High Ancestors among the Hakka Chinese

combined with knowledge of place as fed by traditional understanding of local terrain, especially below the county level, but feeding into the administrative system hierarchy at the county level or beyond.

In Taiwan, genealogical knowledge derived from and shared with the mainland was applied in a very different context, with immigrants from numerous mainland villages and regions forming an agnatically heterogeneous population. This heterogeneity was not simply characteristic of the Minong region as a whole, but rather extended into its major component villages.

Two factors can be singled out as especially important in encouraging this development. The first is precisely recency of settlement: the first Hakka Han Chinese arrived there only 159 years before the Japanese assumed control of the area. Since Hakka migration to Minong continued throughout Qing, by the end of the period a good many families had histories of settlement considerably less than the 159-year maximum. The second factor is that surname heterogeneity was characteristic of Hakka settlements from the outset. It is said that Minong Village was first settled in 1736 by over 40 people comprising 16 different surnames. Six surnames were represented among the immigrants first coming to Longdu in 1737. In Longdu, a local document relates how, after 1738, a guanshi ( 管事 , village manager or village head) was appointed. In the document it is noted that “since persons of mixed surnames lived together and dissension could not be avoided,” one of the

guanshi’s duties would be to “attend to minor and serious incidents occurring

in the village” (MTGCC 1977b: 1215). Later, in 1748, Zhongyun was first settled primarily by persons with the Li and Liu surnames, entering via

Minong and later joined by others such that the pattern of mixed surname settlement was well established and maintained throughout the Minong region.⁸

That this pattern was still characteristic of the Minong area by the end of Qing is shown in Table 1. For this table, which conveys the dimensions of surname heterogeneity and surname representation in Minong, I use data from the 1902 Japanese cadastral survey. In that survey there is identified for every plot of land a proprietor or owner (yezhu 業主 ), which may be an individual or a corporation. In the latter case, a manager (guanliren 管理人 ) is also identified, and a few plots have managers even though they are listed as owned by individuals. Additional individuals are in some cases listed as having use rights due to conditional purchase (dianzhu 典主 ), as owners of

“large rent” rights (dazu 大租 ), or as managers of associations having such large rent rights or conditional purchase use rights.

8　On initial settlement of the different Minong area villages, see MTGCC (1997a: 35-58).

Table 1: Minong Family Heads by Surname, 1902

Family Heads Surname Number Family Heads Surname Number

劉 305 羅 26

Note: This table includes all owners or right holders after eliminating those where residence is indicated as being outside Minong. The total is 2,229 people, down from the gross total of 2,333.

All individuals recorded in the Japanese 1902 cadastral survey of Minong are taken to be family heads, in conformity with the Japanese practice of identifying owners, co-owners, pledge custodians, and managers of land, housing, or other forms of property covered by the survey, large rent rights or conditional purchase use rights.

In generating a master list of names, I combined into one database the names of all these persons and after eliminating non-residents of Minong, I merged the data into a list of unique individuals, which I then subdivided

simply based on surname. Because I used the cadastral survey, and not household data, the numerical breakdown of surnames has as its basis family heads only; as we know, families as corporate entities held rights to property, but the cadastral survey simply identifies each family by the name of its head.

The survey will have missed those in Minong having no rights of any kind to any land, not even to the sites of residential compounds. Generally, the head of each component family in such a compound is listed as a joint owner, such that the landless poor in the vast majority of cases were included in the survey at least as co-owners of compounds, so that very few family heads were missed. In any event, given Minong’s population of about 10, 000, the 2,229 persons included in the cadastral survey without any doubt can be held to be accurately representative of the dimensions of surname distribution. They amount to about one-quarter of the estimated population at that time.

That only 14 out of these 2,229 were women confirms the formidable focus on male authority within late imperial society, but also tells us, obviously, that these cadastral data include about half of the male population such that we can confidently use them for considering surname distribution in a population where patrilineality and virilocal residence were dominant arrangements. The 14 women, by the way, all have among the most commonly encountered surnames – be they natal surnames or those of marriage -- and thus do not influence data concerning different surnames or their distribution.

