THE “LOCALIST TURN” AND THE RISE OF THE GENTRY

The view of the Song as a turning point in history has an intellectual lineage that goes back to Naitō Konan 內藤湖南 (1866–1934). In his view, the Song marked the beginning of the rise of imperial autocracy, a trend which culminated in the Ming-Qing period. Following the demise of the aristocracy during the late Tang and the Five Dynasties, he argued, the principle of bureaucratic recruitment shifted from “an independent birthright” beyond the emperor's control to “a conditional appointment by imperial will.” Whereas the Tang emperor was only a representative of the aristocracy whose power was constrained by aristocratic interests, the emperor from the Song onward exercised absolute power over his officials, whom he himself recruited from the commoners through civil service examinations. Naitō pointed out emphatically the many changes in formal governmental organizations and the dynamics of court politics that led to an “increasing arrogation of state functions in the emperor's hands and a corresponding elevation of his ritual status.”Footnote ⁵

In the 1980s, social historians proposed a competing view. They shifted attention from politics and institutions to the behavioral patterns of the elite and from distribution of power within the government to the relationship between state and local society. The most important proponents of this view include Robert M. Hartwell and his student Robert P. Hymes. In his pioneering article “Demographic, Political, and Social Transformations of China, 750–1550,” Hartwell gave less weight to changes between the Tang and Song than between the two halves of the Song dynasty itself.Footnote ⁶ He described the change as “the demise of a semi-hereditary professional bureaucratic elite… [in the early twelfth century] and its replacement… by a multitude of local elite gentry families.” Whereas the local elite based its power in local society and “viewed a bureaucratic career as only one of a variety of occupational choices,” the professional elite, argued Hartwell, specialized in government service and perpetuated itself in high offices over multiple generations.Footnote ⁷ Hartwell defined the professional elite by its distinctive social behavior, which in many ways resembles that of the ninth-century Tang elite in Tackett's study: its members claimed descent from the great aristocratic clans, established their main residence in the capital region, intermarried with each other regardless of regional origin, and served in high offices generation after generation. It should be noted, however, that the professional elite identified by Hartwell was composed of a fairly limited number of families in the Song which dominated, but did not monopolize, the Song government. Although it provided a “disproportionate number of incumbents” to high offices in the Song government, it never monopolized government offices to the extent that the ninth-century elite did. At all times and at all levels of government, it shared political power with men from local elite families. Even at the peak of its power, for example, it claimed less than one-third of policy-making offices while the local elite families always staffed at least half of these positions.Footnote ⁸

Hartwell's study broke new ground by defining the ruling elite as much by its social behavior as by its claims of ancestry. It is his student, Robert P. Hymes, who has applied this approach fully. In an influential study published four years later, Hymes shifts attention from incumbents of national offices to those men and women who had privileged “access to wealth, power, and prestige… in the local scheme of things” in Fuzhou 撫州, Jiangxi, and combines prosopographical research on their residence and marriage patterns with close analyses of their activities in local society.Footnote ⁹ He notes that the social behavior associated with the professional elite—capital residence, intermarriage with each other regardless of regional origin, and self-perpetuation in government—was characteristic of a much larger number of families than Hartwell had realized, including many from Fuzhou whom Hartwell would have categorized as the “local elite gentry families.” This behavior disappeared, however, in the Southern Song. Therefore, Hymes argues, the transition in the early twelfth century should be best described not as the demise of a specific status group but, instead, the abandonment of one strategy in favor of another. Whereas local elite families in the Northern Song pursued a “national” or “bureaucratic” strategy centered on the attainment of high office, elite families in Southern Song Fuzhou followed a “localist” strategy that concentrated on consolidating their local power base and sought only occasional office as one element of perpetuating their local status. Hymes's evidence comes, like that of Hartwell, from their subjects’ contrasting decisions on residence and marriage. In pursuit of high office, Fuzhou men in the Northern Song who succeeded in the examinations and received an appointment in government usually migrated from Fuzhou to major economic and political centers and contracted “long-distance” marriages with families from other counties and prefectures. Only those who had not yet achieved office stayed in Fuzhou, but they too, Hymes suspects, shared the same ultimate goal of rising to political prominence and their efforts at home—accumulating wealth and building connections—was the first step in that direction.Footnote ¹⁰ This all changed in the Southern Song. Elite families in Southern Song Fuzhou focused more on consolidating their status at home than on rising high in government. Not a single prominent family left Fuzhou after 1127. They sought affines almost exclusively from Fuzhou itself and, most often, from just within their home counties. They engaged actively in organizing local defense, developing infrastructure, promoting local welfare, and constructing local cultural spaces. They developed a strong local consciousness, employing history-writing to carve out an identity for their home locales and for themselves as local men. The Southern Song, therefore, saw “the growing importance of locality in elite social life and self-conceptions” and the separation of elite from state.Footnote ¹¹

Sukhee Lee questions this argument. His study reveals remarkable continuity, not change, in the patterns of elite marriage between the two halves of the Song. He contends that elite families in both Northern and Southern Song Mingzhou arranged a mix of local and non-local marriages. In both periods, Lee shows, approximately half of the elite marriages were between families from the same county while a third were across prefectural borders (64–65). In a review of Hymes's evidence, he points out further that the high incidence of localized marriages in Fuzhou reflects, at best, the tendency in the last decades of the Southern Song or, at worst, perhaps a bias in the sources we have for those years (80–82). These findings are consistent with earlier studies on elite marriage inspired by Hymes. Beverly J. Bossler's study on grand councilor families and the Wuzhou 婺州 (now Jinhua, Zhejiang) elite, for example, also reveals more continuity than change between the two halves of the Song, leading her to argue that that the geography of marriage was primarily a function of political status, with higher-ranking officials in both periods more likely to have affines from outside their home counties and prefectures than their lower-ranking colleagues. The relatively higher incidence of local marriage in Southern Song Wuzhou was almost certainly the result of more marriages of commoners being documented in this period. She also stresses the strong preference in both periods for finding affines nearby, noting that grand councilor families in the Northern Song who contracted marriages with each other, in fact, lived mostly in the capital region and that Wuzhou elite families who went beyond prefectural borders for in-laws usually married southerners from not-too-distant regions.Footnote ¹² The apparent contrast between Northern and Southern Song geography of marriage, therefore, was a product of confusion of ancestral homes with places of actual residence and historiographical biases. Likewise, in a study of southern Fujian, Hugh Clark argues that marriage was “at all times first and foremost a local affair,” though elites in both the Northern and Southern Song also regularly had affines from outside their home prefectures.Footnote ¹³

We should exercise caution, however, before rejecting the “localist turn” paradigm too quickly. Read against Tackett's book on the Tang political elite, Lee's study on Mingzhou brings to the fore the dramatic change between the late Tang and the Southern Song. As late as the waning decades of the Tang, offices at all levels of government continued to be monopolized by a capital elite. Political prominence was so closely related to capital residence that relocation to the provinces was nothing short of a death knell for a family's political future. This contrasts starkly with the Southern Song. For Mingzhou men in the Southern Song, living outside the capital was no obstacle to political success. Many local families, like the Lous 樓 and Shis 史, rose to national eminence and perpetuated their status until the end of the Song. Rather than discrediting the “localist turn” paradigm, Lee's study, I think, challenges us to reconsider and better articulate what it meant, when and why it happened, and what the implications were.

First of all, underlying this persistent scholarly interest in the geography of marriage is the assumption we have inherited from Hartwell that the choice of affines, along with decisions on migration, gives a clue to understanding the character of the ruling elite—whether it constituted a distinct, self-perpetuating status group separate from the layer of wealthy, powerful families which assumed leadership in local society—and how it maintained its political status. To take this assumption seriously means that whether or not families without official rank married locally, therefore, is irrelevant.Footnote ¹⁴ It also means that marriage patterns of the officeholding elite must be interpreted in tandem with other aspects of its social behavior (such as its residence patterns) as a measure of its social distinctiveness and the degree of its attachment to local communities whence they hailed. Second, as social historians looking for long-term trends often remind us, major institutional changes in history, not to mention social and economic changes, do not always coincide with the rise and fall of dynasties, and relying on major dynastic markers—such as the founding of the Northern Song in 960 and the Jurchen conquest of 1127—to define our unit of analysis runs the risk of obscuring important changes. Once we are more precise on defining periods of comparison, focus exclusively on the officeholding elite, interpret its choice of affines alongside its patterns of residence, all evidence points to the mid-eleventh century—not the late Tang or the end of the Northern Song—as a major watershed in history.

Relocation to the capital and the Grand Canal region (an economic hub that provided convenient access to the capital) was a popular choice for many who became officials very early in the Song dynasty, not just those who rose to high ranks, but it began to lose its appeal around the middle of the eleventh century. All six emigrant families from Fuzhou identified by Hymes, for example, produced their first metropolitan degree holder (jinshi) by 1015. These six families, plus the Xus 胥, of whom we know little, accounted for the vast majority of degree holders from Fuzhou before 1040. Four of them established new gravesites in the vicinity of the capital or along the Grand Canal by the early eleventh century, almost immediately following their examination success.Footnote ¹⁵ By contrast, while a far greater number of Fuzhou men from a much larger pool of local families passed the examinations between 1040 and 1127, there is no evidence that any of these newly successful individuals, except Cai Chengxi 蔡承禧, relocated from their home prefecture.Footnote ¹⁶ Wuzhou and Mingzhou records reveal a similar trend. Hu Ze 胡則 (963–1039), jinshi of 989 (and, in fact, the only degree holder of Wuzhou before 1040 for whom detailed biographical data survives), established a residence in Kaifeng shortly after receiving his degree and moved to Hangzhou at the southern end of the Grand Canal upon retirement.Footnote ¹⁷ Wuzhou began to produce jinshi with considerable regularity after 1040, but judging from late Northern and early Southern Song sources, nobody but the famous general Zong Ze 宗澤 (1059–1128) seems to have moved out.Footnote ¹⁸ Unfortunately, Sukhee Lee provides no account of elite migration in his book. But as far as my own reading goes, elite families in Mingzhou made similar decisions on residence. All three families—the Wangs 王, the Ges 葛, and the Shens 沈—who relocated from Mingzhou entered office before 1030.Footnote ¹⁹ By contrast, there is no evidence that any of the Mingzhou men who passed the examinations after 1040 abandoned their ancestral homes.Footnote ²⁰ In Sichuan, too, nearly all those who relocated to the capital and the Grand Canal region entered civil service before 1040, while all but one of the officials who received the jinshi between 1060 and 1127 abandoned their homes in Sichuan.Footnote ²¹ The same trend is also visible among men occupying the top positions in the Song government. Beverly Bossler has noted that grand councilor families found the appeal of Kaifeng declining by the mid-eleventh century. A total of twenty-eight grand councilors in the Northern Song were raised outside the capital region. Fifteen of them came into office before 1050 (indicating that they first obtained an official rank sometime before 1030 at the latest), of whom all but one relocated to Kaifeng; thirteen came into office after 1050, of whom only one moved to the capital and two others, by my own count, migrated to the Grand Canal region.Footnote ²²

