Introduction

Names are a part of our individual, family, and social identities [9]. Attachment to one’s name is a close correlate of self-acceptance and self-identity [7, 45], which have political as well as personal meanings [9, 35]. Conversely, changing names can disrupt personal identity as well as recognition by others [15]. In addition to personal identity, choice of surname may reflect broader views about gender [31], and further, attitudes toward surnames may provide a window into public attitudes toward gender equality [11]. In Japan, attachment to family names is strong. One’s family name, myouji, precedes a given name. Individuals may be referred to by their given name, with or without a diminutive/honorific suffix, in a personal network of family and friends. Yet, it is also fine to use surnames informally to refer to oneself and others. In formal settings, calling others by their surname with an honorific suffix is more common, while people refer to themselves by their surname or full name. As Shin [35, p. 172] points out, “changing surnames means changing the entire name, not a half of the name, by which a person could be perceived as a completely different one.”

The potential impact of changing surnames may be particularly important in a country such as Japan since family names are commonly used to refer to others [45]. The Japanese Civil Code Article 750 states that couples must choose either the husband’s or the wife’s surname upon marriage. The standard practice, occurring in 96 percent of Japanese couples, is for women to change their original surname to their husbands’ after marriage [15]. Although Japanese rarely have middle names, in writing, married women sometimes enclose their maiden name in parentheses, as if to connote the lesser value of their original surname. However, as more women enter the labor force, establish themselves in their careers, and marry later in life, they may prefer to keep their maiden name, not only for their attachment to it, but also for practical reasons such as not wanting to confuse business contacts with a new surname or avoiding the hassle of changing family names on common legal documents such as driver’s licenses and health insurance cards. In December 2015, the case of three women and a married couple who demanded married women’s rights to keep their original surname was denied by the Supreme Court in a 10-to-5 decision. Perhaps unsurprisingly, 10 of the 12 men judges voted for the law’s constitutionality, whereas all 3 women judges voted against it. Japan is one of the few countries worldwide, and the only developed country, that requires the married couple to have a single surname which must be either spouse’s surname. Public opinion about surnames is split. About one-half of Japanese adults under age 60 agree that the law should be changed [15]. The current study examines Japanese attitudes toward married persons’ surnames. Using data from the 2000–2003, 2006, and 2010 Japanese General Social Survey (JGSS), this study focuses on the influence of traditional gender ideology and preferences for the extended family household on surname attitudes.

Given the importance of names for identity, it is notable that there is such a lack of research on the topic. Indeed, Pilcher [30] states that “the sociology of names remains in its infancy” (p. 765). Researchers examining surname choices for married women or children tend to focus more on individuals’ own surname decisions or preferences. There has been limited multivariate research using general population samples on Japanese public opinions about surnames after marriage. This may be because there are few countries, especially developed countries, where women have no legal rights to keep their native surname. In this sense, Japan provides a perfect case for the study of gender, family, and surnames. Japan remains very low on rankings of gender equality [53]. While the proportion of extended households has declined, there is evidence that younger cohorts may experience delayed co-residence a few years after marriage [16]. This study is the first to examine the link between gender ideology, familism, and attitudes toward surnames in Japan. Below we review the background on surname law in Japan as well as potential correlates of surname attitudes before turning to the methods and findings.

Origins of the Shared Surname

The requirement that married couples share a surname, recently upheld by the Supreme Court, originates in the Family Registration (koseki) Law and Family Law of the Meiji era (1868–1912). These laws established i.e., a type of stem family based on primogeniture previously confined to the elite for the entire population [12, 35, 48]. Surnames had long signified individuals’ relationship to power, and unsurprisingly, the state seized control over surnames [35, 51]. The Family Registration Law of 1871 laid out key concepts of family register, most notably, permanent domicile (honseki), which is the location of family registry and different from the permanent address. By 1875, the koseki law required everyone, commoner or noble, to take a surname. As part of nation-state building, family registration aimed to centralize administrative purposes such as taxation and conscription.

