Residence and Kinship
Carol R. Ember
Danielle RussellJune 10, 2022
For the vast majority of societies in the anthropological record, kinship principles formed the basis of how human societies were organized. This module gives a general picture of what we have learned from cross-cultural research about variation in kin groups, rules of descent, kinship terminology systems, where couples live after marriage, consequences and predictors of marital residence, and why descent or residence practices may change over time.
Download: PDF | ePUB (ebook)
Residence and Kinship
Unlike some other species, human adults rarely live in isolation. Families, minimally consisting of at least one parent and one child, are customary in all societies and are commonly the building blocks of larger families and kin groups. But how large these groups are and how they are composed varies considerably. In the Marriage and Family module we discuss explanations and predictors of extended families versus independent families. In this module, we consider explanations and predictors of variation in marital residence patterns, the formation and composition of larger kin groups, and the terms used for kin. We then discuss consequences of variation in kinship patterns and we close with questions for which we have as yet few answers.
Let’s start with a thought experiment. If you are a parent and your children are old enough to leave home, you might worry about how you will manage when you are too old or too sick to do much productive work. You would be particularly worried if there was no government-sponsored program to support you in such circumstances. In either case, you might want your children to live with you or near you. But there is a problem with this wish. Assuming that your children will want to marry or have a partner to live with, some set of children from some set of parents will have to leave home. Every parent faces a dilemma then–who shall go and who shall stay? Societies appear to have created rules to solve this dilemma. These rules are called marital residence rules.
If we look at a sample of societies in the anthropological record, the two most common rules specify the gender expected to stay and the gender expected to leave. A matrilocal residence rule specifies that a daughter stays with or near her family after marriage and her husband moves to where her family resides. A patrilocal residence rule specifies that a son stays with or near his family after marriage and his wife moves to where his family resides. These two marital residence rules account for about 85% of the cases in the ethnographic record, although patrilocal residence is more common as Figure 2 illustrates. Notice that we used the phrase “live with or near” to describe where a couple lives after marriage. This is because there are two common possibilities: the married couple becomes part of an extended family household consisting of two or more constituent families (“live with”) or the married couple forms a new household adjacent to or very near one of their parents (“live near”). There are a few other marital residence patterns that are less common: bilocal residence describes an apparent more-or-less equal choice for the married couple to have two options (usually between matrilocal and patrilocal residence); avunculocal residence describes a pattern where the couple lives with or near the husband’s mother’s brother (“avuncu” derived from the Latin for maternal uncle–avunculus); and neolocal residence (“neo” means “new” in Greek) describes a pattern where married children leave their parents’ homes and live in a new place separate from either set of parents. The rarest residence pattern, duolocal residence (“duo” means two in Latin), describes a situation where the married couple lives separately–the wife and the husband both remain in their natal homes. Note that although duolocal residence may give parents their wish to have all their children stay at home, the rarity of such a residence pattern suggests that it isn’t a very workable solution for most societies.
Notice that in about 95% of the societies in the anthropological record, married couples have lived with or near kin (only about 5% are neolocal). Presumably that is because help from family kin has been needed for survival and well-being. But in many, if not most societies, kinship relationships beyond the family are also important. All societies have ways of reckoning and tracking people you are related to. Kinship terminology is used to describe classes of relatives (such as aunts, uncles, grandmothers, grandfathers, and cousins) and most people can name a considerable number of consanguineal (by blood) and affinal (by marriage) relatives when asked.
Societies differ in the importance of these larger sets of relatives and how much they can be expected to help you or whether they can call on you to help them. In some societies, the family and maybe a few close relatives are the largest effective group of kin. In other societies, kinship is so pervasive a part of social life that membership in a kin group governs your access to resources, whom you can and cannot marry, and to whom you owe allegiance in case of an attack. Some kin groups are very large and can encompass people related to you as many as 14-20 generations back.
Anthropologists describe two main types of kinship principles that form larger groups: bilateral kinship and rules of descent. American society is characterized by bilateral (literally “two sided”) kinship. Think of the people you might invite to a wedding. Besides your own family, you would likely invite your kin on both sides of the family, such as your father’s brothers and sisters, your mother’s brothers and sisters, their children (your first cousins), your father’s parents, your mother’s parents (your grandparents) and likely their siblings. Such a group is referred to as a kindred. Anthropologists refer to kindreds as being centered around a particular individual (ego) because aside from siblings, no other people have the same people in their kindred. See Figure 3.
In contrast, rules of descent result in clear, unambiguous groups of kin. Such rules trace kinship relationships linearly or backwards in time to a known or presumed ancestor. In doing so, some close relatives are omitted. The two most common rules of descent are patrilineal descent and matrilineal descent and these two rules create patrilineal kin groups and matrilineal kin groups respectively (these two rules of descent are both types of unilineal descent since they are based on affiliation through only one gender). With a patrilineal rule, all children (male and female) become affiliated with their father’s kin group, usually named for a male ancestor or a more symbolic name such as the name of an animal. Note that when a unilineal kinship principle is invoked, some close relatives are excluded. For instance, with a patrilineal kin group principle, your mother is usually not included, nor are her brothers and sisters and their children. And even on your father’s side, his sisters’ children are usually excluded since they take their patrilineal membership from their father, not from their mother. By excluding some relatives, there is clarity about who is in your group and who is not.
In societies with unilineal descent, membership in a kin group is typically assigned at birth and if you are asked “to what kin group do you belong?” you can give the name of that group by either the name of an ancestor or by another name (such as the Bear Clan). Matrilineal descent is based on the opposite affiliation principle. Siblings affiliate with the kin group of their mother and their father is excluded along with his brothers and sisters and their children. And, on the mother’s side, her brothers’ children are excluded since they take their matrilineal membership from their mother. Some societies have both matrilineal groups and patrilineal groups. Such societies are referred to as having double descent. Such societies are believed to be transitioning from one form of descent to the other. Each person acquires a matrilineal affiliation as well as a patrilineal affiliation. If you imagine overlaying the patrilineal and matrilineal charts on each other, you can see that unlike bilateral kinship, some close relatives are still excluded from either group.
