Marriage and Family
Families are essential for human development. Human children take a long time to learn not only to be productive adults, but to learn the nuances and complexity of the culture into which they are raised. Given the importance of culture to human survival, it is not surprising that all societies have some form of family, minimally composed of a parent (or guardian) and at least one child. Almost all societies also have the custom of marriage, so a marriage partner is usually part of the basic family unit. Marriage is commonly defined as a socially approved sexual and economic union, presumed to be more or less permanent, and entails rights and obligations between the married couple and any children they might have (C. R. Ember, Ember, and Peregrine 2019).
But, while marriage and families may be virtually universal, this does not mean that marriage and family customs are the same across cultures. Indeed, there is tremendous variability in almost every aspect of marriage, from the rules surrounding whom one can and cannot marry, how marriage is arranged, how couples get married, how many people can be married at a time, how long marriages last, and what conditions allow a marriage to be dissolved. And families vary in size and composition from a single-parent unit to very large multi-generation families and households. In this module, we concentrate on marriage and family customs involving at least one man and one woman because although homosexual behavior is not rare, same-sex marriage is rare cross-culturally. As discussed in the sexuality module, homosexual behavior, even if common, often coexists with heterosexual marriage practices.
Universality of Marriage
There are a couple of known exceptions to the claim that marriage is a universal custom. One exception from the past is the Nayar of the 19th century. The Nayar were a subcaste in India. A more recent example is the Mosuo or Na of Yunnan in southwest China. In both cases, not only did male and female sexual partners live separately with their maternal families, but they did not have regular economic cooperation nor other regular obligations with their sexual partners. And, in both cases, males were frequently absent. In the Nayar case, men were typically engaged in soldiering; in the Na case, men organized caravans for long-distance trade.
Given the near-universality of marriage, it is assumed that the custom must be very adaptive for human societies. There are a variety of theories about what particular problems make marriage adaptive. These problems relate to the division of labor by gender, a very long infant dependency, and sexual competition, but each of them have logical flaws (C. R. Ember, Ember, and Peregrine 2019). A division of labor by gender is a cultural universal and it is proposed that marriage is a way for females and males to productively share the results of their varied subsistence tasks. However, marriage is far from a necessary solution to the sharing problem. Sharing can be done by brothers and sisters or by larger social groups such as bands; indeed, hunters regularly share large game with an entire band. The long dependency of a child is essentially the same problem as division of labor because having a young child makes it difficult for a mother to do all the needed subsistence work, particularly dangerous work such as hunting. But as already discussed, sharing can be accomplished in other ways. Finally, the sexual competition argument points to the greater conflict among human males as compared to non-human primates. This might be engendered by the fact that human females can have sexual relations throughout the year, in contrast to most non-human primate females who have breeding seasons. However, there are also logical flaws to this argument. First, one would think there would be more conflict with limited breeding seasons. Second, other sexual rules could be invented, such as rules regulating rotation of partners.
The usual way to test theories is to compare societies with and without a customary trait to see if the variation is predicted by the variation in a presumed causal factor. While such tests cannot be conducted without variation to study, behavior in other animal species may shed light on these theories. Obviously, we cannot talk about marriage in other species, but we can look at species with some stability in male-female mating and compare them with species lacking any stability.
In a comparative study of 40 mammal and bird species (M. Ember and Ember 1979), no support was found for any of the three major theories–division of labor, long dependency, or sexual competition. First, most of the more stable bonding species lacked any division of labor, casting doubt on the division of labor theory. Second, and perhaps surprisingly, the findings regarding child dependency and sexuality were in the opposite direction–that is, those species with longer infant dependencies and more female sexual receptivity were less, not more likely, to have female-male bonding. The Embers did put forward a new theory which did predict male-female bonding well. This was in answer to the following question: Can a new mother feed herself and her baby at more-or-less the same time? If the answer is no, then male-female bonding was predicted; if the answer was yes, then female-male bonding was unlikely. This theory may explain why most bird species have bonding. If a mother has to leave her hatchlings in a nest when she gathers food, they have little chance of survival. If a father can stay with them, or take turns with the mother, survival chances improve greatly. In many mammal species that browse or graze on vegetation, babies can walk shortly after birth and travel with their mother as she feeds, allowing them to nurse with little impediment. Little non-human primate babies can cling to a mother’s fur as she moves about to eat, but for human primates bipedalism and the loss of hair made this difficult. Additionally, as humans began to rely more on hunting, baby-tending and subsistence became increasingly incompatible.
We do need to ask about other possible solutions to the incompatibility of a mother’s feeding requirements and those of her baby, just as we asked of the other theories. Why couldn’t two women cooperate? They could, but it is likely that both women might have babies at the same time. But more importantly, two women are likely to have twice as many mouths to feed and care for. And neither woman could likely provide game through hunting. What about brothers? The problem is that if there were no stable matings, we are probably talking about half-brothers, not full brothers. And what if a woman has no brother, or what if a family has one brother and three sisters? In contrast, a man and woman in a more-or-less stable union not only share one set of children, but have no conflicts regarding other children to support. Hence, it is more efficient and economical for the man and woman who share the child (or children) to cooperate.
Social Recognition of Marriage
If marriage is a socially-accepted union, societies need a way of recognizing that union. But recognition does not necessarily mean an elaborate ceremony. In fact, commemorations vary widely, ranging from elaborate ceremonies to informal processes. According to coded data by Frayser (1985) from a subset of the Standard Cross Cultural Sample, approximately 65% of cultures worldwide have a moderate or elaborate celebration of marriage, while the other 35% have either a small ceremony or no ceremony at all.