The Family Registration Law reinforced the patriarchal family system by two filing rules: one surname per family register and the shared surname for the father and his sons [35]. In its early stage of enforcement, the same surname rule caused much confusion about wives’ family registration record [35]. The key question was whether wives’ record should be part of their husband’s family register [35]. At that time, wives’ surname change was still a new idea. Indeed, the government’s initial position followed the aristocratic tradition where wives kept their maiden names. By contrast, the Civil Code Drafting Committee’s position was to require wives to change their surname to their husband’s [35]. They considered such a requirement more in line with wives’ subordinate status vis-à-vis their husband’s authority [35]. The “same surname” view prevailed, and the Family Law of 1898 stipulated that wives must take the name of their husband’s family. Over time, koseki became a public record of a family household, listing its members’ given names, birthdates, birthplaces, and relationship (zokugara) based on consanguinity, marriage, or adoption. Guided by primogeniture, the eldest son remained in their parental home after marriage. When other sons got married, they left their home to establish a new i.e., and removed their record from the original register to start a new one. Daughters were expected to marry out of their parental home and erased their record to be entered into their husband’s family registry.

The shared surname rule was established just as hypergamy, i.e., women’s marrying men of higher social status, had the effect of strengthening patriarchy [48]. Women who married up became more dependent on their husband. Ueno [48] argues that the crux of the shared surname rule was to weaken wives’ ties to their native kin. The loss of strong ties to their descent group, reinforced by the single surname rule, would further weaken wives’ bargaining power relative to their husbands and precluded intervention by wives’ parental family in case of abuse by their husband. Certainly, neither the shared nor dual surname practice is intrinsically patriarchal. As a symbol of kinship, surnames are shared by husband and wife in endogamous societies, while they are not shared in exogamous societies [48]. In China and Korea, wife and husband keep surnames of their respective origins, and their children’s surnames are typically determined by patrilineage. This practice has been said to brand wives as outsiders to their husband’s lineage [41, 48, 51]. Nevertheless, Meiji Japan’s shared surname rule, enmeshed with its institutional arrangements (e.g., i.e., primogeniture and koseki), would become the linchpin of the patriarchal family system for over a century to come.

The Single Surname Law in Contemporary Japan

After Japan’s defeat in WWII, the system of i.e. was formally abolished under the US occupation. The Family Law was revised with “the language of equality, individual rights, freedom of choice, and voluntary unions” in Euro-American paradigm [12, p. 5]. Yet, basic elements and structures of the i.e. system were kept intact, especially in rural areas [12, 46, 47]. Post-WWII reconstruction kept family registration (koseki), now revised for there to be one registration (and thus one surname) per nuclear family, which underlies the country’s family systems and gender relations, permeating into daily lives of Japanese [38]. With the idea of family lineage at its core, koseki serves as a birth certificate, i.e., a formal proof of identity as Japanese citizen [19]. Koseki is not required for everyday routines involving self identification, but one must present it upon key life events, such as marriage, childbirth, divorce, and inheritance. Passport application is another example where koseki is required. Proponents of the law stress that married women can use their maiden name as an alias for routines, but many routines require official papers, e.g., driver’s license for driving and insurance card for going to a clinic. For example, Shimoju Akiko [34], a renowned journalist who has been known by her maiden name for decades, recounts the frustrating confusion due to a mismatch between the legal surname on her passport and the maiden name that she ended up using by habit to buy her tickets for international flights. Media reports of women’s perspectives on the same-surname law also include the violation of their privacy. Some women fear the possibility of multiple name changes if they adopt their husbands’ surname upon marriage, revert to their maiden name after divorce, and adopt still another conjugal surname in the case of remarriage [23]. Nugent’s [26] argument on “the male privilege of continuous identity” (p. 500) derived from the normative standards for choosing conjugal and child surnames in the US readily applies to Japan, where with its rising divorce rate, the traditional conjugal surname practice disrupts self-identity much more for women (and their children) not only through union formation but dissolution [51]. Japanese women who have taken their husband’s surname use terms such as sorrow, regret, and loss to describe their feelings regarding the loss of their own surname [45]. However, keeping their native surname, too, adversely affects women in that they “solely bear the burden of double surname”, juggling their original and official surnames [35, p. 172].

