Biodemography and Social Biology, 61:98–110, 2015 Copyright © Society for Biodemography and Social Biology ISSN: 1948-5565 print / 1948-5573 online DOI: 10.1080/19485565.2014.981795
Perceptions of Polygyny: The Effects of Offspring and Other Kin on Co-Wife Satisfaction BROOKE A. SCELZA Department of Anthropology, University of California, Los Angeles, Los Angeles, California, USA The costs and benefits of polygyny have been widely debated and vary according to local sociopolitical context, the level of female autonomy, and economic considerations such as the mode of production. This study aims to understand perceptions of polygyny as a function of household demography, particularly the number of female kin present in the household who can provide labor that is largely substitutable to that of a co-wife. The presence of these helpers is proposed to shift the cost-benefit structure of polygyny, in which having more female kin available is associated with a more negative view of the practice. Interview and census data from 106 Himba women, who are traditional, seminomadic pastoralists, were used to test this prediction. Among married women who reside patrilocally, the presence of more elder daughters was associated with a more negative view of polygyny. Among unmarried women, who reside in their natal homes, it is the total number of adult female kin that predicts perception of polygyny. In addition, unmarried women are significantly more likely to report fights over resources as a source of co-wife conflicts when they have more dependent children, but no such association was found among married women.
Introduction Aside from kin, one of the most constant and intimate relationships women in polygynous marriages often have is with their co-wives. These women regularly live in the same residential structure or compound; they may rely on the same resource base; and, of course, notably, they share a husband and their children are half-siblings. These ties present opportunities for both cooperation and conflict, and the literature is rife with examples of both. However, co-wife relationships exist within more complex family structures made up of both kin and non-kin. Despite the ubiquity of this dynamic, we know little about how cowife relations are affected by other aspects of household demography. In other words, how useful (or harmful) a co-wife is may be a function of the greater social support system a woman can draw from, and her perceptions of polygyny may be equally dependent on her overall level of support. Here I aim to highlight the role of co-wife relationships within the greater household economy. To do so I explore how women’s household structures shape their perceptions of polygyny. In particular, I explore how the number and sex of a woman’s offspring and the presence of adult female kin affect her feelings about polygyny.
Address correspondence to Brooke A. Scelza, Department of Anthropology, University of California, Los Angeles, 341 Haines Hall, Box 951553, Los Angeles, CA 90095-1553, USA. E-mail: [email protected]
98
Effects of Offspring and Kin on Co-Wife Satisfaction
99
The Costs and Benefits of Polygyny Currently, the literature is divided about whether women gain from being polygynously married. The polygyny threshold model, developed as a general model of female mating decisions across species (Orians 1969) but successfully applied to humans (Borgerhoff Mulder 1988), predicts that women will only enter into a polygynous marriage when the benefits of doing so outweigh or are equal to those of being monogamously married. Typically, women’s decisions have been measured as a function of male wealth, with women predicted to choose marriage to an already-married man only if he has sufficient resources to provide for multiple wives. This is most commonly seen in societies where there is stratification of male resources, with wealthier men being better able to provide and therefore more likely to attract multiple wives (Borgerhoff Mulder 1989; Borgerhoff Mulder 1990; Hartung et al. 1982). In addition to the material benefits women can gain through polygyny by marrying wealthy or powerful men, several studies have also cited advantages of cooperative cowife relations, including assistance with household labor and child care (Dorjahn 1988; Irons 1983; Ware 1979), greater mobility (Burman 1996; Anderson 2000), and companionship (Solway 1990; Madhavan 2002). These cooperative benefits may increase productive efficiency (economies of scale), or improve the health and welfare of women and their children. Among some Aboriginal Australians, women reported relying extensively on a cooperative system in which they would reciprocate foraging and child-care duties with their co-wives (Scelza 2008; Meggitt 1965). In these same societies, women tend to report that co-wife relations are generally amicable and indeed necessary to successful production. In another example, Adams, Madhavan, and Simon (2002) show that the odds of infant mortality are reduced when women have more non-kin, notably co-wives, in their social networks. Polygyny has also been shown to have a protective effect on post-neonatal mortality in a national survey from Nigeria (Ukwuani, Cornwell, and Suchindran 2002). In these cases, the emphasis is on the co-wives themselves, rather than on the resources of their husbands. In sharp contrast, a substantial number of studies have shown that women’s health and reproductive success are compromised by being in a polygynous union. Women in polygynous unions may experience lower fertility (Chisholm and Burbank 1991; Dorjahn 1959; Josephson 2002) or higher rates of infant mortality (Gyimah 2009; Strassmann 1997), and their children may be in poorer health (Hadley 2005; Sellen 1999). In some cases, they have also been shown to experience greater psychological distress (Tabi, Doster, and Cheney 2010; Al-Krenawi 1998; but see Patil and Hadley 2008), conflict, and jealousy (Jankowiak, Sudakov, and Wilreker 2005; Meekers and Franklin 1995), and they are at increased risk of contracting human immunodeficiency virus and other sexually transmitted infections (Bove and Valeggia 2009). In addition, children of polygynous unions (particularly those of junior co-wives) may suffer social costs, such as fewer opportunities for education (Bledsoe 1993). Conflicts among co-wives often trigger or exacerbate these costs.
