Aloofness and Intimacy of Husbands and Wives:
A Cross-Cultural Study
John W. M. Whiting and Beatrice B. Whiting
Males and females—their temperaments, their relationship to one another, and their relative status—have been of major interest to Margaret Mead. In Sex and Temperament (1935) and later in Male and Female (1949) she convincingly demonstrated the importance of culture in determining similarities and differences in the behavioral styles of men and women of different societies. Her research has always been comparative but based on her own intensive field research in other cultures. We intend in this essay to test some of the hypotheses that she has generated over the years, as well as some of our own, regarding the male and female roles. Ethnographic reports on a world sample of cultures provide our data. Since these data were collected by ethnographers with other interests, our measures cannot be as subtle as those used by Margaret Mead; but we have tried to choose measures that are both valid and meaningful.
John W. M. Whiting is professor of social anthropology in the Departments of Anthropology and Social Relations at Harvard University.
Beatrice B. Whiting is professor of education and anthropology in the Department of Human Development, Harvard School of Education, at Harvard Unisity.
We have chosen the relationship between husband and wife as our focus in studying males and females. We do this because the family and family life are described in the ethnographic literature on most societies and thus comparable data are likely to be available. More important, however, is that a husband and wife are also father and mother to their children, and their relationship to one another is a crucial feature of the learning environments that shape the behavior of men and women of the next generation.
Again, partly for practical and partly for theoretical reasons, we have chosen to use as our measure whether or not a married couple ordinarily sleep in the same bedroom. Since both the floor plan of houses and the occupants of a household are described, it is not difficult to determine husband and wife living arrangements with a minimum of inference and with reasonably high reliability. Theoretically, since young children almost universally sleep in the same bedroom as their mother, the presence or absence of the father in this intimate space should have a strong influence on their conception of the nature of males and females.
Methodology
The standard cross-cultural sample (Murdock and White 1969) was chosen as the basis for testing our hypothesis. This set of 186 cultures was carefully chosen to maximize the independence of societies and to represent reasonably the languages, subsistence levels, and types of social organization of the known cultures of the world. Judgments as to rooming arrangements as well as to numerous other variables to be used in this study were available on a substantial number of societies in this sample.
In determining whether or not husband and wife were roommates, a combination of codes were used: the Palfrey House code of percent polygyny and rooming arrangements,[1] the Barry and Paxson code (1971) on sleeping proximity, and the Murdock and Wilson code (1972) on household composition.
Polygynous societies usually have alternative rooming patterns. In such societies the first wife ordinarily lives for some time monog-
[1] These judgments were made by the staff of the Laboratory of Human Development (Palfrey House) and were used for a study of residence patterns (Whiting and D'Andrade 1959) and of initiation rites (Whiting 1964). The judgments were made by Roy D'Andrade, David Beswick, and Emily Huntington.
amously with her husband before he takes a second wife. Furthermore, it is rare, even in societies where polygyny is highly valued, for every man in the course of his marital history to have a second wife. If, as is often the case, husband and wife share a bedroom as long as they are monogamous but sleep apart if they are polygynous, such a society could be classified in either the rooming apart or rooming together category. For this study we have arbitrarily chosen the rooming arrangement reported for the monogamous alternative if less than 30% of the families are polygynous. If, however, more than 30% are polygynous, the polygynous alternative was used to classify the culture. The decision was made from judgments by the Palfrey House staff as to husband and wife rooming arrangements both for the monogamous and for the polygynous alternatives.
Since this study focuses on the effect of rooming arrangements upon the child, the female-based family rather than the male-based family was used to calculate the percent of polygyny. Two women married to the same man would constitute two female-based polygynous families but only one male-based polygynous family. The use of the married women as a base yields a higher percentage value of polygyny than the use of the male head of household as a base. This is particularly true in societies in which having more than one wife is a privilege limited to a few chiefs. Such is the case for the Trobriands where only 3% of the adult males are polygynously married but 39% of the females between 15 and 45 years of age are members of a polygynous family. For purposes of this study it is even more significant that more than a third of the children are brought up in pologynous families.
