11. Uxorilocal Marriage among the Bai of the Dali Basin, Yunnan

Yokoyama Hiroko


In the sphere of kinship and marriage among the Han Chinese, at least before Liberation in 1949, some uniformities and shared cultural ideals can be found, despite the markedly broad variations in actual local situations. As for marriage and adoption, virilocal marriage and, in the case of the lack of a son, adoption of a boy from within the same lineage was the ideal for most of the educated Han Chinese. On the other hand, as Arthur Wolf demonstrated, "minor" practices, such as rearing a son's wife, or substituting a son-in-law for a son, were fairly common in some regions, because "Chinese marriage and adoption practices were not the simple reflections of uniform ideals. Rather they were the complex reflections of a variety of forces - demographic, economic, and psychological - that interacted to shape family organization" (Wolf and Huang, 1980: 1).
In this paper I would like to focus upon uxorilocal marriage among an ethnic minority of southwestern China, the Bai, and show how it fits into their sinicized patrilineal kinship system. Through discussing the causes of their uxorilocal marriage, I hope to add to our understanding of this type of marriage within those societies which share the Han Chinese descent system as a whole.


According to the 1990 Chinese National Census, the population of the Bai within China was about 1.6 million. Of these, close to one million live in the Dali Baizu Zizhizhou (Dali Bai Autonomous Prefecture) of Yunnan. Except for a small proportion of the Bai who live in the hill areas, most of the Bai live in the river basins or by the lake on the Yungui Plateau. The Bai language is usually classified as forming its own separate subgroup, or as part of the Yi (Lolo) subgroup of the Tibeto-Burman languages. The Bai have used Han Chinese characters to write their own language in the past. This system of writing, baiwen, however, seems not to have been standardised, and was used less after the middle of the Ming period, except in cases such as recording the words of dabenqu recitations.[1] Instead, the Bai increasingly wrote in Han Chinese to keep important records and historical documents.
From 1984 to 1986 I conducted altogether thirteen months of field research in a Bai village located in the Dali Basin. The research was carried out at three different times under somewhat different conditions. However, except for the first three-month period, I carried out the research while living in the village. The data I use in this paper was mostly collected in the village during that research period and in other follow-up visits in the summers of 1986-88 and 1990. In this paper, I will call the village "Green Village" (Cang-cun).
Although the latitude of the Dali Basin is similar to that of Taipei, its high altitude of about 2000 meters gives the area a very pleasant climate. As the site of the capital of the former Nanzhao and Dali kingdoms, the basin was the political and cultural centre of southwestern China from the eighth to the thirteenth centuries, and the present capital of the prefecture is also located there. The Bai people have been the main inhabitants of the basin at least from around the Nanzhao period. Here they have practiced two-crop agriculture based on wet-rice cultivation. Sideline businesses such as carpentry, stonework, straw-weaving, and various kinds of food-processing have been common. All of these have been instrumental in providing a comparatively high standard of living in the Dali area within Yunnan. The Bai have actively absorbed the advancing Han culture, which enabled them to keep their relatively dominant political, as well as economic, position among the ethnic minorities in the area. This kind of description more or less fits the other Bai in the river basin areas, though the political and historical background of the Dali basin seems to have accelerated the process of sinicization more than elsewhere.

Hsu, Fitzgerald and the Bai

The sinicized features of the Bai are particularly clear in their kinship system. According to Edmund Leach, in Francis L. K. Hsu's famous book, Under the Ancestors' Shadow, ancestor worship in "West Town" was actually "presented as prototypical of that which prevails in Chinese culture as a whole" (Leach, 1982: 125). West Towners, who were called Minchia at that time, are definitely regarded as Bai at present. Leach assumes that Hsu's account of the West Towners was a "syncretic blend" of what Hsu himself had learned as a Han Chinese through his personal experiences since childhood and what he saw in West Town. Leach criticizes Hsu because the Minchia peculiarities of the local culture are hardly ever mentioned in his book, and in this respect Leach judges C.P. Fitzgerald's work, carried out among the Minchia at a similar time and place, as better.
Evaluating their work after my intensive fieldwork in the same basin, I would regard my own position as equidistant from that of both Hsu and Fitzgerald. The emphasis of their studies and their main areas of research are different, which makes the comparison difficult. What is more, the work of both men has both strong and weak points. However, as far as their descriptions of kinship are concerned, I think Hsu's account is more accurate and more detailed than Fitzgerald's.
The Bai kinship system is very much influenced by Han Chinese culture, and it was especially so in West Town at that time. This was the home town of famous Bai businessmen who opened shops in important cities of southern China and abroad, such as Kunming, Wuhan, Shanghai, Hong Kong, Yangon (Rangoon), and Colombo. Today, compared with other ethnic minorities, the Bai are one of the most sinicized groups in Yunnan.