In 1996, the Legislative Council within the Ministry of Justice recommended allowing married women to choose whether to keep their maiden name or take their husband’s surname [50]. Conservative legislators thus far have managed to block any change to the single surname law, arguing that it would undermine Japanese family values. In the 2015 Supreme Court decision upholding the constitutionality of the surname law, Justice Itsuro Terada stated that the practice of single surname is “deeply rooted in our society” and “enables people to identify themselves as part of a family in the eyes of others” [3]. He stressed that women can use their maiden name informally every day, and acknowledged that giving it up can disadvantage women professionally or otherwise [3]. The decision gained much attention domestically and abroad. The tone of media coverage was generally negative, presenting it as anachronistic and discriminatory. The UN Committee on the Elimination of Discrimination against Women had been pressuring Japan to amend the law [49]. The Global Gender Gap Index 2017 ranks Japan near the bottom fifth of gender equality [53]. The surname law has both practical and symbolic implications for women’s social status. It endorses patriarchy and refutes marriage as an equal partnership.

Attitudinal Correlates of Preference for the Single Surname Tradition

Gender Ideology

Traditional gender role ideology is a set of beliefs that supports separate spheres for men and women. Statements such as “husbands should be the primary breadwinner, whereas wives should be the primary homemaker” and “it is more important for a wife to help her husband’s career than to pursue her own career” are congruous with the view that married women belong in the private sphere where they require no official surname of their own choosing. Naming practices are inextricably tied to gendered notions of power and dominance [43] and instrumental in reproducing gender inequality [31]. Pilcher [31] asserts that women are “doing gender” when they take their husband’s surname and “redoing gender” when they keep their own surname. Justice Terada’s comment that married women can use their maiden name for informal purposes captures this point of view. Feminist scholars have long maintained that the private–public dichotomy is fed by and feeds on patriarchy [29, 35]. Recognizing women’s existence primarily in the more hidden, private domain endorses women’s subordinate status. When responsibilities in the private sphere (e.g., domestic labor) are devalued relative to responsibilities in the public sphere (e.g., market labor), allowing women to use their maiden name solely informally reinforces patriarchy and justifies unequal gender relations [35].

Gender egalitarianism, defined as the belief that gender role differences must be minimized in order to attain gender equality [8], is likely to serve as an ideology underpinning the protest movement against the single surname rule in Japan, along with a belief in human rights more generally [24, 35, 36]. Most clearly related to egalitarianism, dual surnames would remedy gender inequality created by women disproportionately losing their names [15, 27]. Advocates of the dual-surname system contend that the system enables married women to attain their public identity independent of their husband. Such identity is necessary for women to fully participate in the public life domain with potential for greater professional achievement [27, 45]. Although certain employers, most notably universities, allow Japanese married women workers to use their maiden names, women are at the mercy of employers’ discretion, further burdening women in the labor market, where they already face severe gender inequalities.

While gender traditionalism and attitudes toward surnames seem related, empirical research on this link is lacking in Japan. We turn to studies from the US, Canada, and Norway for evidence that gender ideology (or feminist identity) predicts individuals’ decisions, attitudes, and plans regarding surnames after marriage. Noack and Wiik [25] find that in Norway, gender-role egalitarianism makes it more likely for married women to keep their maiden name. Stoiko and Strough [37], in their study of students at a large US public university, find that those who consider themselves feminist are less likely to support the traditional surname practice for married people generally. Likewise, a study of Canadian undergraduates finds that women who are concerned with gender equity are less likely to favor traditional naming preferences [21]. Admittedly, gender-role egalitarianism and feminist identity are separate constructs, but they are close correlates, particularly among women [44]. Hamilton et al. [11] find that name change attitudes are closely linked to gender attitudes and may serve as an indicator of gender beliefs. Stoiko and Strough [37] also find that only among men, self-identified feminists are less likely than non-feminists to personally conform to the traditional same surname practice. Yet, Kerns [17], in her analysis of data from a sample of US women, finds no clear link between feminist identification and the decision to keep a maiden name upon marriage.

Views Toward the Three-Generation Family Household

Closely related to but distinct from gender traditionalism is a traditional attitude in support of the stem-family extended household. Co-residence with aging parents in East Asian societies may be seen as a form of filial piety, or a way for adult children to repay their parents through elder care. Among married people living in a three-generation household, men are more likely to do so with their own parents, and women with their in-laws, indicative of patrilocal residence. East Asian cultures influenced by Confucianism emphasize family identification, solidarity, and responsibility [6, 42]. Thus, maintaining a positive family image can be quite important for many Japanese. Hyosatsu, nameplates for houses that typically contain one surname or full name of the male head of household, are often expensively custom-built. One may say that Japanese nameplates symbolize familism, one of the key dimensions of i.e. ideology, which is a social normative belief that the welfare or perpetuation of the family unit be prioritized over the self-interest of its individual members [42]. Indeed, arguments in favor of the single surname law often center on the importance of family ties and continuity for children [27, 45]. Surname sharing can be viewed as a collectivist way of family identification, or establishing “a strong connection with the family to the extent of viewing the family as part of the self” [6, p. 701].

