Some Methodological Notes

This article presents a history of marriage (and divorce) across some twelve centuries. This long time frame is in part a function of the methodologies I draw on to reconstruct the history of a region for which the earliest written sources date to the 1860s and even then are highly limited in nature. These methodologies tend to require a very long-term perspective in order to see changes unfolding. Alongside archival sources—especially, in this case, of the Mill Hill Fathers—this article draws on historical linguistic reconstructions, comparative ethnography, and oral traditions. This interdisciplinary approach is by now a well-established approach in the history of places and times beyond the reach of writing (e.g., Ehret Reference Ehret1998; Fields-Black Reference Fields-Black2008; Klieman Reference Klieman2003; Vansina Reference Vansina2004).

This approach is based on the fact that language is not purely functional. We use words to name and describe our physical environment, but we also use them to talk about our intellectual, aesthetic, and social worlds, just as people did in the past. Reconstructing relevant vocabularies for the now extinct ancestral languages of modern tongues gives us insights into those wider fields of historical experience. The use of historical linguistics to write history relies on the premise that related languages are descended from a “proto-language” and inherited a portion of their vocabulary (along with other parts of speech) from that proto-language. It is possible to reconstruct some of that early vocabulary by knowing the specific relationships between modern languages and the changes they have undergone since diverging from their proto-language. The etymology of an individual vocabulary item offers us an intellectual history for that word. The borrowing of a word from neighbors, the coining of a new word, or the narrowing or expansion in the meaning of a word are all the result of actions by individual members of a speaker community grounded in social, cultural, economic, and political changes.

These reconstructions give us a historical framework. In order to construct a more complete picture of the past, of the contexts in which this vocabulary was spoken, I turn to comparative ethnography and oral traditions. While ethnographic data needs to be used in conjunction with independent streams of evidence (such as from historical linguistics or archaeology) to avoid anachronistic interpretations, it offers important evidence for fleshing out those contexts. The material available for the North Nyanza societies ranges from descriptions by European travelers, missionaries, and colonial officials to texts produced by elite Ugandans, such as Apolo Kagwa, in the early twentieth century, to studies by anthropologists and interviews I conducted during fieldwork. In all cases it is essential to pay close attention to the who, where, and when of their production. Oral traditions are also complicated sources of evidence, shaped as they are by the moments and authors of their transcription. They cannot be taken as literal descriptions of past events, but they provide a wealth of detail about social and political structures. This article draws on both published and unpublished sources, including the “Historical Texts” collected in the 1960s and early 1970s by David William Cohen and Ronald Atkinson in Busoga and Bugwere, respectively.

Beyond missionary records and the oral traditions of states, which remember royal marriages, it is generally not possible to uncover individual marriages in the extant historical records (for an important exception in this region, see Cohen Reference Cohen1977). But because it is a social institution that unites the economic and the cultural, the political and the personal, the legal and the emotional, marriage is a topic ideally suited for this interdisciplinary approach, touching as it does on so many aspects of a community’s life and thus being reflected in the different sources that are available. The history that emerges from these sources is one that emphasizes social ideals more than the reality of daily life. Conflict and contestation are apparent, however, in the creation of new forms of marriage and in new words for divorce and elopement.

People speaking the modern languages descended from North Nyanza inhabit an arc that stretches northward from the mouth of the Kagera River on the Victoria Nyanza and eastward across the Mpologoma River and into the foothills of Mount Masaaba (see map below). These languages are Luganda, Lusoga, Lugwere, and Rushana. Their proto-language, North Nyanza, was descended from Great Lakes Bantu (see figure below), spoken by people who lived from around 500 BCE on what is now the boundary between the modern countries of Uganda and the Democratic Republic of Congo. The dialects they spoke gradually diverged until they became distinct languages. Dialects of one of those languages, West Nyanza, spoken by those living west of the Victoria Nyanza, in turn became the languages Rutara and North Nyanza by around the eighth century CE. By around the thirteenth century North Nyanza gave way to Luganda and South Kyoga, and by around the sixteenth century South Kyoga diverged into Lusoga and East Kyoga (or pre-Lugwere). The modern languages of Lugwere and Rushana emerged in the early nineteenth century when a group of people who came to be speakers of Rushana moved some distance away into the foothills of Mount Masaaba in the wake of famine and conflict (see Stephens Reference Stephens2013).

Historical Geography of North Nyanza Languages

Subclassification of North Nyanza Languages

Note: Modern languages are underlined.

