African feminist scholars have produced work on race and gender that has far‐reaching relevance to understanding relations of power and domination around the world. Scholars such as Oyeronke Oyewumi and Nkiru Nzegwu draw attention to the ways in which African concepts have been distorted by colonial domination. Oyewumi argues that ungendered relational categories have been translated from Yoruba as “husband” and “wife,” imposing a gendered division where none previously existed (Reference OyewumiOyewumi 2000, 1096). Similarly, Nzegwu argues that Igbo family concepts have been distorted such that marriage appears both central to Igbo society and profoundly patriarchal, when in fact the “family” was not defined by the conjugal couple, but by the consanguineal kin group (Reference NzegwuNzegwu 2006, 26). Nzegwu argues that the subjugation of African women was achieved by colonial efforts, supported by African men, to redefine the family. This redefined family is now treated as “traditional,” which supports African men's control over wealth and power within the African context, while also upholding a global, racial narrative that places Africa in the evolutionary past and its women in need of saving.

These problems are not limited to a colonial past, however. Oyewumi argues that first‐world feminists conflate “woman” and “wife” such that the woman in feminism is always the wife, particularly the white wife in a nuclear household (Reference OyewumiOyewumi 2000). Marriage, she contends, is not important to African women in the way that it has been to Western women, and “wife” does not define African women's identity. Nzegwu agrees, arguing that “feminism, a priori, injects patriarchy and the category of gender into cultures” (Reference NzegwuNzegwu 2006, 10), so although “gender” has become a legitimate category through the imperial exportation of Western ideology, it is difficult to apply a feminist analysis to societies with “remnants” of ungendered cultures (10–11).

Maria Lugones's concept of the colonial/modern gender system addresses this issue by viewing gender and race as co‐constituted and inseparable (Reference LugonesLugones 2007, 193), and thus as part of a single system. Race is inextricable from gender in this system, because colonial power was premised on a classification of the world in terms of race, which “replaces the relations of superiority and inferiority established through domination [and] reconceives humanity and human relations fictionally, in biological terms” (190). Thus, domination is not understood in terms of a linear hierarchy, but in terms of structural axes that allow for confronting multiple configurations of domination among the categories of difference these axes create.

The model does two additional things: first, it allows for a consideration of domination and power that is not linked to an evolutionary notion of cultural change (“progress”); and second, by suggesting that gendered dominance is a product of history rather than biology, it provides the grounds for a feminist investigation of “ungendered” cultural concepts or radically different gender systems. This allows for a transnational conversation that considers how domination exists within entirely different systems, opening up the possibility that women have not always and everywhere been oppressed by patriarchal systems, and of finding positives in “customary” organization.

The work of these scholars challenges any notion that third‐world women's lives can be understood simply from the perspective of gender: colonial power created both racial and gender hierarchies that disrupted and rearranged existing power relations along multiple categories of difference, and these hierarchies continue into the present. Yet Nigerian sociologist Maureen Ikeotuonye argues that international attention has focused on “progressive” African feminists who repudiate “harmful” traditions and promote universal “human rights” in terms more recognizable and palatable to the dominant, white discourse (Reference IkeotuonyeIkeotuonye 2016). When first‐world feminists speak out against third‐world “cultural traditions,” they can unintentionally legitimate and perpetuate the terms of women's subordination. In doing so, they help to obscure the cultural concepts that valued women.

A productive, transnational feminist conversation requires that first‐world feminists incorporate these critiques and develop different approaches to understanding gender and power in the third world. This article offers one such effort. I argue that marriage and the family are key sites through which the subjugation of women in Africa can be understood, but that this requires local and historical contextualization. Focusing on Asante marriage, I draw on Nzegwu's concept of a “dual gender” system and Lugones's concept of the colonial/modern gender system to trace uneven and complex shifts in gender relations and marital expectations. This analysis reveals the shortcomings of the common narratives of “progress” and “harmful traditions.”