In Japan, married children with stronger family centric beliefs are more likely to live with their parents [40], although the formation of an extended family household may reflect more practical concerns, such as relieving economic insecurity [28], and have little to do with familism. The proportion of older Japanese living with their adult children has declined sharply in recent decades [54]. However, the growing trend of nuclear families has not erased the i.e. system. As Traphagan [46] argues, “the i.e. is not a spatially restricted entity” (p. 369) and continues to set the normative standard for ideal family relations, particularly among individuals with rural backgrounds.

As with gender ideology, most research on familistic views and attitudes toward shared surnames has been conducted outside the Japanese context. Stoiko and Strough [37] find family tradition to be one of the most common rationales invoked by US university students for supporting women’s surname change after marriage. US women’s decisions and plans to change their surname due to marriage seem to also have much to do with their familistic belief [17, 39]. Sharing a surname can show marital commitment and family solidarity [1, 4, 5]. Indeed, women who take their husband’s surname are seen as more communal [10] while women who keep their own surnames may be judged negatively [32]. For example, Shafer [33] finds that in the US, less educated men are more likely to think that a husband is justified in divorcing his wife if she kept her surname. Meanwhile, in Norway, liberal family values, arguably a negative correlate of familism, are associated with women’s greater likelihood of keeping their native surnames [25].

Hypotheses

This study focuses on individuals’ general attitudes toward married couples’ surnames. Based on previous research, we posit the following hypotheses:

Hypothesis 1

Those with more traditional gender beliefs will more strongly approve the traditional surname practice than those with more egalitarian gender views. Those who support a more traditional division of labor with men at work and women at home will likely see more benefits to sharing a surname.

Hypothesis 2

Preference for the three-generation household, regardless of their actual residential status, is associated with approval of the single surname tradition. To individuals who see the extended family favorably, the image of a whole family sharing a surname may have symbolic importance that overrides some inconvenience or injustice.

Methods

Data and Sample

This study uses the Japanese General Social Survey (JGSS) which started in 2000 modeled on the US General Social Survey (GSS). We analyze data from 2000–2003, 2006, and 2010, when the JGSS asked about attitudes towards surnames of married persons. In each of these years, it also asked about gender attitudes and views on three-generation households, two of the predictors of interest. JGSS collected data for a 2-stage random sample, stratified by regional block and population size of cities or districts, of Japanese citizens with voting rights. The response rate of JGSS has been about 60%, which is lower than the rate typically observed in the US GSS. After deleting cases due to missing values, we obtained an analysis sample of 14,356 respondents (6573 men and 7783 women) for this study.

Measures

Dependent Variable

This study predicts Japanese attitudes toward surnames of married persons. Respondents were asked what a couple should do with their surname when they get married and given four options to select from: 1 =  “The wife should take her husband’s surname without question”; 2 =  “It is preferred for a wife to take her husband’s surname under current conditions”; 3 =  “The couple should claim a common surname, however, it is up to them to decide whose name they will keep”; 4 =  “There is no need for either to change their names”. Choosing an entirely new surname for the couple was not given as an option. We consider 4 as showing no support, 3 as weak support, 2 as moderate support, and 1 as strong support for the traditional surname practice. The original order was reversed for the main analysis (see Analytical Strategy section).

Independent Variables

Traditional gender ideology and preference for three-generation families are our predictors of interest. The JGSS has consistently asked questions on various gender attitudes. We use the two items that were included in all years (and forms) used in our analysis: “A husband’s job is to earn money; a wife’s job is to look after the home and family” and “It is more important for a wife to help her husband’s career than to pursue her own career”. Respondents rated their agreement with these statements on a 4-point scale ranging from “strongly agree” to “strongly disagree”. In 2006 (Form B), gender role attitudes are measured with a 7-point scale instead of a 4-point scale. To make it consistent with other years, we recode data from 2006 to a 4-point scale. Specifically, categories 1 (“strongly agree”) and 2 (“agree”), and 6 (“disagree”) and 7 (“strongly disagree”) are combined. 3 (“somewhat agree”) and 4 (“somewhat disagree”) are unchanged. The midpoint 4 (“neither agree nor disagree”) is recoded to 2.5. The original scale is reversed so that higher scores indicate more traditional attitudes. The average score of the two items is used for the analysis.