Innovative Relationships: North Nyanza in the Eighth to Thirteenth Centuries

People who lived in communities speaking North Nyanza between the eighth and thirteenth centuries practiced a mixed economy, combining crop cultivation and livestock raising with hunting, fishing, and gathering of other wild foods. But they increasingly focused on bananas as a staple crop, a development that led to a new importance for land best suited for the crop (see Stephens Reference Stephens2013). Concern within clans to control that land placed a premium on patrilineal descent (see Stephens Reference Stephens2009). Yet the importance of marriage in determining the lineage identity of children does not appear to have held quite the same importance in these communities as it did for their ancestors speaking West Nyanza (Schoenbrun Reference Schoenbrun1998). The language they used to talk about marriage suggests changes in the roles of women and men (or the perception thereof) in the process of marriage and the relationships it created. North Nyanza-speakers dropped a verb that West Nyanzans had coined for marriage and replaced it with two new and distinct gender-specific ones. The first of these referred to a woman’s act of marrying, *-bayira.^{Footnote 4} The development of an active verb that applied exclusively to women reflected a social change whereby women had more active roles in creating alliances between lineages through their marriages and the children that were expected to ensue. North Nyanza speakers gave women’s relatives a substantial stake in their children. A woman’s children performed essential roles for her patrilineage, especially at moments of transition or ritual danger such as preparing a corpse for burial.^{Footnote 5} The second verb speakers of North Nyanza began using to talk about marriage referred to a man’s act of marrying, *-gasa. This was not a new word for this community as they inherited it from West Nyanza, where it had meant “mate” or “copulate.” Whereas *-bayira signified the roles women played in contracting their marriages, this innovation suggests a diminished social role for men in marriage. This interpretation is strengthened when we consider that North Nyanza speakers abandoned the West Nyanza verb meaning “marry” with reference to men: *-túèr(er)- emphasized the role of men in marrying women (see Schoenbrun Reference Schoenbrun1997), in contrast to the more direct meanings of *-gasa. North Nyanzans thus stressed the sexual role of men in marriage, alongside or over their social and political roles.

The historical linguistic evidence also indicates that marriage could be a space for contesting parental and lineage authority. North Nyanzans retained a pair of words from West Nyanza that described women’s and men’s separate, but related, acts of eloping: *-hambuka for a woman and *-hambula for a man. These are both derived from *-hamba, originally a proto-Bantu verb meaning “seize.”^{Footnote 6} Strikingly, *-hambuka (the female form) is an active intransitive verb, yielding a meaning of “seize oneself,” in contrast to *-hambula, which is a transitive verb, suggesting the seizing of another (see Stephens Reference Stephens2013). Centuries later, missionaries regularly described elopement in terms of men abducting unwilling women or girls in order to force a marriage (see, e.g., Le Veux Reference Le Veux1917).^{Footnote 7} This was at times, no doubt, the case, but this older vocabulary indicates a messier—and so more likely—picture: one in which women and girls played active roles in thwarting the ambitions of their elders—by “seizing themselves” into elopements—when it came to arranging marriages and the important alliances that could result from them.

Not all relationships endured, regardless of how they had been contracted. Women could leave their husbands in an act described as *-nóbá. North Nyanzans derived this from an older West Nyanzan verb meaning “hate.” In so doing, they highlighted affective aspects of marital breakdown, rather than practical and political questions around children and bridewealth repayments. (It is worth noting, as an aside, that their vocabulary for the establishment of marriages and for being married did not reference love or affection.) As *-nóbá referred only to women’s acts of leaving their husbands, we can assume North Nyanza speakers also had a word for men leaving their wives, although we do not know what it was.

The innovation in conceptualizations of the meaning and function of marriage affected more than the creation or dissolution of such relationships. To describe the state of being married, but with reference only to women, speakers of North Nyanza took the proto-Bantu verb “eat” and derived a new noun, *-rya (cl. 5/6).^{Footnote 8} Through the language they used to talk about being married, North Nyanzans presented it as a life stage in which a woman controlled food production and preparation in the household. In so doing, she also controlled the distribution of food and so had greater chance of ensuring that she had enough to eat. This would have contrasted with being a daughter in her parents’ household, where her control over and access to food would have been more limited. This organization of food production at the household level was not unique to people speaking North Nyanza; what is very unusual is that well over a millennium ago they used language to describe marriages that emphasized women’s agency.

After North Nyanza: Continuity and Integration

The dialects of the North Nyanza speaker community gradually diverged until, by around the thirteenth century, they formed two distinct languages, early Luganda and South Kyoga, spoken on either side of the River Nile as it leaves the Victoria Nyanza. For those who spoke South Kyoga, between the thirteenth and sixteenth centuries, this was largely a period of consolidation and adaptation to the particular environment between the vast lake in the south and Lake Kyoga to the north. They experimented with crops, adjusting for different microclimates, and with livestock and fishing (see Stephens Reference Stephens2013). But while they changed their language to reflect some important shifts in their social and familial lives, marriage does not appear to have been a focus for these changes. Instead, they continued to use the vocabulary developed by speakers of North Nyanza, reflecting the continued importance of bridewealth in establishing marriages as well as the active participation of women in their marriages and in the households created by those marriages.

Indeed, in this eastern region, marriage increasingly served not only as a means of reproducing the community and strengthening ties within it, but also of managing relationships with people speaking different languages and with different economic and cultural norms. After Lusoga and East Kyoga (or pre-Lugwere) had emerged from South Kyoga, by around the sixteenth century, the speakers of these languages found themselves living in lands that attracted migrants from across the region. Intermarriage was an important feature of life here. We see the consequences of these developments in changes to vocabulary intimately connected to the household and the production and consumption of food within it. For example, speakers of East Kyoga dropped the older noun they used to name a granary in favor of the Luo word *-deero (cl. 7/8). As Luo-speaking women married into East Kyoga households, they brought with them their knowledge of grain storage and introduced new words into the language through their children. Marriage across linguistic and cultural boundaries remained an important feature of life in the North Nyanza region, including Buganda (Doyle Reference Doyle2013), through the twentieth century, despite some assertions to the contrary (see, e.g., Kalibala Reference Kalibala1946).