Polygyny in Context Whether a woman is able to buffer herself against the potential costs of polygyny and/or take advantage of potential benefits of co-wife cooperation may depend on institutionaland individual-level factors that influence her well-being. Societies in which women are more likely to thrive in polygynous relationships tend to be those in which female autonomy is relatively high, increasing the probability that female choice is possible. This can be
100
B. A. Scelza
framed in terms of women’s self-sufficiency, their access to allies, and the level of power they have in household and reproductive decision-making. For example, when a husband has multiple wives, those women may have greater control over fertility decisions and the ability to lengthen interbirth intervals (Ware 1979). Autonomy also often varies according to structural factors such as postmarital residence norms. In one unusual case in which residence is matrilocal and polygyny is non-sororal (i.e., co-wives are unrelated), co-wives maintain separate residences and are supported by kin, and husbands are expected to share their resources and time equally among their wives (Ottenheimer and Ottenheimer 1979). In this case a husband’s several mothers-in-law help to ensure that he is treating their daughters fairly. Contrasting but complementary evidence for the role of female autonomy comes from studies that show polygynously married women having poorer outcomes than monogamous women in situations where they have low levels of autonomy. Among the Dogon, where male control over female reproduction is particularly strong, women appear to be coerced into polygynous marriages, despite their being detrimental to female reproductive success (Strassmann 1997). Other institutional factors such as the prominence of certain state religions, patriarchal control of resources, and women’s poor political status often reinforce male coercion and further the misalignment of relative power and control between men and women over reproductive and marital decisions. In addition to these institutional factors, variation between women within societies also has a significant impact on how women fare in polygynous marriages. A woman’s social and economic standing in the household often varies across her lifespan, as she gains or loses co-wives, as she begins childbearing, as her children age, and as her relationship with her husband develops (Borgerhoff Mulder 1989; Lamphere 1974; Madhavan 2002; Meekers and Franklin 1995). While the particular factors that moderate women’s standing are locally specific, certain aspects of spousal relations have repeatedly been shown to influence women’s outcomes. Women who are in sororally polygynous marriages often fare better than those whose co-wives are unrelated (Chisholm and Burbank 1991; Yanca and Low 2004). First wives also often have better outcomes than later wives (Bean and Mineau 1986; Gibson and Mace 2007; Josephson 2002). In many contexts, seniority means more influence in the choice of later wives, as well as increased control over labor and resource distributions. This has been shown to result in poorer fitness outcomes for women, as well as nutritional deficiencies among the children of later wives (Gibson and Mace 2007). Finally, the individual relationship that any two co-wives have may affect their health and well-being. In Mali, the junior-senior co-wife relationship is greatly affected by individual personalities. Factors such as loyalty to the conjugal versus the natal household and the degree to which junior co-wives feel protected by their senior co-wife influence the relative levels of competition and cooperation reported in a given household (Madhavan 2002). Polygyny and Social Support Co-wives are only one of a suite of potential helpers women have within the household. Therefore, women may temper their opinion of polygyny depending on how valuable cowives are within their particular household setting. Married and unmarried women live within very different household structures, particularly in societies with patrilocal residence norms. In these societies, unmarried women typically reside with their kin, while married women live mainly with affines, especially early on, before they have children of their own. As women move through their reproductive careers, they can expect to gain increasing amounts of help from their children (Lee and Kramer 2002). There is a significant amount of variation both across and within modes of production in how valuable the
Effects of Offspring and Kin on Co-Wife Satisfaction
101
supplemental efforts of children are to household production, but in general, as children age they are increasingly capable of performing household tasks (Kramer 2005). However, the benefits of elder offspring’s production must be counterbalanced with the energetic costs of pregnancy and lactation, as well as the support of younger children, who remain net consumers. The degree of overlap in productive capacity must also be considered. If there is a strongly gendered division of labor, the labor of daughters may be much more substitutible with co-wives’ than is sons’. Whether the presence of a co-wife and her children enhances or alleviates resource stress may therefore be a function of the overall composition of the household.