The geography of marriage cannot be understood separately from changing patterns of residence. Choosing an affine from outside one's native area had different social meanings for a family which continued to live in its native place and a family which did not. Marriages uniting families from different native places in the Northern Song often involved emigrant officeholders on one or both sides, and the actual residences of the two families in a marriage were usually in close proximity. These marriages, therefore, indicate the emigrants’ weakening ties with their native places and their effort to integrate into the capital-oriented elite circle they had moved into. Hymes has observed thirty-nine marriages between families from Fuzhou and those from outside the prefecture, all but five of which were arranged by the six emigrant families.Footnote ²³ Of these thirty-four marriages, only five were with families from other Jiangxi prefectures (nos. 3, 4, 26, 30, 35 in Hymes's list), while at least twenty-one of them (nos. 5, 6, 8–12, 14, 16–20, 23–25, 28, 29, 31, 33, 36) were with officeholding families who lived in fairly close range of where the Fuzhou men had moved (or would have eventually moved). Footnote ²⁴ The Wangs and Zengs, for instance, sought affines predominantly from along the Grand Canal, and the Les and the Yans from the capital region. Many of these families, such as Wu Anchi 吳安持 of Zhengzhou 鄭州 and the Zhous of Taizhou 泰州, were themselves early Song immigrants to the capital and the Grand Canal regions. This behavior, therefore, reveals a strong tendency before the mid-eleventh century for officeholding families of different regional origins to form a distinct status group centered on the Song capital and Grand Canal regions and separate from native communities of each individual family. It is a sign of the political elite's capital orientation and its separation from the leadership of local society. It reveals a close relationship between political success and capital-oriented social networks, highly reminiscent of the late Tang elite discussed in Tackett's study.

By contrast, Wuzhou and Mingzhou records are notably scarce on the early Song.Footnote ²⁵ Evidence of elite marriages in both prefectures comes almost exclusively from after the mid-eleventh century, when most officeholding families—including politically prominent lineages like the Lous and Shis—decided against relocation. Marriages across prefectural boundaries in these cases, therefore, took on a different social meaning. They did not weaken the elite's ties to its native places; rather, they strengthened its local position by extending its social networks beyond its home prefectures. These networks, in fact, centered on the elite's native places. Wuzhou people who chose affines from outside the prefecture nevertheless tended to marry southerners from not-too-distant places (164–65). In the same vein, the vast majority of marriages in Southern Song and Yuan Mingzhou were within the prefectural borders. Of the eighty-five marriages in the Southern Song that were not, seventy were inside the two circuits of Liangzhe 兩浙; of the sixteen such marriages in the Yuan, fourteen were inside Liangzhe (Lee, 281–306, Appendix 2). What Beverly Bossler has aptly observed in the marriage patterns of grand councilor families, I think, applies to the Song political elite in general, except that the change already began to surface around the mid-eleventh century, a few decades earlier than the fall of the Northern Song. In the Northern Song, she says, “the network appears to have had a single center, a single point of convergence in the capital,” while in the Southern Song “that center had disappeared, and the network had instead numerous regional nodes.”Footnote ²⁶

Thus, the elite behavior Lee observes in Mingzhou, for both the Northern and Southern Song, concerns mostly the period after the “localist turn” had already taken place around the mid-eleventh century. Read together with Tackett's book, it actually underscores the dramatic change between the late Tang and Song: that the political elite ceased to exist as a distinct status group centered on the capital after the mid-eleventh century but became more embedded in the local society from which it had come.Footnote ²⁷ This change seems to have lasting significance. The close relationship between the officeholding elite and its native places continued to be a prominent feature of the Ming and Qing.

Why then did the capital lose its appeal? One line of explanation blames power struggles and the Jurchen invasion for making bureaucratic careers less desirable. Power struggles in the late eleventh century, it is argued, turned court politics into dangerous waters and added to the uncertainty of obtaining high office, while constant military threat from the Jurchens after 1125 tightened Song finances and limited its fields of action, making influence over central decisions “less relevant to an elite family's day-to-day life.”Footnote ²⁸ This line of argument faces several difficulties. It does not explain, for example, why the late Tang elite actively pursued high office and had impressive success in monopolizing political power and perpetuating its status, despite the nefarious struggles that plagued the ninth-century court. Neither does it explain why relocation to the capital seems to have started to lose its appeal, in fact, well before the late Northern Song factional strife and the fall of the capital. More importantly, if the localist turn around the mid-eleventh century signals the beginning of a long-term trend that continued into the Ming-Qing period, it also requires a long-term explanation for why the tide was not reversed in subsequent dynasties—an explanation that short-term events are not likely to provide.

Conversely, the dazzling political success of Southern Song Mingzhou men, like that of the Shis, lends support to the alternative explanation that by the mid-eleventh century capital residence and capital-centered networks were no longer as important for achieving political prominence as they had been in the late Tang. As Bossler observes of the Song grand councilors, “it was not so much that people had lost interest in office, but that the political and social advantages once exclusive to the center had become much more widely diffused.”Footnote ²⁹ The institution of civil service examinations as the major route to office, the expansion of quotas for prefectural examination graduates in the south, the flourishing of print culture, and the proliferation of public and private schools outside the capital in the eleventh century ended the capital's monopoly over political opportunity, making it ever more possible to translate wealth into political status via investment in education. Over time, access to the fruits of economic prosperity seems to have become more important than access to the capital for achieving political success.Footnote ³⁰ Southern examination success certainly made connections increasingly available to those outside the capital, but probably more decisive was the institution of anonymity in the examinations and the explicit prohibition of “circulating scrolls” (xingjuan 行卷), which reduced the importance of having connections.Footnote ³¹

The general trend from the ninth century on, therefore, was the expanding scope of political participation outside the narrow circle of a capital elite, eventually giving rise to an officeholding elite in local society which attained political prominence while remaining very much part of their local communities. The civil service examination system played a pivotal role in this change, and it also had a strong spillover effect. By recruiting a few hundred scholars into the civil service every three years, it encouraged many more to emulate them, bringing about in local society a literati population (shi 士) many times larger than the officeholding elite, who participated in a national culture and prided themselves on their learning. Their status as the shi was certified through enrollment in government schools and participation in the examinations.Footnote ³² These men, with their official rank and educational qualifications, constituted a new stratum of local elite—the “gentry” (jinshen 縉紳) or “literati-officials” (shidafu 士大夫). Their offices, degrees, and studentships brought not only social prestige, but also substantial fiscal and legal privileges.Footnote ³³ These privileges functioned as new markers of social standing in local society which rivaled the importance of wealth and history of local residence. The Mingzhou literati-official Lou Yue 樓鑰 (1137–1213), for example, took pride in his hometown for according respect on the basis of a person's seniority instead of his official rank. “Good feelings among fellow residents are so highly valued,” he remarked, “that a man is always ranked on the basis of his age, be he a commoner or a member of the officeholding gentry” 鄉誼最重, 薦紳韋布, 序必以齒 (Lee, 151 [translation modified]). This remark invites many possible readings. It implies that elsewhere, in Lou's times, official rank probably trumped time-honored local traditions that had valued seniority. With a slight stretch of the imagination, one may also take it to mean that a similar effort to give greater weight to official rank than age was likewise attempted in Mingzhou but successfully resisted in order not to hurt the “good feelings among fellow residents.” However one reads Lou's remarks, it is clear that these new markers of social standing threatened to disrupt age-old practices in deciding the pecking-order distinctions in local society. The relative importance of different kinds of status markers was probably constantly contested in local society. One need only compare the organizing principles of the “Genealogies of Chengdu Clans” (Chengdu shizu pu 成都氏族譜), compiled in the late 1190s, and a similar composite genealogy of the Huizhou clans compiled in the sixteenth century. The former lists only local families with unbroken traditions of civil service under the Song, groups them first by the county where they resided, and then within each group arranges them by the date when they produced their first official in the Song dynasty (qijia houxian 起家後先).Footnote ³⁴ The latter, by contrast, did not limit its coverage to the gentry clans or branches and celebrated the right of settlement, arranging different clans according to when they settled in the prefecture (ju suoqian Xin'an chaodai xianhou 據所遷新安朝代先後).Footnote ³⁵

The rise of the gentry in local society, therefore, was a product of both institutional choices made by historical actors and long-term structural changes. The examination system was capitalizing on long-term structural changes, such as economic growth and the spread of printing technology. These changes, which made education more affordable and accessible outside the capital, had been underway for centuries and were beyond the control of political leaders. As an institutional channel for translating wealth into power, the examination system in the Song did not grant everyone from everywhere equal access to government service. Instead, it replicated—if not amplified—regional disparities in economic development in the area of political participation. It ended the dominance of the capital in the examinations, but replaced it with the dominance of affluent, culturally ascendant regions, including the Zhejiang-Fujian coast and, to a lesser extent, the Chengdu basin of Sichuan. In the Song, it was these regions that benefited immensely from the new opportunities provided by the examination system. For this reason, Mingzhou was anything but typical. As a center of both domestic and overseas trade, Mingzhou was illustrative of the shift. It produced merely fifteen jinshi between 960 and 1040 but as many as 107 jinshi in the following eighty-seven years of the Northern Song. Its performance in the Southern Song was even more impressive. With a total of 747 local men receiving the jinshi degree between 1128 and 1279, it ranked third among all Southern Song prefectures.Footnote ³⁶ The experience in less advanced regions, one may imagine, would have been markedly different. Prefectures in North China, for example, with the exception of Kaifeng, produced only a tiny number of degree holders during the Northern Song.Footnote ³⁷ Northern Shanxi, Iiyama observes, was barely touched by the examination culture and the local literati population was negligible throughout the Northern Song (41).

The tension between upholding meritocratic principles and enlarging the scope of political participation fueled debate at the eleventh-century court.Footnote ³⁸ It was partially and temporarily resolved, ironically, by the Jurchen occupation of North China in 1127. A perfect complement to Lee's study on Mingzhou, Iiyama's book Kin Gen jidai no kahoku shakai to kakyo seido examines how local elite and local society evolved in vastly different historical contexts from the twelfth to the fourteenth century. He begins with the institutions of recruitment in the Jin and Yuan period, explores their social impacts in North China, and enriches his general discussion with a few case studies focused on Shanxi. One of Iiyama's most important contributions, I think, is to demonstrate how the Jurchen conquest changed the geography of examination success in the north and fostered the growth of a gentry class in northern society.