Barry and Paxson coded "the nighttime sleeping proximity of mother and father to infant" (1971:467, col. 1). Their code required evidence as to the whereabouts of the infant as well as the mother and father; in contrast, the Palfrey House code made separate judgments for each diad. The Palfrey House code has been used for overlapping cases. Of the 76 societies for which there were scores by both teams, there was 92% agreement as to whether or not a husband and wife sleep in the same room. This degree of intercoder reliability seemed quite acceptable; therefore, the Barry and Paxson coding was used for those societies of the standard sample not coded by the Palfrey House staff. Finally, 9 cases were not rated
by either the Palfrey House staff or Barry and Paxson but were rated by Murdock and Wilson (1972) as having households in which husbands and wives sleep apart.[2]
The resulting number of coded societies then totals 159. They are distributed over the world as follows: Africa 40, Eurasia 34, Insular Pacific 25, North America 32, and South America 28.
Examples of rooming arrangements reported for the sample will serve to elucidate the code. A familiar pattern for modern industrial societies but one that is also found in many of the simple societies such as the Copper Eskimo is described by Jenness: among these Eskimos, "every inmate has . . . a definite place on the sleeping platform. The woman sleeps in the corner beside her lamp, and the little children lie between her and her husband. Outside of him sleep the older children and any guest whom they may happen to entertain (1923:85). This common arrangement, which was described for 92 societies (57% of the sample), occurs in societies with monogamy or limited polygyny in which members of the nuclear family share a bedroom.
Although co-wives do not ordinarily sleep in the same bedroom, this does sometimes occur. There are 20 examples of this arrangement in our sample (12%). The Kwoma of New Guinea (J. Whiting 1941) are a society in which 55% of adult females are polygynously married. Their houses are rectangular gabled structures with thatched roofs and sides. At one end there is an open porch that serves as kitchen, dining room, and parlor. At the other end there is a room with a small door and no windows that serves as a bedroom for a man, his wives, and their children. The husband sleeps on a raised platform. His wives and the children sleep on mats on the earthen floor.
As suggested above, it is more common for each wife in polygynous families to have her own bedroom. The husband may room with each wife in turn, he may have a bedroom of his own, he may sleep in a hut with his older sons, or he may sleep in a men's clubhouse. It is not unusual for several of these patterns to occur in the same society. The Kikuyu of Kenya have a sequence of rooming arrangements that is typical for many of the societies in subsaharan Africa (B. Whiting Field Notes 1972). When he is first married, a man builds a small wattle and daub house with a pyramidal
[2] Other categories in the Murdock-Wilson code could not be used because of the ambiguity between household and bedroom.
thatched roof. He ordinarily shares this with his wife until she has born a second child. At this point he feels the bedroom is too congested and he sets about building a second house that, if he can afford the bride price, he shares with a second wife until he is displaced by her children. By this time the sons of his first wife may be old enough—6 or 7—to leave their mother's bedroom. Then the father builds a third hut that he shares with his sons.
Societies with more than 30% female-based polygyny in which co-wives sleep in separate bedrooms were classed as rooming apart, whether the husband rotates, has a hut of his own, or sleeps in the men's clubhouse. There are 40 societies (24%) in our sample that fall in this category.
Finally, there are a small number of monogamous or limited polygynous societies in which husband and wife have separate sleeping quarters. Nine of the societies (5%) in our sample are reported to have this pattern. The Rajput of Khalapur who have rooming arrangements similar to the Senapur of our sample serve to illustrate this arrangement. "In Khalapur a man generally slept in the men's quarters which he shared with his brothers and other male relatives. This was built near the cattle corral so that he could guard his livestock against theft and predation." It is here also that he spends most of his leisure time chatting and gossiping with other men. Although the men sometimes eat in the women's house, they eat in private rather than with their wives and children and more commonly they eat in the men's quarters. "The dual household arrangement thus effectively separates the men from the women and young children and prevents an intimate relationship between these two groups" (Whiting and Whiting 1975).