Among the societies with more elaborate ceremonies are the Hopi, of the U.S. Southwest, who practice a wedding ceremony that lasts four days. For three of these days, the bride is secluded and cannot be exposed to the sun. During this time she mainly grinds corn, signifying a symbolic payment of service to the groom’s mother for her son. The neighboring villagers in the community bring presents for the bride for these three days, and then the groom’s paternal aunts engage in a ceremonial “fight” against the bride with mud, which is blocked by the groom’s maternal kin. Finally, the last day consists of a ceremonial washing of the bride and groom to signify “the washing away of all ‘remaining traces of youth.’” The washing is followed by the couple’s hair being tied into a single knot (Geertz and Lomatuway 1987).
In other cultures, social recognition of marriage is dependent on a successful trial period. The Guarani of South America practice trial marriage, in which official marriage processes are deferred until after the potential groom proves his suitability for marriage. Schaden and Lewinsohn (1962) tells us that
Many men have a ‘companion’ before marriage. After speaking to the girl, the boy goes to her parents, not needing permission from his own parents. He takes his companion to the paternal home, where he lives with her for a while and where she cooks together with the boy’s mother. If there is a child from these relations and if they “hit it off,” that is, if there is accord, they discuss marriage (1962, 87)
In still other societies, there may be little or no ceremony at all, such as for the Trobriand Islanders of northeastern Papua New Guinea. Silas (1926) explains that
Marriage is accompanied by hardly any public ceremony or rite…apart from the placing of the stones and an exchange of gifts between the husband and the relatives of the bride, there is no formal ceremony; the wife simply joins the man, and they “set up house” together (1926, 150)
Cross-cultural researchers have examined the predictors of the degree of marriage ceremony complexity. Here are some of the main findings. Societies with more complex celebrations tend to have
Greater social involvement in the marriage (Frayser 1985).
Why? Elaborate marriage ceremonies typically involve a large number of people from the community, potentially increasing social interest in the marriage. Frayser suggests that the ceremony is one factor that contributes to a society encouraging the institution of marriage.
Why? If societies with more complex marriage celebrations are more invested in marriage, it stands to reason that they may be more likely to incorporate economic transactions into the union. Such transactions, such as bride price or dowry, reflect higher involvement in the establishment of a marriage.
Greater sexual regulation of women (Frayser 1985).
Confinement of women’s reproductive potential to one man (Frayser 1985).
A strict taboo on premarital sex for women (Kitahara 1974).
Why? Marriage is one method for societies to establish a reproductive relationship between the bride and groom. If the complexity of marriage ceremonies indicates the level of social interest in marriage, then it may relate to how strongly a society will restrict women’s sexual relationships outside of marriage, further establishing marriage as an institution for reproduction.
Extramarital sex or reproductive issues as grounds for divorce (Frayser 1985).
Why? Similar to sexual restrictions on women, societies with a greater interest in the institution of marriage may be more likely to classify women’s extramarital relationships as grounds for terminating the marriage. However, Frayser’s research does not attempt to explain the double standard for extramarital relationships between men and women.
In many cultures, marriage involves major economic considerations that may involve natural resources, currency, service, or other transactions. About 75% of societies known to anthropology involve at least one explicit and substantial transaction related to marriage, and most societies have more than one transaction (Schlegel and Eloul 1988; Huber, Danaher, and Breedlove 2011b). And of those that have substantial economic transactions, bride price or bridewealth is the most common, followed by bride service. Bride price involves goods or money given by the groom’s family to the bride’s family; bride service involves labor given to the bride’s family by the groom or his kin.
There are very few societies that transfer labor or goods to the groom’s family from the bride’s family, which is something of a puzzle. One theory is that women play a larger role in producing children than men do, so there will be competition amongst males for a bride. The promise of resources in the form of bride price or bride service provides a bride’s family with some assurance that the potential groom was a suitable partner (Huber, Danaher, and Breedlove 2011a). Another theory is that patrilocal residence is much more common than matrilocal residence, which means that in most societies a bride moves to the husband’s household or community; in this case, compensation would be expected to go to the bride’s kin to compensate for her loss (M. Ember 1970; Huber, Danaher, and Breedlove 2011b). The following findings are consistent with these theories:
The more likely that marital residence is with or near the husband’s kin, the more likely there is compensation to the bride’s kin (M. Ember 1970). Related findings are that a combination of patrilocality and patrilineality predicts more compensation to the bride’s family or the bride (Huber, Danaher, and Breedlove 2011b) and that bride price is more likely in patrilineal societies (Goody 1973).
Note that bride price transfers “wealth” horizontally within the parental generation. As long as families have daughters and the daughters marry, every family will receive bride price and pay out bride price when their sons marry. It does not create economic inequality between families. Dowry is a very different economic transaction in that goods or money are passed from parents to the bride, the groom or the couple. (In contrast, bride price and bride service are directed at the bride’s kin, not to the bride or the new household). In other words, with dowry there is a downward passage of wealth. Another type of dowry is indirect dowry. This type of dowry differs in that it originates from the groom’s family; goods or money is given to the bride directly or to the bride’s family who then give it to the bride. Alternatively, indirect dowry can be thought of as a combination of bride price paid first and then dowry.
Equal exchanges between kin groups also occur. In gift exchange, the kin groups of the bride and groom give each other gifts of about equal value. Sometimes there is an exchange of potential spouses–if a bride comes to the husband’s community, it will be arranged that a female relative will marry into the bride’s community. Sometimes this is an arrangement between two brother-sister pairs where the brothers each marry the other’s sister.
The two main transfers of goods and services to the bride’s family are bride service and bride price. Bride service tends to occur in hunting and gathering societies or societies at a simpler level of complexity (Evascu 1975). On the other hand, bride price
is associated with higher female contribution to subsistence (Schlegel and Eloul 1988).
is associated with bride theft (Ayres 1974).
Why? Substantial transactions make it more difficult to marry which increases the impetus to practice stealing a bride (although sometimes this is with her approval).