Having set out the history of the ancestral speaker communities, I will explore, in the next four sections, different developments in marriage in each of the modern speaker communities descended from North Nyanza. I start with the newest, the Rushana speaker community, and the dramatic changes in the preparation of young people for marriage there. I then move on to Lugwere speakers and the importance of marriage for resolving tensions between social groups. I focus primarily on divorce in the discussion of marriage among Lusoga speakers and address the plurality of marriage forms in the section on Buganda. Together these offer clear evidence of the wide range of relationships subsumed under the label “marriage” and some of the ways in which women and men developed new notions of what marriage was, how best to contract such relationships, and how to exercise control over ending them.

Marrying among Strangers: Bushana

In the early years of the nineteenth century, in the wake of conflict and drought, some of the people who spoke East Kyoga were forced from their homes by Yua, ruler of the Balalaka polity (Atkinson n.d.:Text 44; interview with Yokulamu Mutemere, Butebo, December 2, 2004). One group among them moved some fifty miles away into the foothills of Mount Masaaba. Once there they had to contend with new political and social realities, living as they now did among both Kupsabiny and Lumasaaba speakers. But they retained a distinct identity—becoming Bashana—as the language they spoke diverged from that of their East Kyoga ancestors.

As a small community that practiced exogamous marriage, Bashana faced a particular challenge in contracting marriages with their neighbors. Among the Bamasaaba, male circumcision (imbalu) was a fundamental prerequisite to attaining male adulthood and being able to marry (see Khanakwa Reference Khanakwa2011, Reference Khanakwa2014). If Bashana wanted their sons to marry Bamasaaba women then they needed to circumcise them, something they had not done historically. Whether it represents a sociocultural continuation for the Bamasaaba from their Great Lakes Bantu linguistic ancestors (Schoenbrun Reference Schoenbrun1998; see also Ehret Reference Ehret1998) or more recent innovation, imbalu was a central cultural practice for them by the nineteenth century at the latest (Khanakwa Reference Khanakwa2014). By the time the Bashana settled among them, Bamasaaba men had to be circumcised before they married. The Sabiny, by contrast, used male circumcision and female excision to mark the transition to adulthood and both operations were a prerequisite for marriage (Goldschmidt Reference Goldschmidt1986). The Bashana, thus, also adopted female excision in order to facilitate marriages between their daughters and Sabiny men (Edgerton Reference Edgerton1971).^{Footnote 9} The Bashana offer us an example of a community that radically reshaped the way it prepared its young people for marriage—a community that reconceptualized what it meant to be married—as a way of adjusting to the reality of their new social environment.

In addition to new bodily practices, Bashana developed new ways to talk about their marriages during the past two hundred years. While the adoption of circumcision and excision was the product of contact with Kupsabiny- and Lumasaaba-speaking communities, the new vocabulary for marriage was not. The Bashana created two distinct but related verbs: -bimbirra (“marry,” for a man) and -bimbirrikana (“marry,” for a woman). They derived these from a verb that can be traced back to proto-Bantu with the meaning “swell.” In a small number of Bantu languages the nouns “fetus” and “pregnancy” have been derived from this root, and it is a fairly straightforward semantic connection.^{Footnote 10} The semantic leap from swelling to pregnancy to marriage suggested by the Rushana verbs is plausible, although not common. In innovating separate verbs for women’s and men’s acts of marrying, Rushana speakers continued the North Nyanzan tradition of distinguishing between the marriages of their daughters and of their sons.

Marriage in a Diverse Landscape: Bugwere

From the late eighteenth through the nineteenth centuries, Bagwere women and men consolidated political power into three polities. At the same time, they experienced significant linguistic and cultural diversity as different pressures pushed and pulled people from neighboring areas into the lands they considered their own. Oral traditions point to cross-cultural and cross-linguistic marriages: marriages that served to ease tensions between groups, facilitated the incorporation of foreign men (in particular), and created political and social alliances.^{Footnote 11} Perhaps because of this important role for marriage, Lugwere speakers changed the way they talked about it, expanding their vocabulary to include the rather informal phrases ókusuná musaizá and ókusuná mukalí, for “to marry” (for a woman) and “to marry” (for a man), meaning literally “to get a man” and “to get a woman,” respectively. In so doing, they did not distinguish linguistically between women’s and men’s acts of marrying. While marriage processes may have been relatively straightforward in Bugwere, as suggested by “getting a man” and “getting a woman,” they still ideally involved the participation of both lineages and the creation of a contract through the transfer of bridewealth (ókukwa) (interview with Joyce Kawuledde, Bulangira, October 27, 2004; Logose n.d.).

Not all marriages were socially sanctioned, however, and elopement was a common enough occurrence that Lugwere speakers coined a new verbal pair to describe it, again with an active form to describe women’s acts of eloping. The root word was still one of violence: ókubanda, meaning to “force one’s way through” or “split” (Nzogi & Diprose Reference Nzogi and Diprose2011:69).^{Footnote 12} The verbal pair, ókubandúká and ókubandúlá (referring to a woman or a man eloping, respectively) speak to both a possible increase in elopements during these centuries—a time of political centralization and consequently, in part, of conflict and famine—and to important continuities between Lugwere speakers and their North Nyanza-speaking ancestors in the perception of elopement.^{Footnote 13} On the one hand, elopement was a socially violent act because of the absence of a marriage contract in the form of bridewealth and the missed opportunity to create alliances and connections that were crucial to surviving uncertain times. On the other hand, women were perceived as active participants in acts of elopement (see, e.g., interview with Joyce Kawuledde, Bulangira, October 27, 2004). The same pressures of scarcity and political instability that would have led families to desire alliances with other lineages and even other speaker communities could lead young women to seek marriages outside of their families’ approval as they sought to secure their own economic present and future.