Rooming Apart and Polygyny
It is obvious from our coding rules that there will be a strong positive association between rooming apart and polygyny. It would be wrong to assume, however, that polygyny can be taken as an index of rooming apart. Such an assumption would misclassify two of the categories described above—the 20 polygynous societies in which co-wives and their husband share a bedroom and the 9 societies with monogamy or limited polygyny where husbands and wives have separate sleeping quarters. The former have been classified as husband and wife rooming together, the latter as husband and wife rooming apart.
Rooming Arrangements and Husband-Wife Intimacy
In many parts of the world the monogamous nuclear family in which the husband and wife room together is, as a consequence of modernization, replacing traditional domestic arrangements. This is particularly true in societies where polygyny and rooming apart is traditional. In a study of such change among the Yoruba of Nigeria, Robert LeVine, Nancy Klein, and Constance Owen (1967) compared a traditional with an elite group of fathers. "The traditional family tends to be patrilocal, polygynous, and embedded in a residentially localized extended kin group; the elite have isolated, nuclear, monogamous families" (1967:238). LeVine interviewed ten fathers from each group with regard to their relation to their wives and family and found striking differences. Nine of the ten men from the traditional group said they never eat with their wives and children, whereas only one of the ten men from the elite group made this claim. Nine of the men of the elite group said that they had fed or diapered their infants, whereas only two men of the traditional group reported doing so. Finally, all ten men of the traditional group thought that it was not a good idea for a man to be present at childbirth whereas only four of the elite group reported this attitude. All of these differences are statistically significant.
In our study of the children of six cultures (Whiting and Whiting 1975) we found similar contrast. Fathers in the four rooming together cultures eat with their family, interact more with their children, and are more likely to attend their wives in childbirth than are the two societies in which husbands and wives room apart.
To discover whether these findings are generally true, we tested them on our cross-cultural sample. A judgment on eating arrangement was made by the Palfrey House staff on 77 societies in our sample and it was found that in a significant number of societies men and women who room together eat together. In 43 of the 51 (84%) cultures where the husband and wife are roommates, a married man ordinarily eats with his wife and family; whereas this pattern occurs in only 7 of the 26 cultures where husband and wife room apart. The phi value of this association is .57 (p < .001).
In many societies there are spaces set apart that are designated for men's use. The "house tamberan" of New Guinea—forbidden to women, the salaam of the Middle East and North Africa, the man's sleeping platforms of North India, and the men's huts in
subsaharan Africa provide places where men can gather to gossip with other men and, in some cases, also eat and sleep. Men may attend these places regularly or only occasionally. We were interested in discovering whether men who share rooms with their wives are less likely to spend time in exclusively men's groups than are those who room apart. Scores on the presence or absence of these special areas designated for men were made by the Palfrey House staff on 55 of the societies in our sample. Exclusive men's "houses" are present in 18 of the 25 (72%) societies where husbands and wives sleep apart but in only 12 of the 36 societies (33%) where they are roommates. The phi value of the association is .38 (p < .009).
Another estimate of the domestic behavior expected of a husband and father is provided by Barry and Paxson. One of their scales entitled "Role of Father" (1971: 472, col. 14) estimates the amount of interaction between a father and his children. Their five-point scale is so dichotomized that societies scored as "regular, close relationship, or companionship" and "frequent close proximity" were given a positive value while those scoring "occasional," "rare," and "no" close proximity were given a negative value. Separate judgments were made for father-infant and father-child relationship but they were so highly correlated that to present them both would be redundant. The relation of a father to his children during infancy was chosen as the variable to relate to the rooming arrangements score. Fathers have a close relationship with their infants in slightly more than half (26 of 49) of the societies in which husband and wife are roommates but in only one quarter (5 of 20) of the societies in which a married couple sleeps apart. The phi value of this association is .27 (p < .02).
Whether or not a husband is expected to be in attendence when his wife is giving birth serves as another available measure of husband-wife relations. In some societies it is considered dangerous to the woman, in others polluting to the man, for him to be present when his wife is in labor. By contrast, in other societies the husband is expected to be present and even to help deliver the baby. Ratings were made on this variable by the Palfrey House staff on 44 societies in the sample. The husband is permitted to be present in 2 of the 23 societies in which husband and wife sleep in separate rooms and in 11 of the 34 societies in which they room together. The phi value of this relationship is .28 (p < .04).