Dowry is more likely
in complex societies that have social stratification and intensive agriculture (Schlegel and Eloul 1988; Evascu 1975), although such societies are also likely to lack substantial transactions at marriage (Evascu 1975).
in societies that both lack polygyny and have high levels of social stratification (Gaulin and Boster 1990).
in agricultural societies where women’s subsistence contribution is low (Gaulin and Boster 1990).
Customs Regarding Whom One Can or Cannot Marry
All societies have at least one rule about whom one cannot marry, namely, the prohibition on marriage to brothers or sisters or parents (the incest taboo). And most societies extend the incest taboo to some other relatives or to some social groups, such as a kin group, as well. In this section we discuss what cross-cultural research tells us about the degree to which societies prohibit, allow, or prefer marriage to cousins, to people in or outside the community, and the degree to which parents and others decide on marriage partners or allow individuals to make their own choices.
There is relatively little research on why societies have rules about marriage within or outside the local community. Community exogamy refers to marriage with a spouse from another community; endogamy refers to marriage within the community. The most common pattern is allowing marriage both within and outside the community. Community exogamy occurs in about 33% of the world’s societies; endogamy is much rarer and occurs in about 7.5%1 In one study based on modeling and cross-cultural analysis, exogamy was predicted by
Small size communities and greater variation in rainfall (Dow, Reed, and Woodcock 2016).
Why? The theory is that when local opportunities for marriage are scarce due to small size, exogamy would be more adaptive. Rainfall variability is likely to produce uneven productivity between communities and exogamy provides more opportunities for movement to even out resources.
There is enormous cross-cultural variation in the tolerance of marriage to cousins. Some societies frown on cousin marriage so much that even distant cousins are forbidden. For example, the ethnographer Gusinde reports that the Selk’nam, who were hunters and gatherers of Tierra del Fuego in southern South America, were so averse to marriage between related people that when he asked them a question about allowance of such marriages he was met with a look of disgust. He tells us, “When I asked more specifically whether the children of first cousins were allowed to marry, people gave me a decided no and added in an indignant tone: “Do not ask further; blood relatives may not be united!” (Gusinde 1931, 488–89). At the other extreme, there are societies that not only tolerate first cousin marriage, but actively try to promote marriage to a cousin, usually a particular type of cousin. The Komachi, pastoral nomads of southern Iran, preferred marriage with kin who were no more distant than second cousins and about 70% were with first cousins (Bradburd 1990, 115). Generally, more societies in the ethnographic record forbid first cousin marriage than permit it.
What accounts for the allowance of first cousin marriage?
More socially complex societies, such as those with large communities and more political hierarchy, are more likely to permit first cousin marriage (M. Ember 1975).
Why? Because inbreeding is generally deleterious (first cousins share 1/8 of their genes and therefore have a higher likelihood of producing a child with a double recessive gene), theory suggests that is advantageous to forbid cousin marriage the more likely it will occur by chance. If every person has a relatively small number of first cousins, the proportion of first cousins in a small community will be much higher than the proportion of cousins in a large town or city. Therefore allowing some cousin marriage is not as problematic for overall reproduction of the group.
Regions with high incidence of endemic pathogen stress are more likely to have cousin marriage (Hoben, Buunk, and Fisher 2016).
Why? Although this finding seems contrary to the principle that high levels of inbreeding are generally deleterious, there are some pathogens for which recessive genes have been shown to lessen the serious consequences of the disease. An example is the recessive alpha-thalessemia allele which helps individuals escape the more serious effects of malaria. Note that the relationship between pathogen stress, although significant, is very weak.
Relatively small societies that have recently experienced severe loss of population due to introduced disease and that are also relatively small (more than 1,000 but less than 25,000 in the society) are more likely to allow first cousin marriage (M. Ember 1975).
Why? While first cousin marriage may have higher risk of offspring carrying harmful double recessive genes, severe population loss makes it more difficult to find eligible mates in societies that are already relatively small. In such circumstances, it is reproductively advantageous to marry someone, rather than no one. Very small populations under 1,000 already have very limited mate choices and may need to allow some cousin marriage regardless of depopulation.
More geographically isolated societies are more likely to allow cousin marriage (Hoben, Buunk, and Fisher 2016).
Why? Just as it may be difficult to find a non-related mate in a very small population, it may also be difficult in a geographically isolated population to find a mate without allowing marriage to cousins.
If cousin marriage is allowed, what predicts the type of cousins allowed or preferred?
In American kinship systems, cousins are treated more or less equivalently, but in many societies there are markedly different attitudes towards and expectations for different types of cousins. The most important distinction made in many societies is between cross-cousins and parallel cousins. To understand what a cross-cousin is versus a parallel cousin, it helps to think of the term “cross” as meaning related to you by “crossing” gender in the parental generation. Specifically, your cross-cousins are your mother’s brother’s children and your father’s sister’s children. Parallel cousins are related to you through the same gender–mother’s sister’s children and father’s brother’s children. In the diagram below, a circle stands for a female, a triangle for a male, and an equal sign for a marriage. Children from a marriage are symbolized by a downward line from the marriage. Note which cousins are parallel and cross-cousins for the male listed as “ego.”
This distinction mostly matters in societies with patrilineal or matrilineal descent because in such societies one set of parallel cousins is in your own kin group, while cross-cousins are generally not. So, in a patrilineal society, children will be in the same kin group as their father, their father’s brother, and their father’s brother’s children. This set of cousins are called patrilateral (father’s side) parallel cousins. Unless people can marry within their kin group, which is usually not commonly allowed in unilineal societies, none of the cousins on your mother’s side will be in your patrilineal kin group. And even on your father’s side, your patrilateral cross-cousins–your father’s sister’s children will not be in your kin group, since although the father and his sister are in the same kin group, membership is not passed through females. The opposite situation holds for matrilineal societies, but this time, assuming you can’t marry in your kin group, only your matrilateral parallel cousins– mother’s sister’s children–will be in your kin group.
