Kinship, equality, and hierarchy

This work is licensed under the Creative Commons | © Cécile Barraud. ISSN 2049-1115 (Online). DOI: http://dx.doi.org/10.14318/hau5.1.011


Kinship, equality, and hierarchy

Sex distinction and values in comparative perspective

Cécile BARRAUD, Centre Asie du Sud-Est, Centre National de la Recherche Scientifique et École des Hautes Études en Sciences Sociales, Paris

This article deals with the notion of equality in Louis Dumont’s work, specifically relative to ideology and sex distinction in kinship terminology. It provides a new reading of his writings on kinship, analyzing his discussion of South Indian terminology and marriage, the contrasting value of affinity and consanguinity in caste society, and equality of the sexes. Dumont’s works on kinship are used to open a comparative examination of sex distinction in kinship terminology. This article seeks to extend this analysis of equality and hierarchy in kinship to other societies and ideologies, in order to explore the applicability of this approach to understanding of contemporary transformations in kinship relationships.

Keywords: value, Dumont, affinity, gender, comparison

Equality is a recurrent theme in discussions of Euro-American societies. It is both an aspiration and an impossibility which stands in contradistinction to Euro-American notions of power (Dumont 1986; Robbins 1994). As an aspiration, it goes in tandem with the idea of liberty, both linked to the value of individualism, which prevails in these societies. In his comparative works on ideologies, Louis Dumont opposed equality to hierarchy through analysis of the Indian caste system.

He discussed the relationship between hierarchy and equality in nearly all his texts, beginning with Homo hierarchicus (1980a). In his introduction to From Mandeville to Marx, he contrasts hierarchy, as the paramount value of the caste system, with equalitarianism, as one of the main values in “modern” societies (Dumont 1977: 3). He classifies these two systems as individualism and holism. “Holism,” he explains, “entails hierarchy while individualism entails equality,” but this does not [228]indicate that “all holistic societies stress hierarchy to the same degree, nor do all individualistic societies stress equality to the same degree” (ibid.: 4). Furthermore, Dumont explains, “equality and hierarchy must combine in some manner in any social system” (ibid.: 5). He claims that “it is possible for equality to be valued … without its being an entailment of individualism” (ibid.: 5).

This essay grapples with this uncertain definition of equality. Is equality opposed to power, or to hierarchy? If, up to a certain point, equality is opposed to power, is it the same notion of equality which can sometimes be present in a hierarchical ideology? And in the latter case, what is the meaning of this type of equality? The surprising discovery of elements of equality within the hierarchal ideology of the caste system, in Dumont’s study of Dravidian kinship terminology, prompts closer examination.

Dumont’s contribution to kinship studies is substantially less known1 than his work on ideology or caste hierarchy in India. This, however, has not always been the case. His first published piece on kinship, in 1953, the provocative article “The Dravidian kinship terminology as an expression of marriage,” elicited vivid comments from A. R. Radcliffe-Brown (Dumont 1983: 18–35). Since then, numerous authors have engaged with Dumont’s ideas on South Indian kinship. His approach to kinship sparked significant debate among specialists of India and Indian kinship regarding the nature and terminology of Dravidian kinship (see Trautmann 1981; Rudner 1990; Parkin 1992; Pfeffer 1993; Busby 1997; Viveiros de Castro 1998). In the piece Affinity as a value, published in English in 1983, Dumont responded to his critics and put forth his principal arguments on “affinity” in an effort to end years of contentious debate on topic (see Madan 1986 for a review).

Over the past twenty years, Dumont’s ideas on kinship have resurfaced in Amazonian anthropology. In the 1970s, Joanna Overing (1973, 1975) introduced Dumont’s Dravidian model of kinship to scholars in Amazonia. Three decades later, Eduardo Viveiros de Castro (2001) argued for the relevance of Dumont’s arguments on equality between consanguinity and affinity in Amazonia, drawing on Dumont’s notion of affinity as a value to develop his theory of “perspectivism.” His analysis diverges, however, from Dumont’s model, in that Dumont argued that equality between consanguinity and affinity gained its relevance within the hierarchy of the caste system. Recent scholars (e.g., Gregory 2010, 2013) have again turned their attention to the “value question” in India, renewing the debates on affinity by approaching the value of other kin relations.

In this article, I examine a question largely overlooked in debates on “affinity as a value”: that of the equality of the sexes in South Indian terminology, which underlies Dumont’s argument of the equivalent (equal) value of affinity and [229]consanguinity. Dumont defined the notion of equality in opposition to the encompassing hierarchy of the caste system.2 I will examine this argument relative to his later considerations of Western gender equality and kinship, which are relevant to current debates on gender.

Following closely Dumont’s analyses, in this text, I will make use of his own written terms: equality between the sexes and sex distinction. In line with arguments in the field of gender studies, it is first important to point out, however, that the terms “sex” and “gender” always point to a social construct which is culturally defined. While the notion of gender allows for the examination of diverse social representations, symbols, social relationships, and so on, when used in the context of kinship terminology, it is restricted to a category of classification, such that “sex” is a more apt descriptor than “gender.” In this respect, the expression “sex distinction,” like generation, affinity, collaterality distinctions, and so on, are used as criteria of kinship terminological analysis. These criteria or distinctions do not attribute any intrinsic value to kinship terms. Thus, the notion of equality of the sexes in kinship terminology does not imply equal relations between the genders. Following Dumont, I approach the question of equality from the narrow perspective of kinship terminology and, as he does, deal almost exclusively with kinship terminology rather than the kinship practices fundamental to gender studies (Busby 1997; Collier and Yanagisako 1987a). I employ the term “sex distinction,” as opposed to “gender difference,” to refer to relations between the sexes on a strictly terminological level.

Debates in gender studies focused on kinship have yet to specifically examine kinship terminology, although the relationship between gender and kinship, both dichotomous analytical domains, has been widely discussed and criticized (MacCormack and Strathern 1980; Ortner and Whitehead 1981; Strathern 1987, 1988; Moore 1988; 1999; Atkinson and Errington 1990; del Valle 1993). Neither gender nor kinship was capable of imposing the idea that all facts are cultural facts and not natural or precultural givens. At the heart of kinship dichotomies is the analytical division between “domestic” and “politico-jural” domains, which leads us back to twentieth-century kinship theories, frequently discussed by early scholars of gender studies. One of the notions that these scholars questioned is the “biological difference in male and female roles in sexual reproduction” presumed to be “at the core of men’s and women’s relationships everywhere” (Collier and Yanagisako 1987b: 7), such as Meyer Fortes’ (1949) discussion of procreation and child rearing as “natural facts” (Yanagisako and Collier 1987: 31).3 As a social construct, the notion of gender replaced the notion of sex and it became clear that what is meant by “gender” did not refer to the biological difference between men and women. In gender studies, “sex” is finally approached as one aspect of “difference” which includes other types of differences such as class, ethnicity, race, and religion. As “natural” biases and differences were progressively rejected, they have been replaced by the study of “culturally constructed social inequalities” (ibid.: 15), leading to the [230]hypothesis that all social systems are systems of inequalities, kinship included. In this context, it is all the more surprising that the analysis of terminologies was not fully included in the effort to understand inequalities or differences between men and women. Attributing a term to different types of social relationships, kinship terminologies inevitably mark differences between relatives, more or less numerous according to terminological systems, but at minimum distinguishing generation, age, and sex. What then, is the significance of these differences which are not accounted for? Do they constitute inequalities? As modes of classification, kinship terminologies inevitably inform the ways people are classified as relatives, without distinguishing their roles or functions.