Types of Marriage in Ghana

This article is based on ethnographic field research, carried out between 2007 and 2009 in the Ghanaian city of Kumasi. Kumasi is the historical capital of the Asante kingdom, and Ghana's second‐largest city. The Asante are the largest of several Akan‐speaking peoples in Ghana, and have predominantly chosen Christianity when converting from Akan religious belief. Kumasi is the current capital of the populous Ashanti region, and also the historical capital of the pre‐colonial Asante state. It is a center of regional immigration, but the predominant culture of Kumasi is dominated by this Asante history, and the most commonly spoken language is Asante Twi.

The Asante family system is matrilineal, meaning that membership in the extended lineage is granted by birth, and reckoned through the mother. Lineage membership is lifelong: people do not leave their lineage when they marry. Sibling and mother–child relations are the core familial relations, and these relations form the basis of lineage membership and structure; siblings and their mothers often hold property jointly. Marriage is important, but in customary marriage the relationship between husband and wife is secondary to the relationships between people and their matrilineal families.

Marriage in Ghana is quite complex; it is both a personal relationship and a legal one, with formal elements that define things like whom one can marry, and informal elements that define expectations of marital behavior or roles. That marriage combines these different elements is not unique to Ghana, but in Ghana, marriage is further complicated by the multiple types of formal marriage and the different, sometime conflicting, informal expectations that attend these types. I have organized the three main types discussed in this article into the following table:

Type of marriage	Legitimated by	Some Key Expectations
ChristianFootnote 1	Church‐sanctioned officiant	Between individuals Monogamous Divorce not sanctioned
Customary, Akan (This broad category varies by area, family, and group)	Extended family (parents of bride and groom; mother's brothers; lineage heads)	Between families (family approval and participation in wedding ceremonies is required) Can be polygamous Divorce is acceptable and simple to perform if mediation cannot resolve problems
“Ordinance”: marriage by registration with the state	Government worker (at Registrar) or Government qualified religious officiants	Between individuals Monogamous In the past gave wives different access to inheritance, but the law no longer distinguishes between Ordinance and other types of marriage International recognition (this status allows immigration, travel as spouses) Divorce acceptable, but by a time‐consuming court process

In the “legitimated by” column, I have listed the people or institutions that are authorized to perform a marriage. It is important to note that all types of marriage are currently recognized by the Ghanaian state, which has formally incorporated customary law and thus customary marriage into its statutory law. In the past, there were significant differences in the legal status of customary and Ordinance marriages, but that has changed since the 1980s.Footnote ² Customary marriage has long been the most common type of marriage in Ghana. Now that it is legally recognized by the state, it is also by far the most broadly legitimate type of marriage. Since customary marriage is between families (lineages), and families do not accept or recognize Ordinance or Christian marriages that have not also performed customary rites, nearly every marriage is first a customary one, no matter which other types a couple may also enter into.

It is commonplace to deal with the complexity of Ghanaian marriage by describing it in terms of distinct marital types. For example, legal scholar Johanna Bond distinguishes customary from state marriage, claiming that “a woman's rights and duties in marriage vary significantly depending upon the type of marriage into which she has entered” (Reference BondBond 2008, 402). Similarly, many of the Ghanaians I met made distinctions between customary marriage and Christian or Ordinance marriage. However, since the state now extends the rights associated with ordinance marriage to any marriage that can be proven to be recognized by the participants and their community as a marriage—and since it is common for people married by state or church to also be married customarily—a customary marriage does not limit a wife (or husband) to customary “rights and duties.”

Implicit in the distinction between customary and state marriage as entailing different rights and duties is a progressive interpretation of social change that locates “custom” as “backwards” and bad for women. In order to understand gender relations or how women's struggles are affected by their marriage roles, we cannot use marital types as our primary analytical categories. Doing so can obscure the complexities of multiple‐type marriages. What I will show in the second half of the article is how much this view misses, and how recognizable African women's marital struggles can be to first‐world women when considered in more individual detail. But first I will consider marriage as a global concept, and the ideas that equate custom with the oppression of women as the product of global power relations and inequality.

