Sie sind auf Seite 1von 33

Interpretation of family in the social group




l. Definition Of Marriage

Marriage and legitimacy

Variations in familial organization 1-11

Mate selection

The extended family system

The nuclear family system

Principles of preferential mating

The principle of incest avoidance

Ethnocentrism and homogamy

Differentiation of sex roles

Sex dominance in the marital dyad

Studies of marriage


Choice of spouses

The transfer of rights at marriage

The levirate and the sororate

Affinal relationships

Postmarital residence

Alternatives to marriage

Marital stability and divorce


Descent and alliance



Interpretation of family in the social group

The beginning of family formation may be either marriage or parenthood. It should not be concluded from
the fact that sexual intercourse is a prerequi-site for pregnancy that all peoples regard marriage or the
establishing of a man-woman relationship as the first step in family formation. Indeed, according to
Bohannan (1963, p. 73) the matricentric family, consisting of a woman and her children, is “both more
nearly universal and more elementary than is the nuclear family,” consisting of a marital couple plus any
children they may have. In some societies it is thought proper that marriage should precede pregnancy,
while in others the reverse sequence is regarded with favor; in the extreme case marriage is viewed as
irrelevant to family formation. However, it seems safe to assert that in most societies the nuclear family is
thought to be well launched only when both conditions are met.

Cultures also vary according to whether they emphasize marital solidarity over lineal solidarity or vice
versa. Societies with strongly developed extended family systems emphasize lineal solidarity over marital
solidarity. In such societies family formation is scarcely a meaningful concept: since the marriage of a man
and woman and the coming of their progeny represent the carrying on of a continuous line, these events
may signal the establishing of a new household but not the formation of a new family.

In this article the topic of family formation will be treated with reference to the nuclear family. Marriage will
therefore be considered as the focus of the process of family formation, and mate selection as one of its
most problematic features.


From the functional point of view, the family is the one social system that all societies look to for the
replacement of their members. However, from the structural point of view, the word “family” is used to refer
not only to the marital couple and their children but also to the larger kin group; accordingly, it will be
necessary to draw some structural distinctions. The “extended family” includes a nuclear family plus lineal
and collateral kinsmen; to the extent that a society emphasizes rights and obligations among kinsmen who
are not in the same nuclear family, it is spoken of as having an “extended family system.” On the other
hand, a “nuclear family system” is said to exist in a society in which the rights and obligations among those
in the larger kin group are given little emphasis relative to the claims among members of the same nuclear

It should be emphasized that the term “family,” whether it applies to a nuclear or an extended family, is not
equivalent to the term “household”—the aggregate of persons occupying a common dwelling unit, whether
or not those persons are kinsmen. In Western societies the nuclear family is frequently also a household
while the parental couples are in their younger and middle years and before their children attain adulthood.
However, many other arrangements are possible, and some are institutionalized. For example, in South
Africa it has been the practice for decades for the husband-father to be away from his nuclear family for

Interpretation of family in the social group

years at a time. A common type of household among Negroes in the Caribbean and in the United States
consists of a working woman, her children, and her mother. In traditional China the ideal house-hold
included the nuclear family of the head of the household plus his unmarried daughters, his sons with their
nuclear families, his sons’ unmarried daughters, his sons’ sons with their nuclear families, and so on
through all living generations; in practice, however, not many Chinese families could afford households of
such size.

Marriage may be defined as a culturally approved relationship of one man and one woman (monogamy), of
one man and two or more women (polygyny), or of one woman and two or more men (polyandry), in which
there is cultural endorsement of sexual intercourse between the marital partners of opposite sex and,
generally, the expectation that children will be born of the relationship (“polygamy” is the term that
subsumes both polygyny and polyandry). “Homogamy” refers to the marriage of persons of similar
characteristics, which is also known as “assortative” or “assortive” mating; “heterogamy” is the marriage of
persons of different characteristics; and “hyper-gamy” is a marriage in which the husband is of higher
social status than the wife. The term “endogamy” refers to marriage between persons belonging to the
same social group, whereas in “exogamy” the partners come from different groups.

Marriage and legitimacy

By definition marriage is a relationship within which sexual inter-course is legitimate. In general, a woman
who cohabits with a man has a legitimate status in relation to that man only if she is known to be married
to him. Common-law marriage (recognized in the United Kingdom and in the United States) and the
consensual union (recognized in the Caribbean) are forms of man-woman relation-ship that carry less than
full cultural approval and legitimacy. Points of interest to American, as well as to English, courts in
establishing whether or not a common-law marriage exists include: mutual agreement of the man and
woman to take each other as husband and wife; cohabitation and presentation of themselves as a married
couple to friends, neighbors, and the general public; and reputation, that is, the recognition by the
community that the two are husband and wife.

The Caribbean pattern of the consensual union differs from the common-law marriage of Anglo-Saxon
countries in that the former is not a legally recognized marriage. Various writers have held that except for
this lack of legal sanction the con-sensual union carries no social stigma and therefore is quite as
acceptable among the people practicing it as is legal marriage. More recent analyses by Blake (1961) and
by Goode (1960), however, have concluded that there is general recognition among Caribbean societies
that consensual unions are less legitimate and hence less desirable than legal marriages. Goode argues
that whereas legal marriage is recognized throughout Caribbean societies as the legitimate form, there is
variation among the social strata of these societies in the degree of norm commitment, with the
consequence that persons in the lower strata tend to be generally less committed to familial norms than
persons in the upper strata. The more frequent occurrence of consensual unions among the lower social
strata than among the upper is seen as a reflection of the class-linked variation in the degree of
commitment to familial norms.[SeeCaribbean SOCIETY.]
Legitimacy affects the offspring of the marriage as well as the spouses themselves. In asserting what he
called the “principle of legitimacy,” Malinowski (1929) stated that in all societies a socially recognized father

Interpretation of family in the social group

has been regarded as indispen–sable to the child. A legal marriage, then, gives a woman a socially
recognized husband and her children a socially recognized father. According to Zimmerman (1947), the
penalties attached to il–legitimacy vary directly with the power of the extended family; thus, the penalties
are heavy in societies characterized by the extended-family system and light where the nuclear family
prevails. From a sociological point of view, the significance of legitimacy is that it is a necessary condition
for the family to carry out its function of position-conferring. In this sense, the critical meaning of bastardy is
not that the child has low status but rather that he lacks any position and status in his society.

Variations in familial organization

Cultural expectations pertaining to marriage are affected by variations in familial organization. In Western
civilization it appears that the power of the family and the size of the effective kin group (i.e., of the familial
structure) have varied inversely with the complexity of the society of which the effective kin group is a part.
Zimmerman (1947), who extensively analyzed the civilizations of ancient Athens and Rome, reports that in
the early stages of both of these civilizations (i.e., when both societies were relatively simple) there existed
what he calls the “trustee” type of familial organization; whereas in their late (and, to Zimmerman,
decadent) stages, Athens and Rome developed much more complex societies and simpler familial
structures, which he describes as “atomistic.” The kernel of Zimmer- man’s distinction lies in the locus of
power. Where the trustee type of family exists, much power is located in the extended family. The head of
the family, as the responsible center of familial authority, influences the behavior of the family members,
and the extended family feels responsible for the behavior of its members. Where the atomistic type of
family prevails, much power is located outside the kin group in specialized institutions. As the family loses
power, its structure shifts from the extended family system to the nuclear family system. In the process of
making this shift, according to Zimmerman, the divorce rate goes up and the birth rate goes down. Arguing
that there are other lines of development than those of the West noted by Zimmerman, Goode (1963)
holds, as we shall see below, that whether the divorce rate goes up as a society becomes more complex
depends on the nature of the familial structure at the start of the process.

One way of formulating variation in the family’s power and size is to speak of its functioning as a political
unit. Moreover, the family may show variation in other kinds of functioning. In some settings the family is
the basic economic unit that creates and distributes goods and services. In many settings it is the principal
social unit responsible for socializing and educating the young. And in some settings, especially where
ancestor worship is practiced, the family carries out the religious function. In general, as societies become
more complex, specialized societal structures develop for the carrying out of these functions, with the
result that the family loses some of its functions; indeed such a state of affairs is the meaning of societal

Taking account of Asian and African as well as Western societies, Goode (1963) agrees that most family
systems of the world are moving toward a small-family system based on the nuclear family. Because the
traits of non-Western family systems are so varied, however, he believes there will be marked differences
in the direction of this change as the predicted convergence takes place. Thus, in African tribal societies

Interpretation of family in the social group

where matrilineal systems are strong and divorce is common, Goode reasons that urbanization will be
accompanied by a reduction in the conditions that have made divorce easy.

Mate selection

The functional emphasis in modern sociology leads the observer to anticipate that criteria for the choice of
a mate will be related to the roles the mate is expected to enact and, perhaps, that the mate will be chosen
by the incumbent of that social position most influenced by the quality of the mate’s performance. There is
some evidence to support such a set of functional expectations, but of course the empirical world is always
less tidy than the social scientist’s model.

The extended family system

In the extended family system it is common for members of the nuclear family to work in teams of kinsmen.
Under this condition the mate-selective process is frequently a means of recruiting workers, and hence the
members of the extended family have a lively interest in the work-related qualifications of a kinsman’s
prospective mate. Thus it is not unusual for responsible senior members of the extended family to select a
son’s spouse and to employ such famili-ally relevant criteria as the industry and prospective fecundity of a
potential daughter-in-law. For families of higher status, the standing of a girl’s family becomes more
important than her manual skills. Irrespective of status, however, the extended family system makes the
procuring of a mate a matter of moment to a wide circle of kinsmen. It is consistent with this kind of family
organization that mate selection should be a task calling for experienced perception and shrewd
bargaining. Moreover, in order that their plans should not be thwarted by the passions of the young, the
older people institute devices such as early marriage and efficient chaperonage (Goode 1959).