The examination of Dumont’s works both on equality and on kinship terminologies is located at the interface of these questions. The first part of this article considers Dumont’s notion of equality, as presented in his comments on Tocqueville’s thoughts on egalitarian ideology and in his own writings on gender. Having clarified the terms of the discussion, the central part of the article examines the intersection of caste hierarchy and kinship terminology. I analyze the status of affinity and consanguinity, as well as the equality of the sexes in South Indian kinship terminology. I then present my own ethnographic material from eastern Indonesia to further explore Dumont’s analysis of kinship terminology. Finally, I briefly address Dumont’s comparison of kinship terminologies in North India and Western societies, relative to ideology. In conclusion, I suggest some research directions concerning the status of consanguinity and affinity in contemporary societies.

Dumont’s comments on Tocqueville’s notion of equality

Dumont only rarely alludes to questions of equality between men and women. One such instance is found in his response to Tocqueville’s definition of egalitarian ideology (Dumont 1980a:13). What interests Dumont here is Tocqueville’s comparative approach. Tocqueville reflects on the American egalitarian mentality through the eyes of a European, contrasting American ideology “with what he perceived of the hierarchical mentality in the France of the Ancien Régime” (ibid.: 15). Dumont claims that this represents “one of Tocqueville’s very important ideas, which concerns the place of modern political ideology in relation to values as a whole” (ibid.: 14).

Concerning equality, Dumont contends that for Tocqueville, “The first feature to emphasize is that the concept of the equality of men entails that of their similarity” (ibid.: 15). According to Tocqueville, equality thus abolishes differences, such that all human being are alike. Dumont writes,

So long as equality is only an ideal requirement expressing the transition in values from man as a collective being to man as an individual, it does not entail the denial of innate differences.4 … But if equality is conceived as rooted in man’s very nature and denied only by an evil society, then, [231]as there are no longer any rightful differences in condition or estate, or different sorts of men, they are all alike and even identical, as well as equal. This is what Tocqueville says: where inequality reigns, there are as many distinct humanities as there are social categories, … the reverse being true in egalitarian society. (Dumont 1980a: 16)

Dumont then notes that Tocqueville “seems to conflate the social form and the ‘natural’ or universal being” (ibid.). He notes one exception, however, stating that Tocqueville does make clear this distinction “when he contrasts the way in which the equality of man and woman is conceived in the United States and in France” (ibid.). Dumont then quotes this passage in Tocqueville but does not comment further on the relationship between men and women until his discussion later in the piece regarding hierarchical opposition.

Tocqueville and gender equality in America

Let us now turn to Tocqueville’s discussion of equality (Tocqueville 1840: book 3). Following chapters devoted to masters and servants and to “kindred” relations, Tocqueville considers the education of young women and “The young woman in the character of a wife.” Throughout these four chapters, he repeatedly praises American women for their indomitable energy, the firmness of their reason, their fearlessness, courage, and freedom of choice. In chapter XII, titled “How the Americans understand the equality of the sexes,” Tocqueville juxtaposes several dyadic relationships. He states, “I believe that the social changes which bring nearer to the same level the father and son, the master and servant, and superiors and inferiors generally speaking, will raise woman and make her more and more the equal of man.” As François Furet (1981: 32) has pointed out, for Tocqueville, equality of conditions

does not mean that master and servant are really equal, but that they can be, or that the provisional relationship of subordination does not constitute a “state” that completely defines each, because, for example, it can be reversed according to their mutual success. Since the servant can become master, and aspires to become master, the servant is not different from the master,

—they hold the same rights, the servant a citizen like the master. It is important to note here that these dyadic relationships are not equivalent: the servant may become a master, but by nature the son cannot become his own father (see Dumont’s remark above regarding the conflation of “the social form and the ‘natural’ or universal being”). But with regard to men and women, the question of equality and similarity arises differently in Tocqueville. He criticizes those “in Europe, who confounding together the different characteristics of the sexes, would make of man and woman beings not only equal but alike” (Tocqueville 1840: book 3, ch. XII). He asserts that “by thus attempting to make one sex equal to the other, both are degraded; and from so preposterous a medley of the works of nature nothing could ever result but weak men and disorderly women.” He continues, explaining that “nature has appointed such wide differences between the physical and moral [232]constitution of man and woman.” Tocqueville asserted that there were “natural authorities in families” and there was a “natural head of the conjugal association.” This is clearly a question of gendered “functions, duties, and rights” in the marital context, or “marital power” that should not be subverted. Tocqueville claims that, “In no country has such constant care been taken as in America to trace two clearly distinct lines of action for the two sexes, and to make them keep pace one with the other, but in two pathways which are always different” (ibid.). And yet, Tocqueville found that Americans held women in high esteem, noting their “confidence in the understanding of a wife,” and “profound respect for her freedom” (ibid.).

Women pose a problem for Tocqueville, which points to issues still debated to this day concerning equality, similarity, difference, and subordination.5 He concludes chapter XII saying, “Thus the Americans do not think that man and woman have either the duty or the right to perform the same offices, but they show an equal regard for both their respective parts; and though their lot is different, they consider both of them as beings of equal value.”

And yet, he asserts that “whilst they have allowed the social inferiority of woman to subsist, they have done all they could to raise her morally and intellectually to the level of man.” He finally claims that the “singular prosperity and growing strength of that people ought mainly to be attributed … to the superiority of their women” (ibid.), offering no further comments on this assertion.

On the one hand, Tocqueville highlights a tension between the existence of an egalitarian mentality and the dissimilarity of women and men, which reveals the limits of this equality. On the other, he insists on the equal worth of men and women, and, surprisingly, the superiority of women, which simultaneously contradicts equality.

According to Tocqueville, men’s and women’s roles contrast specifically in the context of marriage, their positions as husbands and wives. He views gender differences as limited to their roles given by “nature.” He insists that there is complementarity, whatever their difference (dissimilarity), but finds that this feature does not exclude a certain equality in social worth. This raises the question: Outside the context of marriage and conjugal roles, what, then, prevents the equality of men and women?

Scholars have suggested various interpretations of this passage. Furet (1981: 33) contends that Tocqueville insists on equality of conditions “as a constitutive principle of democratic social order … a norm and not a fact.” Marcel Gauchet (1980: 99) states that “equality as a principle and an ideal cannot account for the differences that manifest themselves at another level; difference, especially the real division of sexes, is not cancelled, but relegated to the background.” He asserts that this is not an ontological difference, but that this “difference is at the same time deprived of its substance and yet assumed so that its presence does not prevent any adult from finding similarity with the other he dominates” (ibid.). In contrast, Dumont, who highlights Tocqueville’s conflation of “the social and the ‘natural’ form, or universal [233]being,” comments that, “In general, … we may grasp here, in Tocqueville himself, how the ideal is made immanent and reified, in the way characteristic of the modern democratic mentality. The fusion of equality and identity has become established at the level of common sense” (Dumont 1980a: 16).