Some societies have ambilineal descent. Like unilineal descent, kin groups are formed with reference to ancestors in the past, but in societies with ambilineal descent, affiliation with a group of kin can be traced through the mother or through the father.
Predictors of Marital Residence
Marital residence is thought to be the basic building block of larger kin group structures because the rule of residence affects who lives together in a community or neighborhood; it is more likely that connections can be traced when members of a potential kin group live together. What does cross-cultural research tell us about what predicts and what might explain variation in marital residence?
Patrilocal Residence versus Matrilocal Residence
The major contrast in residence patterns is between matrilocal residence and patrilocal residence, the two most common patterns. Not only are these two patterns the most common cross-culturally, but these two patterns provide the most contrast on which gender (female or male) is asked to relocate from the home they grew up in to their spouse’s parent’s place of residence. A long-held assumption is that residential variation (matrilocal vs. patrilocal) should be predicted by which gender contributes the most to subsistence. The reasoning is that if males are the “breadwinners” then parents would want their sons to stay home after marriage, but if women are the “breadwinners,” then parents would want their daughters to stay home. As plausible as this theory seems, worldwide cross-cultural research does not generally support it.
- Higher male contribution to subsistence does not generally predict patrilocal residence nor does higher female contribution to subsistence generally predict matrilocal residence (M. Ember and Ember 1971; Divale 1974b, 1975)
However, subsistence contribution is predictive of residence in certain circumstances.
Higher male subsistence contribution predicts patrilocal (versus matrilocal) residence in Native North American societies (Driver and Massey 1957; M. Ember and Ember 1971).
In a worldwide sample of hunter-gatherers, higher male subsistence contribution predicts patrilocal residence and higher female subsistence contribution predicts matrilocal residence (C. R. Ember 1975).
Why? C. Ember suggests that the importance of knowledge of wild species’ locations may give increased impetus for parents to keep their son (or daughter) home after marriage if they do more of the subsistence work. The hunter-gatherer finding and the North American finding are probably related considering that North America has a high proportion of hunter-gatherer societies.
Extremely high male contribution to subsistence predicts patrilocality (Korotayev 2003a)
If subsistence contribution is not a strong predictor of matrilocal versus residence, what is? In cross-cultural research to date, the strongest and most consistent predictors relate to patterns of warfare, particularly whether warfare is internal to the society (that is, occurs between subgroups) or is mostly or purely external (that is, largely or purely with other societies). Here is what cross-cultural research has uncovered:
Patrilocal residence is predicted by internal warfare, whereas matrilocal residence is predicted by purely external warfare (M. Ember and Ember 1971; C. R. Ember 2011; Divale 1974a; Carter Jr. 1977; Adams 1983).
A closely related finding is that higher levels of internal war predict patrilocality, and higher levels of external war predict matrilocality (Divale 1974b).
Focusing primarily on matrilineal descent, Shenk et al. (2019) find that internal war is associated with a shift away from matrilineal descent.
Societies with a history of migration into new territory within the 500 years before the time of description will tend to be matrilocal rather than patrilocal (Divale 1974b, 1975). Note that migration almost always involves offensive warfare by the intruders.
Why may warfare be related this way? Although both the Embers and Divale had the same results–namely, that internal warfare predicts patrilocality, and purely external war predicts matrilocality, they suggest very different theories to explain the relationship. The Embers suggest that warfare patterns influence residence patterns, while Divale suggests that residence patterns influence type of warfare. The Embers (1971), noting the prevalence of warfare in the anthropological record, suggest that parental concerns about defense and protection override their considerations of how much daughters or sons contribute to the economy. The Embers’ theory is that if warfare is internal and at least sometimes close to home, parents would not want their sons to move away at marriage regardless of whether their subsistence contribution was low or high. Presumably, sons would be more reliable defenders than sons-in-law. However, if warfare were purely external, presumably sons-in-laws, having no conflicts of interest between their home village and their wives’ villages, would be willing and able defenders of their in-laws. If defense were not a consideration, the Embers suggest that subsistence contribution would come back into play and would influence residence choices. Indeed, when warfare is purely external, division of labor predicts residence (C. R. Ember 1974).
Divale’s theory is quite different. Given the greater frequency of patrilocal residence, Divale assumes that patrilocal residence is “normal” and that only matrilocal residence needs to be explained. Divale suggests that when related males are localized, they form fraternal interest groups with strong internal loyalties to their own kin and few ties to other communities. This makes them prone to internal fighting when disputes arise. Divale theorizes that matrilocal residence arises when a group tries to move into new territory. If they are going to succeed in their intrusion, matrilocal residence, by scattering related males into different communities, provides a mechanism to minimize internal warfare and increases chances of migrating successfully. C. Ember (1974) questions whether it is plausible that people would know that switching to matrilocality would create internal peace. She points out that while matrilocal societies are more likely to have recently migrated, only about half of migrating societies are matrilocal; the rest are patrilocal, suggesting that migration is not a sufficient condition for matrilocality to develop. In light of Divale’s findings regarding migration and noting two additional findings (matrilocal societies usually have less than 21,000 people, and small size predicts purely external war), C. Ember (1974) suggests that matrilocal societies are likely to come from the pool of small societies that have recently migrated that also developed higher female contribution to subsistence, perhaps because men are heavily engaged in fighting when work has to be done. As an aside, the fact that matrilocal societies have recently migrated suggests that they are more successful in warfare.
