperial gentry has stressed this role as intermediary between the state and rural society. Local elites were brokers: acting between the official world and the local community, participating in both and (in a sense) not belonging totally to either.

F. G. Bailey has written insightfully about brokers on the basis of fieldwork in South Asia. The broker "must make the villagers believe that he can communicate with and manipulate clerks and officials in a way that the ordinary villager cannot." This may require him at times to deliver "an authoritative, if mystifying and unhelpful, lecture" on some particular administrative procedure about which a question has been asked. Above all, he must prove his efficacy, anti this entailed a "complex task of presenting himself to the world." He must wear "a normative mask of devotion to the public weal, which had, however, to be sufficiently transparent to allow clients to see his skill at maneuvers which were normatively condemned."^[109]

This description does not readily fit the gentry or community-rooted brokers

of the Qing, who had the symbolic capital, and sometimes the genuinely high status, to justify their credentials in both arenas. It certainly seems to apply to the new type of village brokers described by Duara, however, and it may apply to the new county elites of Barkan's and Bell's chapters as well. Duara's new middlemen are recognized as necessary, but often, it seems, despised; and as Bailey notes, "the middleman is despised in proportion to the disparity of the two cultures" between which he must operate.^[110] Seemingly the process of state building increased not only the extractive demands of the state but also the cultural distance between it and rural society. Those who served as brokers to the new state were indeed increasingly looked upon with contempt. Respectable members of the community drew back from such tasks; and a new brokering elite stepped in. Where the imperial state had served as a source of legitimacy for local gentry, the Republican state helped both to foster and to delegitimize a new brokering elite.

Thus on the one hand the devolution of state power left locally entrenched oligarchies unchecked by the central state. Such unchecked power was likely to be used abusively, and the repeated complaints against "local bullies and evil gentry" indicate that this was indeed the case.^[111] At the same time, equally disreputable men rose to act as brokers for the expanding governmental activities. The appearance of such unchecked, illegitimate, and exploitative parvenu elites increased social tensions and created constituencies for the radical young intellectuals, described by Averill, who returned home to promote revolution. Because governmental power was most important in North China, both the devolution of state power and the rise of state-brokerage roles should be most important there too. The revolutionary forces grew in the north and the peripheries, where local yamens were a more important focus of elite activity and independent networks of elite cohesion were less developed. That revolution might have been crushed had the Nationalists created a more effective coalition in the rest of the country, but in the core areas the stalemate between Guomindang state making and local-elite organizing left both groups incapable of mounting an effective counterrevolutionary effort.

These changes in modern Chinese elites relate more to context and structure than to behavior and strategy. The behavioral continuities across fundamental changes in twentieth-century economic structures and technology are extremely important. They suggest that the most distinctive characteristics of Chinese elites are to be found in behavioral patterns rather than static attributes such as those associated with the degree-holding gentry. This interpretation suggests new ways to look at the interaction of structure, practice, and change in Chinese society.

Flexible reliance on multiple resources was one of the most important characteristics of late imperial elites, who were never guaranteed enough of

any one resource in education, land, commerce, or military power to be defined by that single attribute. Because they had long since learned to exploit changes in the available resources, they readily adjusted to the new opportunities and challenges brought by the forces of "modernization" intruding from the West and Japan. As a result, many twentieth-century Chinese elites came from the same families that had dominated their localities for generations.

The social resources, constructed in shifting contexts by elites themselves, were the most resilient. Dense networks of human relationships (guanxi ), cultural symbols, and behavioral norms held elites together. Aware of their social superiority, Chinese elites set out to construct webs of influence to protect their status. Such social relationships proved an enduring social resource during the process of China's modern transformation, as preexisting connections facilitated elite control of twentieth-century institutions. This same continuity also insured that modern forms of association were regularly infused with the status- and display-conscious norms of older elites.

Networks were fostered by the larger context of Chinese society. Neither economic integration nor bureaucratic rationalization had progressed to the extent that patrons, brokers, and middlemen could be dispensed with. Such roles were particularly important for activities that crossed arena boundaries, and China was replete with intersecting arenas: between levels of the economic and administrative hierarchy, between state and society, between upper and lower classes. Particularly in the political realm, the combination of formal integration and informal delegation of state power left unusual room for elites to exercise their patronage and brokering roles. These roles existed in dialectical interdependence with the elite's cultural hegemony; hegemony supported their monopoly of these roles and was reinforced in turn by successful performance.

The Chinese elite's exercise of dominance through cultural hegemony, while hardly unique, was certainly striking. Cultural symbols and behavioral norms identified these elites, gave them social cohesion, and advertised their superiority. It is important to understand that the infusion of the norms of gentry society into almost all sectors of the late imperial elite represented more than the hegemony of Neo-Confucian values and ideals. Culture given living interpretation in elite daily life was basic to the ways in which elites defined themselves locally and legitimized their dominance. Elite culture meant literati activities, to be sure, but also included patronage, philanthropy, mediation of social conflicts, avoidance of physical labor, mastery of courteous social intercourse, and public and ceremonial display of affluence. These forms of behavior proved remarkably adaptable to the social requirements of modern life, and many were found among the practices of republican as well as late imperial elites.

These few characteristics—elite flexibility, reliance on networks and other

social resources, and cultural hegemony—can assist us to understand the interplay of unity and diversity among Chinese local elites and the historical processes of elite transformation in late imperial and republican China. Unlike static conceptions of gentry society, they help explain the flexible adaptations that sustained many elite families in the hybrid elite society of the twentieth century. They also allow us to focus on those structural changes that fragmented the modern elite, fundamentally altered its relation to the expanding state, and ultimately left it prey to the revolutionary forces that united a new stronger state above and mobilized the masses below. With that new combination, the cultural hegemony of the old Chinese elites was brought to an end, but it remains to be seen how many of their behavioral norms might still survive.