Bai kinship

The Bai in the Dali basin have a patrilineal kinship system like that of the Han Chinese. Their smallest social unit is the household, which is called haotv in Bai.[2] The Bai word for family, hao is very similar to jia in Han Chinese in meaning. The hao can refer to a social group as small as a household, or as large as a group of several households connected by patrilineal kinship. Before 1949, although the ideal household was that of the large family, it was seldom realised except among a few rich families. The most popular household patterns were either the stem family or the nuclear family.
The Bai patrilineal lineage, benjia and the lineage branch or segment zi, are ritual units which worship their patrilineal ancestors communally. Obviously benjia and zi are derived from Chinese words. Owing to the small number of sons in each generation, as well as other reasons, some hao never grew large in terms of numbers, and consequently never formed lineages. Before 1949, some Bai benjia had citan (ancestral halls) and owned farm land as corporate property, in addition to their rights over the grave sites, but usually the area of this land was so small that it could only pay for the cost of the annual communal ancestor worship. About 10% of the farm land in the villages in Dali may be communally owned, including the land owned by temples and shrines. There were never any cases here of the situation which is found in some parts of Guangdong where most of the farm land is owned communally in the name of the lineage.
For ancestor worship, either by the hao, the zi or the benjia, there are ancestral tablets called xingepai (zuxianpai).[3] The tablet of the hao is usually a wooden plank or a paper scroll, while that of the zi or benjia may be more varied, ranging from a large stone tablet to a cloth sheet. All of these records include collectively the names of a larger or smaller number of ancestors written in Han Chinese characters. Xingepai are not individual personal tablets. The tablets of poor families with no literate members may be very simple with just the names of ancestors for three generations or so. On the other hand, the tablets of large lineages with literate and, well-educated members in the past may trace the ancestors back over twenty generations, and these lineages usually maintain the lineage genealogy in the form of a book.

Uxorilocal marriage

The normal rule of residence after marriage is virilocal, but uxorilocal marriage is also widely practiced. In Bai it is called zome (shangmen). According to my household survey in two of the sixteen sections of Green Village, the percentage of uxorilocal marriage was high, 20.3% in A section and 13.6% in B section (see Table 1).





Uxorilocal marriages

%Uxorilocal marriages













Table 1. Uxorilocal Marriages in Green Village

High percentages of uxorilocal marriage, over 10 or 20% are also reported for Taiwan (Wolf and Huang, 1980:124-25, 318-21), but not as far as I know in mainland China. In Green Village, if a man has no son but only a daughter, he will definitely let her stay home and marry an in-coming husband. The ideal Han Chinese solution to the problem of the lack of a son, the adoption of an agnate, guoji, is known, but it is less popular than uxorilocal marriage. On the other hand, the in-coming son-in-law is treated as a formal heir, like an adopted son.
Sometimes uxorilocal marriage takes place even in the families which have a son, especially when the son has an elder sister and the disparity of age is large. In this case, usually both of the heirs, the real son and the son-in-law, inherit the property. The son-in-law ensures that the family has a heir and somebody to support the household while the real son is still young and his future unpredictable. Some of these cases are also derived from a custom in which the future son-in-law is adopted and moves into the family either a few or several years before the marriage. A real son might be born to the family after the adoption of a son as a future husband for their daughter. Before Liberation, most sons-in-law in Green Village started living in their fathers-in-law's compound prior to their marriage. The same pattern was also known in central China, but seems to have been of rather minor importance there (Fukutake, 1946: 81, 101n.).
Traditionally, a Bai in-coming husband changed his name to that of his father-in-law, and his first name also had to be changed according to the naming system based on generation order among the father-in-law's agnates. This change took place when the young man moved into the father-in-law's house. After Liberation, the change of name was regarded as "feudalistic", and now a man keeps using his original name in family registration and everyday life. However, he is also given another name in his father-in-law's patrilineal descent group and this name is used in descent group activities. Through the change of name, he acquires the full status of a "son" and membership of his father-in-law's patrilineal descent group, and just like a real son, he can maintain the family patriline. In this sense, the Bai in-coming husband can be called an "adopted-son-in-law". In a society with a patrilineal system, however, his status is not simply symmetrical with that of the tong-yang-xi, or adopted-daughter-in-law.