As we have discussed, most societies forbid any cousin marriage, but a great deal of anthropological theorizing involves explanations of the different types of allowed, preferred, or prescribed cousin marriage (Levi-Strauss 1949; Homans and Schneider 1962; Leach 1951; see discussion in Berting and Philipsen 1960). Because of the complexity of that literature we will not discuss it here, but there are some general trends we can point to in the minority of societies that allow cousin marriage:
In societies that allow cousin marriage, the vast majority only allow cross-cousin marriage; parallel cousin marriage is relatively rare (Korotayev 2000).
Of those societies allowing cross-cousin marriage, symmetrical cross-cousin marriage (with both sets of cross-cousins) is much more common than asymmetrical cross-cousin marriage (one side of the family). One estimate is 70% of societies allowing cross-cousin marriage allow symmetrical compared with 30% for asymmetrical cousin marriage (Coult 1965).
Symmetrical cross-cousin marriage is somewhat more likely when economic transactions at marriage (such as bride price or dowry) are absent.
Why? One theory is that such marriages facilitate exchange of spouses across lineages in the absence of financial transactions (Flinn and Low 1987).
Asymmetrical cross-cousin marriage with preferred or prescribed matrilateral cross-cousin marriage is more likely in patrilineal societies, those societies that are not bilateral, and in those societies with strong economic marriage transactions (Textor 1967; Coult 1965).
Preference for father’s brother’s daughter marriage (patrilateral parallel cousin marriage) preference is more likely in patrilocal societies (Flinn and Low 1987).
Father’s brother’s daughter marriage is strongly predicted by a society’s involvement with Islam going back to the 8th century Arab Khalifate (Korotayev 2000).
Why? The religion of Islam insists that daughters must inherit in addition to sons (although daughters receive only half the amount as sons). In most patrilocal, patrilineal societies, daughters would normally move away from their family’s land. But since women in the Arab Khalifate region are typically secluded, women would be unlikely to be able to cultivate their portion of land, leaving it to be controlled by her husband. This would result in a patchwork of small plots of land controlled by different groups. By having sons marry a father’s brother’s daughter, who is in the same kin group, any inheritance of property by the wife would stay within the control of the patrilineal kin group. Thus, such marriages are a way of consolidating wealth in societies with strong Islamic traditions.
Arranged marriage or individual choice
Mate selection takes various forms but generally falls on a continuum from full individual choice to marriages fully arranged by parents or other relatives. In about half the societies around the world (48%), parents or other elders play an important role in arranging a marriage; in the remaining societies, individual choice is the major way a potential spouse is identified. However, sometimes parents still have to approve the choice. Only 31% of societies have full individual choice.2
Hunter-gatherers in recent times overwhelmingly have arranged marriage (Apostolou 2007), strongly suggesting that it was probably the ancestral pattern in human history (Apostolou 2014). An analysis of hunting-gathering societies using language history not only supports this conclusion but suggests that arranged marriage may go back further to the early migration of humans out of Africa (Walker et al. 2011).
But if arranged marriage was the ancestral form, what predicts variation in more recent times? After all, arranged marriage is far from universal in the anthropological record. One important predictor is whether or not women engage in craft specialization or work outside the household–when they do so, there is a lower likelihood of arranged marriage (Hull 1989). This finding is consistent with anecdotal evidence that arranged marriage customs tend to break down with modernization.
We also know that societies with arranged marriage societies are more likely to have
lower rates of premarital sex (Apostolou 2017).
negative attitudes towards female premarital sex (Apostolou 2017).
Why? Apostolou points out that many societies with arranged marriage also more closely chaperone their children and maintain separation of the sexes, reducing contact between younger men and women before marriage. Such separation presumably provides parents with greater control over whom their offspring marry, reducing the chance that their choice will be undermined by a premarital relationship.
low emphasis on love as a basis for marriage (Hull 1989).
Why? Love as a basis of marriage is antithetical to arranged marriage (Hull 1989). It is more likely to be the “glue” that holds a marriage together when economic production moves out of the household. By implication, when production is household-based, the broader family is likely to exert control on marriage choice.
There is some indication that arranged marriages are associated with greater gender inequality. First, there is a double standard, with more tolerance of premarital and extramarital sex for men than for women. Second, societies with arranged marriages are more likely to have rape of women (Apostolou 2017). Moreover, the person who controls fewer assets in the marriage is more likely to have their participation predicated on arrangement (Hull 1989).
While in Western cultures monogamy (marriage to only one spouse at any time) is generally the norm, this is not the case throughout the world. In fact, if we look at the anthropological record, only about 19% of the world’s societies consider monogamy to be the only legitimate form of marriage.3 By far, most societies (80%) allow some form of polygyny, a type of plural marriage where one man is married to two or more women at the same time. Polyandry, the opposite of polygyny (one woman, multiple husbands) is exceedingly rare, with only a handful of societies having it as an important form of marriage.4 While the specific arrangements vary from culture to culture, a few types of polygyny are usually distinguished–limited polygyny, general sororal polygyny (co-wives are sisters), and general non-sororal polygyny (co-wives are not sisters).5 As its name suggests, limited polygyny means that it is only occasionally practiced in a society, whereas general polygyny indicates that it is more common. However, it is important to understand that even where polygyny is generally practiced most men will not have more than one wife at any given point in time. Some men will be unmarried, some men will only have one wife, and some will be currently monogamous but marry polygynously later. Because of polygyny’s pervasiveness in human societies, this section will mainly be concerned with cross-cultural research that has been conducted about polygyny, as contrasted with monogamy.
Many anthropological accounts report jealousy among co-wives. Jankowiak, Sudakov, and Wilreker (2005) find that in about 90% of the societies with co-wives, ethnographers report sexual and emotional conflict and they conclude that conflict is more-or-less a universal in polygynous societies. Polygynous societies appear to have responded to jealousy with certain common customs and rules that include: having separate quarters and kitchens for nonsororal co-wives, having rules about husbands sharing resources equitably, and having rules for sleeping with co-wives in rotation (C. R. Ember, Ember, and Peregrine 2019, 245). Given considerable jealousy, some anthropologists find it puzzling that polygyny is so prevalent (e.g., M. Ember (1974)).