Despite Iiyama's emphasis on how differently literati were treated under the Jin and Southern Song, continuity between the Northern Song and Jin is also noteworthy. Although the Jurchens, especially the imperial Wanyan clan, nearly monopolized all high-ranking policy-making offices, it was the Chinese that satisfied the Jin's staffing needs at the lower levels of bureaucracy. From the very beginning, the Jin implemented a civil service examination system modeled on the Song and Liao practices. Over time, the Jin reformed it, expanded it, and relied on it as the primary route for recruiting Chinese talent into the civil service. Furthermore, since well-established Kaifeng families had fled south and since the highly competitive southern literati were now living under a different regime, scholars in many areas of North China actually stood a greater chance of passing the examinations in Jin times. Thus, Iiyama argues, the scope of political participation expanded and classical education spread in North China in Jin times. This trend, similar to that in South China since the mid-eleventh century, is especially evident if we compare the first and second halves of the Jin dynasty.

In the first half of the Jin, Zhongdu circuit (now Beijing and its environs) was arguably the center of literati culture in North China. Until 1189 it seems to have consistently produced the largest number of graduates in the examinations. However, its success, unlike that of Kaifeng in the Northern Song, reflects its long-standing cultural legacies more than any advantage associated with its status as the capital of the Jin (Iiyama, 75).Footnote ³⁹ It was already the center of literati culture in Liao times. In the early years of the Jin, the dispersal of literati families in Kaifeng left no place in North China capable of challenging Zhongdu's cultural preeminence.

That said, Zhongdu circuit never dominated the examinations to the extent Kaifeng had in the Northern Song. Even in its heyday during the early Jin, it seems to have claimed no more than a third of the degree holders.Footnote ⁴⁰ Its dominance clearly tapered off by the late twelfth century. During the latter half of Shizong's reign (1161–89), the Jin implemented policies highly reminiscent of the Northern Song. The examination system expanded significantly, and the government expressed interest in reviving prefectural schools, though these schools restricted their enrollment to imperial clansmen, relatives of officials, and metropolitan examination candidates (Iiyama, 85–86).Footnote ⁴¹ Estimates show that of the 10,000 or more degrees awarded in the entire duration of the Jin, the bulk were given in the latter half of the dynasty.Footnote ⁴² It was scholars from outside Zhongdu circuit who seem to have benefited most from these expansions. According to the statistics provided by Iiyama, in the 1190s, for the first time in the Jin, Zhongdu circuit lost its position as the top producer of examination graduates. Between 1190 and 1200 more literati from Hedong South (now southern Shanxi) received the coveted degrees than those from Zhongdu circuit. In the meantime, the circuits of Hedong North, Hebei West, and Shandong West (roughly northern Shanxi, western Hebei, and western Shandong today) all closed their gap with Zhongdu dramatically. The percentage of degree holders from Zhongdu circuit dropped to 15%, and it remained at that level until the end of the dynasty (Iiyama, 74–77).

The result was the appearance of a gentry class across North China in the Jin. As Iiyama demonstrates, the size of this gentry population, though only one-tenth of that in the Southern Song, was at least twice—and possibly three or four times—that in the late Northern Song (156–64). Northern Shanxi was a case in point. Situated on the borders between the Chinese, Khitans, and Tanguts, it was best known for its professional military families (jiangjia 將家) since the fall of the Tang and was barely touched by the examination culture of the Northern Song (114–18). Shanxi produced a very small number of degree holders under the Northern Song and Liao, and they came almost exclusively from Taiyuan and places to its south. It was only in the Jin, Iiyama shows, that classical culture spread to northern Shanxi and literati population there expanded dramatically (111–16).Footnote ⁴³ Several literati families emerged during the Jin, including the Lius 劉 of Hunyuan 渾源 and the Yuans 元 of Xiurong 秀容, both of which boasted a tradition in belles-lettres. The Yuans, for example, began as a military family in the Northern Song and produced their first degree holder in 1157 under the Jin. Thereafter, they continued to take the examinations and hold civil offices using facilitated degrees or the yin privilege, even though they did not produce another metropolitan examination graduate until 1221 (110). The Lius were even more successful, having at least two metropolitan degree holders in each generation from their initial success in 1124 until the fall of the Jin (118–19).

In both the Song and Jin, it was clearly the examination system that provided the institutional underpinning for the rise of the gentry in local society. It offered incentives for local elite families to participate in a national culture, opportunities for them to enter government service, and state certifications for claiming themselves as a prestigious stratum in local society. The breakdown of the examination system in the Yuan dynasty, therefore, presented a serious challenge for the local gentry in both North and South China.

Lee and Iiyama both emphasize notable discontinuity brought on by the Mongol conquest and closely associated with institutional changes in the Yuan. They show that scholars in the Yuan period faced an array of obstacles to gaining official rank and maintaining their local status. Although the Yuan state continued to provide ways for local literati to define themselves as a social stratum different from the commoners, such as registering themselves as “Confucian households” (ruhu 儒戶) and enrolling in local government schools, these opportunities were few and the fiscal and legal privileges associated with such status, if any at all, were very limited and insecure and often subject to the whims of local officials (Lee, 216; Iiyama, 335). Moreover, their access to political power also became more restricted than ever before. With classical education no longer a prerequisite for entering government service in Yuan times, scholars now had to compete with military talents, financial experts, and men with other types of specialized expertise, to whom the gate to power had opened wide in the Yuan period. The situation was exacerbated by the breakdown of the examination system. Though a shortcut to high office later in the dynasty, the examinations were not revived until 1314 and graduated only a negligible number of candidates for the rest of the dynasty.Footnote ⁴⁴ What added to the difficulty of translating cultural attainments into political and social status was a new definition of what counted as cultural attainments. When the examinations were resumed, they tested Neo-Confucian moral philosophy rather than belles-lettres, the favorite subject for candidates during the Jin and much of the Southern Song. As the importance of the examinations declined, many scholars began to take up teaching jobs in local academies, clerical posts in local administrations, and managerial assignments under the Mongol nobility who had fiefs (ayimaqs) nearby. Promotion to a higher rank, though possible, was competitive and uncertain, relying heavily on one's ability to secure recommendations from high-ranking officials or Mongol noblemen either at court or in the provinces (Iiyama, 216–89).Footnote ⁴⁵

Very few literati families adapted themselves successfully to these changes. The Duans 段 of Jishan 稷山 in southern Shanxi, discussed at length by Iiyama, was one of the exceptions. Having produced one degree holder in the late Northern Song and three others in the Jin, the Duans continued to attain official rank until the early fourteenth century. Their success owed mainly to two of their strategic moves. On the one hand, they advisedly shifted their focus of learning from literary composition to moral philosophy, and on the other hand, by having family members serve in clerical offices at court, they built up connections in the capital with court officials like Yan Fu 閻復 (1236–1312), through whose recommendations more family members enrolled in the Directorate of Education and received bureaucratic appointments (Iiyama, 359–64).

Most literati families, however, were not so lucky. In Shanxi, Iiyama shows, literati families that produced degree holders in the Jin almost all disappeared from historical records in the Yuan times (Iiyama, 351). In Mingzhou, Lee contends, even prominent Southern Song families, like the Shis and Lous, could not avoid gradual decline, and many struggled to preserve their elite status by embarking on a wide range of career paths, such as clerks, teachers, and sea transport managers (Lee, 216–17). The result was a noticeable change in both the makeup and social character of the officeholding local elite. It was no longer constituted solely by men with a classical education. Instead, they were joined by those who proved their worth in more practical fields of activity, such as military service and maritime transport. Among these men with practical skills, some were born to former literati families, and others, like the Nis 倪 of Mingzhou, came from more modest backgrounds (Lee, 223–24).

The breakdown of the examination system, as Lee and Iiyama both rightly point out, put literati and other aspirant local families in a weak position vis-à-vis the government. It deprived local literati not only of a reliable means of achieving high rank but also of an officially recognized mechanism of defining their group identity and social status (Lee, 207–8). Since fiscal and legal privileges were often granted at the discretion of local officials and gaining political success was ever more dependent on social connections, maintaining an amicable relationship with local and court officials became all the more important. Thus, ironically, as Lee points out, the breakdown of the examination system decreased only the supply of, not the demand for, connections to the state in Yuan times. It turned these connections into “a rare commodity” for the local elite and intensified their need to reach out actively to local officials, which is evident in the lavish praise they heaped on the Yuan state and its local agents in a variety of panegyrics, such as “parting prefaces” (songxing xu 送行序) and “stele in appreciation for departing officials” (qusi bei 去思碑) (Lee, 258–59, 263).Footnote ⁴⁶ They took pride in their connections with Yuan officials by circulating these panegyrics in their literary collections. In the way that Iiyama puts it, whatever privileges the Yuan local elite enjoyed were no longer guaranteed by established institutions and time-honored consensus (kanshūteki 慣習的). For that reason, the process by which the Yuan local elite gained privileges and established itself as local leaders was, at the same time, also a process that empowered the Yuan state (Iiyama, 206).

As notable as they were, these changes should not cloud our vision of the continuity between the Yuan and earlier times. Despite heavy reliance on personal connections, the “diversification of paths to government service” (shusshi keiro no takika 出仕経路の多岐化), to use Iiyama's phrase, recognized the value of distinctive local traditions which nurtured talents in, for example, military service (e.g., in Dingxiang 定襄 county of northern Shanxi; Iiyama, 185–91, 205) and maritime transport (e.g., in Mingzhou; Lee, 222–25). As a result, it offered opportunities for political participation across the empire without privileging a few affluent and culturally ascendant regions, an issue which had constantly plagued the Song and was only partially resolved by political disunity in the twelfth and thirteenth centuries. The long-term trend of expanding political opportunities outside capital circles, therefore, continued in Yuan times and benefited especially those regions which had lacked a strong tradition of classical education (Iiyama, 198, 205).

Secondly, as in the Song and Jin, the expansion of political opportunities in the Yuan created a locally embedded, office-holding elite in the provinces, although not all members of this elite were literati. In fact, government service and local dominance became ever more entangled in the Yuan period. In Dingxiang, for example, at the beginning of the Mongol conquest, the appointment of local men to county and prefectural offices meant little more than an official recognition of whatever dominance these strongmen had already achieved in their home locales in exchange for their willingness to cooperate with the conquerors (Iiyama, 190–93). Even after Mongol rule took hold, as time progressed many of those who pursued a career in the new government continued to be employed near their hometowns as teachers, clerks, officers, tax collectors, and sea transport supervisors. Even for those who set their eyes on national prominence, a solid local base was critical. Many started out on their careers locally and first proved their worthiness to local officials or to Mongol noblemen enfeoffed nearby. Clerical service in the central government and networking in the capital were extremely valuable, but these opportunities were few and highly unpredictable. Even the Duans, Iiyama shows, saw their carefully constructed capital networks collapse following the death of Yan Fu in 1312. Thereafter, the Duans failed to produce a single officeholder during the Yuan. Many of those who did try their luck at the capital often made the trip only after they had gained administrative experience and built up connections near their homes (Iiyama, 233–34).