On the other hand, where the extended family is not highly functional and where the nuclear family system
prevails, it is frequently thought to be inappropriate for members of the extended kin group to exhibit lively
interest in the marital choices of family members, and even the influence of parents is reduced. Under
these conditions the criteria for mate selection are more likely to include attributes having primary appeal
to the nubile pair—physical beauty, sexual attractiveness, and congeniality. The response to one or more
of these attributes comes to be subsumed under the rubric of love. The diminution of relatives’ influence in
mate selection is not, of course, a categorical matter but rather one of degree. By their own religion,
ethnicity, and social status, as well as by their own choice of location of residence and of schools, parents
continue to influence their youngsters’ choice of spouses.

Traditional China provides an example of mate selection carried on by the family for familial purposes.
When a son married, the preferred arrangement was for him to bring his bride into his pa-rental home. The
parents expected the bride to perform two important functions: to bear children, preferably sons, and to
assist her mother-in-law in the performance of domestic chores. As the boy was growing up, he looked to
his parents to provide him with a wife. The parents expected the son to accept whatever bride they chose,

Interpretation of family in the social group

and they condemned vigorously any disposition on the son’s part to make his own marital selection,
especially if the son tried to do so on the basis of love. It was generally agreed that young people of
marriageable age were too inexperienced to have sound judgment in such an important undertaking. Since
most of the bride’s time was to be spent assisting her husband’s mother, functional considerations dictated
that the latter was the most interested party in the marriage; appropriately, therefore, she was usually the
most active person in selecting her son’s wife. Thus, arranged marriages were customary, and it was not
unusual for a young man to meet his bride for the first time at the wedding ceremony. Traditional China
made extensive use of the “go-between,” or marriage broker. This occupation served two useful functions:
marriage brokers made it their business to have extensive and detailed information about marriageable
young people; and they made it possible for families to enter into and break off negotiations without loss of
face (Hsu 1948; Lang 1946; Levy 1949).

With industrialization came pressure for changes in Chinese family law. This was evident as early as the
Boxer Rebellion at the beginning of the twentieth century, and new codes were promulgated in 1930 and
1931 (well before the communist revolution in China) that reflected Western standards—more emphasis
on the nuclear family and less on the extended family, a reduction in male authority, and a closer
approximation to legal equality of the sexes. However, the law retained a feature of Chinese filial piety: the
obligations to one’s parents superseded the obligations to one’s children. In these matters the communist
revolution has represented not so much a break with the past as a continuation of trends already under
way (Yang 1959). Although reliable information on postrevolutionary China is still scanty, it appears that
whereas the communist regime officially deplores both Western and traditional Chinese ways, love
marriages are common, and the influence of the extended family is continuing to wane.

The nuclear family system

As specialized social structures spring up, take over functions from the family, and become societally
important and individually rewarding, the resulting reduction in the functional importance of the extended
family re-moves incentives for maintaining an extended family system. At the same time there are four
functions inherent in the nuclear family that come to the fore as being relevant in mate selection. These
functions are: providing emotional gratification in the marital and parental relationships; providing identity
and a social status in the societal system to individuals who enter the family by birth, adoption, or marriage
—a function to be known here as position-conferring; performing such tasks as cleaning, bringing in
supplies, and disposing of waste products, which may be subsumed under maintenance of the household;
and child rearing, especially with respect to the parental functions of nurturance and control.

Of these four functions emotional gratification is most explicitly recognized in American culture as relevant
to mate selection, and apparently this is so, to an increasing degree, in the middle-class subcultures of
western Europe. There can be little doubt that convictions are widespread in the United States and
western Europe that a couple should be “in love” before considering marriage and that legal codes are

Interpretation of family in the social group

obsolete if they fail to provide for divorce on the ground of chronic marital conflict. Love as a mate-selective
criterion invites idiosyncratic interpretation in the sense that, for instance, one man may be attracted to a
demurely diffident girl whereas another finds the vivaciously extroverted girl irresistible.

As a mate-selective criterion, position-conferring (especially when phrased as status-conferring) evokes

ambivalent responses. In many middle-class settings a girl who is thought to have married for money
rather than for love risks social condemnation (Indian culture, by contrast, has had the tradition that it is
good for a girl to marry into a subcaste of higher standing than her own). If a girl marries for love plus
status improvement, however, she is said to have married “well,” and the durability of the Cinderella legend
suggests that there is little novelty in this theme. The woman’s social status depends so largely on her
husband’s occupational performance that, for her, mate selection is sometimes spoken of as a “mobility
bet.” Such evidence as exists on this matter for the United States indicates that most marriages are
between persons of roughly equal social status.

Although all four of the functions mentioned above are relevant to mate selection, a young couple
considering marriage can usually check the suitability of each other only with respect to emotional
gratification. This may have something to do with the emphasis given love as a criterion. In the premarital
setting of early adulthood the other three functions can usually be no more than the focuses of guesswork.
It is difficult for a young woman to foresee how a particular man will fare in the occupational sweepstakes
and in being a model for their sons. Predictions are similarly difficult for the young man with respect to how
a woman will manage their house and mother their children.

Where marriages are voluntary rather than ar-ranged, there is need of some means for marriageable
young men and women to meet and to select each other. The practice of dating is societally rational in the
sense that it affords this opportunity. On the other hand, dating as a prelude to mate selection has been
criticized on the ground that the leisure-time activities of dating fail to provide an adequate setting in which
to test prospective spouses with respect to maritally relevant criteria, especially with respect to the
functions of house-hold maintenance and child rearing.

In sum, a reduction of functions in the extended family is accompanied by a reduction in the rights and
obligations among extended kin that constitute the extended family system. This reduction in the
significance of blood relationships shifts the emphasis from the extended family to the nuclear family.
Marital solidarity replaces cognatic (both lineal and collateral) solidarity, and love becomes a criterion of
mate selection.

Principles of preferential mating

Let us designate as “ego” a person of reference, that is, a person from whose point of view we shall

Interpretation of family in the social group

consider certain relationships. All societies desig-nate categories of persons whom ego may not marry,
and frequently there are additional categories of persons whom it would be regrettable, but not totally
forbidden, for ego to marry. Usually there are implicit, if not explicit, categories of persons whom it would
be desirable for ego to marry. These negative and positive expectations can be subsumed under the
“principle of incest avoidance” and the “principle of ethnocentrism.” We shall speak of the set of persons
whom ego is permitted to marry in any given sociocultural setting as ego’s field of eligible spouse
candidates or, in shorter form, as ego’s “field of eligibles.” European social scientists use the term “isolate”
to refer to the field of eligibles.

The principle of incest avoidance

Every society has a prohibition against incest, that is, against sexual relations between persons who are
closely related. Although the precise relationships that are viewed as incestuous vary from one society to
another, they regularly include the mother-son, the father-daughter, and the brother-sister relationships,
that is, all heterosexual relationships within the nuclear family except, of course, the marital relationship.
The principle of incest avoidance refers to the set of prohibitions existing in every culture to prevent ego
from marrying someone too close to him in the kinship system.

Just how the principle of incest avoidance works out varies from one setting to another. In traditional China
it was prohibited for ego to marry anyone with the same surname, and in that populous land with few
surnames this rule proscribed hundreds of thousands of otherwise eligible spouse candidates. In northern
India there was a tradition that marriage was not possible with someone re-moved from ego by less than
seven degrees on the father’s side or less than five degrees on the mother’s; a more common rule in India
prohibits marriage between relatives linked to a common ancestor within five degrees on the father’s side
and three on the mother’s (Goode 1963, p. 210). In some societies ego is encouraged to marry a cross-
cousin (e.g., mother’s brother’s daughter) but prohibited from marrying a parallel cousin (e.g., mother’s
sister’s daughter). Prior to 1793 it was illegal in Connecticut for ego to marry the sister of his deceased
wife; but among the ancient Hebrews there was the custom of the levirate, by which a man was enjoined to
marry the widow of his deceased brother if the brother had died without a son. The record shows a very
few isolated cases where persons of opposite sex from the same nuclear family were permitted to marry.
An example is the brother-sister marriage among the Ptolemies of ancient Egypt. Apparently the practice
in these few exceptions functioned to keep power within ruling families
Ethnocentrism and homogamy

Whereas the principle of incest avoidance prevents ego from marrying someone too close to him in the
kinship system, the principle of ethnocentrism prevents his marrying someone too different from him with
respect to a number of social characteristics. In other words, ethnocentrism is a force tending toward
endogamous and homogamous marriages.

Sumner ([1906] 1959, chapter 1) used the term ethnocentrism to refer to the set of attitudes shared by
members of a tribe or other social group to the effect that the members of that group and any others like
them were seen as the center of the civilized world and had, therefore, the correct and desirable set of
social characteristics. Thus, ethnocentric attitudes lead to the condemnation of outsiders to the degree that

Interpretation of family in the social group

they are recognized as differing from one’s group. The minimum degree of social distance on the
Bogardus scale is indicated by an affirmative response to the query as to whether or not the respondent
would be willing to accept a person with a specified characteristic to close kinship by marriage.
Traditionally, the castes of India have been endogamous, as have the subaltern categories of subcaste,
section, and subsection. According to Kapadia ([1955] 1958, p. 118), these endogamous restrictions
limited a Hindu’s field of eligibles to 50 to 300 families. In 1949, however, the Hindu Marriages Validity Act
stipulated that no marriage of Hindus could be invali-dated because of caste or sect differences between
the parties concerned. Expert opinion is divided as to the likelihood that caste endogamy will break down.

In accordance with the principle of ethnocentrism there is evidence that in American society ego tends to
select a spouse similar to himself with respect to race, religio-ethnic identification, socio-economic status,
and other social characteristics. In the United States the most conspicuously homogamous dimension of
mate selection is race. Inter-racial marriages are still prohibited by law in a number of the Southern states;
moreover, even where such laws do not exist, or where they have been repealed, there is little evidence of
enthusiasm for such marriages. Various studies have shown the proportion of racially heterogamous
marriages to be under one per ent.