The social position of uxorilocal marriage

The most distinctive characteristic of uxorilocal marriage in Bai society is probably the fact that it is not especially looked down upon or disfavored. The social status of the Bai uxorilocally married husband is definitely higher than it is in many other cases among the Han Chinese, so that the second or the third son of a fairly well-off family will sometimes marry out to become the adopted son-in-law of another wealthy family. This question of the social status of the in-marrying husband seems to be connected to the Bai people's attitude towards the change of descent group membership and surname.
Generally speaking, the Han Chinese were traditionally reluctant to change their surnames, because it meant a severance from the natal family and descent group, and many of the Han Chinese who married in retained their own surnames. This type of incoming husband could not obtain the status of the heir, and naturally his position in his wife's household was not very strong. However, some of the Han Chinese in-marrying husbands changed their surnames and acquired all the rights and duties of a "son". Although the husband in the latter case seemed to have more rights in the household, his social standing was usually low. Others often regarded him as having been too poor to retain his surname, or would condemn him for his "immoral" decision to desert his own parents and ancestors. On this point, the Bai people have significantly more flexible arrangements and attitudes.
For a Bai man, marrying out as an adopted-son-in-law of another family does not mean a complete rupture of his natural patrilineal relationships. In the matters concerning ancestor worship, he almost exclusively acts as a member of his father-in-law's descent group, but on other ceremonial occasions such as life-cycle ceremonies held by one of his natal agnates, he might act as their kinsman, using his original first name and surname. Moreover, the representative of his natal family or descent group usually plays a special role in the life-cycle ceremonies of his own children and grandchildren. How much the uxorilocally married man retains his relationships with his original patrilineal kin is rather situational, depending upon several factors such as social status and the economic position of his natal family.
A Bai uxorilocally married man apparently belongs to his father-in-law's descent group, but at the same time, his relations with his original descent group continue under the surface. Consequently, one of his sons holds the right to go back to his natal family and inherit property there. This custom is called guizon [guizong]. According to the inhabitants of Green Village, this right of the man to return to his original family is inherent in all uxorilocal marriages, and is preserved until his grandson's generation. However, whether this right is exercised or not depends on many factors. First, the in-marrying husband has to have more than one son or grandson so that the guizon does not leave the family into which he has married without an heir. The economic conditions and the number of the people in both the families concerned, and the extent of the relationships between the in-marrying husband and his natal family, are important factors. Guizon takes place when it seems reasonable and is acceptable to both the families concerned.
The actual incidence of guizon appears to be fairly low, but I witnessed one during my research in Green Village. The guizon arrangement came into effect when the young man took a wife. He started his married life in his father's natal home and received his new first name and surname in accordance with the rules of his new family, in order to use it in ancestor rites. The existence of such a Bai custom demonstrates the way in which uxorilocal marriage is recognised as normal in Bai society. It is going too far to say that the status of the Bai uxorilocally married man is no lower than that of virilocally married man, but his status is not an object of contempt or compassion in Bai society. So, the most typical style of marriage ceremony in Bai uxorilocal marriage is to exchange the bride and bridegroom's houses temporarily so that the bridegroom goes to call for the bride at his natal house. The fact that the Bai attitude towards descent allows double affiliation for the uxorilocally married man to some extent, seems to influence the position of such a marriage in Bai society.

Changing surnames to continue the patriline

In the case of either uxorilocal marriage or guizon, Bai people do not show much reluctance to change their surnames. As a result, the in-marrying husband or father is also placed on the ancestral tablet under his acquired name by marriage. This way, the sonless family can continue their patriline.
In most cases there is no special mention of the uxorilocal marriage on the tablet, and we can only guess the possibility of such a marriage from the names of the deceased couple. The ordinary style of the ancestral tablet is to write down the names of each couple in a row, the man's full name first followed by his wife's natal surname and her personal name. The deceased couple in an uxorilocal marriage are naturally recorded on the tablet under the same surname, but not all the couples with the same surname are married uxorilocally. In Bai society, the rule of avoiding a marriage between men and women of the same surname, tongxing-buhun, does not exist in practice.[4]
So, it is hard to tell whether the deceased man was an adopted son-in-law or not just by looking at the ancestral tablet. The Bai uxorilocal marriage is a useful device to keep the patrilineal genealogy unbroken.