Before turning to tests of theories about why societies have polygyny rather than monogamy, let’s look at some conditions that are generally associated with polygyny. Polygyny is generally associated with
Moderate levels of societal complexity as indicated by (Osmond 1965)
high dependence on animal husbandry (also Ross et al. 2018).
communities with 100-1,000 residents.
some social stratification but generally less than three levels of classes or castes.
nomadic or dispersed settlements.
minimal state-level organization.
Male-oriented social systems
Polygynous societies, particularly those with nonsororal polygyny, tend to be found in societies practicing patrilocal residence (Whiting 1964; White and Burton 1988) and also those having male-biased inheritance (Hartung 1982; Cowlishaw and Mace 1996). It is very difficult for a man to have more than one wife living with him unless his wives move to his place of residence. (Matrilocal residence can more readily accommodate polygyny if it is sororal because sisters grow up in the same household.)
Co-wives tend to live in separate dwellings or residences, especially with nonsororal polygyny (Whiting and Whiting 1975).
Polygyny is one of the most widely researched topics in cross-cultural research. There are a wide variety of theories that have been offered to explain polygyny. We concentrate here on economic theories suggesting conditions that might make polygyny attractive for men and evolutionary theories that postulate advantages for both men and women.
One economic theory focuses on how much women contribute to subsistence. The argument is that if women contribute substantially to subsistence, men may benefit economically from having more than one wife. Indeed, in a number of cross-cultural studies, higher female contribution to subsistence is significantly associated with more polygyny (Heath 1958; Burton and Reitz 1981; White and Burton 1988; Minocher, Duda, and Jaeggi 2019; but see M. Ember 1984). However, this overall finding comes with some qualifications. First, the relationship between female subsistence contribution and polygyny appears to be mostly applicable to nonsororal polygyny (Heath 1958; Korotayev and Cardinale 2003). Second, the relationship appears stronger in societies with simpler forms of agriculture (Osmond 1965). Note that the economic contribution argument fails to explain any advantages for women to be in polygynous marriages, nor does it explain why the opposite form of marriage–polyandry–is not generally found where men do most of the subsistence work.
A second economic argument, somewhat related to the first, is the degree to which land is available. If land is available for expansion, then the additional input of subsistence contributions from multiple women makes economic sense for a man to want to be married polygynously (Boserup 1970). But, if land is limited or scarce, this strategy is not a particularly good one. Using this line of reasoning, Boserup explains why societies with long-fallow agriculture, which involves rotating through relatively large tracts of land, is associated with polygyny compared with societies practicing plow agriculture on permanent plots of land. Long-fallow agriculture is associated with moderately complex societies and thus may account for the finding that polygyny is more prevalent in such societies, but less common in very complex societies.
Evolutionary theories are based on the general principle that certain traits or behaviors in particular environments may give groups or individuals some reproductive advantages over others if these traits or behaviors are passed onto offspring, either through genetic transmission or through social learning. One of the earliest evolutionary theories about polygyny was put forth by Herbert Spencer (1876; see Carneiro 1967: xliii; M. Ember 1974; M. Ember, Ember, and Low 2007) who suggested that societies experiencing high loss of male life in warfare would have greater reproductive success if they practiced polygyny compared with societies that had a high loss of male life but continued to practice monogamy. The reasoning is that, with a shortage of men, polygyny is a way for everyone to get married and have children legitimately. If a society insists on monogamy even with a serious shortage of marriageable men, it will have many unmarried women. Obviously, women can have children without being married, but unmarried women generally face greater economic difficulties when trying to raise children. Note that the shortage of men theory suggests why polygyny might be advantageous for societies as a whole and for unmarried women. In support of the loss of males in warfare theory, general polygyny is found to be more likely with
Why? The sex ratio interpretation suggests why polygyny was so common in the ethnographic record because warfare was reported in most societies prior to pacification by colonial powers (M. Ember 1974). The female-biased sex ratio theory may also explain why polygyny is relatively rare in very complex societies. Complex societies tend to ave specialized or standing armies which means that male mortality in war may only be high among men in the armed forces (M. Ember, Ember, and Low 2007). In contrast, in socially simpler societies, the entire population of able-bodied men may be expected to be warriors.6
A sex-ratio favoring females can also be created by having older men marry much younger women (M. Ember 1984). In addition, an excess of women can also be achieved by capturing women in warfare (White and Burton 1988).
Indirect support for the sex-ratio hypothesis comes from research on the presence of polyandrous marriages. While they are rarely the typical form of marriage in a society, societies practicing polyandry to some degree are more likely to have an opposite sex ratio favoring males rather than females (Starkweather and Hames 2012 – see polyandry section below).
Why? A high male mortality in warfare may be the main cause of a sex-ratio favoring females. M. Ember (1984) notes that both a high male mortality in warfare combined with a discrepancy in age of marriage predict general polygyny more strongly than either condition alone. Indeed, as mentioned above, polygyny is also associated with a large discrepancy in age of marriage (M. Ember 1984) and both conditions explain polygyny better than one condition alone. Ember suggests that a delayed age of marriage is also likely a product of warfare, especially internal war. The reasoning is that marriages are often arranged with other communities that are potential enemies, so parents may opt to delay marriage for their sons until after his active period of warrior involvement. Other causes of sex ratio imbalance in favor of females include migration of men for work or trade.