Thus far, we have examined Dumont’s discussion of the ideal of equality in comparison with Tocqueville’s notion of (gender) equality and “one of his very important ideas, which concerns the place of modern political ideology in relation to values as a whole” (ibid.: 14). To summarize, Tocqueville views equality as rooted in man’s nature and claims that men are alike and thus equal. He distinguishes, however, between men and women, whom he considers different by nature, but of equal value. Dumont is critical of this notion of equality, which he suggests merges equality and identity, reifying the democratic ideal.

Dumont on the relationship between men and women: A hierarchical opposition

In this section, I examine Dumont’s theories on male–female relationships relative to equality and hierarchy. I base my analysis on his discussion of kinship terminology and refrain from making claims regarding gendered social practices in contemporary India.

First, when mentioning status (which he also refers to as “value”), Dumont distinguishes “equistatutory” oppositions from what he calls hierarchical opposition. He developed this contrast over the course of his career, drawing on Harold Conklin’s notion of “paradigm” (Dumont 1979, 1980b: 213–14, 1983: 26–28). In equistatutory oppositions, neither status is superior to the other. Dumont contrasts this relationship to hierarchical oppositions, which he illustrates using the example of the right and left hands (Dumont 1979: 807). Right vs. left is a common binary opposition that organizes ideas and representations in a great number of societies (see Hertz 1973 and Needham 1973, for example). For Dumont, the right and left hands do not occupy the same status or position relative to the rest of the body. They exist thus in a nonequistatutory relation, a hierarchal opposition in which “egalitarianism is not in place” (Dumont 1979: 806). Although right and left do not, therefore, have the same value, this does not suggest the universal superiority of one over the other. This hierarchal relationship may be reversed, but is constitutive of both hands and their relationship to the body.

Dumont draws a parallel between the right and left opposition and the way men and women are distinguished in the Old Testament story of the creation of Eve from a part of Adam’s body (Dumont 1980a: 239–40, 1980b: 225). Adam is the prototypical Man, who represents mankind, which includes both men and women who are identical as human beings. On another level, however, Adam is opposed to Eve, as male is to female and man to woman. Unity on one level thus coexists with opposition on a second level. Dumont (1980a: 240) argues that these “two relations characterize the hierarchical relation” which exists, “between a whole (or a set) and an element of this whole (or set): the element belongs to the set and is in this sense consubstantial or identical with it; at the same time, the element is distinct from the set or stands in opposition to it.” He contends that by subordinating one [234]element to another, the hierarchical principle introduces a multiplicity of levels, which allows for the reversal of status. With regard to the hierarchical unit of male/ female couples and the possible reversal of position at different levels, he says: “The mother of the family (an Indian family, for example)—inferior though she may be made by her sex in some respects—nonetheless dominates the relationships within the family” (ibid.: 240–41). To illustrate this type of reversal, Dumont mentions that in India, while “the priest is superior to the king in matters of religion, … the priest must obey the king in matters of public order,” which he characterizes as “subordinated matters” (Dumont 1980b: 225). He follows with this surprising comment: “One might say, from an egalitarian viewpoint, that traditional societies are made bearable by these reversals” (Dumont 1980a: 241).

Dumont’s approach to male/female relations contrasts with Tocqueville’s. Rather than directly addressing the question of equality (which he introduces only later), in this context he theorizes distinctions which lead to the reversal of status positions. Focusing on the reversal of levels allows Dumont to avoid Tocqueville’s conflation of equality and identity. Tocqueville’s response to the tension between his claims of equality and empirical dissimilarities between men and women (“wide differences” in roles, functions, and destinies) was to introduce the notion of women’s moral value and women and men’s equal worth.

Dumont deals with the relationship between men and women from the point of view of the unity of the couple as a whole. In the postface of Homo hierarchicus, he states clearly: “You may well declare both sexes equal, but the more you manage to make them equal, the more you will destroy the unity between them (in the couple or family) because the principle of this unit is outside them and because, as such, it necessarily hierarchizes them with respect to one another” (ibid.: 240–41, emphasis in original).6

In other passages, Dumont refers explicitly to women’s lower status overall in the Indian context, saying that “woman is in general considered inferior to the man” (ibid.: 117). He also describes this lower-status position in his discussion of marriage, in reference to endogamy (marriage within the caste) and hypergamy (marriage of a woman of a slightly inferior status with a man of a slightly superior status from which the woman’s family draws prestige). Following Dumont, I now turn my attention to the question of endogamy in India and hierarchy, in an effort to address questions of status in relation to marriage and to the male–female relationship.

Kinship in India and hierarchy

In the previous section, I introduced questions of female subordination, as analyzed in Dumont’s work on India, and explored the ways he analyzed hierarchical distinction, both at the logical level (Adam and Eve) and regarding relationships and status within the caste system. I shift now to the question of kinship criteria, [235]including sex distinctions, within the framework of caste, in an effort to offer a new perspective on questions of hierarchy and equality.

Endogamy is one of the principles underlying the separation of castes, which, by definition, appears to exclude hierarchy within the caste itself. The nature of caste hierarchies has been long debated in scholarship on India. Dumont (1980a: 43) sees the separation of castes as “reducible to a single true principle, namely the opposition of the pure and the impure. This opposition underlies hierarchy, which is the superiority of the pure to the impure, underlies separation because the pure and the impure must be kept separated, and underlies the division of labor because pure and impure occupations must likewise be kept separate.” How, then, might we grapple with differences of status within endogamous castes or subcastes? If castes are separate owing to their differences in status, it would seem that a given status would be extended throughout a given caste. This is where Dumont begins his study of marriage.

Dumont addresses several aspects of hierarchy in his discussion of marriage in India. First, he considers the relationship between castes and kinship (including affinity), examining to what extent the ideology of hierarchy influences the field of kinship (ibid.: 112–13). A second aspect, related to the first, is whether or not kinship configurations, specifically kinship terminology—a circumscribed domain—comprise hierarchical oppositions. Dumont’s aim here is to shed light on the relationship between the ideology of caste hierarchy and domains which are not directly determined by it. His writings on kinship introduce theories that he developed more fully later on, concerning hierarchy, comparison, the need to depart from our own ways of thinking in kinship analysis, the limits of generalization, and the importance of reconsidering one’s own analyses in response to discussion and critique.

I will begin with Dumont’s famous article Hierarchy and marriage alliance in South Indian kinship (1957).7 This title makes clear Dumont’s attempt to grapple with the relationship between ideological hierarchy and forms of marriage. I draw on this text to highlight his treatment of sex distinction in the limited but influential domain of kinship terminology (as subsystems of kinship in general) (Dumont 1983: 23). To justify his approach, Dumont often insists on the specificity of kinship terminology, as “a clearly bounded, simply patterned, indigenous system of categories” (ibid., emphasis in original). He asserts that in kinship terminology one finds “the terms in which the people actually think their kinship relationships, and they are for us more important … because the terms in which the people actually think are more important than what they say … about the terms in which they think that they are thinking” (ibid.: 23–24).