Bilocal or Multilocal Residence
A different dimension of residential variation is the degree to which a society follows one pattern of unilocal residence (in which a married couple lives with or near a relative related by blood to one of the spouses–matrilocality, patrilocality or avunculocality) versus regularly following more than one pattern. Following more than one pattern is called bilocal residence residence when two choices are more-or-less equal. In between the two extremes of unilocal residence and bilocal residence is having a frequent alternative residence pattern, such as being predominantly patrilocal, but having matrilocal residence as an alternative or being predominantly matrilocal with a frequent patrilocal alternative. Multilocal residence is a more general term that considers a frequent alternative and bilocality together to contrast with unilocal residence. What predicts multilocal residence?
Severe and sudden loss of population due to introduced diseases predicts multilocal residence (C. R. Ember and Ember 1972, using a worldwide sample; C. R. Ember 1975, using a hunter-gatherer sample)
Why? Elman Service (1962) suggested that sudden and severe loss of population makes it difficult to follow a unilocal rule because the requisite individuals may not be alive. Given the need to live near kin, couples may have to live with other relatives. Notice that this theory assumes that in most anthropologically-described societies people need to live with or near kin for survival.
The more equality in inheritance by females and males, the more likely the society has multilocal residence (C. R. Ember and Ember 1972)
Why? George Peter Murdock (1949, 204) suggested that if either the woman or man can inherit, couples may choose to live with the relatives with more wealth or higher status. (While the relationship is significant, it is fairly weak and when depopulation is controlled it is no longer predictive.)
Migratory band societies (or those lacking much agriculture) are more likely to have multilocal residence (C. R. Ember and Ember 1972; Marlowe 2004; Walker 2015)
Why? Earlier theorists such as Eggan (1966) suggest that hunter-gatherers are likely to need more flexibility because resources seasonally fluctuate or are unpredictable. Consistent with the unpredictability hypothesis, among hunter-gatherers, those with bilocal residence are more likely than those with unilocal residence to have higher variation in annual precipitation (C. R. Ember 1975)
Very small community size predicts bilocal or multilocal residence amongst hunter-gatherers (C. R. Ember 1975). Korotayev (2004) finds the same relationships looking at a world-wide sample that includes all types of subsistence.
Why? Following reasoning by Anderson (1968, 154), Steward (1968, 331), and Lee (1972), C. Ember (1975) suggests that in very small communities, defined as less than 50 people, it is very unlikely to have a fairly equal sex-ratio for those of marriageable age. This means that a community trying to follow a unilocal rule could quickly end up with too many in-marrying spouses or lose too many out-marrying spouses. To maintain a fairly consistent size, multilocal residence may be more adaptive.
Among hunter-gatherers, the more resources are unpredictable, the greater the likelihood of multilocal residence (C. R. Ember 1975)
Why? Implicit in the theory suggested by Forde (1947), Eggan (1966, 1968), Anderson (1968) and Lee (1972) is the idea that alternative residence patterns provide a way for couples to move to other bands where resources are more abundant at any given time. C. Ember (1975) used variation in rainfall predictability to measure resource unpredictability for hunter-gatherers.
Other Residence Forms: Avunculocality and Neolocality
As we noted earlier, the vast majority of societies in the anthropological record had some form of residence where couples live with or near kin. Many scholars have noted that the form of residence that many in the world are used to today–neolocal residence–is probably largely a product of recent times. For example, Goode (1963) thought that industrialization gave impetus to neolocality because it often requires people to move to where jobs are located, but also lessens dependence on one’s own family by increasing economic opportunities for job seekers. While industrial societies do tend to have neolocal residence, Melvin Ember (1967) suggested that kin ties may be weakened before industrialization by the introduction of money as a medium of exchange, particularly when people can earn money through their labor outside of regular subsistence activities. Indeed, both the presence of money as a medium of exchange and the presence of industrialization is a predictor of neolocal residence (M. Ember 1967 on commercial exchange; de Leeuwe 1971 on industrialization). Ember (1967) suggests that the rise of commercial exchange makes it possible for individuals to sell their labor and/or their products in order to buy what they need to live.
Avunculocal residence, where the couple lives with or near the husband’s mother’s brother, is the only major form of residence where couples live with a relative other than a parent. Avunculocal residence is difficult to understand unless you know that it exists in the context of matrilineal descent. We will return to the possible predictors of avunculocal residence after discussing matrilineal descent in the next section.
Larger Kinship Groupings
All cultures recognize some relatives beyond their immediate family, but when scholars of social organization discuss kinship groups, they focus on whether or not societies have formal principles or rules for how such groups are formed, what functions these kin groups serve, the rights and responsibilities associated with kin groups, and the degree to which these larger kin groups play a role in people’s lives.
The principles for larger kin group formation are fairly straightforward. There are three main descent principles–unilineal descent (with matrilineal and patrilineal subtypes), ambilineal descent and bilateral kinship. The functions that kinship groups serve are quite variable and can include the regulation of marriage, access to land and other resources, mutual aid, political and religious leadership, and the provision of allies for raiding, feuding, and warfare. Presumably, the more functions a kin group provides, the more pervasive a kin group will be in everyday life.
In the ethnographic record, unilineal (“one line”) descent is the most common principle underlying larger kin group formation. As mentioned earlier, there are two types of unilineal descent–matrilineal descent and patrilineal descent. A unilineal descent principle affiliates all individuals who derive from a known or presumed ancestor (through one’s mother to a female ancestor for matrilineal descent; through one’s father to a male ancestor for patrilineal descent). This principle results not only in the unilineal kin group having only consanguineal (blood) relatives as members, but this principle also leaves out some close blood relatives (such as one of your parents) as members. Thus, unilineal descent provides a major contrast to bilateral descent, which forms groups laterally through both parents.
In the anthropological record, patrilineal descent is more common than matrilineal descent. In one cross-cultural sample, patrilineal descent is found in 41% of the societies; matrilineal descent is found 17% of the time Murdock and White (1969). Bilateral kinship is almost as common as patrilineal descent. See Figure 8.