With careers in government so unpredictable and so tightly bound to a family's social station at home, it is not difficult to understand why many office-holding families in the Yuan, whether from Shanxi or Mingzhou, seem to have continued to live in their native places instead of relocating to the capital. Their marriage networks, too, were reminiscent of the Song political elite after the mid-eleventh century, centered on their hometowns but extending frequently to nearby counties and prefectures (Iiyama, 192–94; Lee 226–32). The Yuan political practice, in other words, seems to have carried on the long-term trend that began in the Song period: the expansion of political opportunities outside the capital and the attendant rise of a local elite that claimed an exalted social standing on the basis of its connections to the state. The diversification of paths to office in the Yuan times, in fact, gave this elite a more widespread presence across the empire rather than limiting it to only a few culturally ascendant areas like Song dynasty Mingzhou.

This was evidently a major reason that deterred the first few Ming emperors from fully reviving the examination system. Reintroducing southern scholars into the civil service examinations under a unified empire would almost certainly force northern literati out of the jinshi ranks and consequently narrow the scope of political participation to an extent unseen in two hundred years. The metropolitan examination of 1397 attests to this. Initially, all fifty-one graduates for that year were southern literati. The emperor was outraged, and his desire to ensure a balanced representation of different regions in his government was evident in his decision to reissue the roll of jinshi degree holders with sixty-one new graduates, all northerners. Only after 1450, when an acceptable regional quota system had been put into place, did civil service examinations become the chief route to office.Footnote ⁴⁷ A wider geographical scope of political participation was also realized, to a limited degree, by allowing provincial examination graduates (juren 舉人) to hold office. Thus, although the rise of the gentry—locally rooted literati families who had official titles, examination degrees, or student status in local government schools—began in Song times, it was not until the mid-fifteenth century that it extended beyond a few culturally ascendant prefectures. The dramatic increase of metropolitan and provincial degree holders after 1440 that McDermott documents, both in the Shanhe Chengs’ descent group and in Huizhou at large, is also evidence of this change (McDermott, 260–61).Footnote ⁴⁸

GOVERNMENT AND LOCAL ELITE ACTIVISM IN THE SONG AND YUAN

The “localist turn” in the Song period marked the onset of the secular rise of a gentry class in local society. Already in the Song, this was accompanied by the gentry's aggressive moves to affirm a collective identity and expand their role in local governance. They did so by promoting new institutions and practices and repurposing existing ones. In Southern Song Mingzhou, for example, the gentry revived and actively promoted the community drinking ceremony (xiang yinjiu li 鄉飲酒禮) and, by limiting its attendees to “local officeholders and literati” (xiang dafu shi 鄉大夫士), made it a collective ritual exclusively for the local gentry and a public demonstration of their social distinctiveness (Lee, 171). The gentry also claimed for themselves the role of mediators between local society and government, presenting themselves as spokesmen for local interests and advisers on local policies. Government schools, which earlier reformers like Wang Anshi 王安石 (1021–86) had promoted as an instrument of imposing ideological uniformity and as a training ground for capable officials, were touted in many Southern Song writings as a venue for local literati to “deliberate on governing” (yizheng 議政) and offer counsel to local officials (Lee, 163–66). In the Yuan, local literati continued to present themselves as “representative voices of local society.” In the panegyrics they composed, Lee argues, they included subtle exhortations to local officials, “informing them of priorities in local governance” (Lee, 254). Therefore, as Lee underscores, the literati's growing attachment to their native places by no means implies their separation from the state. Rather, the localization of the national political elite was, when viewed from the local perspective, also the politicization and nationalization of the local elite. As new state-certified status markers were introduced into the local scene, complicating the definition of social positions, gentry families asserted their distinctiveness and prestige against other members of the local elite precisely in terms of their connections to the state and their ability to negotiate with it.

The close relationship between the state and the gentry also stands out in Lee's discussion of local activism in Southern Song Mingzhou. He argues that although the Southern Song witnessed an upsurge of communal efforts spearheaded by the local elite, this does not imply a retreat of the state from local society. Rather, the success of these efforts depended heavily on the local elite's cooperation with local government. Not only did the local government, in general, make a greater contribution than local elites in the development of Mingzhou, but the financial and administrative support it provided were also indispensable to many local projects initiated and advocated by local elites. Very few undertakings—including institutions and practices of central importance to the gentry, such as the renovation of schools and the performance of the community drinking ceremony—were accomplished without government support. As Lee aptly puts it, “the rise and spread of local [elite] activism during the Southern Song period… was undergirded by the activism of local government” (198). Both Lee and McDermott show that even landed trusts of kinship groups, including the famous one started by Fan Zhongyan and the charitable estate of the Lous in Mingzhou, had difficulty lasting beyond a few generations without timely intervention of local government in the Song and Yuan times (McDermott, 133–39; Lee, 234).

Why so? The answer, I think, has to do with both opportunity and capability. If it was at all possible to use government funds or someone else's money to advance one's own goals, then why not? The substantial investment that the Mingzhou government made in local projects during the Southern Song certainly demonstrates the state's resourcefulness and interest in sustaining local economic and cultural development, but, from another perspective, it can also be seen as evidence of the Mingzhou elite's exceptional ability in swaying government decisions and channeling government revenues to serve the elite's agendas. Having office-holding native sons, as was increasingly the trend after the mid-eleventh century, certainly helped. This is particularly evident in the construction of the Wujin dike (Wujin qi 烏金碶) that Lee discusses in great detail. In this case, the local elite of Mingzhou first petitioned Shi Miyuan 史彌遠 (1164–1233), a native son and then grand councilor, and Shi, in turn, did his job of urging the prefectural government to take action and cover the bulk of the expenses (Lee, 114–16). With all these opportunities of using government funds for local purposes, private investment was not necessary. It became necessary only when the state was unable to bear the full costs or when the local elite lost its influence over local officials. Thus, it is not difficult to understand why, given the tighter budgets of county administrations and the dynastic crisis, local elite contributions were more conspicuous at county levels and especially in the early years of the Southern Song (Lee, 97, 156). Nor is it difficult to understand why, given the literati's modest political success, insecure social station, and greater cultural distance from local administrators, local projects in Yuan-dynasty Mingzhou received far less financial support from the local government (Lee, 207, 232–44).

The prominent role of Mingzhou government in supporting local development was evidently also a result of the local elite's inability to mobilize funds and manpower by itself. Probably large-scale projects at all times had to depend heavily, even though not exclusively, on government support. As Wei Xian 魏峴, a retired official and sojourner of Mingzhou who took an unusual interest in water management, learned from his own experience, “I often contributed my own money to hire people and dredge the streams, but the resources of private households can never match those of the government” (Lee, 121 [translation modified]). That said, Lee's account of Southern Song Mingzhou portrays a local elite whose heavy reliance on the formal bureaucratic apparatus in carrying out local projects contrasts markedly with what historians have long observed in the Ming-Qing period. The reason, I think, has to do with the absence of social institutions in Song times that the local elite could rely on to mobilize resources in the community without resorting to the state apparatus.

In the Song, the most resource-rich institutions outside the state seem to have been Buddhist establishments, which had owned land and forests and engaged in a wide variety of profit-making activities for centuries. In the early thirteenth century, Mingzhou had a total of 305 Buddhist temples, each owning on average 229 mu of farming land and 293,723 mu of forests.Footnote ⁴⁹ Their relationship with the government and the local elite was complicated. Officials and wealthy local families were sometimes patrons, sometimes predators, and at other times collaborators. Many local families set up merit cloisters or private chapels funded with land or just rents to look after their ancestral graves and provide seasonal offerings; others registered their land under the nominal ownership of monasteries to evade taxes and labor service charges. It was not uncommon, however, for the very powerful to simply seize a temple and its property by force and then have it designated as their grave temple. Many literati associated with monks and studied at monasteries, and the Song government typically used temple grounds to host civil service examinations. These temples played an active role in local development throughout the Song, building and maintaining bridges, roads, dikes, and floodgates, sometimes all on their own, while at other times in collaboration with the local government and the local elite.Footnote ⁵⁰ At the very least, Buddhist establishments were clearly the only corporate entities that had their own fiscal autonomy and acted independently in the local community. Their resources—even lands donated to private chapels—were, at least on paper, beyond the reach of the local elite. After all, the purpose of setting up these grave chapels was precisely to set aside a portion of the family fortune for the long-term upkeep of ancestral graves and sacrifices, entrusting its management to the monks, and thus protecting it from unworthy descendants.

In the absence of effective means for pooling resources and coordinating action, it is not surprising that the local elite in Southern Song Mingzhou would rely heavily on the organizational capacity provided by local government, especially in projects that required routine maintenance. To ensure regular dredging of Dongqian Lake (Dongqian hu 東錢湖) and the Tuoshan Dam (Tuoshan yan 它山堰), for example, the government of Mingzhou not only established landed estates with prefectural funds to cover the expenses but also arranged for the management of these estates. In the Song, the government typically assigned each rural household into one of five grades based on its assets and made each household take turns to fulfill certain labor service obligations appropriate to its grade, such as tax collection and local policing. In the above cases, evidence shows that the same mechanism was employed in assigning managerial duties over the endowed fields. These duties were either entrusted to village servicemen, such as the canton leader (xiangzhang 鄉長), or rotated among households in the top grades (“graded households in nearby cantons with the greatest material resources” [近鄉等戶物力最高者]; “upper households in the canton” [鄉之上戶]). The only non-governmental institutions involved in these projects were “nearby Buddhist monasteries” (jinhu siyuan 近湖寺院), where harvested grain from the endowed fields was stored.Footnote ⁵¹

By comparison, kinship institutions were remarkably weak in the Song and Yuan. As will be discussed later, collective ancestral rites were practiced by small kinship groups by the gravesite and lineage halls did not exist. Although references to less institutionalized charitable action towards kinsmen were abundant, kinship-based charitable estates were few in Song and Yuan times. Where corporate estates of descent groups existed, they were usually very small in size, intended mainly to pay for sacrificial expenses, and under the constant danger of being divided up.Footnote ⁵² Even prominent families, like the Shis of Mingzhou, had neither charitable estates nor corporate sacrificial lands.Footnote ⁵³ The principal form of rural organization was the village worship association. Although it gradually took on some nonreligious tasks after the early thirteenth century, if not earlier, the village worship association remained primarily a ritual organization until at least the mid-Ming and owned minimal property.Footnote ⁵⁴ It is the long-term evolution of these local institutions, from the Song to the Ming-Qing, that McDermott and Du have focused on in their studies. It is to the history of these institutions we will turn now.