The condoning of wife beating provides another index of the
relation between husband and wife. Palfrey House ratings on 39 overlapping cases showed no association between wife beating and rooming arrangements. It is associated rather with independent versus extended households. Wife beating tends not to occur in the latter household type which apparently means that there is safety in numbers. patricia Draper (in press) has presented evidence corroborating this in her discussion of the increase in the frequency of wife beating among the !Kung Bushmen when they move out of their closely spaced camp groups and settle in nuclear households near Herero villages.
Taken together there seem to be two patterns governing the relationship of a man to his wife and children—one in which he is intimate and the other in which he is aloof. In the former he eats and sleeps with his wife and children, gossips with them at the evening meal, helps his wife care for their infant children, and is present and helps her at childbirth. In other words, he is highly involved in the domestic life of his family. In the societies in which the man stays aloof from his wife and children and uninvolved in domestic affairs, he spends his leisure time gossiping with other men and usually eats and sleeps with them as well. He seldom helps his wife with infant care and stays away from his wife while she is giving birth.
Since all of the above measures were available on only a limited number of societies in the sample, it was not expedient to construct an aloofness-intimacy scale. Instead the rooming arrangement score was taken as an index of such a scale.
Association of Husband and Wife Rooming Arrangements with Environmental Variables
The question arose in our minds as to why in some societies husbands and wives are roommates and are more intimate and domestic while in other societies husbands remain more aloof. Practical considerations seem one obvious explanation. Since in those societies where husband and wife room apart two apartments are required, one would expect that for purely economic reasons rooming together would be generally preferred. In fact, 70% of our sample room together. An additional fact in support of the economy hypothesis is that 36 of the 43 societies in which a husband and wife have separate bedrooms are situated in tropical climates where
heating is not a problem and the extra sleeping space generally consists of a simply constructed thatched hut.
But pragmatics explains neither why 25% of the societies do build at least two bedrooms for each married couple nor why there are more than 40 societies (58%) in tropical climates that do not bother to build separate sleeping quarters. In other words, economy explains some but only a part of the variance.
The permanency of residence seems to be another important factor. If a society is nomadic or seminomadic, especially if materials for housing have to be transported, building two houses instead of one evidently poses a problem that is usually avoided. Only 10 of the 63 societies of our sample (16%) that are nomadic or seminomadic (coded B, S, or T in column 1, Murdock and Wilson 1972) have separate bedrooms for husband and wife; whereas 38 of the 93 societies (41%) with permanent settlements (coded P or I by Murdock and Wilson) have such a rooming arrangement. The phi value for this association is .27 (p < .001).
Before considering other functional explanations for rooming arrangements, the possibility that they are the result of historical process should be explored. To do this the distribution of the two types of rooming arrangements have been plotted on regional maps (see appendix, maps 1-6). If, for reasons to be presented later, it is assumed that the arrangement whereby a married couple room apart is a cultural invention that occurred relatively late in the history of mankind, it would be expected that their distribution would show clusters indicating where the pattern might have been invented and then spread by diffusion or migration. It can be seen from the maps that such clusters do indeed occur. There is a large one in subsaharan Africa, one in India, one in Melanesia and New Guinea, and one in the northeastern part of South America. In addition to these four clusters, there are six isolated societies in which husband and wife room apart. They are highland Gheg of Albania, the Rwala Bedouins of Arabia, the Chukchee of Siberia, the Yurok and Pomo of California, and the Mapuche of Chile.
Even the most avid historical determinist would have to admit that the pattern of husband and wife sleeping apart was invented more than once. The assumption that the Ghegs, Rwala Bedouins, and the clusters in Africa and India all were historically connected would not account for the appearance of this pattern among the
Chukchee, the Yurok and Pomo, the cluster in South America, or the Mapuche. This suggests at least five independent inventions and still leaves the functional question unanswered. Can the conditions for such an invention be specified? Can it be explained why the invention spread widely in some parts of the world and remained isolated instances in others?
An explanation can be sought at two levels—the sociocultural and the psychological. We propose that husbands and wives will room apart in those societies where warriors are needed to protect property and that rooming apart has the psychological effect of producing hyperaggressive males.