If we follow Murdock’s (1949, 221–22) theory about how social organization changes or evolves, residence patterns change first and changes to kin group structure follow. Accordingly, we would expect that matrilineal descent would generally follow the emergence of matrilocal residence and that patrilineal descent would generally follow the emergence of patrilocal residence. After all, a matrilocal residence pattern followed over time puts matrilineally related women near each other and a patrilocal residence pattern followed over time puts patrilineally related men near each other. Proximity should facilitate remembering genealogical connections and developing the concept of descent from a common ancestor (Murdock 1949, 59–60; C. R. Ember 2011). But making something easier to form does not explain why such a group is formed. After all, in forming a unilineal group, people are leaving out important blood relatives, such as a parent. So, the question is: What function or functions might a unilineal kin group serve that would favor the emergence of such kin groups? Or, under what conditions will unilocal residence result in the emergence of unilineal descent? Below is what we know from cross-cultural research (most of the studies contrasted unilineal descent with either bilateral descent or contrasted unilineal descent with bilateral and/or ambilineal descent).
Almost all non-state societies with unilineal descent have unilocal residence, but many societies with unilocal residence lack unilineal descent (C. R. Ember, Ember, and Pasternak 1974).
Why? This finding is consistent with the idea that unilocal residence is a necessary, but not a sufficient condition for the development of unilineal descent (C. R. Ember, Ember, and Pasternak 1974; C. R. Ember 2011; Murdock 1949, 59). Note that this relationship becomes stronger if we distinguish societies that are almost entirely unilocal from those with a frequent alternative residence pattern (Korotayev 2004). Korotayev’s finding suggests that spatial contiguity is very important for being able to develop the concept of a unilineal kin group.
The presence of warfare in a non-state society increases the likelihood of a unilocal society having unilineal descent from 72% to 91% (C. R. Ember, Ember, and Pasternak 1974; see also Worchel 1974).
Why? While unilineal groups might serve many functions, the need for clear and unambiguous membership in a group may be particularly important when it comes to who will help you with offense and defense. Unlike bilateral kinship with overlapping membership, in a unilineal kinship system everyone knows what group they belong to and what group they do not. And, in contrast to other types of groups, such as neighborhoods or local communities, unilineal kin groups have the potential to unite people across multiple communities through shared recognition of kinship back in time. Building on suggestions from Service (1962, 11) and Sahlins (1961), C. Ember, Ember, and Pasternak (1974) suggest that competition and the need for allies in non-state societies may be the main catalyst pushing a unilocal society to develop unilineal descent.
Unilocal societies that are predominantly agricultural or that have large communities are more likely to have unilineal descent than unilocal societies that are primarily foragers or that have small communities (Korotayev 2004).
Variation in Unilineal Systems
Although unilineal systems vary considerably in the type of functions performed by unilineal kin groups, we have as yet little cross-cultural research on predictors of specific functions. However, we do have some research suggesting what may explain structural variation in unilineal descent systems. There are three dimensions of variation that have been studied. The first is whether or not a unilineal kin group in a society traces descent back to an ancestor through known links. Unilineal groups with known linkages are called lineages. Some societies have a hierarchy of lineages, with larger lineages tracing ancestry back further in time and thereby including more people. Some societies, such as the Tiv of Nigeria, say that all Tiv belong to one patrilineage tracing back about 14 to 18 generations (Bohannan 1998). Unilineal societies lacking lineages have kin groups in which people believe themselves to be descended from the same ancestor, but the links are not known or specified. Such groups are often named after animals (such as a bear or a wolf). Since membership in a unilineal kin group is acquired at birth, everyone learns the name of the kin group they belong to early in life. The smallest unilineal group with unknown ancestry links is called a clan or a sib. Some societies recognize larger groups of related clans or sibs called phratries. When a whole society is divided into two kin groups without specified ancestry links, these groups are called moieties (from Latin for “half” or “middle”). Moieties can exist on their own, or also include smaller kin groups.1 Another dimension of variation is whether the core members of a kin group live contiguously in a defined territory or are dispersed over the landscape with many members of other groups in their midst.
Unilineal societies with internal warfare tend to have at least one contiguous kin group of core members; those with purely external warfare tend to have dispersed kin groups (C. R. Ember, Ember, and Pasternak 1974).
Why? The theory is similar to the M. Ember and Ember (1971) theory about patrilocal versus matrilocal residence. If warfare takes place between communities within the society, people will want those most likely to defend them to live close by. Since residence is likely to be patrilocal when warfare is internal, kin groups are likely to be contiguous and patrilineal. On the other hand, if warfare is purely external, it may be advantageous to have allies dispersed throughout the society (C. R. Ember, Ember, and Pasternak 1974).
Unilineal societies with internal warfare are more likely to have lineages than unilineal societies with purely external warfare (C. R. Ember, Ember, and Pasternak 1974).
Why? As we have discussed, contiguity facilitates the remembrance of genealogical connection. Lack of contiguity makes it more difficult to remember connections. Since internal warfare favors contiguity of the core members of a kin group, it is also likely to be associated with lineages. However, the relationship between lineages and internal warfare is weaker than between internal warfare and contiguity. Perhaps this is because contiguity is not the same as physical closeness. A contiguous kin group could be spread out over a considerable distance. Research suggests that societies with internal warfare and very low population densities are less likely to have lineages than those with higher densities (C. R. Ember, Ember, and Pasternak 1974).
Lineages to Clans, or, Clans to Lineages?