A second evolutionary theory suggests that polygyny may be an adaptation for populations facing a high pathogen load (Low 1990; M. Ember, Ember, and Low 2007). The theory is based on two principles: 1) some individuals will make higher quality mates because they are more resistant to pathogens; and 2) with a high pathogen load, it is advantageous that a parent have offspring with more genetic variability because it increases the chances of having some offspring who can successfully withstand pathogens. But why polygyny? Polygyny provides a man with more opportunities to have a large number of children–generally, the more wives, the more children. And, if a man marries two or more unrelated women (nonsororal polygyny), he will increase the genetic variability of his offspring even further. Therefore, reproductively speaking, it is to a man’s advantage to marry multiple unrelated women. Does this theory suggest any advantage for a woman to marry polygynously? Yes. Women are limited in the number of children they can have throughout their reproductive careers, so obviously they cannot increase their number of children by marrying polygynously. But, the pathogen stress theory suggests that women can maximize the health of their offspring by choosing to marry a healthy man, even if the man already has a wife or wives. The following findings support pathogen theory:
Why? Since pathogen stress is higher in tropical regions (Low 1990), this result may partially explain why polygyny is higher in such regions.
Research conducted by M. Ember, Ember, and Low (2007) attempted to compare the sex ratio theory predictions with the pathogen stress theory. Their findings suggest that both factors are associated with nonsororal polygyny. However, they found that the complexity of a society introduced a qualification, namely that
High mortality rate predicts nonsororal polygyny in non-state societies, but not in state societies
Why? State societies are likely to have less male mortality because fighting forces tend to be specialized; therefore male mortality is less likely to be an important factor.
Pathogen stress is the only predictor of nonsororal polygyny in state societies
Why? Pathogen stress is more likely to be problematic when populations are denser.
Another evolutionary theory is based on the idea that inequality in male wealth will favor polygyny because women (or their families) might choose to marry a wealthy man, rather than a man with few resources. Research has found that
Higher wealth inequality predicts more polygyny amongst foragers, horticultural and pastoral populations (Ross et al. 2018; Minocher, Duda, and Jaeggi 2019), but not in societies with intensive agriculture. Societies with the highest degree of social stratification have less polygyny (Minocher, Duda, and Jaeggi 2019), which is contrary to the idea that male inequality generally will favor polygyny.
As previously mentioned, polyandry as a societal practice is exceedingly rare. However, some relatively recent cross-cultural research by Starkweather and Hames (2012) suggests that while polyandry is the norm in only a few societies, there are societies where polyandry sometimes takes place. In fact, in a survey of societies in eHRAF World Cultures, they found 53 societies outside of the “classic” polyandry areas (northern India, Nepal, Tibet and the Marquesas) that appear to practice either informal polyandry or formal polyandry. They call these instances “non-classical” polyandry. Informal polyandry is where two or more men are recognized as “fathers” and provide some help to the same woman and her children. Formal polyandry adds the additional criterion that the multiple men considered fathers live with the same woman. Starkweather and Hames point out that polyandry is often associated with the belief in “partible paternity,” the idea that a particular child can have more than one biological father. Their main findings are
non-classical polyandry tends to be found in small scale egalitarian societies that practice food collection and horticulture.
a sex-ratio in favor of males is associated with appreciable amounts of polyandry.
Why? Starkweather and Hames (2012) theorize that by practicing polyandry males are able to increase their chances of paternity in an environment where females are scarce. Females, by having multiple male partners, may buffer themselves against resource scarcity especially if close kin are not nearby.
Divorce, or the severing of marital ties, is found in the vast majority of societies in the anthropological record. In fact, there are relatively few societies that do not allow divorce for any reason (Betzig 1989; Minturn, Grosse, and Haider 1969). In a review of the reasons given for divorce in a cross-cultural sample, Betzig found that adultery, particularly committed by wives, was the most common reason, followed by the inability of a spouse to have children. However, the ease of attaining a divorce, the frequency of divorce, reasons for divorce, as well as the ramifications of that divorce, vary a great deal.
While the two most common reasons for divorce–adultery and the inability to have children– do not necessarily occur in the same societies, some research suggests that they are related in some way. Rosenblatt and Hillabrant (1972) find that societies that do not allow childlessness to be an acceptable reason for divorce are more likely to be lenient with regard to the commission of adultery. Why? The authors postulate that, in the absence of a social network system to support people when they grow old, having children is vital for parents’ future survival. It is important to note that the inability to have a child is often a function of a particular couple’s inability to have a child as a couple, not necessarily the inherent inability of a particular spouse to have a child. Adultery is one mechanism of producing offspring if a couple cannot have children.
Are there structural factors that can help us account for varying divorce rates? Although some research supports the idea that societies with matrilocal residence are more tolerant of divorce (Minturn, Grosse, and Haider 1969), other research does not find the relationship particularly strong when other factors are controlled (Ackerman 1963; Pearson and Hendrix 1979). Ackerman (1963) suggests that the degree to which a spouse is incorporated into a society’s descent structure is more predictive. Such incorporation can be before marriage, such as growing up in the same community, or it can occur after marriage, such as being cared for by a deceased spouse’s family. Ackerman (1963) finds that
in bilateral societies, the greater the likelihood that marriage takes place within the community, as opposed to marrying outside the community, the lower the divorce rate. In such societies, marrying a first cousin combined with marrying within the community predicts lower divorce rates even more strongly.
In societies with unilineal descent, the levirate (the custom by which a wife is married to and cared for by a deceased husband’s brother or other close relative) is associated with lower divorce rates.
Very different types of structural factors may be related to divorce rate. These relate to the degree to which the wife and husband are dependent upon each other for reproductive and economic success. Higher divorce rates are predicted by
more alloparental care (childcare provided by someone who is not a parent) (Quinlan and Quinlan 2007).
Why? Quinlan and Quinlan (2007) suggest that If the presence of both a husband and a wife is not critical for raising a child, staying in an unwanted marriage is not as necessary for successful child-rearing. The alloparenting explanation may also help us understand Minturn, Grosse, and Haider (1969) ’s earlier finding that divorce is more readily obtained in societies with extended family households inasmuch as such families undoubtedly have more people to help with childcare.
the higher the status of females in society (Pearson and Hendrix 1979).