Study Population The Himba are a seminomadic pastoral group living in northwest Namibia. They remain largely isolated from the market economy, relying primarily on pastoral production and small gardens for subsistence. Details of Himba social structure and organization have been published elsewhere (Bollig 2006; Malan 1995; Scelza 2011a). Here I will focus on explaining the most pertinent aspects of women’s social worlds: marriage, co-wife and spousal relationships, reproduction, and household labor distribution. These observations are based on ethnographic and quantitative data collected between 2010 and 2013 in the Omuhonga Basin, a section of the Kunene region located 150 kilometers from the regional center of Opuwo. Himba women enjoy mixed levels of autonomy in reproductive and marital decisionmaking. All marriages are arranged by kin, but as they age, women can exert increasing choice over whom they marry. Second and later marriages are much more likely to be “love matches.” Among the married women in this sample, 43 percent were in “love matches” at the time of data collection. Divorce is common in both arranged and “love match” marriages and can be initiated by either spouse. In addition, women are also relatively free to pursue lovers outside of marriage. This has resulted in a large number of out-of-wedlock births and children born through affairs (Scelza 2011b). In keeping with both the high level of marital fluidity and the low certainty of paternity in this population, bride-price is small compared to many other pastoralist populations in sub-Saharan Africa. When an adult couple marries they reside either with the husband’s kin or at a neo-local camp with close ties to the husband’s paternal kin. In addition to a husband, his wives, and their children, households often include extended kin, typically the mother or sisters of the head of household. Household sizes in the study area tend to be larger than in other areas where Himba reside, and household compounds are also more permanent, partly because of more reliable access to water in the Omuhonga Basin (Bollig 2006). This residential structure results in women being separated from their own kin while living in their marital household, and relying largely on the help of their co-wives, in-laws, and children for tasks that require cooperative production. The Himba are a natural fertility population. In a survey of women’s knowledge and use of contraceptives conducted in 2010 (N = 50) less than a third of women had ever heard of modern contraceptive devices, less than 15 percent had ever used them, and none of the women were currently using them. Average fertility in the current sample is 4.1 (range 0–11, N = 120), but this is likely a slight underestimate, as there are a significant number of stillbirths and early infant deaths, which may be underreported. In a subsample of women for which number of child deaths is more certain (as a result of repeated interviews and checks with spousal interviews), fertility is 4.3 (range 0–10, n = 47). Because I aimed for a sample that was older, in order to include more postmenopausal women (see
102
B. A. Scelza Table 1 Demographic characteristics of the sample population (N = 106)
Mean age (min, max) Mean number of children Current marital status Married Widowed Divorced Never married Number of co-wives 0 1 2 3+ Co-wife order 1 2 3 4
Married (n = 59)
Unmarried (n = 47)
43.28 (19, 84) 4.32
52.68 (20, 88) 3.47
59 (100%) — — —
20 (42.5%) 20 (42.5%) 7 (15%)
19 (32%) 25 (42%) 11 (19%) 4 (7%) 15 (38%) 21 (53%) 3 (7%) 1 (3%)
Table 1), the age distribution of women’s children is also skewed older. For example, the women had on average 1.25 (0–8) daughters and 1.48 (0–7) sons over age 10, and only 0.53 (0–4) daughters and 0.70 (0–4) sons under 10 (ranges in parentheses). Co-wife relations among the Himba have yet to be described in detail. However, an account of Herero marriage published in the 1950s is still largely relevant given the historical and phylogenetic history of the two groups (for details, see Malan 1995). Among the Herero, senior wives are given some ritual and social importance in the household, such as tending to the ritual fire and seniority in milking the holy cows among the herd (Gibson 1958). Senior wives play a critical role in choosing and approving of junior co-wives. The children of senior wives also have primary inheritance rights. Wealthy elder men may have upwards of four to five wives, whereas younger and/or poorer men typically have only one. Because of the high divorce rate and differentials in the ages of husbands and wives, almost all men marry at some point in their lives. Women’s work includes water and firewood collection, grinding maize, milking, souring the previous day’s milk, cooking, and child care. The husband’s livestock are partitioned between wives, and each wife keeps the milk from the cows allotted to her. Women also often own small stock (sheep and goats), and occasionally cows. During the rainy season, women spend large parts of the day in the gardens: planting, harvesting, and performing maintenance and pest control. Gardens are passed matrilineally and may be worked by multiple women within the same family. Many tasks are done collectively, if not cooperatively. Food, firewood, and water may be shared across the household (particularly between kin or co-wives who get along well), and divisions of labor occur in which one woman will watch young children while the other goes to collect water. Children participate in laborious activities from a young age and also help with child care. Boys typically herd small stock, while girls remain in the compound and assist women with their work.