The second dimension of ethnocentric preference and prohibition is that of religio-ethnic identification,
which includes cultural as well as religious elements. Classifying the 1957 population of the United States
into the three major religious categories (Protestant, Catholic, and Jewish), the U.S. Bureau of the Census
found that approximately 94 per cent of the married persons had spouses in the same religious categories
as themselves. If religious endogamy had not been practiced, and if, therefore, matings had been entirely
random with respect to religious affiliation, the proportion having spouses in the same religious category as
themselves would have been about 56 per cent (Winch [1952] 1963, p. 331). There is evidence that in
heterogeneous communities the probability that ego will marry outside his religious category is greater
when his category constitutes a small pro-portion of the community rather than a large proportion. If the
religious category has a highly distinctive ethnic identity (e.g., Catholics who are Spanish-speaking in an
English-speaking community), the probability of ego’s marrying endogamously is increased.
A third dimension of ethnocentric preference is that of socioeconomic status. Commonly used indexes of
socioeconomic status are occupation, income, and number of years of schooling. Numerous studies have
shown that people tend to select their spouses from their own socioeconomic strata with respect to all
three of these indexes (several are cited in Winch [1952] 1963, pp. 336-338). Other characteristics with
respect to which people tend to mate homogamously are age, previous marital status, and location of
residence. Systematic research supports the common observation that young people tend to select young
mates and older people choose older spouses. No doubt it is partially because of this fact that there is a
tendency for people to marry others who are like themselves with respect to previous marital status:
divorced men tend to marry divorcees; single persons tend to marry those who have not previously been
married; and widows and widowers tend to marry each other. Another common-sense observation that has
been supported by research concerns residential propinquity: ego is more likely to marry someone living
nearby than someone living far away (Winch [1952] 1963, pp. 322-324, 339-345).
Since people are not randomly distributed through communities but rather tend to live near and to work
with others of similar social characteristics, one would expect mate selection to be somewhat
homogamous, whether or not there are any sanctions enforcing endogamy. Of course there are sanctions
of varying degrees of intensity: for example, in American culture sanctions are quite intense with respect to
race, less so with respect to religion and socioeconomic status, and virtually nonexistent with respect to
residential propinquity.
Homogamy may also be considered on a more psychological level. For example, there is evidence that
spouses tend to resemble each other in level of intelligence, in values (e.g., religious and aesthetic), and in

Interpretation of family in the social group

attitudes (e.g., toward birth control and toward communism). When spouses are tested by paper-and-
pencil methods, they appear to resemble each other somewhat, but not greatly, with respect to traits of
temperament and personality. However, data gathered by other methods, such as interviews and
projective methods, lead to the contrary conclusion that, at least in such traits as dominance and
dependence, spouses tend to be complementary rather than similar. At present this seeming paradox is
unresolved, although the answer may be that the homogamy apparent in paperand-pencil tests is an
artifact resulting from the effort of people to represent themselves to be as attractive as possible—what is
called the “social desirability” effect
Differentiation of sex roles

The simple fact that only women can bear children causes every society to recognize some differentiation
between the behavior of men and of women. Beyond the behavioral differences that are directly
attributable to anatomy and physiology, however, cultures vary greatly in the degree to which they view
human behavior as being properly sex differentiated.

From a study of 224 societies, Murdock (1937) has found that men tend to engage in such active and
mobile tasks as hunting, fishing, trapping, and lumbering, whereas women tend to specialize in more
sedentary but equally important tasks, such as gathering fuel and fruits and cooking and pre-serving meat
and fish. More generally, it is possible to conceptualize two criteria that distinguish masculine from
feminine tasks. Tasks assigned to men usually require physical exertion and strength, or spatial mobility
and absence from home for considerable periods of time, or both. By contrast, feminine activities are
typically less demanding of great strength, although perhaps requiring a considerable output of energy,
and will involve only a few hours at a time away from home.

Analysis of these differences leads to the conclusion that the sharpness with which a culture distinguishes
between masculine and feminine sex roles will be related to the importance it attaches to tasks requiring
one or both of the two masculine task characteristics. Military activity is one obvious example that involves
both of the masculine criteria; thus it is argued that a highly military-oriented culture will be one that draws
a sharp distinction between properly masculine activities and those that are properly feminine. The
converse inference is that to the degree that a society’s important tasks do not call for either of the criteria
distinguishing masculine activities, there will be no basis for developing highly differentiated sex roles. As
nonhuman power has taken over most of the heavy tasks in the industrial societies, the pro-portion of the
total labor force that is classified as “white collar” has greatly increased. And white-collar occupations,
especially those not requiring travel, can be carried on as well by women as by men. Thus, if Western
cultures have been “feminized” over the past century or so, as some writers have claimed, the present
analysis would interpret such a trend as a consequence of the increased use of nonhuman power.

Sex dominance in the marital dyad

What are the conditions that result in the dominance of one spouse over the other? The opportunity for
dominance exists in a dyad when resources desired by one member are controlled by the other, that is,

Interpretation of family in the social group

when one is dependent upon the other. Resources may be viewed broadly to include both material goods,
such as food, and intangibles, such as a compliment.

Where no organizational feature exists to deter-mine otherwise, it appears that men have usually
dominated women. The reasons for this originate in the two criteria differentiating masculine from feminine
pursuits and in their anatomical and physiological bases. A woman with small children has greater need of
a man to take care of her than the man has need of her. His care may be viewed as a resource, and by
granting or withholding that resource, the man can dominate the woman. This is a state of affairs that has
been remarked by social scientists from Aristotle through E. A. Ross and Willard Waller and is perhaps
best known to con-temporary readers under the rubric of the “principle of least interest”: that is, the person
in a relation-ship who has the least to lose through the termination of the relationship is in a position to
demand more from others and thus to dominate them in exchange for his continued participation. Aside
from this situation of unilateral dependence, other possibilities are mutual interdependence, where the
resources are not available to either one unless they cooperate, and mutual independence, where each
has control over his own resources.

With respect to organizational features, W. G. Sumner and A. G. Keller have remarked that where the
bridal couple lives has bearing on which is the dominant sex and therefore that matrilocal marriage is a
condition favorable to the relative standing of women. In traditional China, the favored pattern was
patrilocal, and a wife was expected to obey her husband; masculine dominance was miti-gated, however,
in the case of adoptive marriage. According to this pattern, a man having no son might seek a young man
(who was usually of somewhat lower social rank) to take the older man’s family name, marry the older
man’s daughter, and live matrilocally.

Studies of marriage

During the second quarter of the twentieth century there was a good deal of concern about the state of the
family in the Western world. There was evidence that divorce rates had risen, that the family had lost
functions to other social structures, that the birth rate had fallen, that certain totalitarian regimes were
trying to bring about the disintegration of the family, and that broken families were spawning delinquent
children. Family disorganization was widely viewed as a social problem; probably for this reason,
numerous studies were undertaken to discover the determinants, or at least some correlates, of what was
variously called marital “adjustment,” marital “happiness,” and marital “success.”

Although these studies did not undertake to dis- tinguish very sharply among the three terms just noted, it
does seem useful to differentiate them as follows. There are two kinds of marital adjustment, one
pertaining to the role and the other to the psyche of the performer. An actor is adjusted to a marital or any
other kind of role to the degree that he knows the expectations that define the role and, under the
appropriate conditions, can produce the behaviors expected. On the other hand, he is adjusted psychically
to the degree that the energy he invests in the role performance is commensurate with the gratification

Interpretation of family in the social group

derived from it. Marital “happiness” refers to the subjective response of the actor to marriage and thus is
related to psychic adjustment; however, one can be psychically adjusted when both output of energy and
input of gratification are low, whereas presumably happiness requires at least a moderately high level of
gratification. The term marital “success” implies the existence of a goal of marriage, and whatever goals
there may be—avoidance of divorce, procreation, personality development of the spouses—seem to be
more clearly conceived by those who write about marriage than by the participants whose behavior the
writers describe.

Much of the research on marriage has been concerned with marital adjustment—that is, both with the
aptitude to carry out the marital role and with the capacity to derive commensurate gratification from the
performance. Kirkpatrick has surveyed a large number of studies and has reported the variables he finds
that have correlated most consistently with what is here called marital adjustment. He has divided these
variables into two sets: those that were clearly operating before the marriage and those that may or may
not have been. Presumably the determinants of marital adjustment are more likely to come from the former
set. Kirkpatrick ([1955] 1963, p. 389) presents the following premarital factors as having shown the
strongest and most consistent association with high marital adjustment: happiness of parents’ marriage;
adequate length of acquaintance, courtship, and engagement; adequate sex information in childhood;
personal happiness in childhood; approval of the marriage by parents and others; adjustment in
engagement and normal motivation toward marriage; ethnic and religious similarity of the spouses; high
social and educational status; maturity (marriage in the late twenties rather than in the teens or early
twenties); similar chronological age of the spouses; and harmonious affection with parents during

Factors that may have become operative during marriage, rather than before, and therefore are regarded
as part of the complex of marital adjustment rather than among its determinants are early and adequate
orgasm capacity, especially of the wife; confidence in the spouse’s affection and satis-faction with degree
of affection shown; equalitarian rather than patriarchal marital relations, with special reference to the role of
husband; mental and physical health; and harmonious companionship based on common interests and
accompanied by a favorable attitude toward the marriage and the spouse (Kirkpatrick [1955] 1963, p. 394).


Every society has rules governing the assumption of the conjugal roles of husband and wife; there are also
discernible rights accruing to and obligations incumbent upon the individuals who assume

these roles. Marriage in all societies thus brings about a change in the jural status of the parties to the
contract. Where marriage is defined by the state, it is possible to describe most of its jural entailments by
reference to one or more legal codes adopted by that state. However, among many of the peoples studied
by anthropologists, the jural tenets governing marriage cannot be ascertained by reference to codes laid

Interpretation of family in the social group

down by a state and hence must be derived from the study of the recurrent patterns of behavior and of folk
models that prescribe ideal behavior.