In this article Dumont first seeks to establish the common character among subcastes “close to one another in the caste hierarchy,” who nonetheless diverge in terms of kinship characteristics (descent, residence, marriage rules, etc.) (ibid.: 36). Dumont’s second goal was to find, within “the internal constitution” of the caste group, “something of the principles that govern its [hierarchal] external relations” (ibid.: 37). He examined whether caste hierarchy was also present in kinship [236]terminology. He finds that while kinship institutions (i.e., internal relationships) show differences, caste hierarchy (i.e., external relationships) reveals closeness in a “common social pattern” (ibid.: 36). Dumont then asks how kinship and caste might be related. He suggests that the relationship is organized at two distinct levels, “irrelevant to each other, on which the interconnection of features could be studied, namely the caste system on the one hand, and the endogamous group on the other” (ibid.: 52).

Dumont questions the notion that status unity exists within endogamous castes or subcastes. His study focuses on the gradation of status within subcastes according to subdivisions in terms of territorial units, lineage, and exogamous units (ibid.: 48). These distinctions illustrate a tendency toward internal hierarchization, founded on and manifest in forms of marriage. He suggests that intracaste hierarchy may be found, for instance, in primary or secondary unions, irregular unions of seniors, marriage of leaders, and of men of high-status subcastes. These examples illustrate how, in South India, “the tendency to graded statuses does not stop at the boundary of the endogamous group, but enters it and finally pervades the whole sphere of kinship, so that in South India kinship cannot be severed from the caste system” (ibid.: 42). Dumont qualifies this statement: in South India the kinship system is (relatively) isolated from caste (there is an “overlap” or “two systems of conceptions” (ibid.: 42 n. 3, emphasis in original). This is not the case in North India, where there is a closer relationship between kinship and caste. Nor is this the case in kinship terminology, as we will see in the following section.

Marriage and affinity

After examining the common characteristics of these groups, Dumont analyzes kinship features in each (succession, inheritance, residence, descent, and marriage rule), which he then groups in a comparative table. Despite the differences between the subcastes, he observes common characteristics of kinship throughout. Beyond a paternal or maternal principle (patrior matrifiliation), there is a combination of paternal or maternal features, which, he suggests, do not constitute a second descent principle (as argued by many Dravidian kinship studies). Dumont therefore seeks to understand the nature of this combination or additional feature, which appears to him to be related to intermarriage and affinity.

Dumont uses the term “affinity” to designate the kinship link created by marriage (parenté par alliance in French; see Needham’s 1960 discussion on this topic). Unlike the notion of “kinship” in English, in its common usage the French term parenté (kinship) includes both affinity and consanguinity (see Dumont 1971 in Parkin 2006). For comparative purposes, Dumont defines blood relatives as all those who are not affines (in contrast with French law, which designates only paternal kin as such). In this respect, consanguinity is not biologically determined, but depends on social recognition (Dumont 1971: 15, in Parkin 2006: 4).

One then must consider marriage “not as a secondary product of other institutions such as descent … (residence, etc.), [for] there is rather an interrelation” (Dumont 1983: 71). To this end, Dumont argues that the study of South Indian kinship must simultaneously consider both the marriage rules and the features of the [237]kinship vocabulary revealed by the analysis. If one expresses a rule as “marriage with the mother’s brother’s daughter,” Dumont explains, “it appears as a rule for deriving a man’s marriage from a relationship excluding any idea of marriage of affinity, i.e. from a relationship of consanguinity” (ibid.). However, if the rule is interpreted as a feature of affinity, it takes on another meaning. It then “expresses a condition … to maintain a certain form of intermarriage”8 (ibid.: 72). He thus claims that the rule is not an instrument “allowing one to deduce a secondary category (a marriage) from a primary category (a consanguinity relationship)” (ibid.: 71). Rather, this regulation “determines one’s marriage by reference to one’s ascendants’ marriage,” which means “reproducing the marriage of one’s father (or one’s grandfather according to the form of marriage) if the rule is maintained” (ibid.: 72). The consequence of the rule in the case of South India is that affinity (kinship through marriage) is transmitted from one generation to the other. It thus acquires a diachronic dimension, which Western systems only attribute to consanguinity.9

On the contrary, in Western systems, while the relationship of a man to his brother-in-law is an affinity relation, the relation between their children is a consanguinity relation, described by the kinship term “cousins.” Dumont easily demonstrates that it is not true in South India, where both children inherit the affinity relation of their fathers. Children are linked to each other by affinity, a relation transmitted from generation to generation, such that “cousins” in Western terms are “brothers-in-law” or potential “brothers-in-law.”

Kinship vocabulary and sex distinction

Having identified the diachronic character of affinity, Dumont claims that one must take into account the interrelationship between marriage rules and kinship terminology in order to understand the role played by the sex distinction. In all the subgroups Dumont studied in South India, kinship terminologies are characterized by an absolute distinction of the sexes in the three central generations and/ or in age groups. More importantly, these terminologies do not comprise specific terms characteristic of affinity, like father-in-law and mother-in-law. There is a single term for both the maternal uncle (mother’s brother) and the father-in-law, who are one and the same person in unions which follow the rule of preferential marriage, in which a man should ideally marry the daughter of his mother’s brother.

In the same manner, there is a single term for the paternal aunt (father’s sister) and the mother-in-law (treating the father’s sister as an affine, in this way, has been [238]considered highly controversial, see, e.g., Dumont’s 1983 response to Radcliffe-Brown). Dumont’s analysis shows that in the three central generations and/or in age groups, kinship terminology distinguishes two categories of parents of each sex: on the one hand, parents connected to the referent or ego of each sex by consanguinity; on the other, parents of each sex connected to ego by affinity. Thus, there is a term for a male ego’s fathers, on the one hand, and a term for fathers’ male affines, on the other (but not maternal uncles). There is one term for brothers and another for brothers-in law (but not cousins). Similarly for women, there is a term for female ego’s mothers on one side and a term for mothers’ female affines on the other (but not paternal aunts), one term for sisters and another for sisters-in law (but not cousins).

Dumont thus proposes that in these three generations, each sex has its own set of kinship terms for consanguines and affines. He asserts, “There is likely to be an affinal content in terms which are generally considered to connote consanguinity, or ‘genealogical’ relationships (such as mother’s brother, etc.). This is obviously so when there are no special terms for affines, for otherwise we should have to admit that in such cases affinity is not expressed at all”10 (Dumont 1957: 25, 1983: 73).

In these South Indian terminologies, at the central generations, when there are two male terms that indicate affinity, the two female terms are completely distinct. “One sees that the distinction of sex and the alliance distinction go together” (Dumont 1957: 26, 1975a: 51, 1983: 74). Where there is no sex distinction, as is the case beyond the third generation, there is no affinity distinction either. Thus it is the combination of sex and affinity distinctions that orders the terminology in these central generations or age groups. Strictly speaking, affinity relationships exist exclusively between persons of the same sex. Dumont insists several times on this feature of the Dravidian terminology, arguing that sex distinction is absolute.