Effects of Offspring and Kin on Co-Wife Satisfaction
103
Methods Sample All adult women currently residing in 24 compounds in the Omuhonga Basin were originally included in this sample (n = 119). Four women declined to be interviewed. One had given birth the night before and was resting. Another two were very elderly. A fourth stated that she was too busy and could not be accommodated at another time. An additional nine women who were interviewed were not asked questions about polygyny. These exclusions occurred because the woman had her co-wife or other affines in close vicinity and full confidentiality could not be guaranteed. This resulted in a final sample of 106 women (Table 1). All interviews were conducted in Otjihimba with the assistance of a local translator. Semi-Structured Interviews Women provided marital and reproductive histories, including number of children, sex and age of each child, current marital status, number of co-wives and co-wife order, and whether each marriage was a “love match” or an arranged marriage. Ages of women and their children were determined by converting traditional Himba year names into numeric calendar years (for methodology, see Scelza 2011a). Women were also asked about their perceptions of polygyny. First, a binary question was asked: “Do you think having a co-wife is a good thing, or a bad thing?” Next, women were asked to list both what co-wives were good for and what kinds of things co-wives fight about. While descriptive data will be presented on women’s responses to the question of what co-wives are good for, only data on co-wife conflicts will be analyzed here. Multiple responses were allowed in both types of free list, and free-response elaborations were recorded, translated, and transcribed. Both married and unmarried women were interviewed in order to understand how perceptions of polygyny differ according to current marital status. Analysis Logistic regression models were constructed to identify variables that were significantly associated with women’s perception of polygyny. The first dependent variable was constructed from a question asking whether women thought polygyny was a “good thing” or a “bad thing.” Key predictor variables were number of sons and number of daughters, broken out into four variables: number of sons under age 10, number of daughters under age 10, number of sons aged 10 or older, and number of daughters aged 10 or older. Ten years was used as a cutoff, as these are the children able to regularly perform productive labor similar to that of adults. In particular, I was interested in labor that was substitutable for women’s work, such as carrying water or firewood, which requires a moderate level of physical strength. Married and unmarried women were analyzed separately. The first analysis, which includes only married women, contains known influences on perceptions of polygyny, including co-wife order, number of co-wives, and women’s age. In the second analysis of unmarried women, covariates related to polygyny were not included, but a variable tabulating the number of female kin over age 10 living in the household was included. Standard errors were clustered by compound to account for overlapping influences within the household. A second set of regressions looked at co-wife conflicts. These were broken into two dependent variables. The first was conflict over resources. Women were coded as noting
104
B. A. Scelza
resource conflict if they gave any response that indicated co-wives would fight over the allocation of goods provided by the husband, the distribution of household labor (e.g., one co-wife is lazy, or there is a lack of reciprocity in turn-taking), or child care (e.g., a cowife refused to take care of a child when the other wife was away from home). The second dependent variable was conflict over sex. Here women cited sexual jealousy either because the husbands spent more time with one wife or because he liked one wife better than the other. The same predictor variables as described previously were used in these analyses, again with married and unmarried women analyzed separately. All analyses were done using Stata 12.0. Informed oral consent was obtained from all participants, and this study was approved by the University of California, Los Angeles Institutional Review Board (10-000238).