Before exploring the sociocultural conditions that produce the need for warriors, it is expedient to detail the psychological consequences of contrasting rooming arrangements. In her brilliant chapter in Male and Female on "Womb-Envying Patterns," Margaret Mead laid the groundwork for the psychological hypothesis that we have adopted. Her description of the Iatmul, a society situated on the banks of the Sepik River in New Guinea, illustrates her concept of womb-envying patterns. This society, though not in our sample, would be classified as rooming apart. Although she does not indicate where husbands and wives sleep, she reports that there are several small men's houses in each village as well as "one large men's house or more, built with the effort of several clans" which is taboo to women and in which "all the important events of the men's elaborate ritual, war making, and debating go on" (1949:93-94). She describes the little boys as "surprisingly feminine, willowy, giving very little premonition of the bombast and high, headstrong behavior that will characterize them as adults. Their identification has been, inevitably, first with women" (1949:95). Later they are initiated into the men's cult where they are taught to "depend for their sense of manhood on a phantasy structure of bamboo flutes, played within leaf hedges imitating man made wombs. . . . [There men] are not peaceful shepherds but bold fierce head hunters . . . capable of magnificent anger" (1949:103-104). The Iatmul, as Mead describes them, exemplifies what we have called sex identity conflict. This hypothesis is derived from Freud's theory of the mechanisms of defense. It holds that a boy who is brought up in a household in which his mother seems to control all the resources will envy her status, covertly practice her role, and as a consequence
develop what we have called a feminine optative identity. If, when he grows up, he discovers that his childhood view of the relative power of males and females was distorted and that in other than the domestic sphere men rather than women control resources, he is placed in a dilemma—do I want to be female or male? This sex identity conflict, the theory holds, can be resolved by behaving in a hypermasculine manner. Furthermore, the greater the contrast between the perception of the relative power of men and women in the domestic sphere and in the sphere of economics and politics, the stronger will be the conflict and the more elaborate the defensive solution.[3]
That sex identity conflict produces hyperaggressive males has been shown in several studies. The relatively high rate of murder and assault in the two father-aloof societies in our study of six cultures (Whiting and Whiting 1975) was interpreted as partially determined by sex identity conflict. This interpretation is further supported because the children in these two cultures behave in a more aggressive-authoritarian and a less intimate-sociable manner than children of comparable sex and age in the other four cultures where their mothers and fathers sleep in the same bedroom.
In a cross-cultural study of the antecedents of crime, Bacon, Child, and Barry (1963) reported a significant association between personal crime and mother-child households. There was no association between the composition of the household and theft. Since their definition of personal crime included murder, assault, and witchcraft, which they interpreted as the intent to murder by magical means, their finding also supports the hypothesis that aloofness between husband and wife is positively related to aggressiveness. Only 22 of their societies overlapped with the sample used in this study. The personal crime score correlated positively (r = .21) but not significantly with the rooming apart score.
Another finding that supports the major hypothesis of this study is that aloofness between husband and wife is associated with warlikeness. Polygynous societies are more likely to place a high value on military glory than are societies with other household types (J.
[3] Studies supporting the sex-identity conflict hypothesis include D'Andrade 1973, Longabaugh 1973, Carlsmith 1973, Harrington 1970, Munroe, Munroe, and Whiting 1973, Herzog 1973, Munroe and Munroe 1973, Burton and Whiting 1961, and B. Whiting 1965.
Whiting 1969). These findings are based on a scale of military glory developed by Slater and Slater (1965). In this scale, items indicating a high value placed on military glory included seeking death in battle, a belief that warfare is the principle road to worldly glory, and that military virtues such as valor, recklessness, and fighting skill are important. The association of this scale with polygyny (phi = .28, p < .01) was interpreted as supporting the sex identity conflict hypothesis. The association between the military glory score and rooming arrangements on the 34 overlapping cases was positive (r = .21) but not statistically significant.