From this, he draws the unexpected conclusion that the “perfect symmetry between the sexes,” which he claims one might call a “superb invention,” “transcribes in terms of equality the segregation of the sexes in social life” (1983: 24). Regarding caste, he adds, “South Indian kinship presents us with a contrast of that sort: something like an island of equality in an ocean of caste” (ibid.: 167), again making reference to Conklin’s distinction between a “paradigm” (equistatutory opposition) and a hierarchy comprised of successive levels of inclusion. He specifies later, saying,

The grid of categories is determined by a specific conjunction of the consanguines/affines opposition with the absolute distinction of (equal) sexes. In other words, the alliance relationship provides the backbone of the classification of relatives. Here is the island of equality, even though the ocean of hierarchy invades it here and there, with the effect of adding status within the framework. (Dumont 1983: 168)

This leads Dumont back to the hypothesis he introduced in Hierarchy and marriage Alliance (1975a, 1983) regarding his attempt to understand whether caste hierarchy [239]penetrates all social domains (kinship, marriage, etc.).11 He asserts that, “within kinship proper, i.e. within the domain defined through the system of kinship categories, hierarchy plays its role within a framework it does not determine, and is thus conceptually subordinate to kinship” (Dumont 1983: 168).

The suggestion that equality in India manifests itself through “the absolute distinction of (equal) sexes” is surprising in light of Dumont’s acknowledgment that sex distinctions are inherently hierarchical (the hierarchical opposition between man and woman—supra Adam and Eve). For Dumont, the hierarchical distinction applies to a particular type of unit (the couple) or relationship (between husband and wife). It must thus be conceptually differentiated from the perfect equality of the sexes in South Indian terminology. The latter is an analytical and conceptual equality that exists within kinship terminology. Gender equality in kinship terminology does not entail equity of social relations between men and women. Nor is it the type of equality Dumont discussed in the context of individualism. Here he defines “equality” as “perfect symmetry between the sexes,” which he opposes to “the segregation of the sexes in social life” (ibid.: 24). I suggest, however, that this is an equality that exists solely outside of any social relation. This type of equality, like that in individualism, separates each individual from the others, given that the two sexes are separated in the South Indian affinity terminology.

As is often the case in Dumont’s theories, these matters are not black or white. Here equality is present without it “being an entailment of individualism.” Equality of the sexes, or the “perfect similitude in the treatment of sexes, in conjunction with the axial distinction” ibid.: 28), is emphasized in that affinity “has equal status with consanguinity, or a value equal to it” (ibid.: vii, emphasis in original). At this point, the kinship terminology does not hierarchize the sexes.

Dumont insists on the perfect equivalence and the equal status of consanguinity and affinity in this system (ibid.: vii and 170). Values are, by definition, considered durable. While Western affinity is ephemeral (see above Dumont 1971: 15, in Parkin 2006: 4), South Indian affinity is permanent, possessing a value equal to that of consanguinity (Dumont 1983: vii). In India, although hierarchy penetrates subcaste endogamy (seen in the different statuses within the subcaste itself), in kinship terminologies, equality of the sexes and of affinity and consanguinity prevails. Returning to Conklin’s paradigm once more and to the equistatutory opposition, Dumont (1983: 26) states that it is “a configuration in which no hierarchical distinction of levels is found.”12

Dumont concludes, “Let me state again that this analysis claims to deal only with the basic level of distinction in the vocabulary—equal status of consanguinity and affinity—and to thus deliver the basic cognitive grid through which the people think their kinship relationships. It is incomplete in the sense that secondary levels of distinction are not explored” (ibid.: 35).

[240]The last section of this article discusses the importance of affinity relationships in other contexts, such as the roles and functions of affines in matters of inheritance that illustrate the central role played by marriage, including gifts, wedding and mourning prestations, the ceremonial role of the maternal uncle, and so on.

In summary, in South India, and within the hierarchical caste system, kinship terminology is organized by a simple distinctive opposition (which Dumont describes as a “paradigm”) which establishes the equal value of consanguinity and affinity, in conjunction with the marriage rule, as well as the perfect symmetry between and equality of the sexes.

These conclusions lead me to reflect upon sex distinctions and the relationship between consanguinity and affinity when it is not a structural opposition, in other kinship systems and terminologies.

Affinity and the sex distinction in another configuration

I will briefly present the example of Kei society, in the Southeast Moluccas of Indonesia, where I conducted fieldwork (Barraud 1979, 1990). Unlike the Dravidian terminology, the kinship terminology here does not show a structural opposition between consanguinity and affinity. Affinity vocabulary is lacking; most of its terms are made from compound consanguinity terms (like the English term “father-in-law,” constructed from the term “father”). In Kei, like South India, there is preferential marriage, but the term “father-in-law” is different from that of “mother’s brother.” Affinity is not a significant opposition in ordering the kinship terminology (each sex treated separately, as in the case of South India, two terms opposing consanguine and affinal relatives in three generations or age classes). In Kei, affinity, unlike consanguinity, is only loosely present in kinship terminology. This does not mean that marriage is unimportant; on the contrary. Rather, it is conditioned by another type of transmitted relationship: that between a brother and a sister. Marriage ties two social units (houses), which are then called “ancestors of the brother,” on the one hand, and “children of the sister” on the other, a relation that cannot be reversed. For generations, the houses continue giving and receiving different types of prestations in all kinds of ceremonies (birth, marriages, funerals, building of boats and houses, etc.; see Barraud 1979, 1990), according to their position relative to the husband or wife in previous marriages. Thus, what we usually call “affine relations,” in Kei, is expressed in terms of the relation between the descendants of the wife’s brother house and the sister’s husband house, that is, through the brother–sister relationship.

The distinction that orders the terminology is the sex distinction, in its “relative” modality (contrary to Dravidian terminology, where sex distinctions are absolute, each sex being called by a distinct term). It manifests itself in the terms used respectively between siblings and husbands and wives (the same term is used by a woman or a man to designate a person of the other sex). These terms do not oppose male and female, whereas in Dravidian terminology each sex is absolutely distinguished in the three central generations.

However, a disjunction appears analytically between this absence of opposition in the use of kinship terms and a hierarchy within the sex distinction itself. [241]This hierarchy is contained in the distinction that encompasses both sexes (there is a single relative sex term for brother–sister and a single relative sex term for husband–wife), thus attributing them a different status with regard to the distinction (Dumont’s “principle of the unit”: 1980a: 240–41).13

What is striking is that this analytical disjunction emerges between actual marriage partners and in the subsequent relationships between their houses for at least two generations after a marriage. The hierarchy inherent in the sex distinction takes the form of different statuses given to the wife’s and the husband’s houses: the wife’s brother’s ancestors being attributed a higher status than the sister’s descendants, who are considered helpers of the former.14 The kinship terminology presents a sort of formal synchronic symmetry of the sexes in the central generation. However, any marriage creates a diachronic dissymmetry of the sexes for the descendants of this marriage (who have different statuses). This is one of the main features of the societies where the brother–sister relationship is central in the analysis of kin terminologies.

Kei society thus presents a contrasting configuration, in which affinity nearly disappears. It is a configuration where a marriage induces a change of status, establishing higheror lower-ranked relationships between intermarried houses. In other words, although there is no distinction of sex in the use of kinship terms, in the social relations between brothers and sisters, and husbands and wives, a sex distinction does exist. Within this distinction, there is a hierarchy, which exists in tandem with differential statuses of marriage relationships.