Results Himba women have generally positive perceptions of polygyny, with 74 percent of women reporting that they thought having a co-wife was a “good thing.” While the majority of women in all marital groups expressed this sentiment, married and widowed women were more likely to think positively about polygyny (76% and 80%, respectively) than divorced (65%) and unmarried (57%) women. However, there were no statistically significant differences between the groups (χ 2 = 2.37, df = 3, p = .498). Among married and widowed women, those with co-wives were slightly more likely to have a positive perception of polygyny (with 79% responding that it was a “good thing”) than were monogamously married women (74%). Among married women, there is a strong negative relationship between number of daughters aged 10 or older and perception of polygyny (Table 2). The more older daughters a woman has, the more likely she is to have a negative view of polygyny. The number of daughters under age 10 and the number of sons (either over or under age 10) had no effect on perception of polygyny. Number of co-wives had a positive and significant effect; having more co-wives was associated with a more positive perception of polygyny. However, cowife order had no significant effect. Among unmarried women, none of the child categories Table 2 Odds ratios (95% confidence intervals) of perception of polygyny as predicted by logistic regression: Effects of kin on married and unmarried women
Current age Number of daughters (
Perceptions of Polygyny: The Effects of Offspring and Other Kin on Co-Wife Satisfaction BROOKE A. SCELZA Department of Anthropology, University of California, Los Angeles, Los Angeles, California, USA The costs and benefits of polygyny have been widely debated and vary according to local sociopolitical context, the level of female autonomy, and economic considerations such as the mode of production. This study aims to understand perceptions of polygyny as a function of household demography, particularly the number of female kin present in the household who can provide labor that is largely substitutable to that of a co-wife. The presence of these helpers is proposed to shift the cost-benefit structure of polygyny, in which having more female kin available is associated with a more negative view of the practice. Interview and census data from 106 Himba women, who are traditional, seminomadic pastoralists, were used to test this prediction. Among married women who reside patrilocally, the presence of more elder daughters was associated with a more negative view of polygyny. Among unmarried women, who reside in their natal homes, it is the total number of adult female kin that predicts perception of polygyny. In addition, unmarried women are significantly more likely to report fights over resources as a source of co-wife conflicts when they have more dependent children, but no such association was found among married women.
Introduction Aside from kin, one of the most constant and intimate relationships women in polygynous marriages often have is with their co-wives. These women regularly live in the same residential structure or compound; they may rely on the same resource base; and, of course, notably, they share a husband and their children are half-siblings. These ties present opportunities for both cooperation and conflict, and the literature is rife with examples of both. However, co-wife relationships exist within more complex family structures made up of both kin and non-kin. Despite the ubiquity of this dynamic, we know little about how cowife relations are affected by other aspects of household demography. In other words, how useful (or harmful) a co-wife is may be a function of the greater social support system a woman can draw from, and her perceptions of polygyny may be equally dependent on her overall level of support. Here I aim to highlight the role of co-wife relationships within the greater household economy. To do so I explore how women’s household structures shape their perceptions of polygyny. In particular, I explore how the number and sex of a woman’s offspring and the presence of adult female kin affect her feelings about polygyny.
Address correspondence to Brooke A. Scelza, Department of Anthropology, University of California, Los Angeles, 341 Haines Hall, Box 951553, Los Angeles, CA 90095-1553, USA. E-mail: [email protected]
98
Effects of Offspring and Kin on Co-Wife Satisfaction
99
The Costs and Benefits of Polygyny Currently, the literature is divided about whether women gain from being polygynously married. The polygyny threshold model, developed as a general model of female mating decisions across species (Orians 1969) but successfully applied to humans (Borgerhoff Mulder 1988), predicts that women will only enter into a polygynous marriage when the benefits of doing so outweigh or are equal to those of being monogamously married. Typically, women’s decisions have been measured as a function of male wealth, with women predicted to choose marriage to an already-married man only if he has sufficient resources to provide for multiple wives. This is most commonly seen in societies where there is stratification of male resources, with wealthier men being better able to provide and therefore more likely to attract multiple wives (Borgerhoff Mulder 1989; Borgerhoff Mulder 1990; Hartung et al. 1982). In addition to the material benefits women can gain through polygyny by marrying wealthy or powerful men, several studies have also cited advantages of cooperative cowife relations, including assistance with household labor and child care (Dorjahn 1988; Irons 1983; Ware 1979), greater mobility (Burman 1996; Anderson 2000), and companionship (Solway 1990; Madhavan 2002). These cooperative benefits may increase productive efficiency (economies of scale), or improve the health and welfare of women and their children. Among some Aboriginal Australians, women reported relying extensively on a cooperative system in which they would reciprocate foraging and child-care duties with their co-wives (Scelza 2008; Meggitt 1965). In these same societies, women tend to report that co-wife relations are generally amicable and indeed necessary to successful production. In another example, Adams, Madhavan, and Simon (2002) show that the odds of infant mortality are reduced when women have more non-kin, notably co-wives, in their social networks. Polygyny has also been shown to have a protective effect on post-neonatal mortality in a national survey from Nigeria (Ukwuani, Cornwell, and Suchindran 2002). In these cases, the emphasis is on the co-wives themselves, rather than on the resources of their husbands. In sharp contrast, a substantial number of studies have shown that women’s health and reproductive success are compromised by being in a polygynous union. Women in polygynous unions may experience lower fertility (Chisholm and Burbank 1991; Dorjahn 1959; Josephson 2002) or higher rates of infant mortality (Gyimah 2009; Strassmann 1997), and their children may be in poorer health (Hadley 2005; Sellen 1999). In some cases, they have also been shown to experience greater psychological distress (Tabi, Doster, and Cheney 2010; Al-Krenawi 1998; but see Patil and Hadley 2008), conflict, and jealousy (Jankowiak, Sudakov, and Wilreker 2005; Meekers and Franklin 1995), and they are at increased risk of contracting human immunodeficiency virus and other sexually transmitted infections (Bove and Valeggia 2009). In addition, children of polygynous unions (particularly those of junior co-wives) may suffer social costs, such as fewer opportunities for education (Bledsoe 1993). Conflicts among co-wives often trigger or exacerbate these costs.