In cross-checking surname data drawn from the cadastral survey with the surnames of individuals in the Japanese household registers, only three

High Ancestors among the Hakka Chinese

additional surnames were discovered dating from pre-Japanese late Qing, each representing one instance of a women’s marriage into the Minong administrative region from the outside, this being precisely a circumstance under which surnames would missed by the survey. It can easily be seen from Table 1 that surname heterogeneity is the condition of every administrative village, with representatives of eleven surnames found in all six. Of the 56 different surnames found in the Minong area, evidence from household registration data as well as from the building sites included in the cadastral survey show all to be fully established in Minong by the end of Qing in that the presence of each surname is in the form of families, not isolated individuals, with at least some men from each of these surnames Minong-born and thus representing the continuity of settled agnatic descent lines.

Thus throughout the Qing period Minong was characterized by mixed surname settlement even at level of its constituent villages, not to speak of the region as a whole. Agnatic heterogeneity was characteristic even as far as each of the different surnames was concerned, or at least those where there were enough men to allow for it. Making up a population of a particular surname would be people from different mainland villages, lineages, and regional descent lines. Such heterogeneity hardly hindered powerful religious or organizational expressions of agnatic concerns. Indeed, it was mainland agnatic culture that in the first instance provided ideals of agnatic solidarity and genealogical knowledge in the form of the widely known and widely shared higher-level decent lines, such that in Taiwan immigrants from different

mainland villages and districts knew the common ancestors around whom they could organize. In Taiwan this genealogical knowledge combined with knowledge of place, that is, place of origin on the mainland as fed by understandings of native place local subcounty spatial hierarchies and also, at the county level and above, by the administrative system hierarchy. Location and genealogy were combined in the organization of ancestral associations. As has been noted frequently in the literature, the importance of Taiwan-born ancestors, or of the ancestors who made the move from the mainland to Taiwan, increased in tandem with the length of Han Chinese settlement on the island, the obvious point being that it took time to produce local-born ancestors, and even more for there to be generational depth sufficient to allow for the kind of intimate segmentation and branching characteristic of the larger mainland local lineage communities. As far as Minong is concerned, the Japanese cadastral survey of 1902 recorded ancestral associations as these had developed during the preceding 166 years, since the first arrival of the Hakka-speaking Han Chinese.

Thus the associations recorded by the survey fall into two large categories according to whether they worshipped mainland or Taiwan ancestral figures, it being understood that categorization of associations says nothing about membership per se, since people could and did own shares in several associations. Among associations with mainland ancestral figures, some focused on national surname founders such as Chen Hu 陳胡 of the ancient Zhou period, or the high Han Dynasty official Xiao He 蕭何 ; others on later

High Ancestors among the Hakka Chinese

eminent figures such as Zhu Xi 朱熹 , the venerable Neo-Confucian scholar, or Liao Guangjing 廖光景 , another Song scholar-official. Most of the mainland figures selected as objects of worship by the Minong ancestral associations are regional founding ancestors (kaijizu 開基祖 ), with some associations, including the largest, focusing on ancestors held to be founders with respect to the entire Fujian/Guangdong Hakka heartland.⁹ Such major regional founding ancestors were (and still are) quite well known in both the Taiwan and mainland Hakka regions, and in Minong, as elsewhere, the names of these founding ancestors are commonly inscribed on tablets centrally placed in domestic ancestral halls together with those of closer ancestors.

People see the names of these ancestors every time they enter the hall, for ancestor worship at particular times, but far more frequently simply as a consequence of everyday activities. Those founding ancestors chosen by Minong associations include regionally famous figures such as Gu Zongyue 古宗悅, Wu Jipu 吳吉普 , Li Huode 李火德 , Lin Pingshi 林評事 , Qiu Xiao 邱烋, Zhang Huasun 張花孫 , Huang Rixin 黃日新 , Wen Jiulang 溫九良 , Liu Kaiqi 劉開七, and Zhong Kui 鍾逵. Many lower level regional ancestors are also represented, including founding ancestors for Mei or Zhenping counties as well some founders of subcounty regions (xiang 鄉 or bao 堡 ).