Another consequence of rooming apart is of interest. Whiting and D'Andrade (1959) found on a cross-cultural sample of 47 societies no instance of an adolescent boy sharing the bedroom with his mother and sisters when his father and mother room apart; whereas he does sleep in the same bedroom in more than half the societies in which his mother and father room together. This finding was interpreted as a response to incest taboos. It is also important that the young man in rooming apart societies usually moves in with his father and older brothers, or enters a bachelor group and lives with them. The Masai "Morans" (junior warriors) are an example of such an arrangement. They move into a "manyatta" (special living quarters for young men) and spend their time protecting the family herds and making retaliatory rustling raids on the herds of neighboring tribes. The rooming apart pattern thus is conducive to the formation of "fraternal interest groups" which have been shown by recent studies to be associated with feuding (Van Velsen and Van Wetering 1960, Otterbein 1970, Ember 1974, and Divale 1974).
That the practice of sleeping apart is associated with a high value placed on military glory, with hyperaggressive males, and with "fraternal interest groups" lends support to the hypothesis that the custom may have been invented and have diffused as a concommitant of the need for warriors, which in turn may be related to the presence or importance of a substantial capital investment that needs to be protected. If this assumption is correct, then the husband aloof pattern should occur more commonly among farmers and herdsmen than among hunters, gatherers, and fishermen since the former have more property to protect. Herds and gardens involve a greater capital investment than a foraging terrain or fishing grounds.
To test the above hypothesis the ratings made by Murdock and White (1969) as to the major base of subsistence were used. The results are presented in table 1. As can be seen, the hypothesis is confirmed. Only one gathering society (the Pomo of California) and two fishing societies (the Yurok of California and the Callinago of the Caribbean) have the pattern of rooming apart. If our interpretation is correct, the remaining 38 societies have little need to train their sons to be warriors.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||
Husbands and wives sleep apart in 6 of the 14 herding societies. All of these societies are either nomadic or seminomadic, which should make such an arrangement less probable.
The interaction between climate and herding as a determinant of rooming arrangements is of interest. Of the 6 herding societies situated in the tropics, all but 1, the Tuareg of North Africa, have a rooming apart pattern, whereas of the 8 herding societies situated in colder climates, only one—the Chukchee of Siberia—have this pattern.
That only 3 of the 13 societies practicing horticulture have the sleeping apart pattern requires some comment. All save one of the societies with this type of subsistence in the sample are situated in the insular Pacific. Husbands and wives sleep apart in none of the 8 societies that are on small islands of which they are in sole possession. The sleeping apart pattern is present in 2 of the 4 horticulture societies situated in the larger islands of New Guinea and Melanesia which are occupied by two or more groups. Although inter-
island raiding certainly occurred between the occupants of the small islands in Polynesia and Micronesia, their insularity, if our hypothesis is correct, may have served as a protection, but this is a posthoc interpretation.
Perusal of the list of societies coded as having agriculture, either of a simple or advanced nature, as a subsistence base is also enlightening. Many of these are American Indian groups who also depend to a large extent upon hunting and gathering, have seminomadic settlement patterns and, of course, dwell in a cold climate. In none of the 9 societies that fit this pattern do the husband and wife sleep apart.
Also in the lists of agricultural societies are those complex stratified societies that have developed a constabulary and/or a professional army as an alternative means of protecting property. If our hypothesis is correct, husbands and wives should room together in these societies. To test this hypothesis the scale of cultural complexity developed by Murdock and Provost (1973)[4] was used. Any society receiving a score greater than 30 on their 40-point composite scale of complexity was considered complex. Of the 21 societies in our sample rated as complex by this criterion, only 2, the Hausa of northern Nigeria and the Senapur of north India have the sleeping apart pattern.
The relationship between rooming arrangements, subsistence patterns, and cultural complexity for the five major regions of the world is summarized in table 2. As can be seen, husbands and wives seldom sleep in different bedrooms except in middle level societies. Of the 48 societies with such a rooming arrangement, only 3 occur with a hunting, gathering, or fishing economy, and only 2 occur where the societies are judged to be at a high level of complexity. It is also clear from table 2 that rooming apart is most likely to occur in Africa where there is the greatest concentration of societies at the middle level of development. The association is in the predicted direction for all the regions with the exception of North America, although it only reaches statistical significance for Africa (p < .001) and Eurasia (p < .05).