For some who followed the socially prescribed path to married life, there was the possibility of being the senior wife, kaidu, in a polygynous household. In 1913, on a questionnaire about marriage customs, the Mill Hill missionary Father Kirk was emphatic in his response to the question whether wives had equal status. “No!” he wrote, “All the women have not precisely the same status; there is such a thing as a principal wife” (Letter from Fr. Kirk to the Bishop, December 22, 1913). He did not, however, elaborate on what the difference in status meant. The position of kaidu in a royal household was less important in Bugwere than in the Soga kingdoms. Nonetheless, she controlled ritual drums and could wield significant influence over her husband.^{Footnote 14} In ordinary households, the kaidu was responsible for performing funeral rituals when her husband died (interview with Joyce Kawuledde, Bulangira, October 27, 2004). The position of kaidu was not new in Bugwere—either in commoner or royal households—but something Bagwere inherited from their North Nyanzan linguistic ancestors.

The arrival of European missionaries in the opening years of the twentieth century contributed to the range of options for marriage by introducing a radically different form: a church wedding that resulted in a supposedly indissoluble relationship. Colloquially described as a “ring marriage”—ow’ómpéta—this foreign form of marriage came with a foreign name. Missionaries in Uganda used the Luganda translation of the Bible in many non-Luganda-speaking areas, and so in Bugwere church marriage had a Luganda name: ókufumbílwá. Catholic and Anglican missionaries alike sought to exert control over marriage among their converts, albeit with varying degrees of success. In order to exert that control, they needed to make sense of what kinds of relationships Bagwere were entering into in the early twentieth century, and how. The Mill Hill missionaries’ questionnaire on marriage customs demonstrates the diversity of forms of marriage in Bugwere and the potential for dissolving a marriage that was unsatisfactory to either partner. Although the questionnaire is from 1913, the forms described therein fit into older regional practices that are reflected in the linguistic evidence, which highlights continuities and changes from North Nyanzan times.

An important function of marriage for Bagwere women and men was creating alliances across patrilineal divides. One way for parents to try to ensure their daughters’ marriages would create useful connections with another family and lineage was to betroth them “while they are still at the breast” (Letter from Fr. Kirk to the Bishop, December 22, 1913). By promising their child in this way, families would have sought to reduce the risk posed by the possibility that she might elope (ókubandúká) or contract an undesirable marriage. Still, such a betrothal did not form an indissoluble contract, according to Father Kirk’s informants, although, “no other man has much of a chance of getting her” (Letter from Fr. Kirk to the Bishop, December 22, 1913). This practice was known as ókukwa, emphasizing the payment of bridewealth on the part of the husband’s family.

Marriage marked by the transfer of bridewealth was known as eirya (interview with Joyce Kawuledde, Bulangira, October 27, 2004). Father Kirk’s interlocutors insisted that the “dowry is always paid in one way or other.” Yet they also acknowledged that when the husband and his family failed to complete the transfer of bridewealth, the woman’s family would fetch her home or “wait till she comes home on a visit; and then she is prevented from going back to the man” (Letter from Fr. Kirk to the Bishop, December 22, 1913). This contradiction gives us a glimpse into the space between social ideals and the reality of daily life. Many factors could intervene to prevent the full payment of bridewealth, including, presumably, poverty, disease, and the reduced leverage of the woman’s family once she lived with her husband. If a marriage failed, the social expectation was that the bridewealth would be returned. A woman unhappy in her marriage could repeatedly leave her marital home, until her husband grew “tired of following her and then goes after her no more.” At that point he would seek the return of the bridewealth given to her family. “If she refuses to return to him he almost always gets back the dowry” (Letter from Fr. Kirk to the Bishop, December 22, 1913).

The status of the woman was the determining factor in the kind of marriage contracted and her security in her marital home. When a man married an “inherited woman”—for example, his brother’s widow—there were no marriage ceremonies, but she could not easily be “sold” to another man. By contrast, when a man married a “slave woman,” “some marriage ceremonies” were performed, but such a wife was vulnerable in times of hardship. Indeed, Father Kirk, noted that she “is readily sold” (Letter from Fr. Kirk to the Bishop, December 22, 1913). Despite the focus of Mill Hill missionaries on ceremonies and transfers of bridewealth as markers of marriage, therefore, the performance of wedding rituals was less a guarantee of protection than was being enmeshed in social and kinship relations, something denied to enslaved women.

The example of Bugwere is instructive in showing how, even in a place of relatively limited economic inequality and small-scale polities, there were different marriage possibilities. This underscores that what we call marriage was a conglomeration of different kinds of often messy relationships between women and men. And in that messiness we can find both historical change and the ability of ordinary women and men to shape their lives, albeit with the constraints of social expectation and economic and political pressures.