Marital Expectations and Conceptual Entanglements

During my fieldwork, I interviewed church officials, lawyers, chiefs, and public servants on the subject of marriage and inheritance, as well as conducted focus groups and a semi‐structured survey with Kumasi residents about their own experiences. I also participated in the daily life of a Kumasi neighborhood for eighteen months between 2007 and 2009. The discussions of marriage that emerged over this period tended to fit fairly neatly into progressive and development discourses: women had little power, and customary (polygamous) marriage afforded fewer rights than state/Christian (monogamous) marriage. Customary marriage remained the most common type of marriage, and people explained it most often as being “between families” as opposed to “between individuals.” For example, Father Peter, an Asante Catholic priest, told me:

Most people saw women as doing more work in marriage and having less claim, social or legal, to the security that one expects to gain through that work. As Father Peter suggests, a wife's ability to get support from her husband (whether divorced or not) can be affected by how her husband's lineage members view her. Many Ghanaians, particularly those active in Christian churches, were ready to lay the blame for women's troubles on the “backwards” traditions of custom.

Consider the following statement from Ophelia, an Asante woman and church marriage counselor who described for me how she and other counselors advised couples on the meaning and values of Christian marriage:

Ophelia casts Africans as “mistaken” about the meaning of marriage: it is not for procreation, but for companionship. Her claim was woven into a larger narrative centered around the core claim that in Christian marriage it is the husband and wife who become one. This narrative stands in contrast to that of customary marriage, in which it is lineal families that are married, and each married person remains “one” with their own lineage. In the interview, Ophelia described counselors' instructions to engaged couples with respect to decision‐making, chores, child care, and finances: all of these were to be shared equally between spouses. Couples should live together, separate from their parents and siblings. This stands in contrast to customary marriage, in which finances are kept separate, husbands and wives can live separately, and specific tasks are divided by gender.

Recognizable in Ophelia's positioning of monogamous marriage as a better type of marriage whose purpose is companionship are two significant global discourses. The first is modernity as development, a discourse of progressive change that emerged in the 1950s and is linked to the earlier colonialist belief that all societies progressed through a linear hierarchy of cultural evolution. Anthropologist James Ferguson argues that in the post‐World‐War‐II era, a new turn to “modernity” reimagined this theory of linear evolution in a narrative of “progress” that describes global inequality as the product of historical time rather than of geopolitical relations of power and inequality. Although this account is inaccurate, it has remained a popular interpretation in Africa because it also “encoded a set of political promises” (Reference FergusonFerguson 2006, 185), namely that “developing” nations could achieve the social status and living conditions of “developed” nations by adopting the “progressive” practices of developed nations. Thus, the “progressive” concerns articulated in human‐rights or development discourses are founded on colonial assumptions about the nature of humanity, evolution, and civilization.

The second global discourse implicit in Ophelia's statement is that of companionate marriage, in which values and assumptions associated with marriage in “developed” nations have become part of a global concept of what marriage should entail. Holly Wardlow and Jennifer Hirsch argue that companionate marriage can be understood as an increasingly common global ideology of marriage as it is conceived within “modernity.” In particular, it combines notions of individualism, consumerist practices in self‐fashioning, and “progressive” gender norms (Reference Wardlow, Hirsch, Hirsch and WardlowWardlow and Hirsch 2006, 14). They suggest:

Although companionate marriage ideals are found globally, companionate marriage in practice is not the same everywhere. Wardlow and Hirsch argue that it is not a homogenizing force, rather it is a marriage concept that interacts with different, local marriage concepts such that actual marriages take on a variety of local expressions. The companionate marriage concept is one that has global scope because of its historical spread through Christian missions and contemporary mass media (6). Wardlow and Hirsch argue that the cross‐cultural similarities in companionate marriage occur because Western beliefs that companionate marriage represents “progress” combine with the global relations of power in ways that make participation in Western notions of progress a necessary condition of participation in global relations.