Marriage entails not only a change in the jural status of the individuals who enter the roles of husband and
wife but also a change in the lawful status of specifiable consanguineal kinsmen of the individual partners.
In fact, it is the linkage of groups as well as of individuals that is crucial to the formulation of the difference
between marriage and its social analogues. Only marriage creates (or maintains) affinal relationships
between the kinsmen of individuals who claim the roles of husband and wife (see Fortes 1959, p. 209).
Even where it is socially admissible for individuals to presume conjugal status—that is, where they may
assume the husband—wife roles without their actions being legitimated according to prevailing jural rules
—this presumption of status does not generate lawful relations of affinity between kinsmen of the
“spouses” concerned.

The importance of affinity to an understanding of marriage is made clear through a consideration of the
nature of kinship. The social relations subsumed under the concept of kinship are of two fundamental
types which, though referable to the biological processes of heterosexual mating and procreation, cannot
be reduced to biology. Those social relationships based on parenthood and descent or, more precisely, on
parenthood and filiation, are generally termed consanguineal relationships. All persons related by socially
defined direct or shared descent are consanguineal kinsmen (P. Bohannan 1963, chapter 4). These “blood
relatives” are distinguished in all societies from affinal relatives, i.e., those whose kinship status is funda-
mentally grounded “in law.” Human mating is everywhere socially regulated, and adult mating for the
purpose of procreation is normally preceded by the creation of jurally derived kinship ties between the
mating pair and between certain of their respective consanguineal relatives. The continuance of publicly
acknowledged affinal kinship depends on adherence to prescriptions and proscriptions delimited by the
particular society under consideration. Whereas many societies make no provision for the legal severance
of consanguineal kinship bonds, they all provide for the severance—“by law”—of those which are based “in

Societies differ considerably with respect to the rules governing the way in which the roles of hus-band and
wife should be assumed, with respect to the specific rights and obligations which accrue to persons in
these roles, and with regard to the behavioral and jural attributes of the other affinal roles created by
marriage. Nonetheless, most anthropologists have regarded the institution of marriage as a universal in
human societies, and many have attempted to provide definitions of marriage sufficiently general to
encompass its various manifestations.

The fact that marriage is closely linked to parent-hood has led many scholars, including Westermarck,
Malinowski, and Radcliffe-Brown, to pro-pose definitions of marriage which center on what Malinowski
termed “the principle of legitimacy.” Thus, Radcliffe-Brown writes: “Marriage is a social arrangement by
which a child is given a legitimate position in the society, determined by parenthood in the social sense”

Interpretation of family in the social group

(1950, p. 5). The general, though by no means universal, acceptance of this formulation is indicated by the
fact that Notes and Queries on Anthropology defines marriage in an essentially similar, but by
implication more limited, manner: “Marriage is a union between a man and a woman such that children
born to the woman are the recognized legitimate offspring of both partners” (British Association for the
Advancement of Science 1951, p. 110).

Edmund R. Leach was among the first to argue that a definition of marriage in terms of legitimacy is too
limited. In his opinion, any attempt at a universal definition of marriage is inevitably “vain,” since the
“institutions commonly classed as marriage are concerned with the allocation of a number of
distinguishable classes of rights” (1961a, p. 107). Leach suggests that in most cases the institution of
marriage serves to allocate rights to either or both spouses; in some cases it serves primarily to allocate
rights to the husband and his wife’s brothers.

Despite Leach’s arguments against a universal definition of marriage, his formulations stimulated two fresh
attempts at universal definitions. Prince Peter of Denmark suggested that in light of Leach’s propositions,
marriage should be defined as “the socially recognized assumption by man and woman of the kinship
status of husband and wife” (Peter, Prince of Denmark 1956). The task of the anthropologist would then be
to ascertain and delineate the particular rights and obligations associated with these kinship roles in the
particular societies being studied.

H. Fischer (1956) called this definition tautological, on the grounds that the Oxford and Webster
dictionaries defined “husband” and “wife” respectively by phrases such as “a married man” and “a married
woman.” In a discussion of Nayar marriage, Gough agrees and reaffirms the heuristic value of a definition
of marriage based on “the principle of legitimacy.” In an attempt to overcome the difficulties inherent in any
formulation which de-fines marriage as a union of “a man and a woman,” and in an attempt to provide a
substantive definition for the concept of legitimacy, Gough suggests that marriage be defined as “a
relationship established between a woman and one or more other persons, which provides that a child
born to the woman under circumstances not prohibited by the rules of the relationship is accorded full birth-
status rights common to normal members of his society or social stratum” (1959, p. 32).

Her effort to refine the older, more general “principle of legitimacy” definition has yielded one which, on
close examination, is equally inadequate. Operating with such a definition, no investi-gator could classify
as married any particular woman who had assumed the jurally recognized kinship role of wife but who
had not borne children. Of course, the conditions under which a child would be accorded “full birth-status
rights” could be elicited by the investigator. However, for any given case, the researcher would have to
await the birth—or perhaps the conception—of a child be-fore he could ascertain whether conditions
entailed in the husband-wife relationship had been violated. Furthermore, Gough’s definition implies that in
any society each person having “full birth-status rights” is the child of a relationship which can be termed
marriage. Among various peoples of the world, “full birth-status rights” accrue to persons born of

Interpretation of family in the social group

relationships which are not recognized as marriage according to prevailing jural rules.

If a universal definition of marriage is to be formulated, it would seem that the one proposed by Prince
Peter should serve as a model. Fischer’s criticism of Prince Peter’s definition may be dis-regarded, since
dictionary definitions are usually unsatisfactory bases for discussions of roles. The roles of husband and
wife must be defined in terms of the essential rights and obligations and the behavioral attributes entailed
in them in any particular society. Gough and Fischer are justified in their concern that confronted with
different forms of mating, the anthropologist employing Prince Peter’s definition would be unable to decide
which institutions should be referred to as “marriage,” as “concubinage,” etc. However, if the statement
were modified so as to define marriage as the jurally valid and socially (or publicly) recognized
assumption of the kinship roles of husband and wife, there would be few or no problems concerning the
distinction between marriage and its socially recognized alternatives. Such a proviso emphasizes that the
publicly acknowledged kinship roles created by marriage—as opposed to its alternatives—derive support
from the juridico-political domain of the society. Of course, there may be more than one jurally valid way of
assuming the roles of husband and wife—as is the case in some present-day African states which
recognize marriages contracted according to one or more sets of “customary laws” as well as marriages
contracted in accordance with legal codes based on European models.

It would appear that the cross-cultural study of marriage must rest on the premise that all societies
recognize kinship roles which are founded “in law” as well as those which are based ultimately on actual,
assumed, or presumed genetic relation-ships. Fundamental to the understanding of the concept of
“lawfully based” kinship is the fact that human mating is everywhere subject to socially derived regulations.
While it is normally expected that marriage will lead to parenthood, the roles of husband and wife need not
be defined by reference to children who will come to be regarded as legiti-mate offspring of individuals in
these roles. The roles of husband and wife should be defined in terms of the rights and obligations which
attach to them, and marriage must be defined as the lawfully or jurally recognized assumption of these

Choice of spouses

In all societies, socially derived limitations are placed on the range of persons from among whom spouses
may be chosen. Regulations which prescribe marriage outside a stipulated group are referred to as rules
of exogamy. Kin groups such as lineages, or territorial groups such as bands or villages, may constitute
exogamous units. Societies possessing corporate unilineal descent groups usually prescribe that a person
select as spouse someone from a descent group other than his own. In some cases the selection may be
made from among persons within the descent group but outside specified degrees of relationship. Among
the Gisu of east Africa, for example, it is the minimal patrilineage, comprising persons who trace their

Interpretation of family in the social group

descent from an ancestor three to five generations removed from the oldest living generation, which
constitutes the exogamous unit.

Every society prohibits heterosexual mating between certain “close” consanguineal relatives. This

prohibition is referred to as the incest taboo, and ordinarily it proscribes mating between relatives who
stand to each other in the relationships of mother and son, father and daughter, and brother and sister. In
many societies the incest taboo is extended to various other kinsmen in the parental and filial generations.
Among some royal or ruling groups, as in dynastic Egypt and in Polynesia, relatives ordinarily prohibited
from mating may be preferred as marriage partners. The mating of close relatives is also permitted in
some societies on specified ritual occasions.

A rule of endogamy exists where the field of possible spouses is limited to persons within an individual’s
territorial and/or social group. The castes of traditional India are the most often cited example of
endogamous groups. Other societies also prescribe marriage among persons of the same social stratum.
Among the Swazi of south Africa, where lineage exogamy prevails and where royalty marries royalty, there
are frequent subdivisions of the royal lineage so as to make possible otherwise prohibited marriages. A
number of studies indicate that in the absence of explicit prescriptions, it is posssible to discover
endogamous tendencies within social or territorial groups of various size and scale.

In addition to proscriptions associated with incest and exogamy, societies usually prohibit marriage
between certain other categories of persons In some instances slaves cannot marry freemen. Where age
sets are a feature of social organization, as among the Nuer, a man may be prohibited from marrying the
daughter of another man in his age set.

Societies which prescribe that a spouse be chosen from among one or more designated categories of
persons have been said to possess closed marriage systems. Those in which such prescriptions do not
exist have been characterized as having open marriage systems. The designation of a marriage system
as “closed” is not meant to suggest total absence of choice in the process of mate selection. This point is
illustrated by Klass (1966), who shows that in Bengal (and in other parts of India), while caste affiliation
delimits the broad category of persons from which a spouse is chosen, a man who must choose husbands
for his daughters or “wards” does so from within a relatively narrow selection of eligible males known to
certain of his kinsmen.