Status and Separation: Busoga

The significance of marriage as a means of creating political alliances and of enabling outsiders to establish themselves in Busoga has long been recognized in scholarship (see, e.g., Cohen Reference Cohen1977). One reason marriages worked to created alliances here, from the sixteenth century, was the understanding that even if the marriage failed, the relationships formed through the transfer of bridewealth and the birth of children survived. A Lusoga proverb captures the enduring nature of these relationships: Obúko byáike: tíbiwá mu kibyá, “the relationship with in-laws is like the smell of fish: it lingers in the dish.” The compilers of Ensambo edh’Abasoga (Proverbs of the Basoga) offer the following explanation: “If you get married, you have a permanent relationship with your in-laws” (Cultural Research Centre 2000b:73). Given the apparently high levels of marital breakdown in Busoga, it may have been particularly important that the alliances formed by marriages could outlive the relationship that created them.

Lloyd Fallers (Reference Fallers1957) estimated that between a quarter and half of all marriages in Busoga in the mid-twentieth century resulted in separation, the vast majority of which involved the repayment of bridewealth. While these marital breakups might easily have been attributed to changes wrought by colonial conquest and rule, Fallers concluded instead that such marital instability was a long-standing feature of Soga society that proved resistant to the introduction of Christianity and literacy. Certainly the archives of the Mill Hill Fathers suggest that divorce—or marital breakup—was already common at the start of the twentieth century. In 1906 Bishop Hanlon described marriages in Busoga as “casual unions” (Bishop Hanlon to Fr. Matthews, March 6, 1906), while five years earlier Father van Term noted that both women and men easily ended their relationships (Anthony van Term to Very Reverend Father Rector, April 12,1901). And, in an ethnographic essay written at his mission station in Iganga, Father Condon commented that “divorce takes place and for very slight causes” (1911:371).

The causes for marital breakup were wide ranging and, significantly, both wives and husbands could initiate a separation. According to Father Burus, writing in 1913 in Bulamogi in northern Busoga, men could send their wives away for “okukyala (visiting), obutwa (poisoning), obuyombi (quarrelling, strife), obunafu (laziness).” In these cases, the woman was able to remarry and her family would repay the bridewealth. However, “when the wife is dismissed for ‘bulogo’ [sorcery],” the bridewealth was not repaid: “The reason is obvious—she cannot find a second husband.” A woman, for her part, could leave her husband if he beat her, insulted her when drunk, kept “getting into cases and thereby making her life miserable,” or did not live with her at home. Burus noted that “a clever wife goes in first instance to her father-in-law to accuse her husband & thence to her own father.” In these cases, the families would seek reconciliation, but divorce with the repayment of bridewealth would occur if reconciliation was not possible (Answers to questions re Native Marriages among the Balamogi, December 29, 1913).

In other areas of Busoga, however, the repayment of bridewealth was apparently not as usual. Father Brandsma noted in 1910 that in Bugabula men could send away their wives and women could leave their husbands for slight reasons and that divorces occurred regardless of the duration of the marriage; it was common for “young women, shortly after their marriage,” for “women married for some years,” and for “older women having children by their husbands.” He was told that while bridewealth would have been repaid in the past, even if children had been born, “nowadays” it was not (Marriage questions in Kamuli district). Father Jackson in Jinja offered a similar picture, although with the proviso that “divorce cannot be said to be common” for older women with children “as it did not happen often, but took place” (Questions about native marriages, Jinja).

It is impossible to know what marital breakup rates were for Basoga women and men between the sixteenth and twentieth centuries. And we must be cognizant of the consequences of the profound disruptions brought by the nineteenth century with the increase in warfare, slave raiding, disease epidemics, and the transformative nature of long-distance trade. The impact of slavery on domestic relationships is clear throughout the missionary accounts from the late nineteenth and early twentieth centuries (see Tuck Reference Tuck, Médard and Doyle2007). It is surely not unreasonable to posit that these developments affected marital stability in Busoga. Nonetheless, we should be wary of attributing too much importance to this period of upheaval. The continuity of high rates of divorce into the mid-twentieth century suggests that it was as much a tradition as was marriage.

Indeed, Lusoga speakers used the same word as their North Nyanza–speaking ancestors to describe women’s acts of leaving their husbands, the verb okúnóbá. In Lusoga, as in North Nyanza, it has a secondary meaning of “hate,” a reminder of the affective aspects of marital breakdown (Cultural Research Centre 2000a; Nzogi & Diprose Reference Nzogi and Diprose2011; Le Veux Reference Le Veux1917). Only a divorced woman who remained without remarrying and set up her own household was thus labeled in modern Lusoga. The nouns nákyéyómbekeiré or nákyéghómbekeiré name such a woman, a noun also used by Lugwere speakers in the form ónákyómbekeedé (Cultural Research Centre 2000a; Gulere Reference Gulere2009; Kagaya Reference Kagaya2006).^{Footnote 15} To talk of men divorcing their wives, Lusoga speakers coined the verb okútwalúmá along with the noun omútwalumé (divorced man). The contrast between the shared, inherited verb for women’s acts of leaving their marriages and the innovation in the language for men’s similar acts is noteworthy, although it is difficult to extrapolate about relative frequency or social perception from the difference.