VILLAGE WORSHIP ASSOCIATIONS

McDermott and Du have both demonstrated that, in Huizhou and Zezhou alike, villages became more organized on their own terms in Song times. But they were organized not on the basis of common descent, but on the basis of common residence and joint participation in the worship of local gods. Du, in particular, provides a compelling account of the appearance of village worship associations as a form of communal organization in Northern Song Zezhou and the expansion of their power in the following centuries. Placing their works within the broader scholarship on religious and social history suggests several related areas of change. First, social and religious life in the village acquired an institutional form unknown before Song times. Collective worship of earth spirits (she 社) had a history going back to antiquity. In the Han dynasty, the government mandated seasonal sacrifices, also called she, to earth spirits at all levels of administration, from commanderies and districts down to administrative villages (li 里). These sacrifices were organized by officials and village officers. As early as the Han, rural residents also started organizing mutual-aid societies (dan 僤) and Buddhist worship associations (yiyi 邑義 or fayi 法義), which had compacts and managers and recruited members on a voluntary basis.Footnote ⁵⁵ These organizations, especially the Buddhist worship societies, flourished after the fall of the Han and many, by the Tang period, also called themselves she 社 and had become formally organized.Footnote ⁵⁶ In the Song, if not in late Tang, these two aspects of she—as communal sacrifices to earth spirits and as a form of voluntary organization for mutual aid or collective worship—began to conflate, leading to the rise of an “inclusive territorial institution” in the countryside based on ascriptive membership and centered on collective worship which, again called she, were often—though not always—coextensive with a hamlet or natural village (Du, 8–10, 254–57; McDermott, 53–58).Footnote ⁵⁷ As a widespread form of village self-organization, these she also provided the building blocks for multi-village worship communities, which began to emerge in the Northern Song.

The formation of these multi-village worship communities resulted from momentous changes in local religious life in Song times. Increasingly each god or spirit was given a distinct identity with close local ties and thereby transformed into a local tutelary deity. Recent scholarship has shown that since the late Tang, and especially during the Song, previously nameless earth spirits which were believed to literally inhabit the soil were now identified as historical or legendary figures in local history. They “shed the violent, morally ambivalent character and grotesque forms” of their older selves and became guardians of local welfare. They were worshipped not because their angry spirits needed to be appeased, but more because they responded promptly to a wide range of local needs and demonstrated their efficacy in protecting the local community.Footnote ⁵⁸ In Huizhou during the Song and Yuan, McDermott shows, earth spirits were increasingly anthropomorphized into historical celebrities with a local connection—such as Cheng Lingxi 程靈洗 and Wang Hua 汪華—who usually came from a military background, founded the village, or at some time protected the village or its general area (56–61). Rain gods in southern Shanxi, argues Du, underwent a similar transformation. Before the Song, local people made prayers for rain to nameless dragons near local rivers and lakes, where they supposedly dwelled (24–25). Come the Song, rain gods were given an identity with local ties, real or fictitious. In some areas, local people worshiped a historical figure like Cheng Tang 成湯, the first king of Shang, as a rain god and claimed that it was near a local river where the king had successfully prayed for rain in history (28–36). Where such local connections were not found, local ties were invented on the basis of efficacy. The cult of dragons continued, but it was now believed that the White Dragon on Mt. Yan (Yanshan bailong 崦山白龍) was particularly efficacious (36–43). In the high tide of imperial promotion of Daoism, the Jade Emperor was also appropriated as a rain god in Shanxi, but again prayers to him in a particular village was believed to be more effective than elsewhere (51–52). These deities were increasingly adopted by local people as their tutelary gods from the Song onward. These changes, I think, paved the way for wide-ranging religious networks to emerge between villages. The greater emphasis on the gods’ role as a guardian of—rather than an threat to—local welfare and the process of “branding,” which differentiated otherwise nameless gods from each other, created a market for different gods, each with his or her distinct identity and associated with a different locale, to compete for patronage with his or her ability to address local needs. Some deities became superstars and attracted a large following in a wide area. These deities did not displace the tutelary god of each individual village (Du, 59–61). But they and the places they were associated with evolved into, to use the words of Richard von Glahn, “focal points for communal identity and welfare” in a wider geographic area and thereby gave rise to multi-village worship communities. The village worship association also acquired a new significance: it became the principal form in which each constituent village participated in the collective activities of these multi-village communities.Footnote ⁵⁹

A related change in the Song religious practices was the growing tendency to shelter these tutelary gods in a small shrine rather than making offerings to them under a tree, near a lake or river, or at an open-air altar (Du, 53).Footnote ⁶⁰ The building of these shrines seems to have flourished after the eleventh century, probably reflecting both local economic growth and a more tolerant religious policy by the Song government.Footnote ⁶¹ Such an activity frequently required cooperation among several natural villages. In Shanxi, for example, the 1076 construction of a shrine dedicated to the Jade Emperor in Fucheng 府城 village was sponsored by fifteen village worship associations within a radius of five kilometers of Fucheng, each headed by a deacon, elder, registrar, or the like—titles borrowed from Buddhist and government institutions. Once constructed, the shrine became the corporate property of those villages participating in its construction, which, in turn, strengthened both the multi-village alliance and its basic organizational units (i.e., its constituent village worship associations). In many cases, it created a multi-village alliance centered on the shrine that basically survived into the eighteenth century (Du, 56–59). Likewise, the construction of the White Dragon Shrine on Mt. Yan also involved several villages. After a Yuan-dynasty renovation, it became the corporate property of four participating village worship associations, forming an alliance that also lasted even until today (Du, 40–42).Footnote ⁶²

These developments did not take place in the absence of imperial authority. On the one hand, the Song court served as a powerful source of legitimacy that loomed large in the entire process. Cheng Tang was a sage-king in the Confucian tradition. The Jade Emperor was a Daoist deity who received generous patronage from the court (Du, 25–26). Cheng Lingxi and Wang Hua, while protecting the local community from the rebels, also demonstrated their loyalty to the dynasty (McDermott, 57–60). The adoption and appropriation of these religious symbols for local purposes, therefore, also reaffirmed the authority of the state in the cultural and religious realms. The court's bestowal of noble titles was also valued as an explicit form of legitimation, especially for those deities who lacked historical ties to the dynasty and its cultural traditions. The strenuous effort of local people to prove that the White Dragon on Mt. Yan had actually received enfeoffment from the Tang and Song court underscores the importance of state recognition (Du, 38–39).

On the other hand, the village worship organization's organizational capacity for mobilizing local resources and coordinating communal activities made it the government's chief target for cooptation after the thirteenth century. In 1206, the Jin made the village worship association the basic unit of its rural administration, requiring its head to assist the village head in registering the local population, collecting taxes, maintaining order, and promoting agriculture. In the Yuan, its duties were further extended to the running of a village school and its shrine received generous patronage from local officials (McDermott, 64; Du, 70–71). The early Ming government continued to have the village worship association function as the basic unit of its rural administrative structure, but it also made more aggressive moves to bring its ritual practices into line with canonical requirements. Then in the early sixteenth century it further attempted to have some Huizhou village worship associations take on new tasks of rural governance, such as running village granaries and providing moral-instruction lectures (McDermott, 180–81; Du, 200–216, 257). Although these early and mid-Ming government efforts had limited success, these policies attest to the village worship association's ubiquitous presence in the countryside and the state's growing interest in coopting this rural institution and prescribing and regulating its functions.

In sum, the Song witnessed the beginning of rural residents organizing themselves based on common residence and shared religious symbols. Village worship associations (she) began to develop when collective worship or mutual aid societies (she), previously based on voluntary subscriptions, acquired an ascriptive and territorial dimension and came to define the community of a village or hamlet which participated in seasonal sacrifices (the she sacrifices) to the same local deity. These village-wide associations seems to have developed greater organizational capacity from the eleventh century onward, because the rise of supra-village tutelary gods and the growing trend of worshipping them in shrines, rather than in the open air, required the participating villages to marshal local resources for the construction of shrines and to coordinate their religious activities. The tutelary god and the supra-village shrine where he or she was sheltered, then, also served as a binding force, both spiritual and economic, for an enduring alliance among the component village worship associations. As such, these village worship associations and multi-village alliances that were built on top of them, as an organizational framework for local life in Song times, evolved primarily as a result of local, rather than imperial, forces. They also enjoyed remarkable independence from the gentry, who, after all, had not yet become a recognizable social stratum in Shanxi and many other areas at the time. This arguably made their heads and managers another type of local elite—as opposed to the gentry—whose power was rooted in local institutions rather than in official titles or degrees conferred by the state.

But the Song was only the beginning. From the Song to the first half of the Ming, the village worship associations seem to have controlled little, if any, corporate property besides their communal shrines, which, in turn, seems to have limited their sphere of activity largely to local religious affairs, despite the government's repeated attempts to have them perform some duties of rural governance. It was only after the mid-sixteenth century that the village worship association embarked on a sustained expansion in its resources, responsibilities, and coercive power and gradually evolved into a property-holding, self-governing body in the countryside.

In Zezhou, Du contends, some associations began to own corporate property and engage in moneylending by the mid-sixteenth century (226). No later than the late Ming, the village worship association began to expand its sphere of activity beyond religious affairs and claimed the power to police its members. In 1651, after a large-scale renovation of its local shrine dedicated to Cheng Tang, the worship association of Guoyu 郭峪 market town in Zezhou set up two landed estates, tightened regulations on the use of its collective property, and threatened to enforce them with regular policing. Three years later, it assumed responsibility for the upkeep of the town's defensive wall and its attached living spaces. Through “public discussion of the entire association” (heshe gongyi 闔社公議), it publicized a set of rules regulating the use and management of these structures, asserted its power to collect rents and impose fines, and declared these revenues its corporate property. This provided a stable source of income to cover the association's routine operating expenses (137–40). The authority of village worship associations underwent an even more dramatic expansion in the mid-Qing. In the mid-eighteenth century, many village worship associations no longer relied merely on voluntary donations to finance public projects but instead imposed levies directly on their members based on the acreage of their landholdings. Thenceforth, “shares in the association” (shefen 社分) came to denote a bundle of rights and obligations associated with landownership and was transferred during land transactions. After the mid-Qing, village worship associations regularly claimed authority in a wide range of activities in local life, including farming, sericulture, mining, lumbering, gambling, begging, funerals, and proper family conduct. They elaborated their procedures of assembly and deliberation and, asserting supremacy over competing channels of village authority, issued rules on how the government's village officers should be selected. In what may be considered a confirmation of the self-governing powers the village worship association had acquired in Zezhou, the Qing government, from the late eighteenth century on, gradually abandoned the lijia 里甲system and turned to village worship associations for carrying out many of its tasks (215–30).