The most frequently cited closed marriage systems are found among the indigenous societies of Australia.
Some of these societies, for example the Kariera, practice what anthropologists term “symmetrical cross-
cousin marriage,” wherein pairs of local groups engage in the “simultaneous or nearly simultaneous
exchange of women” (Leach 1961a, p. 59). The male members of the two groups concerned exchange
their “sisters” for “wives.” Ideally, a male ego marries his mother’s brother’s daughter, who may at the

Interpretation of family in the social group

same time be his father’s sister’s daughter and the sister of his own sister’s husband.

Among the Murngin of Australia is found a type of asymmetrical cross-cousin marriage wherein marriage
with the mother’s brother’s daughter is preferred and marriage with the father’s sister’s daughter is
proscribed. In this society and others practicing matrilateral cross-cousin marriage, a localized descent
group gives wives to one or more other such groups and receives wives from a different set of such
groups. In Murngin society there are descent groups which are allied through ties of kinship and ritual.
Moreover, each pair of such allied groups stands in balanced opposition to another similar pair with which
it exchanges women on a nonexclusive basis. Since men do not marry within their own moiety, any male
ego and his mother’s brother are in opposite moieties. Ego’s grou receives wives from and gives
prestations to his mother’s brother’s group. Ego’s mother’s brother’s group receives wives from and gives
prestations to the group with which ego’s group is allied. This latter group is the one containing ego’s
mother’s mother’s brother, who, of course, stands in the relationship of mother’s brother to ego’s own
mother’s brother. It can be said, therefore, that in Murngin society the “mothers’ brothers” stand in the
relation of “wife givers” to their sisters’ sons (see Leach 1961a, pp. 68-72).

Claude Lévi-Strauss, Edmund R. Leach, Louis Dumont, and others have discussed the economic and
political implications of this and other forms of “cousin marriage.” Leach (1961a, pp. 54-104) has shown
that where matrilateral cross-cousin marriage prevails, there exist permanent status differences between
wife-giving and wife-receiving groups and has demonstrated that the marriage system is not insulated from
other domains in the society. In fact, he argues that marriage alliance in such situations is but one of
“many continuing relationships between paired local descent groups.” Political and economic relationships
are reflected in and sustained by the system of matrilateral cross-cousin marriage.

In open marriage systems, the only group of persons unequivocally proscribed as marriage partners are
those to whom the incest taboo is extended. There are no normative prescriptions relating to groups from
which spouses should be chosen. Nonetheless, many studies indicate that demographic, ecological, and
sociological factors enter into the choice of spouse. Age, residential propinquity, class, religion, ethnicity,
education, and occupation have been isolated as important determinants in the choice of marital partners.
Likewise, parents and peer groups are often instrumental in delimiting for each individual the field from
which a spouse will be chosen.

The transfer of rights at marriage

Marriage involves the allocation of rights and obligations among the parties to the agreement. A number of
anthropologists have attempted to classify the various rights which are known to be allocated at marriage
in different societies.

In discussing the jural element in marital and other kinship relations, Radcliffe-Brown (1950, p. 12)

Interpretation of family in the social group

distinguishes personal rights (jus in personam) from possessive rights (jus in rem). A right in
personam confers on an individual or a group the power to order the performance of certain duties by
another individual or group. Rights in rem constitute claims on an object or person such that any
encroachment on the object or person constitutes a violation of the “possessor’s” rights. In most societies
husbands and wives have personal rights in each other: either spouse may claim certain duties of the
other. It is also common to find that a husband has “possessive” rights in relation to his wife. Her
seduction, her abduction, or her murder would constitute a serious infringement of her husband’s rights.

In an important contribution to the literature on marriage, Laura Bohannan (1949) distinguishes two
classes of rights in females which may be allocated at marriage. Rights in uxorem (rights in a woman as
wife) are distinguished from rights in genetricem (rights in a woman as mother).

In her discussion of Dahomean marriage, Bohannan shows that rights over a woman’s sexual powers and
certain of her domestic services were transferred from a woman’s patrilineage to the man or woman who
made the appropriate bride-wealth payments. In most of the “types” of Daho-mean marriage, rights to any
children a woman might bear during the course of her marriage were also transferred from a woman’s
patrilineage to that of her husband. Distinct classes of marriage payments were necessary to the transfer
of each of these two classes of rights. However, in certain “types” of marriage, rights in genetricem were
retained by the woman’s natal patrilineage; this might occur in cases where a lineage was faced with a
shortage of male heirs and one of the daughters of the lineage was given in marriage to a man who agreed
to make all the bride-wealth payments except those which would have given him jural authority over
children of the marriage. Moreover, the marriage of a woman of the royal lineage never involved the
transfer of rights in genetricem to the lineage of her huband (L. Bohannan 1949).

Even though it is usually rights in women which are in the forefront of marital negotiations, Leach has
pointed out that marriages also serve to allocate rights in and over men (1961a, pp. 107-108). He suggests
that a marriage may serve to do the following:

(1) To establish the legal father of a woman’s children.

(2) To establish the legal mother of a man’s children.

(3) To give the husband a monopoly of the wife’s sexuality.

(4) To give the wife a monopoly of the husband’s sexuality.

(5) To give the husband partial or monopolistic rights to the wife’s domestic and other labor services.

(6) To give the wife partial or monopolistic rights to the husband’s labor services.

Interpretation of family in the social group

(7) To give the husband partial or total rights over property belonging or potentially accruing to the wife.

(8) To give the wife partial or total rights over property belonging or potentially accruing to the husband.

(9) To establish a joint fund of property—a partnership—for the benefit of the children of the marriage.

(10) To establish a socially significant “relationship of affinity” between the husband and his wife’s brothers.

Leach thus focuses attention on rights in and regarding children, sexuality, domestic and economic
services, and property. In the last instance, he suggests that marriages may establish between groups of
men mutual interdependencies which could entail any of the above rights as well as others of a political

Where there are corporate kin groups, the allocation of rights at marriage is usually effected by and
between at least two such groups. In the case of first marriages, it is usual that the groups into which the
husband and wife were born are parties in this rearrangement of social relations.

Where recruitment to the corporate kin groups is based on patrilineal descent, normally the rights over a
woman’s sexuality and procreative capacities that are held by her natal group are transferred

at marriage to the groom and his natal group.

Thus, whereas prior to marriage any sexual offense against a woman constitutes a violation of rights held
by her kin group, after marriage such an offense is an infringement of the groom’s rights. Similarly,
whereas children born to a woman out- , side marriage would fall under the jural authority of her natal kin
group, those born after marriage are subject to the authority of, and have rights in, the groom’s kin group.

Total rights over the bride’s domestic and economic services are seldom transferred from her natal group
to her husband or his kin group. The woman herself, as an adult member of the society, may retain some
control over the dispensing of these services. Often her kin group retains the right to call upon these
services. Among the Yoruba of Nigeria, for example, rights in the bride’s sexuality, rights over her
procreative powers, and partial rights over her domestic services are acquired at marriage by the groom
and his patrilineage. However, a woman maintains control over her economic powers and resources, and
her natal lineage retains the right to call upon her domestic services in certain circumstances. She is called
upon to buy and prepare food at times when deities associated with her lineage must be propitiated, and
on the death of a member of her lineage, she is expected to be of service in various ways.

This raises another point: in most societies possessing corporate patrilineages, a married woman does not
usually relinquish all her rights in her natal lineage. She may retain some proprietary rights therein, and
she usually remains under the religious protection of her lineage ancestors. Moreover, a woman’s lineage

Interpretation of family in the social group

may have the right to reclaim control over her sexual and procreative powers should there be a breach of
the marital agreement on the part of her husband. While these statements are generally true, there are
some patrilineal societies in which a married woman becomes virtually “absorbed” into her husband’s
lineage. According to Gluckman (1950), a married woman among the Zulu of south Africa had virtually no
rights outside her husband’s lineage; once a woman was married, her natal lineage forfeited virtually all
authority over her.

Whatever rights are transferred to the husband or his lineage may be temporarily or permanently
reallocated by him or his lineage. The most common example of this is the practice of “wife-lending” found
among the Kipsigis and others. The fact that a man may permit another to have access to his wife’s
sexuality is proof of his monopoly over her sexual capacities. In some societies, a man who is impotent
may choose a sexual partner for his wife in order that she may bear children. Where this is so, the
husband is the lawful father of the children, even though he is not the genitor. Where a female is permitted
to assume the role of hus-band, she bestows her rights of sexual access to her wife on a man of her own
or of her wife’s choice.

In matrilineal societies, rights over the procreative capacities of women are held in perpetuity by their kin
groups while partial or total rights in their sexuality are transferred at marriage to their husbands.
Customarily, the husbands also attain rights to the domestic services of their wives. Among the Bemba of
east central Africa, for example, a husband has monopoly over his wife’s sexuality, but the children of any
marriage belong to their mother’s matrilineage and are therefore under the jural authority of the adult
males of that group. A wife keeps her husband’s house and con-tributes her labor to his agricultural

Marriages and the exchanges of goods and/or services occasioned thereby are sometimes processual
events extending over considerable periods of time. The rights and obligations entailed in the marriage
may be allocated in serial fashion, the timing of their transfer being dependent on the transfer of the
appropriate goods and services. In such cases, the exchange of goods and services may commence
during the period of betrothal and continue even after the formal transfer of certain rights has taken place.

Where goods and services are exchanged as part of the marriage procedure, certain of these may be
regarded as necessary prestations without whose exchange a transfer of rights will not take place. Others
are contingent prestations which, although part of the contract, are not essential to the ex-change of jural
authority and the assumption of marital rights and obligations. As Fortes says, they constitute the “means
of winning and preserving the goodwill of those with the power to transfer marital rights” (1962, p. 10).