Statements like Ophelia's present a challenge to scholars seeking to center African perspectives in their analysis, because they seem deeply informed by racist colonial accounts of Africans as “backwards” and African cultures as less “advanced.” As Africanist scholars Maurice Vambe and Khatija Khan caution, we should not assume “subaltern perspective[s] are free from … dominant values,” pointing out that this would require that “ordinary people have extraordinary capacity to detect and resist domination at all times” (Reference Vambe and KhanVambe and Khan 2013, 310). Yet it would be insufficient to dismiss Ophelia's views as merely the product of domination. Her perspective is certainly linked to both colonial history and global relations of power, but it is also local, Ghanaian, and Asante. In short, her perspective must be contextualized in terms of the complicated entanglements produced between colonial and local power. Thus, I begin an exploration of Asante marriage by considering the historical emergence of companionate marriage as it was produced in conceptual struggles between the colonial/modern and Asante dual gender systems.

Contemporary Ghanaian gender relations began to take shape during the early colonial period, in the first decades of the twentieth century. Jean Allman and Victoria Tashjian identify this as a particularly understudied era in Asante/Ghanaian history, but one of considerable significance to gender relations (Reference Allman and TashjianAllman and Tashjian 2000). During this period, struggles between Asante and colonial authorities as well as between women and men produced effects that Allman and Tashjian call “gender chaos.” The period began when the British imposed the first Marriage Ordinance in 1884 and a colonial court system. The Ordinance attempted to redefine marriage (in particular, as monogamous), and the court system differentiated between men and women, particularly by favoring men's marital rights and claims to property over women's. Asante women and men resisted the British colonial government's redefinition of marriage, labor, and property, resulting in an era where nothing about the relations between men and women or the meaning of marriage could be assumed. I expand on this below, but first I briefly outline some key aspects of the dual‐sex system and marriage in the precolonial Asante state.

The Asante state was hierarchical and imperialist. It was founded in Kumasi in the seventeenth century, and expanded aggressively, conquering neighboring territories and controlling trade routes to the coast, where Europeans had established trading forts and later the Gold Coast colony. The Asante both sold and kept conquered people as slaves. Among those who resided within the Asante state, only citizens could be fully free and only people born to an Asante mother could gain full citizenship. Thus, the matrilineal family was the basis of state‐recognized personhood and foundational to the structure of society. For those who could never be full citizens, however, there were various ways of gaining partial citizen status. This partial status contributed to ongoing hierarchies, creating different classes of people within the state. Over time, the apparatus and authority of the state became increasingly separated from everyday life (the society), but Asante law and political structure continued to draw conceptually on the matrilineal family and community (Reference McCaskieMcCaskie 2002).

In Europe, the division of public (political) and private (domestic) spheres was strongly gendered, the political domain of men separate from and subsuming the private domain of women, which Nzegwu refers to as a mono‐sex system. But in West Africa, gender systems were “dual‐sex” in that they divided authority and tasks between men and women such that both men and women were required in all domains of society (Reference NzegwuNzegwu 2006, 199–233). Within the Asante system, gender was not as significant in defining access to power as statuses that conferred citizenship or lineal chieftaincy rights. For example, female members of royal lineages held the positions that could make a male lineage member a chief (or remove him from office). As in many West African societies, the roles typically divided between men and women were interdependent and spanned social, political, and religious contexts. This produced what Nzegwu identifies as a key feature of the dual‐sex system: gender complementarity, which divides men's and women's roles such that both are necessary (222). Functionally, this means that men and women must work together toward mutual goals, and gives a lived reality to the concept that both men and women are necessary to the success of a society. In the Asante state, as in much of West Africa, both public and private arenas of community‐building and daily sustenance required the participation of men and women. Thus, women's contributions were necessary in all social contexts.

Asante marriage between free citizens organized the obligations of husbands and wives and their extended families into labor and exchange relations. Wives were obliged to provide sexual exclusivity, work for the farm, cook food, and share the products of their labor. Husbands in turn were obliged to clear fields, provide sexual satisfaction (which included having children), provide physical and spiritual protection, educate their children, and provide wives with the fresh produce used to cook meals (Reference Allman and TashjianAllman and Tashijan 2000). Both men and women were embedded in their lineage, their local community, and their spouse's and father's lineages through multiple ties of mutual aid and moral responsibility that produced interdependence. Men's brothers were expected to continue meeting a man's obligations to his wife or wives and children should the man die and the brother succeed him, and siblings of both parents had ongoing obligations to the children. The frequently repeated statement that “marriage is between families” describes two lineages brought together into a mutually beneficial relationship mediated by and focused on, but not limited to, the married couple.