Why? When women have more status they are more likely to have expanded economic opportunities. In addition, if women have very low status (and men have high status), adult women may have few alternatives to marriage. Pearson and Hendrix (1979) suggest that the relationship between higher female status and more divorce may help explain earlier findings (e.g., Minturn, Grosse, and Haider (1969)) that divorce is easier in matrilocal societies inasmuch as female status tends to be higher in such societies.
Finally, the degree to which spouses choose their marriage partners may help us understand divorce rates. Broude (1983) finds that the following predict less divorce:
permissive attitudes towards premarital sex
individual choice in a marriage partner
Why? Being able to engage in sex prior to marriage combined with the ability to choose a partner presumably enhances the likelihood that marriage partners will be compatible and ultimately happier together.
Though marriage primarily involves developing a social bond between those who are getting married, how their lives are shaped hinges heavily on the type of household they will live in. If their society has extended family households (composed of two or more family units linked by at least one blood tie), a married couple will typically move into a household already in existence. Extended family households range from small two-unit families, to large families with 3-4 generations of related individuals. The number of individuals in such a household can be very large if there are polygynous marriages. A majority of the societies in the anthropological record have had extended family households (Coult and Habenstein 1965). The minority of societies have independent family households. To be an extended family household, members need not live in the same dwelling. Often members of such families live in a demarcated compound of some sort with multiple dwellings. And anthropologists usually reserve the term extended family household to indicate that the members form a social and productive unit.
What predicts variation in family household form? Extended family households are more likely in societies where
Agriculture or fishing are dominant forms of subsistence and communities are sedentary (Nimkoff and Middleton 1960).
Why? Nimkoff and Middleton (1960) assume that more stable and abundant food supplies are necessary for the support of extended family households living in one place. They make the further assumption that agriculturalists and fishing people have more stable food supplies. As pointed out by Pasternak, Ember, and Ember (1976), the problem with this reasoning is that most hunter-gatherer societies have multi-family bands that cooperate with each other, so clearly they can support multiple families. In addition, while extended families are associated with agriculture and sedentariness, the associations are very not very strong (Pasternak, Ember, and Ember 1976).
Social stratification is present (Nimkoff and Middleton 1960).
Why? The assumption is made that societies with social classes are likely to have private property and such property would create interest in keeping land together rather than dividing it into smaller and smaller pieces.
Societal complexity is mid-range (Blumberg and Winch 1972).
Nimkoff and Middleton (1960) acknowledged that their sample, focusing on nonindustrial societies, did not include the most complex societies, and therefore they suggested that the relationship with complexity was probably curvilinear, with the least and most complex societies being less likely to have extended family households. Blumberg and Winch found support for the curvilinear hypothesis both with data from a nonindustrial sample and then also with data from a country sample. Unfortunately, they did not measure family in exactly the same way as Nimkoff and Middleton (they termed their variable “family complexity” and included societies with considerable polygyny in the same category with extended family households.
“Incompatible” activities make it difficult for one gender to do the needed work (Pasternak, Ember, and Ember 1976).
Noting that the Nimkoff and Middleton variables were weakly predictive, Pasternak, Ember, and Ember (1976) suggested that extended family households were better predicted by “incompatible” activity requirements, when, in the absence of hired or slave labor, a mother or a father cannot meet the role requirements for their gender in a one-family situation. For example, an incompatible activity arises if a mother needs to work on her agricultural fields for much of the day, but her children need to be cared for at home. Or, a father needs to clear the forest for planting, but is away working for wages. The hypothesis is not only strongly predictive, but also predicts well in both agricultural and non-agricultural societies.
Consequences of Marriage and Family Forms
Since the family is the major context in which children learn and grow, it would not be surprising if the type and form of family did not have important impacts on human development. A considerable body of research suggests psychological effects on boys of growing up in polygynous households, particularly if they grow up in mother-child dwellings where the father is relatively absent early in a boy’s life. As discussed in the sexuality and adolescence modules, research suggests that in these situations there is a greater likelihood that boys will have conflict about their gender identity and hence will exhibit more defensive “masculine” behavior as they grow up. As part of defensive masculinity, women are often derogated. Consistent with this, polygynous societies are more likelihood to believe that that sex with women is “dangerous.” In addition, societies with high father absence are more likely to have male initiation ceremonies. In such ceremonies, males are often asked to demonstrate their “masculinity” by showing little pain during genital surgeries or in special tests of fortitude.
Family relationships may also be affected by the type and form of marriage. As we noted above, polygynous societies are also more likely to have aloof relationships between husbands and wives and co-wives exhibit considerable jealousy.
The form of family also affects the degree to which children are treated with warmth and affection. Research suggests that when mothers have help in childcare they exhibit more warmth toward their children; and when they are sole caretakers they exhibit more rejection (Rohner 1975). Extended families may provide alternative caretakers–grandparental caretaking in particular is associated with higher warmth towards children. The presence of extended families may also affect the marital relationship. Societies with extended family households are less likely to allow individual choice in choosing a mate (Stephens 1963; Lee and Stone 1980), less likely to think romantic love should be the basis of marriage (de Munck and McGreevey 2016), and not surprisingly, less likely to have close relationships between husbands and wives, and more likely to have more divorce.
There also may be broader societal consequences. Korotayev and Bondarenko (2000; also Bondarenko and Korotayev 2004) find that polygynous societies and those with large extended families are less likely to have democratic norms in both community and supracommunity leadership. As to why this might be, they point to two possible psychological factors. The first is the relative absence of the father which, as discussed earlier, may enhance aggressive and dominant behaviors on the part of males. The second may be related to that finding that polygyny is associated with lower warmth toward children presumably because mothers have little help in childcare (Korotayev and Bondarenko 2000). There are many negative outcomes in adulthood associated with low parental warmth, including low self-evaluation, less generosity, a negative worldview, and more hostility and aggression. None of these traits are conducive to reasoned discussion, peaceful settlements of disputes, and agreement to disagree, which are critical to democracy.