The most general terms used to describe prestations entailed in the marital contract are those of
bridewealth (or bride-price) and dowry. The former refers to gifts presented by the groom’s kin group to
that of the bride, and the latter describes gifts made by the bride’s kin group to that of the groom. The
dowry is the more familiar to Westerners, since for centuries it has been a part of the marriage contract in

Interpretation of family in the social group

Europe. However, both bride-wealth and dowry have been reported for various parts of the world.

Throughout history, the transfer of rights at

marriage has been enshrined in ritual and ceremony. This is a correlate of the fact that marriage
transactions are always “publicly” acknowledged. The ceremonies which take place in effect call forth “the
public” to bear witness to the lawfulness of the transactions. The sanctions which emanate from the jural
domain of the society are strengthened by the incorporation of rituals associated with the religious realm of
the society.

Concurrent marriages. The transfer of rights at marriage and the rituals associated with this transfer signify
the assumption of new roles by the parties involved. In societies which permit polygyny or polyandry—
marriages entailing a plurality of wives or of husbands, respectively—one of the partners to a marriage
assumes the role of co-wife or co-husband along with the role of husband or wife.

In polygynous marriages, the husband usually acquires the same categories of rights in each of his wives.
In patrilineal societies, a man is the legitimate father of all his wives’ children, even though his rights over
the wives’ sexuality may be assigned to or “usurped by” other men. The children of polygynous marriages
may or may not have equal claims on their father’s property. In any case, each wife considers herself the
guardian of her children’s rights within the family created by the polygynous marriage.

Where polyandry is practiced, by definition a man does not have exclusive rights in his wife’s sexuality. He
may or may not have claims over the children which she bears him. Among the Sinhalese, rights over the
wife’s sexuality are partially vested in the first husband. The sexual rights of the other husbands are
exercised with the consent of the first husband and the wife. A husband has claims over those of the wife’s
children whom he has fathered, and the children have legitimate claims over the property of their
respective fathers. All the children have equal claims to the properties owned by their mother.

Among the Nayar of south India, a ritual marriage ceremony, called the talī rite, bestowed upon a group of
men of appropriate caste the right of access to a woman’s sexuality. The completion of the tali rite marked
a girl’s transition to woman-hood. Thereafter, when she attained appropriate age, she could begin to enter
into relationships, termed sambandham unions, with a number of men, for whom she might bear children.
Rights over a Nayar woman’s procreative powers were retained by her matrilineage, which had jural
authority over her children. Nonetheless, in order for a child to have “full birth-status rights” in his mother’s
lineage, he had to have an acknowledged father. A man acknowledged the paternity of a child by bearing
certain expenses associated with its delivery. This man could be any one of those with whom the mother
had entered into a sambandham union. In cases of doubtful paternity, a woman’s current “visiting
husband” could be forced by an assembly of persons in the neighbor-hood to make the birth payments.
“But if no man of appropriate rank could be cited as potential father, woman and child were expelled from

Interpretation of family in the social group

their lineage and caste” (Gough 1959, p. 30).

The levirate and the sororate

In many societies, an individual may assume the role of husband or wife in order to secure rights for a
kinsman. Where the “true” levirate prevails, upon the death of a husband, it is the duty of one of his
brothers to marry the widow, and any children born to the union are counted as the progeny of the
deceased man. Certain of the “ghost marriages” found among the Nuer resemble the levirate. A man could
marry a woman “to the name of” a brother who died childless, and the offspring of the union would be
designated as children of the deceased. These practices differ from the custom of adelphic widow
inheritance, wherein a man marries his deceased brother’s widow and bears children who are counted as
his own. Where the “true” sororate prevails, the husband of a barren woman marries her sister, and at
least some of the children born to the union are counted as those of the childless wife. The term “sororate”
is also used in reference to the custom whereby, upon the death of a wife, her kin supply a sister as wife
for the widower. In the latter case, however, any children born to the woman are recognized as her own.

Affinal relationships

The transfer of rights at marriage not only signals the couple’s assumption of new conjugal roles but also
serves to establish or perpetuate affinal relationships between consanguineal kinsmen of the spouses.
Often associated with affinal roles are behavioral attributes commonly subsumed under the categories of
“joking relationships” and “avoidance relationships.” Radcliffe-Brown (1952, pp. 90-116) has argued that
the respect implied in avoidance practices and the formalized disrespect demonstrated by joking
relationships are expressions of alliance or consociation. The actors in roles characterized by joking or by
avoidance have divergent interests which could generate conflict between them and thereby under-mine
the bases of their common interests. The institutionalization of avoidance and joking serves to minimize
the chance of the development of openly hostile relations between the parties.

The most widespread of the avoidance practices are those which restrict contact between a husband or
wife and the mother-in-law and/or father-in-law. Such restrictions on contact may also extend to actual or
classificatory brothers or sisters of the father-in-law or mother-in-law. Among the patrilineal Swazi of south
Africa, a wife is prohibited from coming into face-to-face contact with her husband’s father and those of his
male relatives of the same generation resident in the compound. A man behaves in similar fashion toward
his motherin-law, but the likelihood of such contact is minimized by their residence in different compounds
and often in different villages.

Joking relationships most commonly exist between a man or woman and affinal relatives of opposite sex in
the spouse’s generation. These relationships are characterized by the use of intimate names, the use of
language otherwise considered lewd or abusive, and, in some cases, by indulgence in sexual play.

Interpretation of family in the social group

Affinal relatives are often expected to give assistance to one another in times of exigency. In many
societies where political functions are vested in roles defined primarily by kinship criteria, affinal relatives
serve to minimize open conflict between their respective consanguineal kin groups. They might serve, as
among the Tiv of Nigeria, as emissaries of peace in cases of latent or open conflict between two lineages.

The linkage of individuals through marriage leads to the creation of new groups or, in NadeFs terminology,
to the creation of new sets of bounded social relationships and thereby constitutes a phase in the
developmental cycle of kin groups. As Radcliffe-Brown has pointed out, the eventual result of most
marriages is that new sets of individuals are linked through common descendants.

Ultimately, the fission of kin groups can often be traced to relations generated by marriage. This process is
evident in many societies where lineages (or, for that matter, ramages) are a feature of social organization.
When adult members of a lineage segment occupy a common residence along with their spouses and
children, the process of incorporation of additional coresidents through marriage often eventually leads to
the founding of households in other locations. In the course of time, the founders of such households and
their descendants may come to form new lineage segments.

Postmarital residence

In some societies, spouses are expected to live together throughout the period of their marriage; in others,
they may be members of separate domestic groups and only visit each other’s residences.

The “residence rules” outlined by anthropologists refer to situations in which husbands and wives are
members of the same domestic unit. Neolocal residence predominates when couples establish in-
dependent domestic units after marriage. Residence is characterized as virilocal when most couples in a
society join a domestic group in which the husband resided prior to marriage or in which he rather than the
wife has proprietary or other claims. Residence is called uxorilocal when couples join the domestic group
to which the wife was attached prior to the marriage or in which the wife rather than the husband has
claims. The above terms may be compounded with others to describe more precisely the nature of the
domestic group joined by the couple. Thus, viripatrilocal residence refers to domicile in a domestic group
whose core includes the groom’s father. Uxorimatrilocal residence refers to domicile in a group whose
core includes the bride’s mother. The term avunculocal is used to describe residence in a group whose
core includes the groom’s mother’s brother.

Data collected by Goodenough (1956) and J. L. Fischer (1958) among the Nakanai of New Britain show
that the classification of postmarital residence patterns is not as straightforward as some might assume.
Their data also illustrate that there is no simple correlation between particular residence rules and
particular rules for recruitment to descent groups. Goodenough shows that in this matrilineal society, a
man takes his bride to live in the village in which his father resides. The couple lives there so long as the

Interpretation of family in the social group

groom’s father is alive, and they may remain after the father’s death, particularly if the father is without
sisters’ sons who would be his jural heirs. More often, however, after the father’s death, the couple moves
to the residence of the husband’s matrilineage, in which he has hereditary land rights. A man whose father
is deceased takes his bride to live with the group which includes the man who acted as father-surrogate at
the time of the marriage. Goodenough shows that even where ideal residence patterns suggest one or
more prevailing modes of residence, the actual choices which couples make may depend on economic
and other factors.

Fischer, who also worked on the island of Truk and who arrived at a classification of residences
significantly different from Goodenough’s, has suggested that residence be elicited for individuals than
rather for married couples. He suggests that every person in a household has a “kin sponsor”

and that his relationship to this sponsor most appropriately describes the residence pattern for that

While Fischer’s suggestion has some merit for the classification of residence patterns for entire
populations, attention cannot be shifted from the fact that in most societies the major spatial arrangements
of individuals are associated with marriage. Moreover, the kinds of rearrangements which do occur have
important implications for many kinds of social relations. It has been shown, for example, that the study of
the developmental cycle of domestic groups touches on virtually all aspects of social structure and that
postmarital residence patterns are crucial to the understanding of the development of domestic groups
(Goody 1958).

Alternatives to marriage

Marriage is a process or event signifying the assumption of the roles of husband and wife in accordance
with jural tenets prevalent in the society or stratum of society to which the parties belong. In contemporary
societies, marriages are contracts which must be formally legitimized by the state. A state may provide that
for purposes of inheritance, or for other specified purposes, persons who are not legally married to each
other but who share a common domicile and who otherwise demonstrate a claim to conjugal status may be
accorded some or all rights associated with legal marriage. Similarly, a state may choose to recognize
marriages contracted according to rules formulated prior to its existence by some or all of the groups which
constitute it. Such is the case in various parts of the world where formerly autonomous or semiautonomous
political entities have come together to form modern nation-states.