The causes of uxorilocal marriage

The causes of the prevalence of Bai uxorilocal marriage are many. Some people might ascribe it to the Bai cultural characteristics, and suggest that after all they are not Han Chinese. I think the case of the Lemo is suggestive in this connection. They are a sub-group of the Bai in the north who are supposed to be less sinicized than the Bai in the Dali Basin, but who rarely resort to uxorilocal marriage in the case of having no son, for they regard it as shameful. The property of a sonless Lemo is usually inherited by his brothers or nephews (Minzu Wenti Wuzhong Congshu Yunnansheng Bianji Weiyuanhui, 1981: 193; Yunnan Bianjizu, 1991: 76). Naturally we should be careful in using the case of the Lemo to infer the situation in the Dali Basin before its sinicization, but this case hardly supports the argument for an indigenous Bai origin for the high incidence of uxorilocal marriage, and indeed suggests the opposite.
I myself think it more productive to consider the Bai situation within the continuum of variation among the Han Chinese, because the Bai in the Dali Basin share the basic Han Chinese ideology of patrilineal descent. Although their deviations from the ideal patterns of the Han Chinese are conspicuous in some ways, all of these deviations can be found separately among particular groups of Han Chinese elsewhere. The degree of sinicization of the Bai at the time of Liberation, at least among their upper stratum, was of such an extent that it was not hard for them to become Han Chinese. So, the question is: what factors drive them to choose uxorilocal marriage rather than the ideal pattern in patrilineal ideology of the adoption of agnates?
The factors concerning the problem of uxorilocal marriage in the patrilineal system can be divided into four kinds: (1) factors affecting the incidence of the lack of a son (heir), (2) factors inhibiting the uxorilocal marriage, (3) factors strengthening the importance of the continuation of the individual family, and (4) factors promoting uxorilocal marriage. The situation of the Bai in the Dali Basin shows an inclination towards uxorilocal marriage in relation to all four of these kinds of factors.
The first three kinds of factors are related to the composition and the function of the patrilineal kinship organization. Compared with the nuclear family, the probability of the lack of a son in a joint family should be lower, and strong lineage organizations favor adoption over uxorilocal marriage. In the Dali Basin, there have been very few joint families and lineage organization is weak. On the contrary, the domestic unit of the individual family is relatively independent and important in Bai society, which might lead to greater significance being attached to the continuation of the individual family as a unit of organisation in the subsistence economy, rather than the possession of communal property and of ancestor worship.
As for the fourth set of factors, I can think of the following reasons which might have favored the development of uxorilocal marriage: (a) the importance of female labor, (b) high population density and need for an efficient method of adjusting population imbalances, (c) lower marriage expenses in uxorilocal marriage, and (d) the position on the "frontier" of the Han Chinese culture, where the Han Chinese are minority late-comers.
The position on the frontier may be related in two ways with the incidence of uxorilocal marriage. Firstly, the ideal patterns would be less prevalent in the marginal areas. Secondly, the Han Chinese in a frontier environment, with an excess of males, may have developed a trend toward uxorilocal marriage, which may have exerted an influence on the non-Han peoples in the process of their sinicization. To the Bai in the Dali Basin, the closest Han Chinese are the ones whose ancestors came to the Basin during the Ming dynasty as garrison troops. Uxorilocal marriages are common among the Han Chinese villages in the Dali Basin and through the examination of the genealogy of a dominant Han lineage, I have been able to trace its incidence at least from the end of the eighteenth century onward. Thirdly, for the Bai, a Han in-marrying husband, as an agent of Han Chinese culture, would be a good son-in-law to have as a member of their group. For a Han Chinese man lacking the land to make a living, marrying into a Bai family would be one solution.
In any case, the high incidence of uxorilocal marriage among the Bai in the Dali Basin can perhaps be better understood by observing their patrilineal system in relation to the case of the Han Chinese than by searching for it in the peculiarities of Bai culture. And in particular, it may be through paying attention to the marriage practices in the frontier areas, as the Han Chinese spread, that we will be able to discern the causes of variations within Han Chinese society itself, and the dynamics of the process by which this society expanded.


1. Dabequ is the Bai art of story-telling to the accompaniment of a folk guitar.
2. I have used the recent Bai-yu pin-yin wen-zi (Bai alphabet) to romanize Bai words, but have omitted indications of tones.
3. For Bai words which are direct loans from Han Chinese I have added the standard Chinese pronounciation in brackets. For other Bai words, the Chinese meanings are given in parentheses.
4. Although the Bai people do not practice tongxing-buhun, that idea of avoiding a marriage between men and women of the same surname seems to some extent to be known. In some cases of husbands and wives with the same surname, the surnames of the wives sometimes seems to be written using a different character with the same pronounciation on purpose. The intent is to avoid being able to trace the same surnames of hubands and wives on the same tablets.

Return to Contents page

Updated 4 June 2020