LINEAGE ORGANIZATIONS

In Huizhou, by contrast, the village worship association's institutional dominance ended in the mid-Ming because of the rise of the corporate lineage. Some of the core ideas and practices of the corporate lineage apparently had an origin in Song times, just like the village worship association, but the development of the corporate lineage followed a very different path.

If the formation of the village worship association was driven, above all, by religious faith and the efforts of the commoners, the evolution of kinship institutions in the Song dynasty was promoted mainly by the gentry and motivated primarily by social and economic concerns: “a fear of the disaster of inequality” within the descent group and the persistent pressure of downward mobility resulting from the privatization of land after the mid-eighth century and the age-old practice of equitable partible inheritance (McDermott, 109, 149). The solution was to establish some form of shared property within a large kinship group beyond the household level. McDermott compares two distinct models of managing shared property and organizing large kinship groups, shaped both by the perennial dilemma between limited resources and growing membership and by new economic and social conditions from the Song period. The dominant type of large kinship organization until the eleventh century had been the large communal family (yihu 義戶), which required all its members to have proximate—if not shared—residence and that their “dining, education, and other social activities were to be communal and tightly regulated.” It was, therefore, “a self-contained community,” which required its members to completely “surrender their private property and much of their social autonomy to rule by a senior figure” and in return promised physical and material security. It also valued economic equality over productivity and focused on the retention of labor. All these features bespoke its origin in the wake of the collapse of the Han dynasty and its flourishing in the dying years of the Tang (117–19). As time progressed, however, sustained peace made physical security less of a concern; in Song times, civil service and other opportunities required successful members of the family to be more mobile, and the extension of arable land became more important than the retention of labor in the increasingly crowded countryside of southeast China. In 1050, Fan Zhongyan proposed a new approach to the management of shared property in response to these new social and economic conditions. His proposal was to establish a charitable estate (yizhuang 義莊)—or “lineage trust,” as McDermott terms it—as a separate property-holding unit without requiring comprehensive control over its beneficiaries’ private wealth, residence, and other social activities. By separating the corporate property from the private and refocusing its activity from the retention of labor to the management of a corporate estate, the trust thus provided a loose and inclusive framework that “limited its members’ claims on the trust's property” and simultaneously limited the trust's claims on its members. Despite their fairly insignificant number in the Song and Yuan dynasties, McDermott argues that it was the lineage-trust model of kinship-based shared property management that gradually gained popularity in south China from the late twelfth century onward (114–15, 118). He shows, for example, that the lure of the large communal family faded for the Wangs of Wukou and their relations in the mid-twelfth century and for the Jins of Xiadong in the mid-thirteenth century. The Wangs constituted a large communal family which had broken up in 1024. In 1158, when a member of an offshoot branch attempted to strengthen ties within his kinsmen, it was Fan's model that he decided to follow (140–58).

With some hindsight, Fan Zhongyan's model of a land trust is typically viewed as a precursor to the trust-based lineages of Ming and Qing times. But two important differences are worth noting. First, Fan's goal was fairly limited and that of some of his followers even less grand (146). Their concerns were primarily economic. The mission of Fan's trust was to provide “its members with no more than a portion of their food and clothing needs, of their funeral and wedding costs, and of their other living expenses, and it devolved most decisions about their daily life onto their separate branch, segment, or household units” (118). While Fan had offered these benefits to all his kinsmen, as time progressed, some trusts limited their goal to no more than ensuring the basic subsistence of their poor members in order to maintain the balance between membership numbers and resource supplies (146). As McDermott points out, Fan's rules allowed most members to escape “with no specified obligation other than to be educated and to attend certain annual performances of ancestral worship, if even that” (120). They expressed little concern for ritual practices or any other extra-economic ties that could unite the trust's members into a tightly organized group. Conversely, Fan's heirs continued to have Buddhist monks look after their ancestors’ graves and chapel (122).

In fact, ritual practices central to the development of a trust-based lineage after mid-Ming continued to be ignored or deemed inappropriate in Song times. Collective ancestral worship in the Song and Yuan Huizhou and elsewhere was usually performed at the graveside by small groups of kinsmen, roughly the size of a traditional mourning group (wufu 五服), who visited the graves of the Prime Ancestor and the four immediately preceding generations at seasonal occasions (99). Worshipping ancestors in a detached building attracted few officials and was off-limits to ordinary families without official rank. Even Zhu Xi reserved the practice as a privilege of scholars and officials and, still more or less following canonical requirements, limited the worship to the Prime Ancestor and a person's four immediately preceding generations. In Huizhou, McDermott points out, merely four ancestral halls in the Song and eleven in the Yuan are recorded (103).

Second, those who were attracted by Fan's model and established land trusts for their kin were typically members of the gentry, men who had a Confucian education and often held office.Footnote ⁶³ Therefore, the growing popularity of lineage trusts in the Song and Yuan bespoke more the gentry's interest in strengthening ties in a large kinship group than a new kinship orientation in broader segments of society. Gentry families after the eleventh century also showed an interest in uniting same-surname groups on the basis of a putative First Settler or Prime Ancestor. The different Li groups in Sichuan, for example, compared their genealogies in the twelfth century and tried to reconcile conflicting accounts of their ancestral origin in the belief that they were actually descended from a common ancestor. Since the settlements of these Li groups, however, were so dispersed geographically—often in different prefectures and sometimes different circuits—and the effort to unite their genealogies seems to have concerned only literati families, these moves probably expressed the elite's desire for networking and were thus quite different from the widespread Ming trend for same-surname groups to merge into a single organized lineage.Footnote ⁶⁴ Recent scholarship on genealogy-writing and ancestral halls leads to similar conclusions. Patricia Ebrey and Robert Hymes both trace genealogies of the late imperial style back to the Southern Song. But they also argue that genealogy writing in the Southern Song and Yuan was mainly a practice of the literati and officials, who used it to establish social status and build political alliances as well as to provide practical information on who participated in graveside ancestral worships and where.Footnote ⁶⁵ In Huizhou, a text from no earlier than the Yuan and quite possibly the mid-Ming reports that “the poor and the rich or titled in his lineage participated in separate performances of ancestral worship rites” (McDermott, 106).Footnote ⁶⁶ In his case study of Fujian, Michael Szonyi argues that until the mid-Ming ancestral halls were constructed mainly by the gentry and descendants of prominent officials as a marker of their social status which set them apart from the rest of local society. They “stress exclusivity not commonality.”Footnote ⁶⁷

McDermott's detailed account of lineage rites and property in the Song and Yuan reveals how reluctant the gentry was to build a detached building for ancestral worship (100–103) and how frequently the gentry efforts of setting up a lineage trust failed in just two or three generations (133–58). The development of trust-based lineages in the Ming, by contrast, was driven by a different social group and a different constellation of social forces. Since the 1990s, a community of scholars, known to some as the “South China School” (Huanan xuepai 華南學派), have adopted a distinctive methodology in studying late imperial and modern China, which combines conventional historical research and fieldwork and draws attention to how written records were produced and used in the local setting. Focusing primarily on Guangdong and Fujian, these scholars—Zheng Zhenman, David Faure, and Michael Szonyi in particular—refute the “trickle down” theory that attributes the widespread growth of lineages in the late imperial period to the influence of Confucian discourses and elite model of proper behavior. Instead, they view lineage organization as a result of strategic attempts by local people, including many commoner families, who appropriated state-sponsored discourses and flexibly interpreted them to meet their own needs.Footnote ⁶⁸ Oftentimes the formation of a lineage began with registering land as a corporate trust under a single progenitor in order to fulfill and limit the descent group's shared military and fiscal obligations to the Ming state or to keep inalienable assets under collective ownership. Starting from the early sixteenth century, when ancestral halls spread through local society as a popular place to perform sacrifices, these corporate trusts were welded to these halls, transforming them into powerful property-holding institutions. These ancestral halls, unlike the gentry ones of earlier times, were “built not by members of the educated elite alone but by individuals from a wider spectrum of local elites, including merchants, moneylenders, and large landowners who did not necessarily hold degrees.” They stressed inclusivity, underwrote the elite's local leadership, and served as an “instrument for the social regulation through ritual of one's kin.” Footnote ⁶⁹ The rituals performed in these halls also deviated frequently from canonical requirements and incorporated many preexisting local religious practices.Footnote ⁷⁰ In sum, the spread of lineage institutions in the Ming and Qing—the establishment of lineage trusts and the building of ancestral halls alike—was shaped mainly by local and non-gentry forces, rather than by the visions and deliberate efforts of the educated elite or the state. In Huizhou, as McDermott also notes, the commoners’ right to build an ancestral hall was recognized only ex post facto by the court, a generation later (178–79). These findings have complicated the relationship between the Song kinship practices and the Ming-Qing lineage organizations. Rather than viewing the Song kinship-based estates and genealogy writing as the direct precursor of Ming-Qing lineage institutions, they suggest that all the Song provided was a set of ideas and experiments on kinship-based landed estates which would have never become a reality had they not been triggered by a confluence of social and political conditions in the Ming.

McDermott's study of Huizhou supplies plentiful evidence that corroborates this view. The development of lineage institutions in Huizhou began in the first half of the fifteenth century not with compiling genealogies or building ancestral halls, but with the founding of landed trusts modeled on Fan Zhongyan's charitable estate. These landed trusts were set up by rich and ambitious local men who often boasted military training or experience but had no office or degree (8, 235). The Doushan 竇山 trust of the Cheng lineage in Shanhe district, to which McDermott devotes half of his book, was formally established in the mid-fifteenth century by a grandson of a retired Ming general. He was a “local boss,” but the basis of his social position was “not an official degree, title, or post,” but “his local wealth and his family's military past.” His decision to establish the trust was evidently prompted by the family's registration as a military household (junhu 軍戶), which entailed a permanent obligation to provide military service to the Ming dynasty (245–47). The construction of ancestral halls did not concern the Chengs—or the Huizhou families in general—until at least the end of the fifteenth century, when they began to explore more commercial uses of their collective properties and tried to enforce stricter observance of ancestral rites (107, 179, 245, 306).