Not all marriages were free marriages. Slaves and pawns could be married to free people, placing them in a subordinate position to the free spouse, and a free person could become a pawn to their spouse during the marriage. Pawning was a temporary status, in which the pawn's labor acted as interest and surety for a loan. A pawn often lived with and worked for the lender until the debt was paid back by their lineage; the lineage could choose to offer any member as a pawn, and in the pre‐colonial era, young, unmarried pawns—male or female—were common. However, a wife who already lived with a lender was an obvious choice for a pawn; once she was pawned, the benefits of her labor accrued to her husband, and he could make demands on her that were previously the right of her lineage.Footnote ³

The distinctions between these different types of customary marriage became blurred during the colonial era, which has had lasting effects on the meaning and practice of “customary” marriage. The colonial government outlawed pawning during the early twentieth century, along with slavery. This created a radical change in the availability of free labor (Reference GrierGrier 1981), and had a significant impact on gender and labor relations (Reference MikellMikell 1989; Reference Allman and TashjianAllman and Tashjian 2000). But because female‐pawn marriages reflected European expectations about marriage—with spousal co‐residence and the wife's work unremunerated and owed to the husband—the practice of pawning wives was relatively invisible to the colonial government. Thus, the Asante continued the practice of pawning wives, even as it became difficult to pawn single men and women; the practice was “feminized” and blended into customary marriage (Reference Allman and TashjianAllman and Tashjian 2000, 70–73).

By the 1940s, Asante researcher T. E. Kyei had described a set of ritual statements in which the groom assumes responsibility for the bride, making him liable for her debts, as “the time honoured Asante marital pledge” (Reference KyeiKyei 1992, 31) and the foundation of the husband's authority in marriage:

This description of customary marriage is indicative of the significant changes to marriage that occurred during the period of colonial rule. Despite the claim that it is “time honoured,” the incorporation of debt obligation into the marriage vow suggests that this custom traces back to pawn marriage rather than free marriage. This change was not simply a matter of colonial power imposing a new system of marriage over Ghanaians; it is instead an example of how local men used the power afforded to them by the colonial/modern gender system to shift the relations between Asante men and women and produce a new form of Asante male authority.

The entanglement of British and Asante gender relations and marriage concepts did not happen immediately, but in varied struggles that were not limited to the political sphere. In particular, the changing focus on agriculture, new markets, and the imposition of wage labor had significant cumulative effects on gender relations. Between 1905 and 1930, cocoa became the dominant export from the Gold Coast, which effected sweeping changes to Asante social structure (Reference MikellMikell 1989). Men and women migrated to build cocoa farms, with the effect of separating women from their lineages (with whom they had the right to co‐own property) and creating long‐term investments in their husbands' cash‐crop farms. Women did not have the right to co‐own property with their husbands, and there was (and remains) no legal assumption of marital co‐ownership; all marital property was, and is, assumed to belong to the husband unless a wife has documentation to show otherwise (Reference ArchampongArchampong 2006).

People's access to land through marital and kinship ties and the demands they could make on conjugal labor shifted considerably during this period, particularly for wives and husbands who migrated in order to develop cocoa farms away from the places in which each had family land (Reference OkaliOkali 1983; Reference Allman and TashjianAllman and Tashjian 2000, 60–70). When agriculture was focused on subsistence crops, women's investments of labor were seasonally returned to them in the foodstuffs their husbands provided. Their ongoing relations (and often ongoing residence) with their matrilineal kin and their limited investments in their husband's property had contributed to an interdependent rather than dependent status between wives and husbands. But in this new era of cash‐cropping, migration, and changing marital expectation and law, the interdependence of male and female roles in marriage was undermined, and women's position became significantly more precarious.

Increasingly, the obligations of marriage were mediated by money: women labored on their husbands' cash crops, and men provided “chop money” to buy food rather than relying on fresh farm produce. Cocoa farms take years to mature, thus women's marital labor became disassociated from their husbands' obligation to support them. This left wives vulnerable to major losses should their husbands die or divorce them before the farm matured, or if a husband simply denied a wife a share of the farm's profits.