For further comparison with South India, we can briefly underline a point concerning Kei ideology (Barraud 1990; Alès and Barraud 2001). In South India, while trying to understand the relationship between ideology (the pure and the impure, caste and hierarchy) and kinship features, Dumont finds in kinship terminology what he calls equality of the sexes and equal status between consanguinity and affinity. At the same time, he discovers that differences of status (hierarchy) enter the caste through marriage. In Kei, by contrast, values are defined in reference to two sets of ideas, which, for simplicity, may be described as the internal organization of the society (subordinate level) and the relationship with the world beyond society (superior level). Marriage must be considered at both levels. Maternal affines (the mother’s brother and their ancestors) have a higher status at the subordinate level of ideology. Sisters destined to marry outside the house are called “women strangers” from birth. These out-married sisters, their husbands and descendants, are associated with the superior level of the world beyond society. In this case, marriage and sex distinction have a dual status.

In Dumont’s terms, “equality” would be an inappropriate description of this configuration. South India displays in its kinship terminology an equistatutory opposition between affines and consanguines for each sex. South Indian kinship terminology includes distinct words for each sex, which receive equal treatment. This is not the case in the Keiese terminology, where a single term designates both sexes. The sex distinction here is manifested through the relation.[242]

North Indian vocabulary

This point may be illustrated through a comparison Dumont (1962) highlighted between North and South India. Unlike Southern Indian kinship terminology, a strict central opposition does not characterize kinship vocabulary in Northern India. There is here no positive marriage rule, but only marriage prohibitions. Upon analysis, only one feature appears systematically: the relationship between siblings of the opposite sex. Dumont further develops these points in an article comparing North India and South India (Dumont 1966). He notes the empirical importance of relationships between affines, manifest in the relatively large number of affinal terms, the difference in status between marriage partners, and (like in the South) the important ceremonial functions of some affines (ibid.: 114). His central argument is that the affinity arrangements depend largely on status and are not directly linked to kinship terminology, which, he claims, has a “more modest place in the whole” (ibid.:114). Dumont argues that “the ordering of affinal relatives is … at any rate, largely a function of status, i.e. of caste, instead of a function of kinship” (ibid.). He goes so far as to say that “there is thus no ‘kinship system’ in a strict sense, as an extraneous principle pervades the kinship field and deprives it of its internal consistency in favour of a more complex arrangement” (ibid.). Dumont contends that this also “reduce[s] the function of kinship in social life” (ibid.: 110). To his regret, he acknowledges: “We must own up to yet another unpleasant recognition: the ordering of affinal relatives is, in [Uttar Pradesh] at any rate, largely a function of status, i.e. of caste, instead of a function of kinship” (ibid.: 114).

Dumont is, as ever, cautious, limiting the comparison to the logical arrangement of the kinship terms in each context.

Hence one hypothesis: the center of gravity of a kinship system lies in the terminology only when it is simple and quite systematic, the outright expression of a largely non-conscious construction of the mind. There are other cases where the link between terminology and behavior is much looser, signaled by the relatively unsystematic character of the terminology, and where the center of gravity of the kinship system is to be sought elsewhere. (Dumont 1966: 103)

Kinship terminologies in French and English

Dumont’s comparison once again goes further. Indeed, in order to clarify his analysis of this very complex and confusing North Indian kinship terminology, he relies on a brief comparison with French kinship vocabulary, drawn from common language. He searches for its axis, or a principle that orders the system, with respect to the arrangement of the three elementary kinship relationships (descent, sibling relationships, and marriage/affinity) in comparison with the South Indian vocabulary (Dumont 1962: 32–36, particularly n. 16).15

[243]In the French system, he says, there are, strictly speaking, only two affinity terms, gendre (son-in-law) and bru (daughter-in-law). Other affinal terms are compound words, composed of consanguine terms like père (father) or frère (brother) and a prefix marking affinity. In French, beau-père (father-in-law), beau-frère (brotherin-law), and so on, do not constitute distinct categories with respect to “father” and “brother.” Dumont thus asserts that among “modern Westerners, affinity is subordinated to consanguinity, for my brother-in-law—an affine[—]becomes an uncle, a consanguine relative, for my children. In other words, affinity is ephemeral, it merges into consanguinity in the next generation” (Dumont 1983: vii). There are no positive marriage rules, there is no repetition of the marriage relations in successive generations. And sexes are equal in the ideology, both ideally and in terminology.

Note that in English, the vocabulary is even more limited: son-in-law is composed of “son,” daughter-in-law of “daughter,” while French has gendre and bru. “Whereas French still has special terms in certain cases, English only has secondary determinants. Comparatively … the dichotomy in English usage corresponds to a sort of apotheosis of consanguinity and a concomitant devaluation of affinity” (Dumont 1971: 14, in Parker 2006: 4).

However, Dumont sees in this system a sort of central distinction, between direct line and collateral line within descent: the terminology distinguishes father and uncle, brother and cousin, son and nephew, and so on … and also the sexes (Dumont 1962: 35). This does not organize in any way the status of affinity, which is residual. Thus kinship consists of consanguinity only and introduces, says Dumont, “an opposition within a substance, descent” (ibid.) between direct and indirect line.


In summary, in South India, in the context of caste hierarchy, kinship terminology exhibits an equal treatment of the sexes in conjunction with the axial distinction, equivalence in value of consanguinity and affinity, with respect to the marriage rule.

In Kei society, the axial distinction is the sex distinction; there is a sort of symmetry between the positions of both sexes at ego’s generation (brother–sister) and a dissymmetry of the brother–sister relation, which is revealed at the next generation after a marriage occurs. There is dissymmetry and hierarchy in the relation between two generations in conjunction with marriage; this is a hierarchy, which is reversed at the superior level of ideology.

In North India, the hierarchy of the caste enters the field of kinship, particularly in affinity relations, such that the two sexes are not equal or equivalent in this regard. There is no strict opposition in the terminology and no axis, but the sibling relationship serves as a guiding principle, which organizes kinship.

In Western systems, where egalitarian ideology dominates, there is no strict central opposition in the kinship terminology. At this level, sex distinction is clear in the non equivalence of consanguinity and reduced affinity. Taking into account the fact that kinship in Western societies nowadays does not order the whole of social life, one can say, however, that with regard to value and permanence, only consanguinity is transmitted and affinity is not valued.16

[244]And finally Dumont remarks:

The irony of the situation is that, in this matter, we Westerners, egalitarians as we are, practice subordination—the relation between consanguinity and affinity is exactly what I have defined elsewhere as a hierarchical relation—while South Indian people, who live in a hierarchical society, look at the two entities with equanimity, and make a simple, straightforward, symmetrical distinction between them, where we maintain a hierarchical distinction. (Dumont 1983: vii, emphasis in original)

Concluding remarks

In Western societies, where equality is an ideal, if not a reality, hierarchy is largely unsuspected and underestimated. I am not an expert on French kinship or marriage, but it seems to me that their relation to ideology engenders three conclusive remarks.