What We Don’t Know
While marriage is universal as a custom, some societies have few unmarried individuals; others have many more. Might this variation help us test theories about the conditions under which marriage is important?
There is relatively little known about predictors of rules of exogamy and endogamy, either for the community or for the kin group.
What might explain why matrilocal societies generally lack substantial transactions to the groom’s kin in contrast to considerable transactions to the wife’s kin with patrilocality?
While there is considerable research on the type of marriage and the conditions associated with it, aside from some research on aloofness/intimacy there is relatively little known about the quality of the marital relationship.
Exercises Using eHRAF World Cultures
Explore some texts and do some comparisons using the eHRAF World Cultures database. These exercises can be done individually or as part of classroom assignments. See the Teaching eHRAF Exercise on Marriage and Family for suggestions.
This topical summary should be cited as:
Ember, Carol R., Benjamin Gonzalez, and Daniel McCloskey. 2021. “Marriage and Family” in C. R. Ember, ed. Explaining Human Culture. Human Relations Area Files https://hraf.yale.edu/ehc/summaries/marriage-and-family, accessed [give date]
- Nepalese Marriage Ceremony. 2012.
Photograph by Krish Dulal, distributed under a CC BY-SA 3.0 license.
- Mousou bride, Yunnan. 2016.
Photograph by Rod Waddington, distributed under a Pixaby license.
- Hopi woman dressing hair of unmarried girl. 1900.
Photograph by Henry Peabody, public domain.
- Maasai family photo. 2015.
Photograph by Looremeta92, distributed under a CC BY-SA 4.0 license.
- Thai brideprice. 2008.
Photograph by Tainscough, distributed under a CC BY-SA 3.0 license.
- Biertan divorce house. 2006.
Photograph by Alessio Damato, distributed under a CC BY-SA 3.0 license.
- Independent family household. 2020.
Photograph by Kingofkings_Lj, distributed under a Pixaby license.
- Bride price
A substantial transfer of goods or money from the groom’s kin to the bride’s kin before, during, or after the marriage.
see: Bride Price
- Bride service
The groom, before or after the marriage, gives work services to the bride’s family.
The children of siblings of the opposite gender (i.e., the children of a woman and her brother are cross-cousins to each other).
A substantial transfer of goods or money from the bride’s family to the bride, the groom, or to the couple.
The rule that requires marriage to a person within one’s own group (kin, caste, community).
- Extended family
Two families that are connected by at least one blood tie that form a single social and/or economic unit. Extended families often include people from 3 or 4 generations.
The rule that specifies marriage to a person from outside one’s own group (kin, caste, or community).
- Formal polyandry
A type of polyandry where a woman’s husbands are all considered to be her children’s fathers, contribute to their wellbeing, and live with their wife.
- Incest taboo
A prohibition on whom one can and cannot marry or engage in sexual relations based on kinship. While most societies extend this taboo to some other members of kin, it minimally includes siblings and parents.
- Indirect dowry
Similar to dowry except that the goods or money originate from the groom’s kin and they are either passed to the bride directly or passed indirectly via her family.
- Informal polyandry
A type of polyandry where a woman is able to have multiple simultaneous husbands who are all considered to be fathers of her children and who contribute to their well being, but do not live together.
- Independent family
A single-parent, monogamous, or polygamous family that constitutes its own social and/or economic unit.
- Limited polygyny
A type of polygyny where only a limited number of men, usually those of greater wealth or social status, have multiple wives simultaneously.
- Matrilocal residence
A pattern of marital residence in which couples typically live with or near the wife’s parents.
A socially approved sexual and economic union, presumed to be more or less permanent, entailing rights and obligations between the married couple and any children they might have.
type of marriage limited to only one spouse at any given time.
- Nonsororal polygyny
type of polygyny in which a man is married to two or more women who are not sisters.
- Patrilocal residence
pattern of marital residence in which couples typically live with or near the husband’s parents.
- Parallel cousins
The children of siblings of the same gender (i.e., the children of a woman and her sister or of a man and his brother are parallel cousins to each other)
- Partible paternity
The belief that a child can have more than one biological father.
Two or more women are married to one man at the same time.
- Sororal polygyny
A type of polygyny in which a man is married to two or more women, specifically sisters.
- Unilineal descent groups
Kin groups formed on the rule of descent, which stipulates that an individual’s membership is assigned at birth through the line of descent of either the mother (matrilineal) or father (patrilineal).
These figures are based on data from the Standard Cross Cultural Sample which were coded by Broude and Greene (1985); these data on marriage were retrieved from D-PLACE variable SCCS739. Of the 148 societies in the Standard Cross Cultural Sample which Broude and Green coded 31.1% have full individual choice, 17.6% have individual choice that requires parental approval, 3.4% have individual choice that also require parental arrangement, 18.2% have either individual choice or arranged marriage, 16.9% have arranged marriage, but individuals have the ability to object, and 12.8% have fully arranged marriage.↩︎
The figures in this paragraph and in the Figure are based on data from the Standard Cross-Cultural Sample (Murdock and White 1969); these data on polygyny were retrieved from D-PLACE variable SCCS211. The Standard Cross-Cultural Sample (SCCS) is a sample of 185 societies. Figures from the SCCS also informed the included pie chart below.↩︎
There are only 2 polyandrous societies in the SCCS.↩︎
Of the various types of polygyny, limited is most common, followed by nonsororal polygyny, and then sororal polygyny.↩︎
A cross-species study of mammals and birds M. Ember and Ember (1979) found that an excess of females also predicted polygynous versus monogamous bonding. Some studies have not found support for the sex ratio theory, but measures of polygyny were different (Hooper 2006; Quinlan and Quinlan 2007) or variable was imputed (Minocher, Duda, and Jaeggi 2019).↩︎