We measure preference for three-generation households with the statement: “it is desirable for three generations (older people, their married children, and grandchildren) to share a home”. Available responses to the question are “desirable” and “undesirable”, and we measure this variable dichotomously (1 = desirable; 0 = undesirable).

We control for age, gender, education, rural background, marital status, and extended household status as they are logical correlates of attitudes toward married persons’ surnames and the predictors mentioned above. Age is measured in years. Older individuals are likely to give stronger support to the single conjugal surname practice. Age is also likely to be associated positively with gender traditionalism and approval of three-generational households. Gender is measured with a dummy variable (1 = female, 0 = male). Under the single surname requirement, wives are more likely than husbands to undergo surname changes, and women may thus find the issue more pertinent than men to their lives. Although it is unclear whether women are more or less likely than men to support the same surname at each threshold, women tend to hold less traditional gender attitudes. Education is measured on a scale from 1 =  “less than senior high school diploma” to 6 =  “college degree and more”. Education, to the extent that it promotes the notion of liberalism, is likely to be associated negatively with each of the attitudinal predictors and the dependent variable.

Rural background is measured with a four-category variable based on rural residence at age 15 and at present. The categories on exposure to rural living include: 1 = did not live in a rural area at age 15 or at present, 2 = live in a rural residence at present (but not at age 15), 3 = lived in a rural residence at age 15 (but not at present), 4 = lived in a rural residence at age 15 and now. As discussed earlier, i.e. ideology continues to exert influence in rural areas [12, 46]. Compared to those with an urban background, those who grew up or now live in a rural region are likely to hold more traditional gender attitudes and be more supportive of multigenerational households. Meanwhile, having lived in a rural region may also directly shape individuals’ perspectives on what the good surname practice is (e.g., name keeping vs. name changing).

Marital status is measured with a dummy variable (1 = married; 0 = single). By law, married people have experienced the surname change for themselves or their spouse, and may be more receptive to the norm of wife’s name changing upon marriage. Married people may also be more gender-traditional and have a stronger familistic belief. For a similar reason, we control for extended family status, defined as living in a family household with three or more generations related by descent or marriage (1 = live in an extended family household; 0 = otherwise). Finally, we include dummy variables for survey years.

Analytical Strategy

We estimate ordered logit models to fully take advantage of the ordinal data on endorsement of the traditional single surname practice for married couples. Three thresholds are used to partition the observations. In other words, the contrast is made in three ways: (1) “no support” versus “weak, moderate or strong” support; (2) “no or weak” support versus “moderate or strong” support; (3) “no, weak, or moderate” support versus “strong” support. For each contrast, the first group is designated as a reference group. As shown below, this model choice helps reveal differing effects of some variables (e.g., attitudes toward the three-generation family household) across levels of support for shared surnames.

An ordered logit model estimates the log odds of giving support at the level exceeding each threshold, thus involving three equations. In a preliminary analysis, the parallel-slope was imposed. In other words, the model produced the same set of slope coefficients across the equations (though with varying y-intercepts). As part of diagnostics, we estimated a generalized ordered logit model using STATA’s gologit2 program [52] and identified a serious violation of the parallel-slope assumption for some covariates. We thus report below results from a generalized ordered logit model that selectively impose the parallel slopes, or a partial proportional odds model.