The economic shifts that favored men were compounded by the politics of the colonial government and its courts. The colonial government had set out to change Ghanaian marriage, which they saw as immoral and detrimental to women. Although the main goal was to eliminate polygamy, which they viewed as tantamount to slavery and an impediment to economic development (Reference MikellMikell 1989, 69), the more immediate effect of colonial and Christian efforts to reform marriage was a period of “gender chaos.” Gender chaos was evidenced by “some women's refusal to marry, by chiefs in several towns ordering the detention of unmarried women, by contentious child‐custody cases, and by the dramatic upsurge in the number of divorce, adultery, and inheritance cases—the latter often pitting wives against deceased husbands' matrikinFootnote ⁴ in struggles over ownership of cocoa farms” (Reference Allman and TashjianAllman and Tashjian 2000, xxvi). Colonial courts and Native TribunalsFootnote ⁵ responded to the emerging crisis in gender relations in uneven and ambiguous ways, but both favored men, particularly men of status (Reference BerryBerry 2000). Colonial courts refused to hear women's cases on property rights, instead referring these to Native Tribunals, which were run by male chiefs (Reference MikellMikell 1989, 116). Allman and Tashjian argue women used adultery claims—which had historically been made between men—to frame claims to their economic rights, for example, by asserting that a husband who failed to meet his customary obligations of support was not entitled to a wife's sexual exclusivity. These claims were not very successful (Reference Allman and TashjianAllman and Tashjian 2000, 177). In the shifting balance of power between women and men through the colonial period, women lost much of their ability to claim reciprocal rights in their marriages, while also failing to gain court‐supported independent control over property.

The most common type of marriage during the era was contracted customarily, but the blurring of the line between pawn and free marriage had changed what “customary” marriage entailed. With sources of free labor restricted by the colonial government and the migration pressures that attended the development of cocoa farms, taking on the wife's debts gave her husband a claim to her labor and gave him authority over decisions such as where they lived. But the inequality of Asante customary marriage is not a feature of matrilineal marriage or “custom” per se; rather, it is a consequence of the entanglements of Asante and British marriage laws and civil practices. Men's ability to demand labor, sexual exclusivity, and ownership of all the property produced during the marriage was transformed by the colonial imposition of gender as a crucial category of difference and inequality. This was accomplished in practice through court, government, and interpersonal processes that recognized the legitimacy of men's traditional claims on women while refusing to recognize the legitimacy of women's claims on men.

Marriage in Context

The struggles between the colonial/modern and dual gender systems have produced a complicated situation where there exist multiple types of marriage that can be contracted simultaneously, with attendant and sometimes contradictory marital expectations. Ophelia, the marriage counselor who told me that Africans are “mistaken” about the purpose of marriage, had described the expectations of Christian marriage as being opposed to those of customary marriage: it is for companionship; husbands and wives become one and merge their responsibilities, lives, and finances. This is opposed to a customary marriage in which spouses remain emotionally, and often physically and financially, closest with their lineage members.

The kind of opposition that Ophelia describes reflects the mythology of the colonial/modern gender system, and the notion of “progress” from an African past to a global future. It is important to question the validity of this opposition and what it accomplishes, because the assumption that “past” custom and “modern” law are significantly opposed also influences scholars and policy makers who help define that global future in terms of “universal” human rights. Thus far I have developed an argument that the opposition is false, that it obscures the ways in which contemporary Ghanaian gender relations and oppressions have developed through historical processes. In this section, I will develop the local context in which these global discourses resonate and have meaning by using ethnographic descriptions of three women's marital experiences. In these three different marriages, women's overlapping experiences of gender relations and social limitations give a sense of what is missed when simply opposing an imagined “customary” marriage to an imagined “companionate” one.

The marriages I observed did not exhibit the sort of clear distinctions that were implied by Ophelia, nor those implied by legal scholars like Johanna Bond, who supposes that ordinance marriage affords women significantly more rights than customary marriage (Reference BondBond 2008). Despite having different types of marriage, the women highlighted in this section have experienced similar difficulties, and are faced with options that are limited in similar ways by the larger social context in which their marriages take place.