Unions other than lawful marriage are known to have existed in stateless societies as well as in states

Interpretation of family in the social group

which did not make the legitimization of marriages their official concern. Yet it seems particularly
characteristic of modern societies that there are individuals who, for various reasons, assume some or all
of the obligations and rights associated with the roles of husband and wife without entering into legal
marriage. Reference has already been made to the fact that one of the crucial ways in which such unions
differ from marriage is that they do not create lawful kinship ties between consanguineal relatives of the

These “consensual unions” occur in different frequencies in different modern societies. In the United
States, in the Caribbean, and in other areas where such unions occur with relatively high frequency among
certain socioeconomic classes and/ or ethnic groups, research has centered primarily on family
organization, and consensual unions are often regarded as but one aspect of over-all “family instability.”
The result is that while many hypotheses have been offered to account for the matrifocal or matricentric
family which, in some areas, is one structural correlate of consensual unions, few students have offered
hypotheses which explicitly attempt to account for the origin and/or persistence of such unions.

M. G. Smith (1962) has presented a wealth of statistical data in support of his hypothesis that specific
mating patterns underlie the various forms of family organization in the Caribbean. He has demonstrated
that the pattern of consensual mating underlies the matrifocal family in that area. However, he does not
deal with the origin and persistence of the mating patterns themselves. Nevertheless, the data suggest that
demographic and economic factors are important determinants of these patterns. For example, where the
sex ratio is altered by the necessity that males migrate to find work, women often enter into extramarital
unions with single or married men who remain behind. Such alliances may or may not entail co-residence.

Consensual unions may constitute a stage in the development of domestic groups and as such are not so
much alternatives as preludes to marriage. In parts of the Caribbean where great prestige is attached to
church marriages followed by festivities requiring the outlay of large sums of money, couples often assume
the roles of husband and wife by mutual consent until such time as they can afford a religious marriage
ceremony. Thus, many couples establish a common domicile and bear children before they enter into
matrimony “before the eyes of man and of God.”

This raises an important point. Even though, in most parts of the modern world, marriages may be
contracted without religious ceremonies, historically marriage was the concern of religious institutions
before it became the official concern of the state, and most religious doctrines still include prescriptions
and proscriptions regarding marriage. Where the influence of religious tradition is particularly strong, civil
marriages may be regarded as little more than alternatives to or complements of “true marriage.” One of
the consequences of this, as evidenced in parts of the Caribbean, is that couples enter into extramarital
relationships until such time as they can finance the religious and convivial ceremonies as well as fulfill the
legal requirements for marriage.

Interpretation of family in the social group

Marital stability and divorce

The ambiguities entailed in the concept of marital stability have been succinctly stated by David Schneider:

Stability may be defined in terms of the change of rules or expectations over time or in terms of the degree
to which the rules or expectations are conformed to. Stable marriage may be defined as stable jural
relations irrespective of conjugal relations, as stable conjugal and jural relations, or simply as stable
conjugal relations. (1953, p. 56)

Thus, divorce, defined as the lawful dissolution of jural ties established at marriage, may occur relatively
infrequently even though separation and other breaches in conjugal relations occur relatively frequently. In
traditional Nuer society, the jural bonds established by marriage were stable; divorce, signified by the
return of bride wealth, was rare. On the other hand, conjugal separation was relatively frequent. Max
Gluckman (1950) was one of the first anthropologists to deal with the factors which contribute to the jural
stability of marriage in preindustrial societies. His data on the Lozi and the Zulu led him to the hypothesis
that the stability of jural relations established by marriage is correlated with the presence of patrilineages.
He argued that where the “principle of father-right” prevailed, as among the Zulu, there was a complete
and final transfer of women into their husbands’ lineages (from which their children obtained their legal
rights), and he suggested that this fact accounted for the virtual absence of divorce in such societies.

In a reconsideration of the Gluckman hypothesis, Fallers pointed out that not all patrilineal societies
provide for the absorption of women into their husbands’ lineages. He suggests that where women retain
rights in their natal patrilineages, patriliny contributes to the jural instability of marriage by dividing the
loyalties of spouses. Fallers (1957) found among the Busoga a relatively high incidence of divorce, which
he attributed in part to the fact that loyalties to natal lineages undermined the bonds established at

Leach (1961a, pp. 114-123), Cohen (1961), and others have contributed to the discussions of marital
stability begun by Gluckman, Schneider, and Fallers. However, there is yet to be undertaken the
quantitative and comparative analyses required for a definitive statement on the determinants of stability in
the jural aspects of marriage. Whether the aim is to isolate the determinants of differential rates of divorce
within a single society or to account for the differences in the divorce rates reported for various societies,
care must be taken to insure that the data utilized are in fact representative of the populations discussed.
Moreover, more attention must be given than has been in the past to the limits of the utility of numerical
data, which, at best, can be considered reliable for relatively short time spans.

The separation of spouses is usually taken as an index of instability in conjugal relations. However, it
should be obvious that separation can only be taken as indicative of the disintegration of conjugal bonds
when the coresidence of spouses is a societal norm. Even in these cases, separation does not always
signal instability in conjugal relations. Among the Yoruba, it is common to find women living and working in

Interpretation of family in the social group

one place while their husbands live and work in another. So long as these women are not known to have
committed adultery, and so long as they fulfill certain responsibilities to their husbands and their husbands’
lineages, their conjugal relations are not necessarily impaired.

The distinction drawn by Schneider between stability in conjugal relations and stability in the jural aspects
of marriage relations is useful in the analysis of marriage in contemporary societies. For example, it would
be useful to make such a distinction in discussions of marriage patterns in the Caribbean and in the United
States. As has been pointed out, among some of the lower-class populations in these areas, consensual
mating is common. Not all the parties to consensual unions are persons who have never been legally
married. In fact, where the economics of divorce are a deterrent to the lawful dissolution of marriage,
consensual unions are often an alternative to divorce and remarriage. Hence, the jural relations
established at marriage are often maintained even though conjugal relations are completely or partially

Most of the societies whose marriage systems are described in the anthropological literature are now
incorporated into independent states. The very existence of these states signals changes in the rules
regarding the establishment of marital contracts, since all contemporary states reserve the right to define
what types of union constitute legal marriage.

There is general agreement that the economic and demographic changes taking place in the “developing
areas” are also effecting changes in traditional marriage systems. However, Goode (1963) has pointed out
the difficulties involved in isolating cause—effect relationships between changes in a society’s family
patterns, including marriage, and

changes in its economic organization. Considerable refinement in research strategies is necessary before
it will be possible to state with confidence the extent to which, the precise ways in which, and the specific
points at which the spread of industrial technologies and the growth of cities impinge upon or serve to
undermine traditional family structures and marriage patterns.

Some of the studies of marriage found in the anthropological literature provide convenient points of
departure for investigations of changes in the rules and behavior associated with marriage in different parts
of the world. However, it is obvious that analyses of changing patterns of marriage require the collection of
a larger body of quantifiable data than is available in most existent anthropological studies of marriage.
Whereas most of the marriage systems described in the anthropological literature lent themselves to
representation in terms of mechanical models, such models are becoming increasingly inadequate as
representations of particular systems and as bases for comparative studies. The rules governing the
establishment of marriage contracts, the factors influencing the choice of spouses, the rights and
obligations entailed in conjugal roles, and the behavior of persons in these roles are sufficiently variable in
any one system to require partial or total representation by means of statistical models. With the
construction of such models, we can begin the assessment of the directions and rates of change in

Interpretation of family in the social group

marriage systems and the isolation of the specific variables which account for these changes.


All societies prohibit marriage with certain relatives, but some societies complement this prohibition by
prescribing, or preferring, marriage with other relatives. In this way two kinds of cousins are sometimes
distinguished, marriage being prohibited between those who are children of siblings of the same sex
(“parallel cousins”), while it is prescribed between children of siblings of opposite sex (“cross-cousins”).
This disposition is generally accompanied by exogamy. This article attempts to sum up recent
developments in the theory of cross-cousin marriage.

Descent and alliance

The expression “marriage alliance,” in which “alliance” refers to the repetition of intermarriage between
larger or smaller groups, denotes what amounts to a special theory of kinship, a theory developed to deal
with those types of kinship systems that embody positive marriage rules, though it also affords certain
general theoretical insights regarding kinship. Two points may be noted at the outset: (1) The combination
of the positive marriage rule with exogamy, or at the very least with a prohibition against marriage between
parallel cousins, is essential to the type of system under description here; a preference for marriage with
the father’s brother’s daughter, as found among some Islamic peoples, is a quite different phenomenon.
(2) The approach here presented is essentially common to several writers, though an element of personal
interpretation is inevitable.

In the initial stages of kinship studies, the re-construction of fanciful marriage rules (or mating
arrangements) as having supposedly existed in the past was widely used in order to explain seemingly
strange ways of classifying relatives (kinship terminologies). This practice has brought discredit, in the
eyes of some, to the study of both marriage rules and terminologies. In 1871 Lewis Henry Morgan made
two assumptions: (1) terminology reflects behavior, and hence, (2) if a terminology cannot be understood
from present behavior, it must be because the behavior it reflects belongs to the pastQuite apart from the
difficulty of reconstructing past behavior, anthropological thought in this matter is still ethnocentric. The
underlying assumption is that all peoples entertain the same ideas about kinship; their classifying of
relatives in different ways is, therefore, due to differences in behavior. Fully excusable in Morgan, such an
assumption is less so today.

W. H. R. Rivers recognized the link between an actual marriage rule (symmetrical cross-cousin marriage)
and a certain type of terminology (often called “bifurcate merging”). For Rivers, the marriage rule was the
cause, the terminology the effect, and he saw his task as explaining the marriage rule itself. [See the
biography of RIVERS.] Once again, terminology reflects behavior, and again historical speculation is
called in, this time to discover the “origin” of one item, which is in fact essentially a normative trait. In our

Interpretation of family in the social group

time the different features of a kinship system are, in practice, often considered in isolation or are
hierarchized according to what is assumed to be their degree of reality or determinativeness. This
tendency, if not found in such crudity as in the past, still exerts considerable pressure even on the best
minds, and that it constitutes a major obstacle to the understanding of certain kinship systems can be
shown by the example of Australian kinship, a classical subject for kinship theory. In Australian section
systems, de-scent is overstressed; the reasons that may elsewhere justify this emphasis are here
misplaced, for it prejudices the consideration of other elements in the system.