Since many societies have lineages as well as unilineal descent groups with unknown links (clans, phratries, and moieties), a natural question arises as to whether there is a typical evolutionary sequence from one type of kin group to another type–in other words, which type generally comes first? One theory is that lineages come first, but, as population grows, links tend to be forgotten which produces larger groups such as clans, phratries, and moieties (Titiev 1943). A second theory reverses that causality and proposes that groups with unknown links form first and that lineages form later, especially if internal warfare develops as the societies grow (C. R. Ember, Ember, and Pasternak 1974). Without historical evidence, is there any way to test these theories? One possibility is to compare societal population sizes of the various combinations–societies with lineages only, societies with lineages and at least one unilineal kin group with unknown links, and societies with only unilineal kin groups with unknown links. If the first theory were correct, societies with only lineages should have the smallest populations. If the second theory were correct, societies with lineages should be larger. A test of these theories (C. R. Ember, Ember, and Pasternak 1974) supports the second theory–societies with only unilineal kin groups with unknown links tend to be the smallest, societies with lineages and unilineal groups with unknown links tend to be in the middle, and societies with only lineages tend to be the largest in size.
What explanation might there be for this result? If warfare is the main catalyst for forming a unilineal kin group in a unilocal society, such groups are likely to be absent if there is little or no fighting, but if fighting erupts, it is easier to form a group with assumed links rather than precise links. Small groups are also more likely to be able to maintain internal harmony, but as population increases, warfare is more likely to become internal. Internal warfare may provide the impetus for keeping one’s closest relatives nearby which would lead to the greater likelihood of remembering genealogical connections via lineages (C. R. Ember, Ember, and Pasternak 1974).
Transitions from One System to Another
In classic work on social organization, Murdock (1960, 221–22) laid out principles theorizing how social organization evolves over time. He argued that major changes first begin with a shift in residence pattern. Second, residence patterns alter the form that extended families take as well as the composition of the group of kin that live together in a community. Third, residential arrangements affect the form of consanguineal kin groups, and fourth, kinship terminology changes to reflect new patterns of social organization.
As we noted earlier, patrilineal descent is much more common in the recent past than matrilineal descent. But what is true recently does not necessarily reflect past history. In fact, some evidence suggests that matrilineal descent may have been more prevalent in the past than it is today. For example, Shenk and colleagues (2019) find that colonialization, a market economy, or an economic transition to pastoralism or to intensive agriculture, all predict a greater likelihood of transitioning away from matrilineality.2 Larger population size and internal warfare also predict such shifts. In addition, Shenk et al. (2019) find that transitions from matrilineality are about three times more likely than from patrilineality, which is consistent with the idea that matrilineal descent was more prevalent in the past.
Studies of linguistic closeness have been used to infer past history and make inferences about the likely traits possessed by societies long ago. Scholars looking to answer these questions have studied language families that currently have variation in kin group principles. Studies of Bantu language-speakers in Africa, suggest not only that the Bantu originated in western Africa and migrated east and south, but that those matrilineal groups that acquired cattle subsequently shifted to patrilineal descent (Holden and Mace 2003; see also Jones 2011). Jordan and colleagues (2009), studying the Austronesian language family of the Pacific, suggest that proto-Austronesian people were originally matrilocal and only later developed patrilocality in some branches. Today, patrilocal residence is more common than matrilocal residence among Austronesian-speakers and a considerable number of societies have bilocal residence. Jordan and colleagues (2009), like Holden and Mace (2003), find that transitions away from matrilocal residence were more common (estimated twice as common by Jordan et al.) as transitions away from patrilocal residence. Jones (2011) uses both of these language family histories to suggest that the histories are consistent with C. Ember’s (1974) idea that if a society is small and matrilocal (and perhaps also matrilineal) they may be fairly successful expanding into new territory, but as they grow in size, internal fighting may begin to occur, which may propel a shift to patrilocality (C. R. Ember 1974).
Once we know something about the conditions favoring residence and unilineal descent, we can make predictions about whether descent will be matrilineal or patrilineal. The main determinant will be the form of residence predicting the analogous type of descent. Patrilocal societies will tend to have patrilineal descent; matrilocal societies will tend to have matrilineal descent (Textor 1967; Surowiec, Snyder, and Creanza 2019). And, since internal warfare appears to predict patrilocal residence, patrilineal descent will be more likely to go with internal war. However, there is one other residence pattern that we need to consider–avunculocal residence.
Societies with avunculocal residence all have matrilineal descent (M. Ember 1974). So, when a couple resides with the husband’s mother’s brother, they are living with or near the husband’s closest matrilineal relative in the parental generation. And, if the husband has brothers, they will be nearby too as will other male matrilineal relatives. But why live avunculocally, not matrilocally? Murdock (1949, 207) theorized that the same conditions that produce patrilocal residence might favor avunculocal residence. We previously noted that the presence of some internal warfare is a strong predictor of patrilocal residence (M. Ember and Ember 1971; Divale 1974b) as opposed to matrilocal residence, which is predicted by purely external warfare. But, as M. Ember (1974) points out, avunculocal societies, in contrast to matrilocal societies, are characterized by internal warfare. This suggests that if internal warfare occurs in a previously matrilocal, matrilineal society, switching to avunculocal residence may provide a way of localizing matrilineally-related males for defensive purposes. The main question becomes why such a society doesn’t immediately switch to patrilocal residence. Part of the answer lies in the fact that unilineal descent is important in providing an unambiguous set of male relatives as defenders, so if the society were previously matrilineal, a switch to avunculocality would provide a clear set of defenders. But another part of the answer may relate to the likelihood that internal war, in addition to external war, might dramatically increase male mortality. If so, women might be likely to have children from different fathers. A switch to patrilocality/patrilineality might then be more difficult because half-brothers, having different fathers, would be in different patrilineal groups, whereas they would be in the same matrilineal group (M. Ember 1974).
It is widely believed that the terms we use to refer to relatives will tell us something about the social organization of a society in the present and even about the past. As noted earlier, kinship terminology is used to describe classes of relatives such as aunts, uncles, grandmothers, grandfathers, and cousins. Although there is a very large number of ways that different societies could classify relatives, it turns out that there are not that many different patterns of doing so. And these patterns are commonly found with certain types of residence and kinship systems. We concentrate our discussion of kinship terminology for cousins and the parental generation. Figure 10 represents a common way that anthropologists show kin relationships. Ego is the reference point.