Findings

Descriptive Statistics

The JGSS has used a standardized question about attitudes concerning surnames of the married on a fairly regular basis. Based on the available JGSS data from 2000 onward, Japanese attitudes about surnames were stable. Table 1 shows support for the same surname practice by gender. The table also reports gender-specific statistics for our key predictors, traditional gender attitudes (mean) and support for three-generation family households (percent). Japanese attitudes are evenly split on the issue of surnames. The proportion of individuals who support the traditional practice of the wife changing her surname to her husband’s stayed constant in the period 2000–2010. It is not surprising that the patriarchal single surname rule gets moderate public approval in a country where most men and a sizeable minority (40%) of women prefer a traditional division of labor between spouses [14, 20]. A similar proportion of men and women disapprove of the same conjugal surname practice. However, there is a gender difference for those who give at least some level of support. Men are more or less evenly split across the categories of “weak”, “moderate”, and “strong” support. Women, by comparison, are most likely to show only weak support. On average, men score slightly higher than women on traditional gender attitudes. Despite the upward trend of nuclear families, over 60% of respondents find three-generation family households desirable, although more men than women share this viewpoint. See “Appendix” for summary statistics of the study variables.

Table 1 Summary statistics of the key study variables by gender
Full size table

Predicting the Level of Endorsement of the Same Surname Practice

Table 2 summarizes results from a generalized ordered logistic model with the JGSS data from 2000–2003, 2006 and 2010. As noted above, we use a partial proportional odds model [52]. Informed by our preliminary analysis, we impose parallel slope restrictions on variables on rural background, extended family status, and survey years 2002, 2003, and 2010.

Table 2 Generalized ordered logit model of support of the shared conjugal surname, JGSS 2000–2003, 2006 and 2010
Full size table

As expected (Hypothesis 1), gender traditionalism positively predicts support for the practice of women taking their husband’s surname upon marriage at all levels (p < .001). Moreover, the log odds of either giving the practice “at least weak” or “strong” support is larger than the log odds of giving “at least moderate” support. Note that giving “at least weak” (i.e., “weak,” “moderate,” or “strong”) support is the equivalent of disapproving of the most progressive view on conjugal surnames (“There is no need for either to change their names”). That is, gender traditionalism better predicts strong attitudes, whether approving or disapproving, surname sharing, than the middle-of-the-road attitude leaving it up to each couple to decide whether to share a surname.

The effect of attitudes toward three-generation households is significant, but its direction varies across thresholds of support for the same surname. Preference for three-generation households positively predicts “at least weak” support for the surname practice, in line with Hypothesis 2. However, this residential preference negatively predicts “strong” support for the shared surname. Its effect on “at least moderate” support is negative, but quite small. Preference for the traditional family form may be in line with acknowledging the view that sees at least some need for married couples to share their surname, but it seems also incongruous with the rigid view insisting that the wife take her husband’s surname, reminiscent of a situation where a son-less father who owns a business desires both a three-generation household and mukoyoshi (adopted son-in-law) who would be married to his daughter and take her surname, which is quite common in Japan [2].Footnote 1

The control variables have significant effects on support for the shared surname practice at one or more levels. Older people are more likely to give the same surname “at least moderate” or “strong” support, although age has no effect on “at least weak” support. Education negatively affects support for the same surname. The negative effect of education is significantly (p < .001) larger at the highest than the lowest or middle threshold of support for the shared surname. Compared to individuals who were urbanites at age 15 and are now urbanites, individuals with at least some rural background, especially those who lived in a rural region at age 15 and do so at present, are more likely to support the traditional surname practice at each threshold. Married people are more likely to support the same surname at the lowest threshold but are less likely to support the same surname at the highest threshold. Extended family status significantly predicts support for the same surname consistently across all thresholds.

Finally, controlling for the other variables, support for the same surname rose in 2010. The effect of the survey year 2010 is significantly higher not only compared to the year 2000, but the other four survey years. The exception is the difference in the effect on moderate support between 2006 and 2010. (The effects of survey year are not tabled but available upon request.)

Conclusion

In December 2015, the Japanese Supreme Court upheld the same surname law, thus rejecting the right of married women to keep their surname legally (unless their husbands agree to change their own). This decision not only highlights the court’s conservative history since its inception after WWII [22], but also reflects the varying public opinions on married people’s surnames and corresponding cultural divisions on normative gender relations. This study set out to understand how traditional gender ideology and preference for multigenerational households are associated with attitudes toward married persons’ surnames in contemporary Japan. We find that those with more traditional gender attitudes are more likely to endorse the same surname practice. This is consistent with research from Western countries that shows that egalitarian and feminist views are associated with rejecting traditional surname practices [21, 25, 37]. In comparison, preference for three-generation households affects surname attitudes in a more nuanced way. Whereas this residential preference suppresses the liberal view that insists on separate surnames, it seems to also temper the most conservative position on conjugal surnames. Previous research suggests that familistic beliefs encourage traditional surname practices [17, 39]. This may be due to the assumption that wives who take their husbands’ surnames are more committed to their marriages and more communal [4, 5, 10]. While this aligns with our finding that those who think a multigenerational family is desirable are more likely to prefer that couples share a surname (weak support), it does not explain why those with an extended family orientation are less likely to think the wife should take her husband’s surname without any question (strong support). It may be that those who favor multigenerational families see the practical benefits of a shared household [28] as well as the potential for close family relations [46] and thus they are more flexible about how couples achieve a shared surname.