[4] This scale represents the sum of the following ten 4-point scales for which 4 represents the most complex value: writing and records, fixity of residence, agriculture, urbanization, technological specialization, land transport, money, density of population, political integration, and social stratification.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
If one assumes from the above evidence that rooming apart occurred relatively late in the history of mankind, it is interesting to speculate as to possible mechanisms that might have led to its invention. First, it is unlikely that there were in the distant past ethnopsychologists who understood the hypermasculine defensive syndrome. It also seems unlikely that the knowledge that better warriors could be produced by adopting the rooming apart pattern was discovered entirely by blind trial and error. The sexual division of labor practiced by middle level societies today suggests a possible mechanism. Transhumant herdsmen such as the Samburu (J. Whiting Field Notes 1968) leave the women, young children, calves, and a few milking cows in high country during the dry season while the men and junior warriors take the rest of the herd into the lowlands. It is in the lowlands that conflict with neighboring tribes occurs. Herds must be protected from rustling raids and counter raids must be carried out. The Samburu believe that they would be considerably hampered in these activities if they had to protect their families as well as their herds. Middle level gardeners also leave their wives and children behind when they are acting as sentinels or carrying out retaliatory raids.
The widespread belief that contact with women is dangerous or weakening for a warrior before battle is another mechanism that may have induced those middle level societies who most frequently engaged in war to adopt the rooming apart pattern.
The increase in polygyny with agriculture may be another reason for the rooming apart pattern in middle level societies. In rudimentary agriculture, especially when it was introduced as an adjunct to herding, women are the main gardeners. When women play an important role in gardening and agricultural products are an essential part of the diet, it is expedient for a man to marry more than one wife and the rooming apart pattern is likely to be adopted.
It is our interpretation of history that when, as a consequence of transhumance, sentinel duty, and the sex taboo on warriors before going to battle, the rooming apart pattern was adopted, it had unanticipated consequences of inducing sex identity conflict in the sons—a result that was adaptive as long as all young men in each generation had to become warriors.
That the rooming apart pattern was given up when the culture
became sufficiently complex also needs to be explained. One interpretation is that rooming together is the most convenient form for family life and it will be assumed when the need for rooming apart is no longer present. Another possible explanation is that in middle level agricultural societies women rather than men are likely to be chiefly involved in gardening. This makes polygyny advantageous and increases the likelihood of rooming apart. With the advent of the plow, which is utilized by all of the complex societies of our sample except the Hausa, men rather than women take over the responsibility for food production and polygyny loses its advantage. Again it is only the Hausa among the complex societies of our sample where more than 30% of the families are polygynous. Such a decrease in the utility of polygyny could be another mechanism that might impell a reversion to monogamy. The effect of modernization in East Africa as well as many other parts of the world clearly attests to the shift toward monogamy and rooming together.
In sum, then, it is seldom customary for husbands and wives to room apart. They do so only under special circumstances: when the climate is benign and, as a consequence, building extra bedrooms is not very costly; when they live a settled life and thus do not have to move camp frequently; and when they have a substantial capital investment to protect and inadequate help from the state.
The production of junior warriors, however, is not the only or perhaps even the most important consequence of a husband and wife sleeping in separate bedrooms. From our evidence it is probable that the husband who does not sleep with his wife remains aloof from her and the children in other spheres of domestic life. It is likely that he does not eat with them, that he does not help with the care of young children, that he does not help his wife give birth. It is also probable that he spends most of his time in the company of other men in some space that is not frequented by women. This aloofness is also transmitted to children. The social behavior of children brought up in cultures with the rooming apart pattern is less intimate and more aggressive than that of children brought up in cultures in which the fathers are more involved in domestic affairs.
References
Bacon, Margaret K., Irvin L. Child, and Herbert Barry, III. 1963. A Cross-Cultural Study of Correlates of Crime. Journal of Abnormal and Social Psychology 66:291-300.
Barry, Herbert, and Leonora Paxson. 1971. Infancy and Early Childhood: Cross-Cultural Codes 2. Ethnology 10:466-508.
Burton, Roger V., and John W. M. Whiting. 1961. The Absent Father and Cross-Sex Identity. Merrill-Palmer Quarterly of Behavior and Development 7:85-95.