To end a marriage, Basoga women and men, of course, had first to enter into one. They used vocabulary they inherited from North Nyanza to talk about the process of marriage, whether it was the verb okúbáyírá to describe a woman’s act of marrying (but with the new form, okúbáyízá, for a man), or omúgole to describe the bride (again with some innovation in also using it to mean bridegroom, in contrast to a gender-specific meaning in North Nyanza), or okúkwa to describe the payment of bridewealth. There were, therefore, important continuities between the ways in which Basoga men and women conceptualized formal marriage processes and how their linguistic ancestors had understood them. Nonetheless, Lusoga speakers also coined new words to reflect social and cultural changes to their marital practices. Some of the innovations were related to Christian marriage and so can be dated to the early twentieth century and missionary activity. As in Bugwere, the verb okúfúmbírwá to describe a woman’s act of marrying came with the Luganda translation of the Bible. In addition, Basoga innovated the verb okúgáítá, meaning “to wed in church.” They derived it from a Lusoga verb meaning “to add; to put together,” reflecting the Christian concept of marriage as a joining of two individuals.

When it came to more informal—or less socially sanctioned—forms of marriage, Lusoga speakers abandoned the vocabulary used by their North Nyanza- and West Nyanza-speaking forebears. They started to use the verb okúpáála to refer to elopement. This is a gender-neutral verb, referring to both a woman eloping and to a man doing so. Not only was this a break from their linguistic ancestors, but it also marked Basoga apart from Luganda and Lugwere speakers. Given the apparent permanence of obúko, or “in-law relationships,” even in the case of divorce, elopement may have offered an attractive alternative to formal marriage for both women and men who sought to avoid such a commitment. Social and economic status must also have been a factor in the decision to elope. A man without the means, or whose kin was without the means, to pay bridewealth would have welcomed the opportunity to avoid paying it or at least to reduce the amount of goods requested; once the woman had moved to her husband’s home, her family would have been in a much weakened negotiating position.

On the other hand, for women whose families were in a strong position to negotiate during marriage arrangements and who married through the formal route of the payment of bridewealth and accompaniment to the marital home, there were certain advantages. The most obvious of these was for those women who held the position of kairu or kaidu, “senior wife,” in the marital household.^{Footnote 16} As we saw above, this status existed during the time North Nyanza was spoken and it persisted in Bugwere and Buganda, as well as in Busoga (see Stephens Reference Stephens2013). The woman who held this position wielded significant authority in her marital home. As Father Brandsma noted, “she rules the other wives” (Marriage questions in Kamuli district). In Jinja, Father Jackson listed a range of responsibilities for her: “She takes charge of the household, manages his other women, looks after the slaves, helps him to settle cases that occur within his home affairs.” Furthermore, after propitiating at the household shrines, it was with her that her husband “must first go and have intercourse before any other of his women” (Questions about native marriages, Jinja). Writing in 1913, Father Burus offered a detailed description of the kairu in Bulamogi and the significant authority wielded by this woman in her marital home:

These descriptions highlight the kairu’s economic control and her importance to the spiritual well-being of the household. Because ancestral spirits as well as nature spirits and deities could ensure the success or failure of crops and bring disease or health to members of the household, her control of household shrines and propitiatory ceremonies was an important component of her power.^{Footnote 17} In the position of kairu we have a clear example of how a woman’s social status affected the kind of marriage she might enter into and the leverage she had to be able to maintain in her marital home throughout her life. Women whose kin were not able to negotiate such a position would have faced a harder time in their marriages, with less control over the key resources of the household.

Many Kinds of Marriages: Buganda

Perhaps nowhere in the societies speaking North Nyanza languages did marriage encompass such a wide range of relationships as in Buganda, or at least, nowhere else did people develop such a detailed taxonomy for them. This was a product of both conservation and innovation. Luganda speakers used verbs they inherited from North Nyanza, notably òkùwayira (“to marry,” for a woman) and òkùwasa (“to marry,” for a man). (The latter could also be used in the passive mood to apply to a woman: òkùwasibwa.) Other words were coined by speakers of Luganda, including the verb òkwògereza (meaning “to entice away,” “to woo,” or “to ask in marriage” [Snoxall Reference Snoxall1967:267]), which was derived from the verb òkwògera (“to speak,” “to say,” “to talk”). According to Julien Gorju of the White Fathers in his ethnography of the region, this was a marriage of “fiançailles,” that is, a marriage that followed formal engagement, and the nouns for “fiancée” (òmwògerezwa) and “fiancé” (òmwògereza or òmwògerezi) were derived from the verb. In Gorju’s opinion, this constituted “proper” or formal marriage (1920:303).

Another new word in Luganda was òkùfumbirwa (“to marry,” for a woman), which was derived from the verb òkùfumba (“to cook,” “to boil”). While òkùfumbirwa is a passive construction, its derivation yields a literal meaning of “to be cooked for.” The anthropologist Lucy Mair drew on this term and its etymology to make the case that a Muganda man expected his wife to be “a good housewife,” with the particular qualities of “diligence and cleanliness” (1940:13). But the verb suggests a more nuanced interpretation, namely that marriage offered the possibility of greater social status for some wives as well as assistance with household work, such as cooking. This points to the marked differences between forms of marriage and domestic relationships in Buganda, where some women held a status, marked in the language, in their marital homes that was denied to others.

