![Sample our Humanities journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5939/TOC-Subject-Sample-2021-Humanities.jpg"})
ABSTRACT
In pre-industrial Taiwan, an uxorilocal marriage, in which a man moved in with his bride’s family, was a familial strategy used to continue family lineage and to enhance family farm labor. We examine the prevalence and circumstances in which a family would call in a man for one of their unmarried daughters. Using data from the Taiwan Historical Household Registers Database (THHRD) from 1906–1945, we identify the individual-level factors (including parental status, sibling status, household heads’ occupations, and the capacity of the family labor force) and a community-level factor – the prevalence of uxorilocal marriages by region, which are predictive of uxorilocal marriages. Our analyses first show that women without siblings and women with only female siblings were more likely to adopt the uxorilocal form of marriage. In addition, the effects of siblings’ status were moderated by the presence or absence of parents. For women without any male siblings with at least one parent, especially a father, residing in the household, the likelihood of having an uxorilocal marriage was higher than for those without any parents. Second, an uxorilocal marriage was less common in families with more young family members in the labor force to fulfill the manpower needed for farming. Third, uxorilocal marriage was more likely to occur in families living in the poorest socioeconomic conditions, especially those families in which household heads did not own land and had to sell their labor for agricultural production. Our findings imply that the adoption of uxorilocal marriage varied not only from place to place but also from time to time; it was conditioned by the modes and the means of labor production.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1. See Taiwan Sifa Renshi Bian (臺灣私法人事編, Citation1990), which is a collection of local regulations. The published information in this document-type book was based on the materials that Rinji Taiwan Kyūkan Chōsakai (りんじたいわんきゅうかんちょうさかい; 臨時台灣舊慣調查會) collected in 1910.
2. Using the vocabulary developed by Arthur Wolf, most marriages in traditional Taiwan can be classified as one of three types: major, minor, or uxorilocal (Wolf, Citation1974; Wolf & Huang, Citation1980). ‘Major’ marriages represent the normative form of marriage in which, upon marrying, a woman leaves her family of origin to live with her husband’s family. What Wolf calls the ‘minor’ marriage type is when the family of the future husband adopts a daughter at a very young age; they raise her from childhood together with her future husband. Minor marriages formed a large share (up to half) of all marriages in some parts of Taiwan, including our study population, and were also not uncommon in China. In major and minor marriages, post-marital residence is patrilocal, that is, the bride goes to live with the husband’s family and is absorbed as a member of his household. However, in major marriages, the bride is transferred as a young adult, while in minor marriages, the bride is transferred as an infant or small child. In contrast, in uxorilocal marriages, the groom moves in with the family of the bride. Freedman described the groom of an uxorilocal marriage as ‘a sort of male daughter-in-law’ (Citation1957, p. 122).
3. As a consequence, in the present research, we do not include ethnicity as a predicting factor in explaining the occurrence of uxorilocal marriage in colonial Taiwan.
4. Wolf and Huang (Citation1980) pointed out that uxorilocal marriage not only guaranteed the continuity of the wives’ families without male heirs, but also maintained the family lineage of husbands who were too poor to have a virilocal marriage to find themselves wives.
5. Lin et al. (Citation2014) reported some special cases and illustrated that to maintain their family lines, prostitutes in a red light district in colonial Taipei implemented uxorilocal marriage to call in husbands for their adopted daughters. These husbands were usually immigrants working in the area near the red light district.
6. Chuang et al. (Citation2013) reported that the grooms of uxorilocal marriage usually did not have their parents present at the marriage ceremony.
7. The research by Wang et al. (Citation2008) revealed a contradictory finding that the proportion of males aged 15–64 in a household had no association with the application of uxorilocal marriage among women.
8. Unlike previous research (Lin et al., Citation2014,; Wang et al., Citation2008), our results generalize better to all of pre-industrial Taiwan, and our larger coverage improves both the internal validity (statistical precision) and external validity (generalizability) of our results.
9. In their manuscript, Dong et al. (Citation2015) used the abbreviation, CTHRD, to name the Taiwan Historical Household Registers Database (THHRD).
10. Dong et al. (Citation2015, p. 1073) also pointed out ‘in the CMGPD-LN [China Multi-Generational Panel Dataset – Liaoning, 1749–1909], CMGPD-SC [China Multi-Generational Panel Dataset – Shuangcheng, 1866–1913], and KMGPD-TS [Korea Multi-Generational Panel Dataset – Tansung, 1678–1888], many children were missing’ in contrast to the Taiwan Historical Household Registers Database. Additionally, in the same manuscript, Dong et al. (Citation2015, p. 1073) compared the quality of the Taiwanese data and the NAC-SN [Japanese Ninbetsu-Aratame-Cho Population Register Database – Shimomoriya and Niita], and mentioned that ‘[r]records of births and deaths in the NAC-SC are also incomplete, but the problems are much less serious than in other registers [CMGPD-LN, CMGPD-SC, KMGPD-TS]’.
11. shows the cross-table of birth cohorts and marriage cohorts.
12. Lin et al. (Citation2014) and Wang et al. (Citation2008) focused on urban-rural diversity in the implementation of uxorilocal marriage in the pre-industrial Taiwan but found few differences. In our study, we instead choose a study design that eliminates most regional differences due to our multilevel design (and our household level controls for sociodemographic differences across regions). We argue that the regional diversity in the prevalence of uxorilocal marriage was caused by factors other than simple urban-rural differentiation, most of them related to the household level context.
13. In regression models we only include marriage cohorts, as it is very strongly correlated with birth cohort, but we present the descriptives of birth cohorts.
14. Note that both variables refer to kin in the household, and that they refer not to the survival status of the kin but to their residency in the household. This explains the low number of sisters in the household, as they were often already married before the uxorilocal marriage of the indexed women.
15. As presented in Appendix 1, birth cohorts and marriage cohorts are highly associated with each other. The Spearman’s rank correlation coefficient is 0.91. Therefore, to prevent the issue of multicollinearity, we omit ‘birth cohort’ from the analytical models.
16. Please note that the proportion of uxorilocal marriage is based on women’s birth communities. To examine the effects of family structure on the application of uxorilocal marriage, we can only observe women in our research sites. In general, uxorilocal marriages are observed in birth communities, while women with virilocal marriages occurring in our research sites can be married in or married out. Therefore, the prevalence of major marriage can be overestimated and that of uxorilocal marriage is underestimated.
17. As mentioned above, Taiwanese life expectancy at birth was shorter than 50 years during the colonization (Barclay, Citation1954). Therefore, it is not surprising to see almost 48% of women lived in households with single parents or no parents at their first marriage.
18. Li et al. (Citation2003) suggested that uxorilocal marriage had ‘practical’ and ‘preservative’ functions. These two terms are the same as the ‘contingent’ and the ‘institutional’ functions that Wolf (Citation1989) used.