Polygyny in Context Whether a woman is able to buffer herself against the potential costs of polygyny and/or take advantage of potential benefits of co-wife cooperation may depend on institutionaland individual-level factors that influence her well-being. Societies in which women are more likely to thrive in polygynous relationships tend to be those in which female autonomy is relatively high, increasing the probability that female choice is possible. This can be
100
B. A. Scelza
framed in terms of women’s self-sufficiency, their access to allies, and the level of power they have in household and reproductive decision-making. For example, when a husband has multiple wives, those women may have greater control over fertility decisions and the ability to lengthen interbirth intervals (Ware 1979). Autonomy also often varies according to structural factors such as postmarital residence norms. In one unusual case in which residence is matrilocal and polygyny is non-sororal (i.e., co-wives are unrelated), co-wives maintain separate residences and are supported by kin, and husbands are expected to share their resources and time equally among their wives (Ottenheimer and Ottenheimer 1979). In this case a husband’s several mothers-in-law help to ensure that he is treating their daughters fairly. Contrasting but complementary evidence for the role of female autonomy comes from studies that show polygynously married women having poorer outcomes than monogamous women in situations where they have low levels of autonomy. Among the Dogon, where male control over female reproduction is particularly strong, women appear to be coerced into polygynous marriages, despite their being detrimental to female reproductive success (Strassmann 1997). Other institutional factors such as the prominence of certain state religions, patriarchal control of resources, and women’s poor political status often reinforce male coercion and further the misalignment of relative power and control between men and women over reproductive and marital decisions. In addition to these institutional factors, variation between women within societies also has a significant impact on how women fare in polygynous marriages. A woman’s social and economic standing in the household often varies across her lifespan, as she gains or loses co-wives, as she begins childbearing, as her children age, and as her relationship with her husband develops (Borgerhoff Mulder 1989; Lamphere 1974; Madhavan 2002; Meekers and Franklin 1995). While the particular factors that moderate women’s standing are locally specific, certain aspects of spousal relations have repeatedly been shown to influence women’s outcomes. Women who are in sororally polygynous marriages often fare better than those whose co-wives are unrelated (Chisholm and Burbank 1991; Yanca and Low 2004). First wives also often have better outcomes than later wives (Bean and Mineau 1986; Gibson and Mace 2007; Josephson 2002). In many contexts, seniority means more influence in the choice of later wives, as well as increased control over labor and resource distributions. This has been shown to result in poorer fitness outcomes for women, as well as nutritional deficiencies among the children of later wives (Gibson and Mace 2007). Finally, the individual relationship that any two co-wives have may affect their health and well-being. In Mali, the junior-senior co-wife relationship is greatly affected by individual personalities. Factors such as loyalty to the conjugal versus the natal household and the degree to which junior co-wives feel protected by their senior co-wife influence the relative levels of competition and cooperation reported in a given household (Madhavan 2002). Polygyny and Social Support Co-wives are only one of a suite of potential helpers women have within the household. Therefore, women may temper their opinion of polygyny depending on how valuable cowives are within their particular household setting. Married and unmarried women live within very different household structures, particularly in societies with patrilocal residence norms. In these societies, unmarried women typically reside with their kin, while married women live mainly with affines, especially early on, before they have children of their own. As women move through their reproductive careers, they can expect to gain increasing amounts of help from their children (Lee and Kramer 2002). There is a significant amount of variation both across and within modes of production in how valuable the
Effects of Offspring and Kin on Co-Wife Satisfaction
101
supplemental efforts of children are to household production, but in general, as children age they are increasingly capable of performing household tasks (Kramer 2005). However, the benefits of elder offspring’s production must be counterbalanced with the energetic costs of pregnancy and lactation, as well as the support of younger children, who remain net consumers. The degree of overlap in productive capacity must also be considered. If there is a strongly gendered division of labor, the labor of daughters may be much more substitutible with co-wives’ than is sons’. Whether the presence of a co-wife and her children enhances or alleviates resource stress may therefore be a function of the overall composition of the household.