Carlsmith, Lyn. 1973. Some Personality Characteristics of Boys Separated from Their Fathers During World War II. Ethos 1:466-477.
D'Andrade, Roy G. 1973. Father Absence, Identification, and Identity. Ethos 1:440-455.
Divale, William Tulio. 1974. Migration, External Warfare, and Matrilocal Residence. Behavior Science Research 9:75-134.
Draper, Patricia. In press. !Kung Women: Contrasts in Sexual Egalitarianism in the Foraging and Sedentary Context, Toward an Anthropology of Women (Rayna Reiter, ed.). Monthly Review Press.
Ember, Carol R. 1974. An Evaluation of Alternative Theories of Matrilocal Versus Patrilocal Residence. Behavioral Science Research 9:135-150.
Harrington, Charles. 1970. Errors in Sex-Role Behavior in Teen-Age Boys. Teachers College Press.
Herzog, John D. 1973. Initiation and High School in the Development of Kikuyu Youth's Self-Concept. Ethos 1 (4):478-489.
Jenness, Diamond. 1923. Physical Characteristics of the Copper Eskimos. Report of the Canadian Arctic Expedition 1913-18 , vol. 12, pt. B. F. A. Acland.
LeVine, Robert A., Nancy H. Klein, and Constance R. Owen. 1967. Father-Child Relationships and Changing Life-Styles in Ibadan, Nigeria, The City in Modern Africa (H. Milner, ed.). Praeger.
Longabaugh, Richard. 1973. Mother Behavior as a Variable Moderating the Effects of Father Absence. Ethos 4:456-465.
Mead, Margaret. 1935. Sex and Temperament. William Morrow & Co.
———. 1949. Male and Female. William Morrow and Co.
Munroe, Robert L., and Ruth H. Munroe. 1973. Psychological Interpretation of Male Initiation Rites: The Case of Male Pregnancy Symptoms. Ethos 1:490-498.
Munroe, Robert L., Ruth H. Munroe, and John W. M. Whiting 1973. The Couvade: A Psychological Analysis. Ethos 1:30-74.
Murdock, George P., and Caterina Provost. 1973. Measurement of Cultural Complexity. Ethnology 12:379-392.
Murdock, George P., and Douglas R. White. 1969. Standard Cross-Cultural Sample. Ethnology 8:329-369.
Murdock, George P., and Suzanne Wilson. 1972. Settlement Patterns and Community Organization: Cross-Cultural Codes 3. Ethnology 11:254-295.
Otterbein, Keith F. 1970. The Evolution of War. HRAF Press.
Slater Philip E., and Dori A. Slater. 1965. Maternal Ambivalence and Narcissism: A Cross-Cultural Study. Merrill-Palmer Quarterly of Behavior and Development 11:241-59.
Van Velsen, H. H. E. Thoden, and W. Van Wetering. 1960. Residence, Power Groups, and Intra-Society Aggression. International Archives of Ethnology 49:169-200.
Whiting, Beatrice B., 1965. Sex Identity Conflict and Physical Violence: A Comparative Study. American Anthropologist 67 (6), part 2.
Whiting, John W. M. 1941. Becoming a Kwoma. Yale University Press.
———. 1964. Effects of Climate on Certain Cultural Practices, Explorations in Cultural Anthropology (Ward H. Goodenough, ed.), pp. 511-544.
———. 1969. The Place of Aggression in Social Interaction, Collective Violence (James F. Short, Jr., and Marin E. Wolfgang, eds.). Aldine-Atherton.
Whiting, John W. M., and Roy G. D'Andrade. 1959. A Cross-Cultural Study of Residence from Infancy Through Marriage. Paper presented at the annual meeting of the American Anthropological Association, Mexico City.
Whiting, Beatrice B. and John W. M. 1975. Children of Six Cultures: a Psycho-Cultural Analysis. Harvard University Press (in press).
Appendix
The following maps illustrate the location of societies in the major regions of the world. Husband and wife rooming arrangements are indicated as follows: Apart
| ||||||||||||||||||||||||||||||||||||||||||||||||
| ||||||||||||||||
| ||||||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||||