An example of a class of kin terms in mainstream American culture is the term “aunt.” “Aunt” is used to refer to a person’s mother’s sisters, father’s sisters, and the wives of parents’ brothers. Similarly, the term “uncle” is used to refer to a person’s mother’s brothers, father’s brothers, and the husbands of parents’ sisters. Moreover, all the children of the people called “aunt” and “uncle” are referred to by the same term “cousin.” This by itself does not tell us much about American social organization until we contrast it with other kinship terminologies. Note that American kinship terminology is not just characteristic of American culture–it occurs widely and anthropologists refer to it as Inuit terminology (formerly Eskimo terminology.) As a point of contrast, let’s compare Inuit terminology with Hawaiian terminology as shown in Figure 11. In Hawaiian systems on the bottom (named after traditional Hawaiian culture) the same kin term for your mother (#3), represented by the color red, will be used to also refer to your mother’s sister (#3) and your father’s sister (#6); the same kin term used to refer to your father (#4), represented by the color gold, will be used to also refer also to your father’s brother (#5) and your mother’s brother (#1). The contrast between the two systems suggests that the nuclear family is much more important in societies with Inuit terminology because the terms given to the nuclear family are not used for anyone else. On the other hand, when we realize that where large extended families are present, kinship terms for mother, father, brother and sister are much more likely to be used to refer to others, suggesting the greater importance of other relatives. But these two systems (Inuit and Hawaiian) of terminology are similar in two important ways. First, the mother’s and the father’s side of the family are treated the same with respect to kin terms. In other words, the terms outside the nuclear family are symmetrical (the same on both sides of the family). Second, all types of “cousins” (parent’s sibling’s children) are referred to similarly.
In striking contrast to Inuit and Hawaiian terminology systems are the Crow, Omaha, and Iroquois systems (See Figures 12 and 13). In all three systems, some terms of reference differ for two sides of the family–that is, they are asymmetrical. For example, the terms for father’s sister (#6) and mother’s sister (#1) are different from each other in all three systems. Second, Iroquois, Crow and Omaha make important distinctions between “cousins”. The most common distinction is between parallel cousins (mother’s sister’s children and father’s brother’s children: #9-10 and #14-15) who share “brother” and “sister” terms with siblings, and cross-cousins (your mother’s brother’s children and father’s sister’s children: #7-8 and #16-17), who do not. In Iroquois terminology (Figure 13), cross-cousins on both the mother’s and the father’s side are referred to in the same way. In Crow and Omaha systems (Figure 12), cross-cousins on the father’s side and the mother’s side have different terms of reference. But, perhaps surprisingly to people with other kinship systems, Crow and Omaha systems merge terms across generations for one type of cross-cousin; Crow and Omaha differ in what side of the family has merging terms. In Crow terminology, the terms for cross-cousins are merged on the father’s side of the family–for example, father’s sister (#4 in red) is referred to by the same term as father’s sister’s daughter (#16) and father’s brother (#5) is merged with father’s brother’s son (#17). In Omaha terminology, the terms for cross-cousins are merged on the mother’s side of the family–for example, mother’s brother (#1) is referred to by the same term as mother’s brother’s son (#8).
Murdock (1949, 136) has argued that proximity acts as a social equalizer. If people live with or near kin, such as mother and mother’s sister in matrilocal residence, there is a greater likelihood that they will share the same kin term. If they also have matrilineal descent, they will participate in the same group, making it less likely that the same kin term will be given to father’s sister, who is not only not likely to be living nearby, but will not share the same matrilineage. There is also the principle of analogy: If your mother’s sister is referred to as “mother” then her children are likely to be referred to as “brothers” and “sisters.”
Understanding kinship terminology is fairly complex, but as you will see, these systems may reflect different underlying social organization, which begins to make the systems more understandable. Here is what we know about the relationships between kinship terminology systems and residence and kinship systems.
Societies with Inuit terminology tend to be bilateral (or have no apparent kin groups), have independent families, and practice neolocal residence and monogamy (Murdock 1949; Textor 1967; Goody 1970)
Why? Inuit terminology, by differentiating the terms mother, father, brother, and sister from all other terms, is consistent with independent family life and living separate from kin (neolocal residence). The symmetry of terms on both sides of the family is consistent with the symmetry of bilateral descent, which does not give priority to the mother’s or to the father’s side of the family.
Societies with Hawaiian terminology also tend to have bilateral descent (or have no apparent kin group), but they have extended families, bilocal residence, and do not allow cousin marriage (Textor 1967; Murdock 1949; Goody 1970; Köbben, Verrips, and Brunt 1974)
Why? Like Inuit terminology, Hawaiian terminology is symmetrical and is consistent with bilateral kinship. But it is also associated with bilocal residence, which means that you might live with (in an extended family) or near kin on either side of the family. With bilocal residence and extended families, you are likely to live in the same household with one or more of your parent’s siblings and their children. Living together increases the likelihood that these other kin will act similarly towards you and therefore be referred to with similar kin terms. As noted above, Hawaiian terminology groups together your mother with other women in your parents’ generation and father with other men in your father’s generation. And children of the people you call “mother” or “father” will be called “brothers” and “sisters. Extending brother and sister terms to cousins is consistent with not allowing any cousin marriage.
Societies with Iroquois, Crow, or Omaha terminologies tend to have unilineal descent and people tend to live in communities structured by unilineal descent (Textor 1967)
Why? Unilineal descent implies that one side of the family (mother’s or father’s) is more important and all three systems have at least some different terms on the two sides of the family–for instance, mother’s sister and father’s sister are referred to differently in contrast to Inuit and Hawaiian terminology systems. Note that in the three systems (Iroquois, Crow, and Omaha) of kin terminology, the terms for mother, father, brother, and sister are all used to refer to other relatives, consistent with the greater importance of other relatives.