9　The Hakka heartland of course includes adjacent areas of Guangxi. Nevertheless, the local genealogies and the genealogical summaries included in many of the introductions to association account books reflect the fact that most Minong Hakka traced their descent from long-established lineages in Guangdong, especially those in Zhenping or Mei counties. The common understanding in all of these documents is that within the Hakka heartland movement to Guangdong was from earlier settlements in Fujian; in the context of mapping a hierarchy of founding ancestors, Guangxi is not relevant.

But far less common are associations in Minong taking as an ancestral focus a mainland ancestor whose agnatic sphere of influence, so to speak, was confined to that of a village level founding ancestor, that is the founder of a mainland lineage still otherwise restricted to one community at the time of migration to Taiwan. Given what appears to have been the relatively few people from any particular mainland village who might find themselves once again living close by on Taiwan, the mainland ancestral focus usually represented an appeal to an agnatic ideology of social intimacy precisely so as to establish such intimacy among people with the same surname but from different mainland communities and regions.

Yet such lower levels of agnatic identification could be accommodated, especially by the larger associations, those appealing to major regional founding ancestors. Almost all association account books include a list of the original “share names,” by which a share is identified either by the name of the first shareholder or by an ancestor he chose to honor. In some associations, especially those drawing members from relatively many villages from within the south Taiwan Hakka region, there may be indicated on top of a share-name the name of the shareholder’s village or town in Taiwan. In others, however, the similarly placed identification is what is referred to as his “hu, 戶 ” or mainland place of origin. This term, which literally means “household” has, in the mainland and Taiwan Hakka regions the extended meaning of “a number of households with the same founding ancestor” (Zhong and Feng 1971:

High Ancestors among the Hakka Chinese

58-9).¹⁰ In other words, this term conveys a specified geographic identifica-tion that asserts genealogical linkage.

The ancestral corporations in the 1902 records taking as their focus Taiwan-related ancestors can also be placed into two broad categories. First are those dedicated either to laitaizu 來臺祖 (Ancestor who came to Taiwan) or to a post-arrival ancestor who founded the agnatic line in a particular community, that is to a local kaijizu; it is obvious enough that under most circumstances a laitaizu will also be a kaijizu, but a kaijizu can also be several generations removed from his laitaizu, in that some of the founders of new local agnatic lines had moved from elsewhere on Taiwan. In the second category are generationally closer Taiwan-born ancestors for whose worship land was usually set aside during the process of family division, such that the corporation shares owned in the first instance by the brothers or (in a few cases) by the paternal nephews were acquired not by purchase, but according to the terms of the family division contract.

Ancestral corporations formed through the purchase of shares by men belonging to different families are known to the Hakka as “public ancestral

10 This term derives from the old Ming dynasty taxation system, whereby “official registry as a fiscal household [hu 戶 ] was a social honor for families to cling to as long as they could”

(Dardess 1997: 75). According to Szonyi’s analysis, this Ming era state-imposed fiscal household was in its own right a major factor behind lineage formation and development in the Fuzhou region (Szonyi 2002: 56-89). Through time and family division, the number of families within the hu would increase thus turning it, de facto, into a unit of agnatic organization which outlived the Ming system that gave it birth. The Minong material shows that the transformation of what originally was the Ming fiscal hu into an agnatic unit went even beyond the local lineage, however, for in the Minong ancestral associations the hu is a unit of agnatic reckoning in most cases figured based on territorial rather than genealogical identification. For the purposes of a high-level association it is assumed that a hu identifies a genealogically coherent population of agnates, although, genealogies or sections of genealogies will sometimes identify a descent line as a hu.

estates” (gong chang 公嘗 ) or as “association share ancestral estates” (huifen chang 會份嘗 ); the former term refers to the quality of a new corporation as being open through share purchase to any man qualified on the basis of sharing descent from that corporation’s chosen ancestral focus; all with such a qualification are eligible to purchase shares, and those who do not are excluded from corporation membership. Corporations of this kind include all focusing on mainland ancestors, and some of those focusing on laitaizu or kaijizu.