Study Population The Himba are a seminomadic pastoral group living in northwest Namibia. They remain largely isolated from the market economy, relying primarily on pastoral production and small gardens for subsistence. Details of Himba social structure and organization have been published elsewhere (Bollig 2006; Malan 1995; Scelza 2011a). Here I will focus on explaining the most pertinent aspects of women’s social worlds: marriage, co-wife and spousal relationships, reproduction, and household labor distribution. These observations are based on ethnographic and quantitative data collected between 2010 and 2013 in the Omuhonga Basin, a section of the Kunene region located 150 kilometers from the regional center of Opuwo. Himba women enjoy mixed levels of autonomy in reproductive and marital decisionmaking. All marriages are arranged by kin, but as they age, women can exert increasing choice over whom they marry. Second and later marriages are much more likely to be “love matches.” Among the married women in this sample, 43 percent were in “love matches” at the time of data collection. Divorce is common in both arranged and “love match” marriages and can be initiated by either spouse. In addition, women are also relatively free to pursue lovers outside of marriage. This has resulted in a large number of out-of-wedlock births and children born through affairs (Scelza 2011b). In keeping with both the high level of marital fluidity and the low certainty of paternity in this population, bride-price is small compared to many other pastoralist populations in sub-Saharan Africa. When an adult couple marries they reside either with the husband’s kin or at a neo-local camp with close ties to the husband’s paternal kin. In addition to a husband, his wives, and their children, households often include extended kin, typically the mother or sisters of the head of household. Household sizes in the study area tend to be larger than in other areas where Himba reside, and household compounds are also more permanent, partly because of more reliable access to water in the Omuhonga Basin (Bollig 2006). This residential structure results in women being separated from their own kin while living in their marital household, and relying largely on the help of their co-wives, in-laws, and children for tasks that require cooperative production. The Himba are a natural fertility population. In a survey of women’s knowledge and use of contraceptives conducted in 2010 (N = 50) less than a third of women had ever heard of modern contraceptive devices, less than 15 percent had ever used them, and none of the women were currently using them. Average fertility in the current sample is 4.1 (range 0–11, N = 120), but this is likely a slight underestimate, as there are a significant number of stillbirths and early infant deaths, which may be underreported. In a subsample of women for which number of child deaths is more certain (as a result of repeated interviews and checks with spousal interviews), fertility is 4.3 (range 0–10, n = 47). Because I aimed for a sample that was older, in order to include more postmenopausal women (see
102
B. A. Scelza Table 1 Demographic characteristics of the sample population (N = 106)
Mean age (min, max) Mean number of children Current marital status Married Widowed Divorced Never married Number of co-wives 0 1 2 3+ Co-wife order 1 2 3 4
Married (n = 59)
Unmarried (n = 47)
43.28 (19, 84) 4.32
52.68 (20, 88) 3.47
59 (100%) — — —
20 (42.5%) 20 (42.5%) 7 (15%)
19 (32%) 25 (42%) 11 (19%) 4 (7%) 15 (38%) 21 (53%) 3 (7%) 1 (3%)
Table 1), the age distribution of women’s children is also skewed older. For example, the women had on average 1.25 (0–8) daughters and 1.48 (0–7) sons over age 10, and only 0.53 (0–4) daughters and 0.70 (0–4) sons under 10 (ranges in parentheses). Co-wife relations among the Himba have yet to be described in detail. However, an account of Herero marriage published in the 1950s is still largely relevant given the historical and phylogenetic history of the two groups (for details, see Malan 1995). Among the Herero, senior wives are given some ritual and social importance in the household, such as tending to the ritual fire and seniority in milking the holy cows among the herd (Gibson 1958). Senior wives play a critical role in choosing and approving of junior co-wives. The children of senior wives also have primary inheritance rights. Wealthy elder men may have upwards of four to five wives, whereas younger and/or poorer men typically have only one. Because of the high divorce rate and differentials in the ages of husbands and wives, almost all men marry at some point in their lives. Women’s work includes water and firewood collection, grinding maize, milking, souring the previous day’s milk, cooking, and child care. The husband’s livestock are partitioned between wives, and each wife keeps the milk from the cows allotted to her. Women also often own small stock (sheep and goats), and occasionally cows. During the rainy season, women spend large parts of the day in the gardens: planting, harvesting, and performing maintenance and pest control. Gardens are passed matrilineally and may be worked by multiple women within the same family. Many tasks are done collectively, if not cooperatively. Food, firewood, and water may be shared across the household (particularly between kin or co-wives who get along well), and divisions of labor occur in which one woman will watch young children while the other goes to collect water. Children participate in laborious activities from a young age and also help with child care. Boys typically herd small stock, while girls remain in the compound and assist women with their work.