Societies with Crow terminology tend to be matrilineal and matrilocal and, if cousin marriage is allowed, it tends to be allowed only with one set of cross-cousins (Textor 1967; Goody 1970; Stewart and Jones 1972).
Societies with Omaha terminology tend to be patrilineal and patrilocal and, if cousin marriage is allowed, it tends to be allowed only with one set of cross-cousins (Textor 1967; Goody 1970).
Why? As explained earlier, a major difference between Crow and Omaha terminology is that their terms are merged across generations on the less important side of the family. Since the Crow are matrilocal and matrilineal this makes sense when you realize that merging across generations is on the father’s side (the less important side), whereas for the patrilocal and patrilineal Omaha this merging is on the mother’s side (the less important side). The most marriageable set of cross-cousins will be those that do not share a term for a parent or a brother or sister. For Crow systems this would be mother’s brother’s children (matrilateral cross-cousins) and for Omaha systems father’s sister’s children (patrilateral cross-cousins). Scholars believe that societies switch to Iroquois terminology when they begin to allow marriage to both sets of cross-cousins (Iroquois terminology does not have terms merging across generations).
More complex kinship terminology systems are more likely where there is cousin marriage and where descent is unilineal (Rácz, Passmore, and Jordan 2019)
Why? Complexity, in this context, refers to the number of dimensions used for terminology distinctions. For example, Hawaiian primarily uses just two dimensions–generational difference and gender difference and is the simplest. Where cousins are referred to as siblings, cousin marriage may reflect the extended incest taboo and would thus be prohibited (Rácz, Passmore, and Jordan 2019, 749–50).
Consequences of Residence and Descent
From what we have learned about kinship structures so far, it is reasonable to imagine that females in a matrilocal/matrilineal society would have tremendous advantages, whereas males would have tremendous advantages in a patrilocal/patrilineal society. From a woman’s point of view, staying home near one’s parents and other close relatives after marriage would appear to be much more appealing than moving to your husband’s place where you know few people. And the reverse should be true for men moving to your wife’s place. What is somewhat surprising is that researchers have not found large differences between matrilocal/matrilineal societies and patrilocal/patrilineal societies in the relative status of women. As discussed in the Gender module, matrilocal residence and matrilineal descent have only modest contributions to predicting higher status of women. Perhaps this is partly because of what anthropologists call the “matrilineal puzzle.” The matrilineal puzzle refers to the fact that even in matrilocal/matrilineal societies, women generally do not have that much political authority because typically it is the males of the matrilineal kin group that have political authority and in matrilineal societies political succession typically passes from a woman’s brother to his sister’s son. Thus, in the matrilineal system the line of authority (through male matrilineal relatives) and the line of descent (through female matrilineal relatives) diverge. In contrast, in patrilocal/patrilineal societies, the line of descent and the line of political authority converge. In both systems, political authority is generally in male hands. In terms of domestic authority, women in matrilineal societies tend to only have more authority when neither their brothers nor their husband dominate in domestic matters (Schlegel 1972, 135).
The matrilineal puzzle is also consistent with other differences between matrilocal/matrilineal and patrilocal/patrilineal societies. One of the most important is how far away individuals move at marriage. Because males (such as mother’s brothers) tend to play an important role in their kin groups after marriage, they tend not to move too far away even when they live matrilocally. Indeed, many matrilineal societies have multiple kin groups in one community so a man might only need to move to another place in the same village. Patrilocal/patrilineal societies, in contrast, are more apt to have community exogamy, where women have to move away to another community and in-marrying spouses from other communities move in often from considerable distance (M. Ember and Ember 1971, 581; Adams 1983).
But while residence and descent do not seem to confer large differences in women’s status, this does not mean that they do not make a difference. Indeed, women in matrilineal societies and/or matrilocal societies have relatively higher status than women in other societies (Pearson Jr and Hendrix 1979; Whyte 2015, 133, 171) and women have more power in marriage when residence is matrilocal and descent is matrilineal (Warner, Lee, and Lee 1986). Recent research also suggests that girls from matrilocal households also have higher educational enrollment (Bau 2021). That being said, in no societies that we know of do women generally have higher status than men.
Variation in residence and kinship are also linked to type of marriage and how it is arranged. Patrilocal societies are more likely to have polygyny, where men marry two or more women (Whiting 1964; Korotayev 2003b). As noted in the Marriage and Family module, high male mortality in warfare is a predictor of polygyny and since patrilocal societies typically have internal war, we might surmise that male mortality may be greater in such societies. Moreover, polygyny is generally incompatible with matrilocal residence since women who are not sisters typically grow up in different places and generally it is not practical for a husband to live in more than one place. If matrilocal societies have polygyny, it is likely to be sororal polygyny, which is when a man marries sisters. Indeed, sororal polygyny is associated with matrilocal residence (Aberle 1961; Textor 1967), although matrilocal societies generally tend to be monogamous. Patrilocal societies are much more likely to give the bride’s kin substantial goods or services at marriage (M. Ember 1970) and a combination of patrilocality and patrilineality predicts more compensation to the bride’s family or the bride (Huber, Danaher, and Breedlove 2011).
Variation in residence and descent also appears to influence norms about sexual behavior and how much individual choice a person has in whom they marry. Societies with patrilineal descent and patrilocal residence tend to be more intolerant and punitive with regard to premarital sex as compared with matrilocal/matrilineal societies (Goethals 1971; Martin and Voorhies 1975; Horne 2004). Goethals (1971) (see also Horne 2004) points out that when a society is matrilocal and matrilineal, premarital pregnancy is not so problematic since a woman stays with or near her natal family after marriage and she would get needed support from her family. As for individual choice in marriage, the presence of unilineal descent appears to increase the likelihood that parents and other relatives will arrange marriages (Stephens 1963).