A recent government poll showed 42.5% of respondents support the dual surname option for married couples, a 7% increase from the previous poll from 2012, energizing opponents of the current law [18]. As the country’s family structure becomes more diverse with the rising rates of divorce, remarriage, international/inter-ethnic marriage, de facto marriage, etc., the limitations of the Family Law, and koseki should become self-evident to most Japanese. When that time comes, the highest court may at last strike down the same conjugal surname rule. Japanese may soon need to reject the preconception that members of the same family must share a surname [48]. As Ueno [48] points out, Japanese emphasis on patrilineage of children, along with the rising divorce rates for couples with children, has led to an increase in the number of separate-surname families. A typical example is where divorcées who revert to their maiden name get custody of their children who go by their paternal surname. Another scenario concerns international marriage. Japanese married to non-citizens can retain their surname without requiring their foreign spouses to change their surname. The dual surname practiced under the current legal system, however, can be grounds for discrimination against children whose parents keep their separate surnames. These children, similar to children born to unmarried parents, may be labeled illegitimate, a term so demeaning that its use is discouraged to describe any child in many societies. Children of international marriages may also experience difficulty gaining access to social services (e.g., immunization) due to the institutional inability to respond to more diverse family forms and identify children by their parents with dual surnames [Osawa 2010, cited in 41].

This study has a few limitations. First, although our study is based on national data, the Japanese General Social Survey’s sampling frame is limited to citizens with voting rights. Because naturalization is rare in Japan, this sampling method excludes both foreign-born residents and native residents of non-Japanese descent (e.g., Korean and Chinese). Yet, foreign nationals comprise an increasing proportion of the population. Their exclusion from the survey is unfortunate in that they may hold more accommodating attitudes toward surnames. Second, due to its use of pooled cross-sectional design, this study does not allow us to make causal inferences. Longitudinal data would allow more disentanglement of the relationship between various attitudes related to gender ideology, extended family household, and surnames. It would also be informative to investigate how experiencing marital separation, divorce, and remarriage affect individuals’ attitudes toward the surname law and related social norms with longitudinal data. Third, our measure of attitudes toward surnames focuses on shared surnames, with two options for the wife changing her name, one option that leaves open which spouse changes their name, and one option for no change. However, it could be interesting to see how people feel about married couples choosing an entirely new surname or taking on a combined name. One US study suggests that wives who keep their own last name are judged more harshly while wives who use a hyphenated last name are seen as similar to wives who take on their husband’s last name [33]. Fourth, due to data limitations, we are unable to consider birth order (e.g., eldest son) of the respondent, which is a possible correlate of familistic belief and support of the traditional surname practice for the married.

In short, focusing on the shared surname rule as a married women’s issue is incomplete. Japanese are marrying late or not at all, and the married are more likely than ever to experience divorce and remarriage. The surname issue should concern individuals’ rights to legally choose their own name regardless of marital status, and also protection from any stigma associated with certain surname situations. After all, the surname has multiple functions, and its role for self-identity matters as much as its role as a symbol of kinship. As one plaintiff in the surname lawsuit lamented soon after hearing the court’s decision, “I won’t be able to die as Kyoko Tsukamoto” [13], older women, married or widowed, often express strong wishes to be buried in the grave of their parents and other ancestors when they die. Ancestral gravestones typically bear a patrilineal name, which suggests that desire for maiden-name keeping, too, is intertwined with the traditional gendered family norm in the minds of (especially older) women. As Pilcher [31] argues, women who keep their original surname can be viewed as “redoing gender”. Future research on attitudes toward surnames in Japan and elsewhere will benefit from expanding the context of surname practice.