Serwaa,Footnote ⁶ a professional woman living in Kumasi who by all outward appearances was doing quite well, had a “three times” marriage to David: it was a customary, Christian, and ordinance marriage. David was also a professional, and their children were all university educated. They lived in a large (although incomplete), single‐family home in suburban Kumasi. As an educated, middle‐income, nuclear family, they appeared to fit easily into a narrative of modernity and “progressive” companionate marriage, which equates wealth, education, and success with type of marriage.

However, in our conversations, Serwaa complained regularly about her husband's emotional and financial neglect, and as I got to know the family, both she and her daughter revealed a history of abuse. Serwaa often threatened to leave David, but she never did. Serwaa's daughter told me that it was because the church would not let her mother preach in the villages if she were divorced, suggesting Christian rules of marriage kept Serwaa from leaving a harmful situation. Indeed, all of the ministers I interviewed across multiple denominations said they would never condone divorce, only temporary separation, even in situations of violence.

But Serwaa herself rarely gave the church as the reason she stayed with her husband. Sometimes she said it was because David would not care for her younger children, and she feared for their spiritual safety if she left. Her explanations to me changed over time, but this one resonated with the statements made to me by an Asante priest, Father Osei, who had told me that if a woman divorces, there is no recourse for her to hold the man responsible for his children. Father Osei blamed matrilineal kinship, because the man knows that the children are “for the woman” and so refuses to care for them, an explanation that locates the problem in custom and positions the Church's rejection of divorce as a solution. But although difficulties with post‐divorce child support may be expressed (or experienced) in matrilineally specific ways, they are hardly unique to matrilineal societies.

Serwaa sometimes said she stayed with David because of their jointly owned house; she was afraid she would lose ownership if she left him. This explanation resonates with common assumptions that African women lack access to property, assumptions that, in Ghana, have led to policy initiatives to increase women's rights to marital property. Yet later she explained that she had bought the property herself and had been the sole owner. Only recently had she changed the papers to include David's name in an attempt to encourage him to help her finish building it. Later I asked Serwaa directly about some details of her marriage, and learned that although she had told me that David contributed nothing to the development of her property and the daily expenses of the house, in fact he frequently did contribute a small amount—less than she wanted and usually after she had already paid for something in full—but nevertheless, enough that divorcing him would probably constitute a significant decrease in her financial security.

Serwaa's response to her abusive marriage was difficult to understand by itself. But as I looked at other women's experiences, some patterns emerged, particularly that women are disadvantaged in all formal (and many informal) channels of conflict‐resolution: customary courts and Christian ministers treat women and children similarly, making children apologize to parents and wives to husbands during conflict resolution, and preferring subtle, indirect cues to any direct acknowledgment of an elder's or husband's wrongdoing (Reference NaveNave 2016). State divorce is a long and expensive process, and support for women with custody or maintenance claims is similarly limited by a woman's resources and her ability to spend a long time waiting for an outcome.

The difficulties facing women who divorce are known to women like Serwaa, who are considering divorce in part through their knowledge of women who have been through divorce. This knowledge affects their understanding of what is possible and what is likely in their marital conflicts and resolutions. Such knowledge can come from women like Nana Adwoa, who worked as a petty trader and had six grown children and many grandchildren. She had been customarily married as a young woman, but in the 1970s, while her children were still young, her husband went to the head of her family (abusua panin) and asked for a divorce. Nana told me that the ease of divorce is different for men and women:

Nana Adwoa did not know her husband was going to divorce her, and once she found out, she felt she had no control over the outcome. The decision was made between men: her husband and her abusua panin. There was no recourse for her to hold her ex‐husband responsible for his six children: customary authorities were unable to help her, and when she went to seek help from the social welfare office, she found a man from her hometown who told her where to bring her case, but

The difficulty of resolving family issues through state court processes was a common theme during my research in 2007–2009. Court cases often take years and are expensive, and I found that people did not trust that a finding in their favor could be enforced. Opinions like Serwaa's and Father Osei's—that fathers will not care for the children of divorced wives—are informed by observations of women's difficulties in winning child support from state or customary courts, and illustrate why a woman might prefer to solve a marital problem rather than end a marriage. For women invested in continuing their marriages, the Christian prohibition on divorce may offer a certain amount of leverage.