Thus, unlike the village worship association in Zezhou, which began as a ritual organization and acquired corporate property later on, the lineage formed first as a property-holding unit and then took on ritual activities. But these two institutions resembled each other in one way: at a later stage of their development, they both claimed sanctioning powers and evolved into a form of self-government in the countryside. For lineages in Huizhou, the opportunity presented itself in 1567 when the Huizhou government decided to have community pacts “set up on the basis of already existing lineage organizations (zu 族) or subdistrict administrative units (tu 圖),” instead of village worship associations (McDermott, 196). Lineages and their leaders, as McDermott has demonstrated in an earlier article, seized this opportunity and used community pacts (xiangyue 鄉約) to bolster their own power. Community pacts have a long history dating back to the Song dynasty Neo-Confucian scholars Lü Dajun 呂大鈞 (1031–82) and Zhu Xi 朱熹 (1130–1200). “Whereas Lü sought to create a mutual aid organization for village commoners” with his community pact, McDermott argues, “Zhu was more anxious to stress the central role of scholars in rural life and emphasized the moral, at least as much as the material, well-being of Chinese villagers” by inserting much more etiquette and ceremony into his redaction of Lü’s pact. The community pact received little attention from the Ming court and its officials, at least after the death of its founding emperor in 1398. But starting from the late fifteenth century, there was a revived interest among local officials in the community pact as a way of imposing control and restoring order in the countryside. These officially drafted and promoted community pacts, often drawing on the ideas of Wang Yangming 王陽明 (1472–1529) and Wang Tingxiang 王廷相 (1474–1544), differed from the Song texts by “giving a larger role to the state as initiator and implementer of their rules” but had limited success in their implementation. But local officials’ backing of community pacts opened the door for their adoption in the lineages.Footnote ⁷¹

From the mid-sixteenth century on, as McDermott and a modern Chinese scholar have observed, community pacts were “initiated, enforced, and led by gentry and elders” in many lineages across South China. The rules of these pacts, in contrast to the officially drafted versions, were in fact, no different from family rules—referred to variously as cigui 祠規 or jiafa 家法 among other terms—that also began to proliferate. As the title of one such text from Huizhou—“The Community Pact and Family Rules of Wentang” (Wentang Chenshi xiangyue jiafa 文堂陳氏鄉約家法)—suggests, they were considered one and the same. These pacts and rules “stressed filial piety to the lineage and largely ignored the demands of the state.” To enforce them, lineages set up their own structure of authority, including, for example, positions of pact heads (yuezheng 約正), pact assistants (yuefu 約副), prosecutors (sisong 司訟), and pickets (sijiu 司糾).Footnote ⁷²

With the implementation of these pacts and rules, lineage leaders sought to restrict the government's role in lineage affairs. Taxes were discussed in the lineage's monthly meetings to ensure that they would be paid on time and not become an excuse for government interference. They required that all conflicts within the lineage be handled by household heads, pact heads, and deputies, and discouraged their members from filing lawsuits in government courts. They strictly barred the state from interfering in the lineage's management of its own human and material resources, such as late payments of rent, robberies, and general unrest.Footnote ⁷³ The developments in the mid-sixteenth century, therefore, transformed the lineage from a trust-based group of shared economic interests into a tightly organized self-governing body that was capable of exercising a wide range of sanctioning powers.

One may ask why the Huizhou government turned away from the village worship association to the lineage as the organizational interface between the state and local residents. The short answer is that by the mid-sixteenth century, the lineage—not the village worship association—had become the dominant village institution in Huizhou. This dominance was hard won. And McDermott has painstakingly reconstructed the lengthy and contentious process through which the lineage rose to predominance at the expense of its institutional rivals: Buddhist establishments and village worship associations. As has been discussed, Buddhist establishments had, for many centuries, owned farmland and forests, engaged in moneylending, and looked after ancestral graves for local families. Village worship associations that had developed in Song times, on the other hand, provided an inclusive form of village organization based on territorial affiliation and the building blocks for an extensive network between villages. Situating his study in this wide institutional context, McDermott shows that over the fifteenth and early sixteenth centuries, individual village worship associations “had commonly, though not universally, fallen under the domination of a single lineage,” which restricted association membership to its own kin and forced out non-kin members. In the meantime, lineages also aggressively replaced village worship associations as the building blocks for multi-village alliances. To do so, the lineage claimed the association's tutelary deity—for example, Cheng Lingxi and Wang Hua—as its Prime Ancestor and then seized the community shrine as a lineage property, despite the Chengs’ and Wangs’ clearly insignificant role in the earlier history of the cults. In this regard, the widespread anthropomorphization of earth spirits since the late Tang actually paved way for this mid-Ming development (200–210). Moreover, from the latter half of the fifteenth century and usually after they had successfully seized power from the village worship associations, lineages waged protracted battles against local Buddhist establishments. By the late sixteenth century, many larger Huizhou lineages had successfully forced the monks to return land grants, ancestral graves, and their surrounding structures to the descendants of their benefactors, thus transforming chapel lands into lineage trusts. Thus, over the first two centuries of Ming rule, there was “a sea change in the ranking of village institutions” (179): lineages replaced village worship associations and Buddhist temples as the dominant institution in rural Huizhou, descent ties replaced territorial affiliation as the dominant principle of rural organization, and many of Huizhou's villages became single-lineage villages where a single kinship group had achieved organizational dominance. The 1597 decision of the Huizhou government, therefore, was a recognition of the lineage's victory after two centuries of battle.

This trajectory of the lineage's rise to organizational dominance was not unique to Huizhou. Local residents in Leizhou 雷州, Guangdong, had worshipped a “thunder ancestor” (Leizu 雷祖) as the rain god in the Tang. Usually portrayed as a grotesque-looking, beast-like creature, with horns and wings, the thunder ancestor began to be identified by the early eleventh century as a certain Chen Wenyu 陳文玉, a putative native of Leizhou who had passed the civil service examinations and carved out a successful bureaucratic career. No later than the mid-seventeenth century and, quite possibly, as early as the Wanli reign (1573–1620), the Chens in the villages near the shrine started claiming themselves descendants of the thunder ancestor, turned the shrine into their ancestral hall, and managed its attached fields as a lineage trust. Likewise, in Gaozhou 高州, Guangdong, the Fengs 馮 in Old City village started organizing their kinsmen into a tighter kinship group in the early seventeenth century and by mid-Qing they had constructed a story tracing their descent to Lady Xian (Xian furen 冼夫人), focus of a local cult and wife of a certain Feng Bao 馮寶, and made Lady Xian's shrine in the Old City village, as opposed to the Fengs’ ancestral hall, the focus of their seasonal offerings.Footnote ⁷⁴

The victory of the lineage over its rivals was never absolute. In Leizhou, another Chen group in Bangshan 傍山 village lost a lawsuit as late as 1784 when they tried to claim a nearby shrine dedicated to the thunder ancestor as their ancestral hall and the shrine's attached fields as their lineage property. In Gaozhou, over the course of the Qing, Lady Xian's shrine in the prefectural seat continued to be maintained and its attached fields owned by a multi-surname worship organization.Footnote ⁷⁵ In some places like Zezhou, corporate lineages never fully developed. There, as Du's account makes clear, efforts to bind kin into a more tightly organized group centered on a free-standing ancestral hall and a landed trust surfaced after the mid-sixteenth century, but they reflected no more than the ideals of individual gentry figures and received little support. The lineages those individuals established usually had a weak organizational structure, possessed minimal collective property, and rarely survived the death of their founder. Throughout the Ming-Qing period, ancestral worship in Zezhou was typically performed at the graveside. Genealogy writing and ritual preparations were placed under the charge of a “gravesite sacrifice association” (musihui 墓祀會) or an “ancestral worship association” (zuzonghui 祖宗會). By contrast, it was the village worship association dating from the Song times that continued to provide an effective means of local organization all the way into the Qing. Even single-surname villages participated in public affairs in the form of village worship associations rather than lineages (Du, 192–96).

It is also important to note that even in places like Huizhou and Fujian where lineage organizations were more developed, shared religious symbols and common worship activities remained a powerful mechanism of local organization. In Huizhou, even after individual village worship associations each fell under the domination of a single lineage, the lineage and the village worship association remained conceptually different and each lineage typically maintained both an ancestral hall and an association hall (shewu 社屋). Whereas the ancestral hall was for the “master of the clan” (jiazhu 家主), the association hall was dedicated to the “master of the village” (cunzhu 村主). Supra-village religious symbols (e.g., those in the Shuanglie Shrine 雙烈廟) and the joint performance of worship rites continued to create a ritual space jointly participated in by different surname groups living in different villages, each representing themselves with the tutelary gods it had sheltered in its association hall. Likewise, in a multi-surname village, common worship of the same tutelary gods (e.g., those in the Cundi Shrine 準提廟) continued to provide an organizing node for different descent groups in the village, each of which may have had its own lineage hall and lineage-sponsored association hall.Footnote ⁷⁶ Similarly, on the basis of their extensive fieldwork on the Putian plains in Fujian, Kenneth Dean and Zheng Zhenman have argued that a “nested hierarchy of temples” continued to demarcate both village-level and multi-village communities, even though the organization of religious activities in some single-surname villages was based on branches of the lineage.Footnote ⁷⁷

Thus, from the Song to mid-Qing, local society became increasingly organized around resource-rich and tightly organized social institutions, first based on territorial affiliation and shared religious symbols and then also on claims of common ancestry. Over the course of the Ming and Qing, these institutions evolved into property-holding corporate entities with elaborate organizational structures and possessing sanctioning powers over their members. These institutions provided organizational frameworks in local society that competed, intersected, as well as collaborated with each other. It is evident—in particular, from our discussion of the lineage's phenomenal success in Huizhou and the gentry's failure to replicate it in Zezhou—that the rise of these social institutions was driven mainly by local forces, rather than resulting from government mandate or successful promotion by the educated elite. In contrast to the gentry, whose privileged local status depended heavily on degrees and official titles awarded by the state, leaders of these institutions derived their power from the economic and organizational resources they controlled. These two sources of power and, correspondingly, the two types of local elites are useful analytical constructs for articulating patterns of change in local society after the Song dynasty. In reality, however, and especially from the late Ming on, the boundaries between these two realms were blurred.

THE GENTRY, LOCAL INSTITUTIONS, AND LOCAL ELITE ACTIVISM IN THE MING AND QING

As lineages and village worship associations evolved into powerful social institutions with substantial material and organizational resources in the Ming and Qing, they provided the institutional underpinnings for their leaders to take on a more active role in local society. The leadership in neither type of these institutions was constituted exclusively by men of gentry status, but the role of gentry figures in these institutions was always prominent from the Ming on.