A woman who married through arrangements labeled as òkùfumbirwa, òkwògereza, òkùwasibwa, or òkùwayira was known as an òmùwase (“wife”). Those who entered into sexual and domestic relationships with men by other means were not granted this respect. A brother, for example, might have “given” (òkùgabirwa) his sister to a friend, or a woman’s family might “add” (òkwòngera) another woman as a “gift” to a son-in-law (Gorju Reference Gorju1920:302). In another example, a patron might make a “pure and simple donation of a woman to a client, to a friend.” The latter would result in òbùfumbo òbùwumirizê (Gorju Reference Gorju1920:300). From the woman’s perspective, this process was called òkùwumìrizibwa, and she was known as an òmùwumirizê, not an òmùwase.^{Footnote 18} While Gorju described this as “pure and simple donation,” in fact this kind of arrangement involved a client laboring for the woman’s father, brother, or, in the case of an enslaved woman or other dependent, her master (Audrey Richards [1899–84], Field notes). This form of marriage contract was more commonly found in societies farther south in Africa that traced descent matrilineally (see, e.g., Phiri Reference Phiri1983; Saidi Reference Saidi2010). In the patrilineal context of Buganda, it enabled a wealthy—or relatively wealthy—man to have his son-in-law labor for him, something that would not have been socially acceptable had bridewealth been paid.

Other words also focused on the means by which a woman entered a household. These included the nouns òmùgulè (“the bought one”), òmùnyagè (“the seized one”), òmùsikìre (“the inherited one”), èndobòlo (“the share”), and ènvùmâ (a low-ranked female slave) (Le Veux Reference Le Veux1917:1954*). According to Père Le Veux of the White Fathers, an òmùgulè or òmùkazì omùgulè was “a female slave who had been bought” (Reference Le Veux1917:231). Gorju, meanwhile, noted that the verb òbùgule referred to a father buying a woman/wife (“une femme”) for his son (Reference Gorju1920:299). Fathers often would have assisted sons in paying bridewealth, especially for a first wife, but the existence of this form of transaction emphasized the distinction between buying a wife as a slave and paying bridewealth to a woman’s family and lineage.

Unfree women could also enter households after having been kidnapped during warfare. The act of kidnapping was known as òkùnyaga and the woman as òmùnyagè. She was initially referred to as ènvùmâ (a low-ranked female slave), but her status could be regularized if her master chose to do so by “y’okwebûla obuko” or “confessing to the in-laws,” after which she was known as “wa busa” or free (Gorju Reference Gorju1920:298). While we should not overstate the flexibility of slavery in Buganda, or the ease with which enslaved women and girls (as well as men and boys) might gain familial status, it is evident that the distinction between an enslaved woman and a formal wife was not always very clear. Indeed, this was something that caused confusion for missionaries seeking to establish the validity of the various relationships their converts had entered into, as we will see below.

These forms of relationships created through gifting, purchase, or kidnapping surely increased, if they did not originate, as a result of the prolonged conflict and social dislocation that accompanied the expansionary wars launched by Buganda’s rulers in the early eighteenth century. The commodification of relationships was also exacerbated by the dramatic expansion of long-distance trade (in slaves, ivory, and guns, among other goods) from the late eighteenth century on. An elite, but nonetheless powerful, example of this commodification comes from John Hanning Speke’s Journal of the Discovery of the Nile (Reference Speke1864). In Karagwe in 1862, he wrote, as the group pitched their “tents after sunset by some pools on the plain, Dr. K’yengo arrived with the hongo [toll, tribute] of brass and copper wires sent by Sŭwarora for the great king Mtésa, in lieu of his daughter who died” (1864:200). Such was the premium on trade goods in the mid-nineteenth century that brass and copper wire could replace a daughter as a gift in prestation to the king of Buganda.

We should be wary, however, of seeing only change. The complex forms of relationships that can be gathered loosely under the term “marriage” were not all the product of the powerful forces that dominated the eighteenth and nineteenth centuries in Buganda and beyond. Kidnapping and the purchase of slaves as well as the dramatic increase in the numbers of “wives” attached to royal and chiefly households (with women numbering into the hundreds in some cases) can be related to the expansionary warfare that began with the reign of King Kateregga and Queen Mother Nabusa Nabagereka (see Musisi Reference Musisi1991; Médard Reference Médard2007).^{Footnote 19} By contrast, forms of domestic relationships such as those in which the woman was referred to as èndobòlo (“the share”) or òmùsikìre (“the inherited one”) formed part of older social practices (see Stephens Reference Stephens2013). The retention of North Nyanza terms such as òkùwayira and òkùwasa similarly points to important continuities in the meaning and structure of marriage, even in the midst of upheaval.

“Real Marriage”: Missionaries and Determinations of Marital Standing

The Mill Hill Fathers in Uganda were working in a context in which marriage looked very different from what they had known in their communities in northern Europe. They often found themselves unable to determine the marital status of their converts, a necessary prerequisite to baptism. Because of the supposed finality of this determination, and the transcription of that information in church records, many Ugandan Catholics wanted to ensure that this was done on their terms and to their satisfaction. Men in polygynous households asserted their right to choose which wife they remained with once baptized, while their wives often complained vigorously—both about being chosen and not (e.g., L. Litoff, Case of Kibatu-Muwebwa, January 22, 1901).