First, the devaluation of affinity in kinship terminology must be examined with regard to the contemporary desire to give both sexes equal treatment. Empirically, there are today a growing and infinitely expandable number of affinity terms used in contexts where affinity is no longer present. I am thinking of terms which designate the spouse of ego’s mother or father, in French beau-père (stepfather) and belle-mère (stepmother), or the children of ego’s spouse: beau-fils (stepson), belle-fille (stepdaughter), who are neither affines nor consanguine relatives for the children but who take the place of kin.17 Given the valorization of consanguinity relative to affinity, could the use of affinity terms (beau-père, belle-mère) by children (parent–child relationship, a filiation relationship) lead to the confusion of relations and a further devaluation of affinity? Furthermore, does consanguinity itself not lose its value, becoming a kind of second-rate consanguinity? To follow Dumont’s questioning, one can ask what causes this ordering of affinal relatives and the devaluation of affinity in Western societies (Dumont 1966: 110).

My second point is a comment on filiation (Dumont 1962: n. 13) in France. Does speaking of filiation (in French) produce confusion between its two meanings? On the one hand, the parent–child bilinear relationship (“filiation” in English), and, on the other, filiation as “descent,” that is, understood as “transmission of membership in an exogamous group,” unilineal and even patrilineal in many Western societies, connected to a genealogical arrangement tracing a relation back to ascendants. This confusion is also reported by Agnès Fine regarding the parent–child [245]relationship (specifically “parenthood”) and the English “filiation.” When the term is used in the first sense (parent–child relation) and in the contexts described in the first point (beaux-parents, terms which in French connote affinity), there is a collusion of meanings between affinity and consanguinity. The term parents in French, meaning “relatives” in English, becomes a catch-all term in which no distinctions are made. “The second notion tends to cover the first, which means that the parent–child link tends to occupy all that contains filiation” (Fine 2001: 46). It also contributes simultaneously to a confusion with affinity, owing to the use of the same terms in second marriages. Affinity is recognized only through a preference for consanguinity (parent–child relationship) at the same time as consanguinity is reduced to the parent–child relationship (parenthood).18

Finally, my third point is that in all examples of kinship terminologies presented here, except the French one, the sex distinction operates in conjunction with another principle to organize the kinship vocabulary relative to marriage and affinity. On the other hand, in Western kinship terminologies, sex distinction, although present, is not able to organize a consanguinity–affinity relation, since affinity is an attribute of consanguinity and does not significantly oppose it. The sexes are ideologically equal (if not in reality), but sex distinction has no effect in the construction of elementary kinship oppositions. One could say that marriage and affinity in Western systems do not have much to do with la parenté (“kinship” in English, in its broadest definition), on the restricted level of the construction of elementary kinship oppositions. Their status on the ideological level is also unclear (an issue which preoccupies contemporary debates regarding marriage, if not affinity strictly speaking—see footnote 17—which, at this point, is beyond the scope of my own research).

And again I give the floor to Dumont, on comparison.

While with us, only consanguinity is inherited, in South India affinity is transmitted also from one generation to another, making it an exact counterpart of consanguinity. In saying this, we conceptually move from our own conceptual system to the one that is studied, and our comparison is complete. We subsumed under a formula our conception and their conception … we have not achieved the universal, but only a step toward it, because there is in the world a wide variety of kinship systems based on other principles. (Dumont 1992: 244, emphasis in original)

[246]After rejecting dichotomies, and “natural” differences, the editors of Gender and kinship propose a premise “that social systems are systems of social inequalities.” This follows the question of how attributes and characteristics of people are “culturally recognized and differentially evaluated” and its corollary: “what are a society’s cultural values?” (Yanagisako and Collier 1987: 39–40). The key phrase here, however, is “values entail evaluation,” meaning that “all things and actions are not equal” in a system of social relationships, and values create inequalities (ibid.).

This is where the advantage of Dumont’s theory of hierarchy becomes clear. As we have seen, hierarchy does not create inequalities but distinguishes levels in which the values attributed to elements can be reversed. The logic of equality excludes anything that is not of equal value, whereas hierarchy includes multiple levels of elements of different values, without exclusion. The separation of sexes in kinship terminology does not attribute a priori a value to one sex or the other. Similarly, the affinity–consanguinity distinction, as criteria of analysis, does not attribute a value to one or the other. Only the analysis of the entire terminological system allows for the attribution of status. The combination of a limited number of terms in a terminology does not account for “natural” differences and even seems to erase them, as we have seen with the term “relative sex” in the Keiese terminology. Equality is not in place, according to Dumont’s use of the term (in his discussion of the equality of sexes in South Indian terminology), which points exclusively and significantly to an equal treatment of the sexes as a formal kinship category by contrast with the caste hierarchy.

Once more, Dumont’s comparative method is at work here (see also Dumont 1975b: 156–57). The notion of “equality” has been presented first in comparison with Tocqueville’s idea that equality abolishes differences and implies similarity except in the case of men and women: dissimilarity of women but equal value with men, in the egalitarian mentality. Dumont argued that Tocqueville confused equality and identity, which the latter locates in human nature. In contrast, Dumont’s approach to hierarchy distinguishes equality, identity, and difference, distinctions that remain at the heart of contemporary discussions of gender as a “social construct.” In this respect, a closer reading of Dumont’s work can shed light on questions of equality and gender relations. This was my aim in this essay through the examination of kinship terminology. This analysis limits the scope of the question, illustrating the place of notions of equality within the hierarchical context of the caste system. The challenge we now face is to determine to what extent and in what contexts a reversal is feasible. The present analysis has revealed the presence of equality within hierarchy, leading to the question: How might hierarchy be conceivable within egalitarian contexts?

In 1988, in her book Feminism and anthropology, the British anthropologist Henrietta L. Moore wrote, “The different theoretical positions within feminist anthropology are best demonstrated through a consideration of the debate which dominates the subject: is sexual asymmetry universal or not? In other words, are women always subordinate to men?” (Moore 1988: 13).

In this article, I have proposed an answer to these questions. Yes, there is always sexual asymmetry. No, women are not always subordinate to men. And the value subordination (not to be confounded with the either/or implied in formulations of equality vs. domination) that takes place in the social manifestation of the sex [247]distinction is not uniform, but rather unfolds in diverse ways, which include reversals of hierarchal relations.


This article has benefited from long discussions with my colleagues Ismaël Moya and André Iteanu. The English version was greatly improved with the assistance of Chelsie Yount-André, to whom I am deeply grateful.


Alès, Catherine, and Cécile Barraud, eds. 2001. Sexe relatif ou sexe absolu? De la distinction de sexe dans les sociétés. Paris: Éditions de la Maison des Sciences de l’Homme.

Atkinson, Jane Monnig and Shelly Errington, eds. 1990. Power and difference: Gender in Island Southeast Asia. Stanford: Stanford University Press.

Barraud, Cécile. 1979. Tanebar-Evav: Une société de maisons tournée vers le large. Paris: Éditions de la Maison des Sciences de l’Homme.

———. 1990. “Wife-givers as ancestors and ultimate values in the Kei Islands.” Bijdragen tot de Taal-, Landen Volkenkunde 146: 2–3.

———. 2001. “La distinction de sexe dans les sociétés: Un point de vue relationnel.” Esprit 273 (mars–avril): 105–29.