In the weakest form, the gentry participated actively in the affairs of these institutions and cooperated with them. In the absence of powerful lineage organizations, the village worship association was an indispensable means of marshalling local resources, organizing public projects, and promoting community solidarity in late imperial Zezhou. Even gentry leaders who were troubled by religious festivals dedicated to the local tutelary deity recognized the strength of the village worship association in these areas. Some gentry families (e.g., the Chens of Guoyu) did distance themselves from the local worship association, but many others (e.g., the Zhangs from the same town) participated actively in its affairs and joined the community's religious assemblies (hui 會) and festivals (saishe 賽社). Evidence from the mid-seventeenth century shows that gentry leaders who proposed a local project consulted the leaders of village worship associations and relied on their financial and organizational resources (Du, 265). The relationship between the gentry and corporate lineages could be even closer. Gentry families had long recognized that perpetuating their political status and solidifying their local power base were mutually reinforcing. Kinsmen with gentry status provided political protection for lineage interests, whereas a lineage with abundant material resources, in turn, could afford superior educational facilities for their members, increase their odds of success in the examinations, and thereby improve the chance of perpetuating their gentry status.Footnote ⁷⁸ As McDermott has pointed out, from very early on, aspirant members of the Chengs—those in the third branch—pursued a two-pronged strategy of “maintaining official representation in government and a consistently strong leadership at home” (291). They pushed their descendants towards official careers, but chose one from their ranks, who usually had no degree whatsoever, as the resident manager of the lineage's trust (356). Thus, although men holding advanced degrees and offices were usually posted away from home and rarely involved directly in the management of lineage trusts, their connections with these trusts were far from inconsequential.

Indeed, the relationship between the gentry and local institutions often went beyond strategic accommodation and cooperation for mutual benefits. Descent groups that had great success in the examinations and government enjoyed a competitive edge that made them more likely than others to dominate a lineage or village worship association. Their competitive edge was both demographic and sociopolitical. Du shows, for example, that the Yuans 原, which claimed the majority of positions in the managerial corps of Xiajiao 下交 village's worship association, not only produced a considerable number of degree holders and high-ranking officials, but also had a much larger population than other surname groups in the village (164). McDermott observes a similar phenomenon in his discussion of the Doushan trust of the Shanhe Chengs. After the death of its founder in 1454, the trust was, in theory, placed under shared management of its five constituent branches, but in practice almost always dominated by the lineage's third branch, which enjoyed consistent success in the civil service examinations and, at the same time, expanded demographically at a much faster pace than other branches in the lineage. Examination success, as McDermott puts it, was “virtually essential for the flourishing of any large lineage and the dominance of any of its branches from the middle of the fifteenth century onward” (260–61).

From the sixteenth century on, the distinction between local power derived from state-certified status and that from institutional leadership became even more blurry, as men of gentry status and the leaders of local institutions began to overlap significantly. This development resulted, in particular, from the growing ranks of men holding lower-ranked degrees, such as district school and Imperial University studentships. These degrees became easier to obtain after 1537, when the Ming court made some of them available for purchase to all who could afford them, further deepening the mutual infiltration between the gentry and the local commercial elite.Footnote ⁷⁹ These degrees, which rarely led to examination success at higher levels or official assignments, were essentially markers of status which not only exempted their holders from onerous labor service duties but also certified their membership in the local gentry. In the Doushan trust, for example, the relationship between gentry status and lineage leadership became all the more intimate—and the third branch's dominance all the more absolute—after the late sixteenth century when the Doushan trust fell under the control of the “Culture Group” (siwen 斯文), which was constituted exclusively by men holding at least—and usually only—lower-ranked degrees (McDermott, 356). Although the adoption of an examination degree as a requisite condition for lineage leadership served the third branch well in its competition with other branches, arguably it also reflected a general recognition that having degree-holding leaders represent the lineage would allow for more effective negotiations with the local government. Likewise, in Zezhou, men with low-ranked degrees, like Licentiate Zhang Duoxue 張多學 in Guoyu, evinced no interest in a bureaucratic career but instead served as a manager of an association (sheshou 社首) for many years (Du, 130, 139). Wealthy merchants such as Wang Chongxin 王重新, who made generous donations to local projects and became an association manager, purchased low-ranked degrees and urged his sons to study for the examinations (Du, 141–42). Unlike the Doushan trust, however, examination degrees seem never to have been a requirement for leadership in a village worship association.Footnote ⁸⁰

In some places, the gentry could be far more aggressive. In the mid-Qing Foshan, for example, gentry leaders established their own corporate institution and expanded its power at the expense of existing local institutions. In the mid-sixteenth century, the only recognizable collective body in Foshan was the Eight Tu (batu 八圖), a management structure of town affairs derived from the old lijia institutions and centered on collective worship of the Northern Emperor in the Beidi Temple. Not surprisingly, the property of the temple constituted the largest existing source of wealth for public purposes in the town and was under the control of the Eight Tu. In the closing years of the Ming, however, the gentry of Foshan began to have a stronger presence in the town. With donations, they revived a defunct local school in 1624 and erected a meeting hall (huiguan 會館) for themselves three years later. The establishment of the meeting hall, later known as Dakui Hall (Dakui tang 大魁堂), signaled the gentry's independence from the Beidi Temple, where they had previously held meetings, and began to give the gentry network the look of a formal organization. The gentry's relationship to the Beidi Temple came full circle a century later, when they successfully forced the Eight Tu to relinquish its special rights to the temple and its attached property and thus relegated it to a purely ritual organization. A “watershed in town affairs,” as Mary Buckus Rankin calls it, this victory gave the gentry access to the considerable income of the temple as well as to its symbolic capital. Almost at the same time, a petition was sent to the magistrate in the name of “all gentry leaders of the town,” contesting the Huo lineage's claims on the property at Foshan's main wharf. The petition was resolved, again, in the gentry's favor, confirming—if not establishing—that shops and other structures at the wharf were the town's collective property. These developments in the mid-eighteenth century gave the increasingly powerful gentry establishment access to lucrative sources of revenue for financing community projects they favored, including a charitable school in 1744 and a community granary in 1795, among others. The Dakui Hall became a core institution in Foshan, providing an organizational structure for gentry leadership, with its own managers, and a forum for deliberation of town affairs and negotiation between the gentry and other local interests. It also managed public funds, organized public projects, and supervised the operation of the charitable granary and other local institutions.Footnote ⁸¹

Thus, several centuries of institutional evolution provided the gentry in Ming and Qing times with enormous tangible and intangible resources that allowed them to take on a more active role in local society than their Song and Yuan predecessors. These changes stood out most clearly when one compares privately organized local defense in the Southern Song, the late Ming, and the mid-nineteenth century. In the Southern Song, private militias were few and relied heavily on the leadership and resources of individual families. In much of Fuzhou, as Robert Hymes has observed, “no indigenous, effective, permanent, private force emerged” throughout the Southern Song. The elite's involvement in local defense was usually sporadic and provisional, mainly in response to immediate dangers of local disorder, and each militia's activity was “strongly localized within [its leader's] own county or the area around his residence.” The only enduring militia organization was the “Two Societies of the Dengs and Fus” (Deng Fu ershe 鄧傅二社) of Jinxi county, which seems to have survived from 1127 at least to 1261 and, at times, operated in several counties of the prefecture. But, as its name suggests, the Two Societies was largely a “cooperative enterprise” of two prominent local families. It was an association of two militia units organized by the Dengs and Fus respectively, each of whom built and manned a private fort close to home and disbursed their family's wealth to hire militiamen from local families living nearby but usually having different surnames. Both the overall leadership of the Two Societies and the commandership of each constituent unit were expected to pass on from father to son within the same family.Footnote ⁸² In Mingzhou, too, there was only one sure case of privately organized local militia in the Southern Song, which was led and funded by the Li family and seems to have been no more than a provisional response to the Jurchen invasions (Lee, 88).Footnote ⁸³

In late Ming Shanxi, worship associations in some market towns began to take on a prominent role in organizing local defense. The construction of the Guoyu and Diji 砥洎 forts was a case in point. Although gentry figures reputedly advocated—and even boldly claimed full credit for—the construction of these forts, the projects were obviously funded and organized by worship associations in these towns and the forts considered a collective property of the townspeople (Du, 143–46, 150–52). Admittedly, village or town worship associations did not always play an active role in late Ming local defense. In much of late Ming Shanxi, local defense continued to rely on whatever resources individual families could provide, just as in the Southern Song, and typically protected no more than an individual village or market town with little sign of higher-level coordination. Even many expensive projects, including the construction of fortified structures, were funded by the private wealth of individual families and manned by their retainers (jiading 家丁). Since lineage organizations never developed fully in this area, sometimes even pooling resources from a larger kinship group beyond the household level was not feasible.

This contrasted markedly with the mid-nineteenth century when non-governmental local institutions had evolved into property-holding corporate entities with coercive power and provided enormous material and organizational support that shored up gentry activism. First, the proliferation of lineages and village worship associations provided gentry leaders with a ready means of social organization and mobilization. In the mid-nineteenth century, local defense systems often grew out of existing lineage and village organizations, often by superposing an added layer atop these organizations to coordinate activities and pool resources from a vast area. Near Guangzhou, during the summer of 1841, just after the incident at Sanyuanli 三元里, the gentry quickly organized a well-ordered militia by constructing a new “community school” (shexue 社學) which served as a headquarters for twelve or thirteen multi-village alliances in surrounding market towns, which in turn coordinated a total of more than eighty villages.Footnote ⁸⁴ These villages and multi-village alliances, however, were more than loosely articulated, informal frameworks of gentry cooperation, as was once surmised. Instead, a recent study of the nearby East Mountain community suggests that they were more inclusive village or supra-village associations based on common residence and collective worship, though they may have come under the domination of powerful gentry lineages by the mid-nineteenth century.Footnote ⁸⁵ In Hunan, the lineage was the basic unit from which militia were raised and organized, and these autonomous lineage militia forces were then aggregated and placed under the leadership of gentry figures who were bound by teacher-disciple ties and classmate bonds.Footnote ⁸⁶

Second, the rise of various types of corporate institutions also provided gentry figures with convenient access to a vast body of economic resources, including not only profits from these institutions’ collective property but also revenues they could requisition with their private taxing powers. In a more abstract sense, several centuries of institutional evolution and economic progress provided the gentry in the mid-nineteenth century with a complicated toolkit of financial management: the very idea of setting up a corporate organization to manage collective property and finance local initiatives, the gentry's experience in managing these organizations and negotiating with the government on their behalf, and the widely accepted practice for these organizations to exercise private taxing (tax-farming) powers. It then causes no surprise that gentry figures moved swiftly in the mid-nineteenth century along these lines. Their typical course of action in funding local defense was to establish a corporate institution (often in the name of a literary society [wenshe 文社], public office [gongju 公局], militia bureau [tuanlian ju 團練局], or community school [shexue 社學]) and fund it by setting up trusts and giving it the power to impose levies. To support the Shunde 順德 militia near Guangzhou, for example, the gentry immediately set up an Azure Cloud Literary Society (Qingyun wenshe 青雲文社), to which at least several thousand mu of land was attached, and a militia bureau which not only recruited militiamen but also collected a tax from the silk spinners.Footnote ⁸⁷ As Philip Kuhn has observed, these special taxes provided a more important source of funding for local militias in the mid-nineteenth century than personal wealth and lineage-owned resources. The taxing authority of these gentry-led institutions became “a buttress of localism” and further augmented the power of the gentry by deepening their infiltration of the taxing process as a whole.Footnote ⁸⁸