Not all relationships would have been as messy as those that the Mill Hill missionaries wrote about to their superiors, requesting advice. The cases that made it into the archives were necessarily those that were complicated or contested in some way. We should not, therefore, read them as representative of marriages in Uganda in general. But they reflect the complexity of marriage and the significant flexibility and mutability in domestic and sexual relationships, all of which, as we have seen, were long-standing characteristics of marriage in the region. They also reflect the real consequences for individuals and families of the violent dislocations of the previous decades. The following cases encapsulate both the way in which the events of the nineteenth century had distorted domestic relationships and the long history of diversity in such relationships.

The first case, concerning a woman named Yatuwa Amanyi and a man named Alevi Zavuga, was reported by Father Biermans from Nagalama in Buganda in October 1902. “When he was a boy of about 15 years,” Biermans reported, Alevi’s father “gave him a slave girl,” namely, Yatuwa. “When they were grown up they agreed, it seems, properly to stay together as husband & wife, liking each other very much.” Alevi, therefore, “gave his father mwenge [banana beer] & 2000 shells,” in order to thank him “for having given him Yatuwa amanyi, and at the same time to show him that he liked her & wanted to keep her as wife. She herself told me that she agreed to all this.” Some time later, after the birth of two children and the death of one of them, Yatuwa left Alevi for about a year. “He did not want her anymore when she had kunoba’d [left the marriage].” But once he had been baptized in the Catholic Church, Alevi changed his mind and “went to fetch her back; she agreed & lived with him” (Case Alevi Zavuga, Nagalama, October 8, 1902).

Here, then, was a case in which an enslaved girl had been given by her master to his son, but they had formalized their relationship, something that, paradoxically, had enabled her to leave him. Now that they wished to live together again as wife and husband in the eyes of the church, Father Biermans felt he needed to determine whether they were ever “properly married” and whether the fact of her leaving him before either one had been baptized rendered them divorced. His opinion in regard to the first question was that “though rather doubtful,” they had been married, “for although she was a slave girl & given to Zavuga, she agreed to be his wife, & agreed him to take mwenge & bintu [i.e., bridewealth] to his father; moreover they liked each other very much.” Nonetheless, given that she had left him and he, “not liking her anymore, when she had run away” did not “fetch” her for a year, Biermans considered that they “were divorced & not any longer husband & wife” (Case Alevi Zavuga, Nagalama, October 8, 1902). For the missionary, the key change in status was that Alevi had since been baptized and so could not contract a new marriage outside of the church. The couple were thus instructed that once Yatuwa had been baptized, they could get married in the church.

Other situations resulted in less happy conclusions. In 1898 Father Kerstens reported the case of a recently baptized man named Lauki who told the missionary “that he did not want his wife, she was a thief.” On further questioning of the couple, Kerstens determined that “she was not properly married to him. She is a munyoro woman, stolen by Lauki in the Bunyoro war. They remained together and had children, but she says she does not love him.” According to Kerstens they did not have a “Buganda marriage and whether they promised each other some thing is difficult to work out.” He considered that as they had three children, and had seemingly agreed to live together, they had “both consented to it.” Despite this, “both seem to prefer to be free again.” He encouraged them to reconcile, “especially for the children they have.” But the woman, whom the missionary does not name, was adamant that she wished to leave Lauki: “she is a stolen woman she says and wants to go back to her father” (Letter from G. Kerstens to Bishop, Nagalama, March 2, 1898). Whereas Kerstens sought to reconcile this couple, albeit unsuccessfully, other missionaries were more ready to conclude that kidnapping could not result in marriage, no matter the intervening developments (Marriage Cases, Nyenga, December 11, 1901).

The cases also highlight the ways in which women were able to assert themselves and leave unsatisfactory marriages in order to contract better ones, such as happened with Musibika we Nakabidzi. According to her case as reported by Father L. Litoff from Nyenga in 1901, Musibika had previously been married to a man named Kalulwe, who had paid bridewealth of 10,000 cowrie shells, one goat, two barkcloths, and four calabashes of beer. Kalulwe then married another wife, and “Musibika left Kalulwe because he after some time made the second woman the mukuru [elder, i.e., senior wife].”^{Footnote 20} He “came once for her,” but “she refused to return to him & so he (Kalulwe) charged Salongo her brother 1,000 Sh, 3 mbugo [barkcloths], mwenge [beer] (quantity not known), all which things” were paid to him. Having left Kalulwe, Musibika then married EbwayaKyalo, who only gave her brother “4 bita of mwenge,” although the brother “charged him” the same bridewealth as the first husband had paid. While in other cases missionaries used the incomplete payment of bridewealth to assert that couples were not married, Father Litoff did not raise that concern here. His concern was rather whether Musibika and Kalulwe were divorced, and thus whether she had been in a position to marry EbwayaKyalo at all. He determined that the divorce was valid, “because Musibika left him & he (Kalulwe) received recompense thereby acknowledging the divorce” (Case of EbwayaKyalo-Musibika, Nyenga, January 22, 1901). As this and other cases show, missionaries responded to different situations in a variety of ways. While their self-interest often shaped their determinations, they were no doubt also motivated by a desire to help new Catholics find ways to make their sometimes complicated situations fit the new order of things.