It is commonplace to link women's subjugation to particular marital practices—claims that polygamy or child marriage or widowhood rites are damaging to women abound—but these difficulties are better understood as a result of the contingency of their status as wives; the inequality of the colonial/modern gender system permeates all aspects of social life, and though the particular options available in a given type of marriage may vary, women are disadvantaged in all of them. This was highlighted in a brief interview with Wanda, who experienced these problems despite having no formal marriage type. Wanda was an impoverished street trader who lived with the father of her children and experienced significant domestic violence. Her father wanted her to leave the man because he had not performed the customary marriage rites (which is to say, they were not formally married). But a combination of poverty, social expectations, and the small security she gained through living with his relatives kept her with him:

I had known Wanda for most of my stay in Kumasi. This interview took place about midway through. In the remaining nine months, she did not go to live with the wealthy friend nor find a place of her own. Wanda's difficulties were not caused by having a customary marriage: since they had not completed the rites, she did not have one. But similarly, Serwaa's difficulties navigating an abusive and economically dependent marriage were not made less by her state‐registered marriage and maximized marital rights. Nor can Wanda's and Serwaa's decisions to stay be easily reduced to marital or customary expectations. Although Serwaa was a devout Christian, the church's prohibition on divorce did not factor into her own explanations of why she stayed, and her actions were focused on drawing her husband into a greater emotional and financial commitment to their shared lives. Wanda's reasons to stay were likewise multifaceted, although more recognizably driven by her lack of alternatives.

If we take the view that marriage types, or particular marital practices, are directly correlated to particular gender relations or to the relative equality of women, we can miss important local realities. Here, Serwaa, Nana, and Wanda each had very different types of marriages, with different attendant rights. However, it is clear that their experiences do not diverge as strongly as these differences might suggest. The possibilities that shape the lives of women and men, wives and husbands, are not dependent on particular marriage rules, but rather on a whole, interacting system of marriage rules and expectations, cultural values, and gender relations. The complexity of their lives and the commonality of their experiences is obscured when we assume a meaning of marriage that is independent of its local history or lived experience.

Contextualizing Gender

Marriage is a relationship in which large‐scale social relations are enacted at an interpersonal level. Those that can be described in “companionate” terms are often seen as better for women: monogamous relationships between individuals, entered into by personal choice. Implicit in companionate marriage, though, is a narrative of “progress” that assumes that the global evidence of gendered dominance is the product of natural differences between women and men, and that customs that produce unequal gender relations in one place must also therefore produce them in another. When these assumptions are unexamined in first‐world feminism, they can distort the realities of third‐world women, who may find themselves caught between different systems of gendered domination.

Centering third‐world perspectives requires more than including third‐world narratives. It requires contextualizing these narratives in analysis that challenges and exposes the assumptions of “natural” global systems of domination. Lugones's model of the colonial/modern gender system provides one such analytical tool by giving a historicized account of global gender dominance that rejects the “natural” differences of gender and race. This enables us to question how various systems of oppression interact. In the case of Asante marriage, I have shown that customary and Christian/Ordinance marriage do not simply represent opposing “cultural” or “modern” beliefs; rather, these marriage types have been produced through the entanglements of dual and colonial/modern gender systems in ways that have allowed forms of Asante dominance to become gendered and incorporated into everyday marital expectations. Rather than treating specific marriage customs as linked in an evolutionary progression from oppressive to equal, we can instead consider how these customs interact with both local and global concepts, ideologies, and power relations in a given context.

In embracing the challenges that third‐world scholars have posed to first‐world feminist theory, we have an opportunity to work toward a different center: one that assumes third‐world perspectives are valuable not only for the ways they can illuminate third‐world experiences, but also because they have novel and important insights into the problems created for all women by the colonial/modern gender system.