Effects of Offspring and Kin on Co-Wife Satisfaction
103
Methods Sample All adult women currently residing in 24 compounds in the Omuhonga Basin were originally included in this sample (n = 119). Four women declined to be interviewed. One had given birth the night before and was resting. Another two were very elderly. A fourth stated that she was too busy and could not be accommodated at another time. An additional nine women who were interviewed were not asked questions about polygyny. These exclusions occurred because the woman had her co-wife or other affines in close vicinity and full confidentiality could not be guaranteed. This resulted in a final sample of 106 women (Table 1). All interviews were conducted in Otjihimba with the assistance of a local translator. Semi-Structured Interviews Women provided marital and reproductive histories, including number of children, sex and age of each child, current marital status, number of co-wives and co-wife order, and whether each marriage was a “love match” or an arranged marriage. Ages of women and their children were determined by converting traditional Himba year names into numeric calendar years (for methodology, see Scelza 2011a). Women were also asked about their perceptions of polygyny. First, a binary question was asked: “Do you think having a co-wife is a good thing, or a bad thing?” Next, women were asked to list both what co-wives were good for and what kinds of things co-wives fight about. While descriptive data will be presented on women’s responses to the question of what co-wives are good for, only data on co-wife conflicts will be analyzed here. Multiple responses were allowed in both types of free list, and free-response elaborations were recorded, translated, and transcribed. Both married and unmarried women were interviewed in order to understand how perceptions of polygyny differ according to current marital status. Analysis Logistic regression models were constructed to identify variables that were significantly associated with women’s perception of polygyny. The first dependent variable was constructed from a question asking whether women thought polygyny was a “good thing” or a “bad thing.” Key predictor variables were number of sons and number of daughters, broken out into four variables: number of sons under age 10, number of daughters under age 10, number of sons aged 10 or older, and number of daughters aged 10 or older. Ten years was used as a cutoff, as these are the children able to regularly perform productive labor similar to that of adults. In particular, I was interested in labor that was substitutable for women’s work, such as carrying water or firewood, which requires a moderate level of physical strength. Married and unmarried women were analyzed separately. The first analysis, which includes only married women, contains known influences on perceptions of polygyny, including co-wife order, number of co-wives, and women’s age. In the second analysis of unmarried women, covariates related to polygyny were not included, but a variable tabulating the number of female kin over age 10 living in the household was included. Standard errors were clustered by compound to account for overlapping influences within the household. A second set of regressions looked at co-wife conflicts. These were broken into two dependent variables. The first was conflict over resources. Women were coded as noting
104
B. A. Scelza
resource conflict if they gave any response that indicated co-wives would fight over the allocation of goods provided by the husband, the distribution of household labor (e.g., one co-wife is lazy, or there is a lack of reciprocity in turn-taking), or child care (e.g., a cowife refused to take care of a child when the other wife was away from home). The second dependent variable was conflict over sex. Here women cited sexual jealousy either because the husbands spent more time with one wife or because he liked one wife better than the other. The same predictor variables as described previously were used in these analyses, again with married and unmarried women analyzed separately. All analyses were done using Stata 12.0. Informed oral consent was obtained from all participants, and this study was approved by the University of California, Los Angeles Institutional Review Board (10-000238).
Results Himba women have generally positive perceptions of polygyny, with 74 percent of women reporting that they thought having a co-wife was a “good thing.” While the majority of women in all marital groups expressed this sentiment, married and widowed women were more likely to think positively about polygyny (76% and 80%, respectively) than divorced (65%) and unmarried (57%) women. However, there were no statistically significant differences between the groups (χ 2 = 2.37, df = 3, p = .498). Among married and widowed women, those with co-wives were slightly more likely to have a positive perception of polygyny (with 79% responding that it was a “good thing”) than were monogamously married women (74%). Among married women, there is a strong negative relationship between number of daughters aged 10 or older and perception of polygyny (Table 2). The more older daughters a woman has, the more likely she is to have a negative view of polygyny. The number of daughters under age 10 and the number of sons (either over or under age 10) had no effect on perception of polygyny. Number of co-wives had a positive and significant effect; having more co-wives was associated with a more positive perception of polygyny. However, cowife order had no significant effect. Among unmarried women, none of the child categories Table 2 Odds ratios (95% confidence intervals) of perception of polygyny as predicted by logistic regression: Effects of kin on married and unmarried women
Current age Number of daughters (