Corporations created as an outcome of family division are known as “private ancestral estates” (si chang 私嘗 ) or as “sacrificial ancestral estates” (xie shi chang 血食嘗 ), both terms reflecting the fact that shares in such corporations are obtained as one procedure of succession to the family estate as based upon patrilineal decent line: the estate is private because succession rights to a share are the only basis for receiving one; it is “sacrificial” because it is created so as to protect and subsidize what is already the ongoing worship of a close ancestor.

The high ancestors, be they national or regional, who framed the rituals of small ancestral corporations in Taiwan were precisely those to whom the large urban-based ancestral halls in China were dedicated. If the urban ancestral halls were an important concern of the imperial degree-holding and merchant elites during the Qing era, the fact that ordinary farmers could bring to Taiwan knowledge of these high ancestors identifies kinship as a major factor with respect to rural-urban and elite-commoner social and cultural integration in late imperial times. My brief visits to Meixian City in Guangdong Province during

High Ancestors among the Hakka Chinese

the past two years came at a time when rural lineage ancestral halls were being refurbished, as were several urban halls dedicated to high ancestors.

All had been destroyed or put to other uses during the Cultural Revolution or at an earlier phase of the Maoist era. Among the urban halls newly revived were those of the Huang surname, dedicated to Huang Rixin, the Liao surname, with Liao Guangjing as the ancestral focus, one with a focus on Li Huode, worshipped by the Li surname, and yet another dedicated to Yang Yunxiu, taken as a focal ancestor by the Yang. As noted above, all of these high ancestors also served as foci of worship for Taiwan ancestral associations.

The vast majority of what had been Meixian City’s many ancestral halls remain unrecovered, but during brief interviews, it emerged that every high ancestor I could think of, as based upon my Taiwan research had had his ancestral hall in Meixian City.

During Qing what is now Meixian was the seat of Jiaying Prefecture, which was composed of four counties. In China’s traditional examination system, tests given at the county level merely qualified those passing to take the prefectural exams, which awarded the lowest level and most common degree, shengyuan or xiucai. So it is not surprising that the various ancestral halls in Meixian tended to cluster in the vicinity of the examination compound, for one of the services provided by these halls was to make space available where examination candidates with the appropriate surname could spend a few days and nights prior to taking the tests. These candidates generally came from rural lineages located in the prefecture’s different

counties. Also from different parts of the prefecture were the merchants and degree-holders who would present themselves at the prefectural seat ancestral hall bearing their surname and the tablets of their high ancestors. These urban ancestral halls were supported by contributions from the lineages bearing the surname, from wealthier private individuals, and through the selling of space where a family could place the ancestral tablets of close kin. The tablets in an urban hall did not comprise a complete statement of patrilineal genealogy, very much unlike the arrangement of tablets in many rural lineage ancestral halls, where lineage membership through birth guaranteed the men tablet space after death. In the urban halls, the high ancestor maintained his status precisely because the position of his tablet was not placed based on close genealogical connections with the mass of tablets arranged around his. Thus, these ancestral halls made common surname an urban asset for they in fact served to “departicularize” agnatic ties by framing them within very large-scale regional and administrative contexts.

It is not surprising that the tight-knit rural lineage in China has been an object of anthropological fascination since the mid-20^th century, for this was a period when the lineage as a cultural and social form was heavily represented in the ethnographic literature, especially in British social anthropology and

在文檔中客家人的祖先：今昔、城鄉與全球 (頁 22-37)