In writing about Australian kinship systems, authors vie with each other in stressing that in

symmetrical cross-cousin marriage arrangements, double descent is always present or implied. This is
unobjectionable in itself, but in the literature it is accompanied by a bias which makes itself obvious by
repetition, whether it be in B. Z. Seligman’s at-tempt to reduce the “type of marriage” to “forms of descent”
(1928, p. 534), however strange the latter forms may appear, or in Radcliffe-Brown’s overemphasis upon
descent, or in Murdock’s out-bidding of Radcliffe-Brown in this respect. Radcliffe-Brown was not content
with finding an underlying matrilineal exogamy in his classic Australian patrilineal systems and with seeing
in what is now called “double descent” a widespread principle of Australian kinship. He claimed that his
second kind of exogamous group actually “existed,” whereas he had only inferred it (1931, pp. 39, 439);
the point is insisted upon by Goody (1961, pp. 6 ff). It is perplexing later on to find Murdock opposing
Radcliffe-Brown, while praising the same discovery in others; but the crux of the matter is that in Murdock’s
opinion Radcliffe-Brown had not gone far enough in stressing descent and descent groups, for Radcliffe-
Brown had maintained, at another level, the primacy of individual relation-ships and marriage rules over
the arrangement of groups (Murdock 1949, pp. 51 ff.).

Actually, the hypothesis of underlying matrilineal exogamy among the Kariera and Aranda accounts for the
allocation of alternate generations to different groups. Among them, the patrilineal group is conceived not
as a unity over a continuous series of generations but as a duality made up of two alternate generation-
sections, called by different names and following different marriage rules (the grandson falling back, so to
speak, into the grandfather’s section). This is the simple, concrete sociological fact, widespread in
Australia. If we take this for granted, together with intermarriage between the named sections, we can in
each case draw a simple diagram of the whole tribe. In Figure 1 the sign [=] denotes intermarriage in both
directions, the letters A, B, etc., represent patrilineal groups, and the numbers 1 and 2 are used for the two
alternating generation-sections in each patrilineal group. The system of Ambrym (Balap) is easily
represented in the same fashion (Deacon 1927). All three systems represent variations on the same
theme, the number of patrilineal groups being respectively two, four, and three, the number of sections
four, eight, and six. Each of the three systems may be conceptualized as forming a single whole through a
regular chain of intermarriage and patrilineal descent. The differences in the arrangement follow
necessarily from the numbers of groups (for details, see Dumont 1966). I do not pretend that a second
unilineal principle cannot be said to underlie these systems, but only that the above is a simpler view of

Interpretation of family in the social group

them. Let us now turn to the general theory that, like the above analysis, recognizes intermarriage as a
basic element in those systems which possess a preferential or prescriptive marriage rule.


Since 1949 the Lévi-Straussian theory has been tested and has undergone partial modifications and
developments. To mention only the major themes, we have first the clear-cut distinction, advocated by
Needham, between prescription and preference in marriage rules. He claims that prescription alone has
“structural entailments” in the total social system, and that Lévi-Strauss has dealt only with prescription or
at any rate should have done so (Needham 1962). “Prescription” is here defined more as the characteristic
of a system than as simply a marriage rule: it involves the combination of a rule prescribing some relatives
and prohibiting others, a corresponding terminological distinction, and a sufficient degree of observation of
the rule in practice (Needham 1958a, p. 75; 1958b, p. 212). The advisability of the distinction has been
challenged by R. B. Lane (1962, p. 497). At first sight the distinction seems justified, and there is no
objection to isolating a clear-cut type of “prescriptive alliance.” That there is a danger of underestimating
the importance of other types is apparent from the exacting criteria by which the author excludes the
recognition of forms of patrilateral intermarriage as “prescriptive” in his sense (Needham 1958b). These
latter forms, like preferential marriage in general, do have “structural entailments” of a kind, as we shall
see. Moreover, the two forms are not easily distinguishable; the distinction, so presented, is more one of
levels than of systems (for a recent clarification of this question, see Maybury-Lewis 1965).

The main development has probably been a refinement of the concept of alliance and the substitution of a
more structural for a more empirical notion. At the start the theory, although anchored in the notion of
complementarity, was in large part concerned with the exchange or circulation of women between the
major exogamous components of the society. To begin with, three authors have asserted that the units
which may be said to ex-change women are, in concrete cases, smaller than the exogamous units. In 1951
Leach sternly insisted—with empirical, if somewhat dogmatic, good sense—that the agents arranging
marriages are as a rule the males of the local descent groups, as distinct from the wider exogamous units
and from the “descent lines” used in terminological diagrams and often unwittingly reified by the analyst
into actual groups (see Leach 1961, p. 56; cf. Needham 1958a). Quite logically, Leach went on to criticize
the assumption that a matrilateral marriage rule should necessarily result in the groups intermarrying “in a
circle,” an idea which Needham, on the other hand, tried to refine (1958a; 1962). A criticism from Berting
and Philipsen may also be noted: to be meaningful, they suggest, the “marriage cycles” must be limited in
number, and the people themselves must be aware of them (Needham 1961, p. 98). While such “alliance
cycles” (Needham) do meaningfully exist in some cases, their existence does not exhaust the function or
meaning of marriage alliance. On this all our authors agree, for Lévi-Strauss (1962, p. 333) himself
recently recognized—if my interpretation is correct—that “conscious rules” have emerged from recent
research as more important than their results in terms of “exchange.” Leach had pointed out that, in the
absence of cycles, the basic relationship is “one of the many possible types of continuing relationship

Interpretation of family in the social group

between paired local descent groups” (1961, p. 101). Elsewhere, while marriage alliance does not result in
a system of exchange at the level of the group as a whole, it is an integral part of the system of categories
and roles as conceived by the people studied (Dumont 1957, pp. 22, 34).

Needham has gone furthest in submitting LéviStraussian structuralism to criticism from the in-side and in
referring the “mediating” concepts of exchange and reciprocity back to that of (distinctive) opposition
(1960, p. 103). The more fundamental “integration” is not that of groups but rather that of the categories as
it occurs within the social mind: the marriage rule is part and parcel of this system of ideas. Like everything
else, social relationships are defined by classification. Studying the “symbolic order” of the Purum and
others, Needham (1958a) found that asymmetrical inter-marriage, although it could not function with less
than three intermarrying or “alliance groups,” can be dualistically conceptualized (wife-givers and wife-
takers) in accordance with an over-all dualist scheme. Here are found “structural entailments” different
from the group arrangements on which attention had first focused. The expression “marriage alliance” thus
covers both the general phenomenon of mental integration and the particular phenomenon of group

In its restricted field this truly structural theory alone transcends the bias inherent in our own culture. Such
expressions as “cross-cousin marriage” are technically useful but basically misleading. Real understanding
is reached when the marriage rule understood as marriage alliance is seen as giving affinity the diachronic
dimension that we tend to associate only with descent and/or consanguinity. By this means we are able to
transcend the limitations of thinking based upon our own society and make comparisons in terms of the
basic concepts involved (consanguinity and affinity).

Much remains to be done. Certainly the implications of marriage alliance for status, economy, and political
organization (i.e., the physiology of the system) should be worked out (Leach 1961, chapter 3). But even
regarding the morphology, our analyses are as yet imperfectly structural; we still take too much for granted
in the study of terminologies. Before attempting ambitious (re)Constructions, the basis in comparative data
must be strengthened and extended, and we must obtain a clearer view of the limits of the logical
integration of features, or conversely, of the plasticity and tolerance of systems, which can in some cases
go so far as to deny in effect the ideological primacy postulated above in principle.

Blake, Judith 1961 Family Structure in Jamaica: The Social Context of Reproduction. New York:
Free Press.

Bohannan, Paul 1963 Social Anthropology. New York: Holt.

Goode, William J. 1959 The Theoretical Importance of Love. American Sociological Review 24:38-47.

Interpretation of family in the social group

Goode, William J. 1960 Illegitimacy in the Caribbean Social Structure. American Sociological Review
25: 21-30.

Goode, William J. 1963 World Revolution and Family Patterns. New York: Free Press.

Hsu, Francis L. K. 1948 Under the Ancestors’ Shadow: Chinese Culture and Personality. New York:
Columbia Univ. Press.

Kapadia, Kanailal M. (1955) 1958 Marriage and Family in India. 2d ed. Bombay: Oxford Univ. Press.

Kirkpatrick, Clifford (1955) 1963 The Family as Process and Institution. 2d ed. New York: Ronald

Lang, Olga 1946 Chinese Family and Society. New Haven: Yale Univ. Press.

Levy, Marion J. 1949 The Family Revolution in Modern China. Cambridge, Mass.: Harvard Univ.

Malinowski, Bronislaw (1929) 1962 Marriage. Pages 1-35 in Bronislaw Malinowski, Sex, Culture and
Myth. New York: Harcourt.

Murdock, George P. 1937 Comparative Data on the Division of Labor by Sex. Social Forces 15:551-553.

Sumner, William Graham (1906) 1959 Folkways: A study of the Sociological Importance of Usages,
Manners, Customs, Mores, and Morals. New York: Dover. → A paperback edition was published in
1960 by New American Library.

Winch, Robert F. (1952) 1963 The Modern Family. Rev. ed. New York: Holt.

Winch, Robert F. 1958 Mate-selection: A Study of Complementary Needs. New York: Harper.

Yang, Ch’ ING-K’UN 1959 The Chinese Family in the Communist Revolution. Cambridge, Mass.:
M.I.T. Press.

Zimmerman, Carle C. 1947 Family and Civilization. New York: Harper.

Interpretation of family in the social group