We have already noted that the type of warfare may influence whether societies are matrilocal or patrilocal. Recall that matrilocal societies tend to have purely external warfare. More broadly, Jones (2011), pointing to work by Paige (1974) and Swanson (1969), suggests that matrilocal and patrilocal societies have different patterns of societal solidarity in how they are organized politically–matrilocal societies are generally more communal in nature, whereas patrilocal societies are more factional and more likely to permit special interests to flourish.
What We Don’t Know
What conditions favor ambilineal descent? Do the same conditions favoring multilocal residence predict ambilineal descent?
Do different types of unilineal groups (lineages, clans, phratries, moieties) have different functions?
Do dispersed unilineal kin groups enhance internal cohesion?
What factors explain the breakdown or weakening of unilineal systems?
Does unilineal kinship predict higher status for the elderly?
Do women in matrilocal, matrilineal societies have better physical and mental health outcomes than in patrilocal, patrilineal societies?
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 Residence and Kinship for suggestions.
This summary should be cited as:
Ember, Carol R., Anj Droe, and Danielle Russell. 2022. “Residence and Kinship” in C. R. Ember, ed. Explaining Human Culture. Human Relations Area Files https://hraf.yale.edu/ehc/summaries/residence-and-kinship, accessed [give date].
Wedding Party Photo at Meiji Shrine Tokyo. 2011.
Photograph by Sgroey, distributed under a CC-BY-SA 4.0 license.
Iroquois longhouse. 2017.
Photograph by Marina Markel, distributed under a CC-BY-SA 4.0 license.
Maasai collect wood. 2017.
Photograph by Yulia Avdeeva, distributed under a CC-BY-SA 4.0 license.
Photograph by Susan Hubbard, distributed under a CC-BY 2.0 license.
Photograph by Wilses, distributed under a CC-BY-SA 4.0 license.
- Ambilineal descent
The rule of descent affiliating an individual through their father or through their mother with a group of kin who conceive of themselves as related to known or presumed ancestors.
- Avunculocal residence
A pattern of marital residence where couples typically live with or near the husband’s mother’s brother.
- Bilateral kinship
A kinship system where individuals are affiliated with both their mother’s and father’s kin relatively equally.
- Bilocal residence
A pattern of marital residence where two residence patterns (usually matrilocal and patrilocal) are about equally frequent.
- Clan (sib)
The smallest unilineal descent group whose members believe that they descend from a common ancestor, but genealogical connections are not known.
- Community exogamy
Exogamy refers to a pattern where people marry individuals from a different group. In this case, marriage would be with people from a different community.
In relation to ego, their mother’s brother’s children and father’ sister’s children. The link between a parent and their sibling is of the opposite gender. See ego.
- Double descent
Two unilineal rules of descent (matrilineal descent and patrilineal descent) are present. Individuals affiliate with two sets of kin groups–one through the mother and the other through the father.
- Duolocal residence
A pattern of marital residence where the married couple lives separately–the wife and the husband both remain in their natal homes.
The point of reference in a genealogical diagram.
- 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.
An ego-centered group of kin who are related to ego bilaterally (on both the mother’s and father’s side of the family). See ego.
Unilineal descent groups that have known links to a common ancestor.
- Marital residence rules
Cultural rules that specify where married couples should live after marriage.
- Matrilineal descent
The rule of descent that affiliates individuals with kin of both sexes related to them through women; at birth an individual affiliates with their mother’s kin group.
- Matrilocal residence
A pattern of marital residence where couples typically live with or near the wife’s parents. Some anthropologists use two different terms for such residence patterns–matrilocal when there is also matrilineal descent and uxorilocal when matrilineal descent is absent.
- Moieties (singular, moiety)
Unilineal descent groups that divide the society into two kin groups that lack specified ancestry links.
- Multilocal residence
A residence pattern where alternative marital residence patterns are common; includes both bilocal residence and a unilocal pattern with a frequent alternative
- Neolocal residence
A pattern of marital residence where couples live apart from kin (usually with some distance).
- Parallel cousins
In relation to ego, their mother’s sister’s children and father’s brother’s children. The link between a parent and a sibling is of the same gender. See ego.
- Patrilineal descent
The rule of descent that affiliates individuals with kin of both sexes related to them through men; at birth an individual affiliates with their fathers’s kin group.
- Patrilocal residence
A pattern of marital residence where couples typically live with or near the husband’s parents. Some anthropologists use two different terms for such residence patterns–patrilocal when there is also patrilineal descent and virilocal when patrilineal descent is absent.
- Phratries (singular, phratry)
Unilineal descent groups of supposedly related clans or sibs.
- Rules of descent
Cultural rules that specify how a person affiliates with a group of kin descended from a known or presumed ancestor.
- Unilineal descent
Unilineal descent stipulates that membership in a kin group is assigned at birth through descent links of one parent (either mother or father). Matrilineal descent groups are formed by links through mothers; patrilineal descent groups are formed by links through fathers.
- Unilocal residence
A marital residence pattern specifying one set of relatives that the couple lives with or near (patrilocal, matrilocal, or avunculocal)
Although Lévi-Strauss (1969, 75) believed that moieties reflected underlying dualism of the human mind, there may be a simpler demographic explanation. Moiety societies tend to be fairly small (populations less than 10,000) and therefore may only have two groups. Analysis of the population size of the largest unilineal group with unknown links shows that there is no significant difference between the average size of clans, phratries and moieties (C. R. Ember, Ember, and Pasternak 1974).↩︎
Consistent with this, matrilineal descent is more likely to be found with horticulture, rather than plow agriculture (Aberle 1961; Surowiec, Snyder, and Creanza 2019) and matrilineal descent is less likely when cattle are important (Surowiec, Snyder, and Creanza 2019).↩︎