Busby, Cecilia. 1997. “Of marriage and marriageability: Gender and Dravidian kinship.” Journal of the Royal Anthropological Institute (N.S.) 3 (1): 21–42.

Collier, Jane Fishburne, and Sylvia Junko Yanagisako, eds. 1987a. Gender and kinship: Essays toward a unified analysis. Stanford: Stanford University Press.

———. 1987b. “Introduction.” In Gender and kinship: Essays toward a unified analysis, edited by Jane Fisburne Collier and Sylvia Junko Yanagisako, 1–13. Stanford: Stanford University Press.

del Valle, Teresa, ed. 1993. Gendered anthropology. London: Routledge.

Descombes, Vincent. 1999. “Louis Dumont et les outils de la tolérance.” Esprit 253 (juin): 65–82.

Dumont, Louis. 1957. Hierarchy and marriage alliance in South Indian kinship. Occasional Papers 12. London: Royal Anthropological Institute of Great Britain and Ireland.

———. 1962. “Le vocabulaire de parenté en Inde du Nord.” L’Homme 2 (2): 5–48.

———. 1966. “Marriage in India, the present state of the question: III. North India in relation to South India.” Contributions to Indian Sociology (O.S.) 9: 90–114.

———. 1971. Introduction à deux théories d’anthropologie sociale. Paris: Mouton EHESS.

———. 1975a. Dravidien et Kariera: L’alliance de mariage dans l’inde du sud et en Australie. Paris: Mouton.

[248]———. 1975b. “On the comparative understanding of non-modern civilizations.” Daedalus 104 (2): 153–72.

———. 1977. From Mandeville to Marx: The genesis and triumph of economic ideology. Chicago: University of Chicago Press.

———. 1979. “The anthropological community and ideology.” Social Science Information 18 (6): 785–817.

———. 1980a. Homo hierarchicus: The caste system and its implications. Translated by Mark Sainsbury, Louis Dumont, and Basia Gulati. Chicago: University of Chicago Press.

———. 1980b. “On value: The Radcliffe–Brown Lecture in Social Anthropology.” Proceedings of the British Academy 66: 207–41.

———. 1983. Affinity as a value: Marriage alliance in South India, with comparative essays on Australia. Chicago: University of Chicago Press.

———. 1986. Essays on individualism. Chicago: University of Chicago Press.

———. 1992. “Rétrospection. L’enseignement de l’inde: le tout et les parties.” Religio, Bulzone Editore (remise du prix européen Amalfi): 239–51.

———. 2006. Introduction to two theories of social anthropology: Descent groups and marriage alliance. Translated and edited by Robert Parkin. New York: Berghahn.

Fine, Agnès. 2001. “Vers une reconnaissance de la pluriparentalité.” Esprit (mars–avril): 40–52.

Fortes, Meyer. 1949. The web of kinship among the Tallensi. London: Oxford University Press.

Furet, François. 1981. “Préface.” In De la Démocratie en Amérique, Alexis de Tocqueville, 7–71. Paris: Garnier-Flammarion.

Gauchet, Marcel. 1980. “Tocqueville, l’Amérique et nous.” Libre 7: 43–120.

Gregory, Christopher A. 2010. “Siblingship as a value in Middle India.” In The anthropology of values: Essays in honour of Georg Pfeffer, edited by Peter Berger, Roland Hardenberg, Ellen Kattner, and Michael Prager, 3–18. Delhi: Dorling Kindersley.

———. 2013. “The value question in India: Ethnographic reflections on an ongoing debate.” Hau: Journal of Ethnographic Theory 3 (1): 116–39.

Hertz, Robert. 1973. “The Pre-eminence of the right hand: A study in religious polarity.” In Right and left: Essays on dual symbolic classification, edited by Rodney Needham, 3–31. Chicago: University of Chicago Press.

MacCormack, Carol, and Marilyn Strathern, eds. 1980. Nature, culture and gender. Cambridge: Cambridge University Press.

Madan, T. N. 1986. “Reviewed work: Affinity as a value: Marriage alliance in South India, with comparative essays on Australia, by Louis Dumont.” American Journal of Sociology 91 (4): 1016–18.

Moore, Henrietta L. 1988. Feminism and anthropology. Cambridge: Polity.

[249]———. 1999. “Whatever happened to women and men? Gender and other crises in anthropology”. In Anthropological theory today, edited by Henrietta L. Moore, 151–71. Cambridge: Polity.

Needham, Rodney. 1960. “Descent systems and ideal language.” Philosophy of Science 27 (1): 96–101.

———, ed. 1973. Right and left: Essays on dual symbolic classification. Chicago: University of Chicago Press.

Overing, Joanna. 1973. “Endogamy and the marriage alliance: A note on continuity in kindred-based groups.” Man (N.S.) 8 (4) : 555–70.

———. 1975. The Piaroa, a people of the Orinoco Basin: A study of kinship and marriage. Oxford: Clarendon Press.

Ortner, Sherry B., and Harriet Whitehead, eds. 1981. Sexual meanings: The cultural construction of gender and sexuality. Cambridge: Cambridge University Press.

Parkin, Robert. 1992. “Dispersed alliance and terminological change in South Asia.” Journal of the Anthropological Society of Oxford 23 (3): 253–62.

———. 2006. “Introduction.” In Two theories of social anthropology: Descent groups and marriage alliance, Louis Dumont, vii–xxvii. Translated and edited by Robert Parkin. New York: Berghahn.

Pfeffer, Georg. 1993. “Comments: Dispersed alliance, terminological change and evidence.” Journal of the Anthropological Society of Oxford 24 (1): 49–63.

Robbins, Joel. 1994. “Equality as a value: Ideology in Dumont, Melanesia and the West.” Social Analysis 36: 21–70.

Rudner, David. 1990. “Inquest on Dravidian kinship: Louis Dumont and the essence of marriage alliance.” Contributions to Indian Sociology (N.S.) 24 (2): 153–74.

Sahlins, Marshall. 2013. What kinship is—and is not. Chicago: University of Chicago Press.

Strathern, Marilyn, ed. 1987. Dealing with inequality: Analysing gender relations in Melanesia and beyond. Cambridge: Cambridge University Press.

———. 1988. The gender of the gift: Problems with women and problems with society in Melanesia. Berkeley: University of California Press.

Théry, Irène, and Anne–Marie Leroyer. 2014. Filiation, origines, parentalité: Le droit face aux nouvelles valeurs de responsabilité générationnelle. Paris: Ministère de la Famille.

Tocqueville, Alexis de. 1840. Democracy in America. Vol. 2. Translated by Henry Reeve. http://www.gutenberg.org/files/816/816-h/816-h.htm.

Trautmann, Thomas R. 1981. Dravidian kinship. Cambridge: Cambridge University Press.

Viveiros de Castro, Eduardo. 1998. “Dravidian and related kinship systems.” In Transformations of kinship, edited by Maurice Godelier, Thomas R. Trautmann, and Franklin Tjon Sie Fat, 332–85. Washington, DC: Smithsonian Institution Press.

———. 2001. “GUT feelings about Amazonia: Potential affinity and the construction of sociality.” In Beyond the visible and the material: The Amerindianization of society in the work of Peter Rivière, edited by Laura M. Rival and Neil L. Whitehead, 19–43. Oxford: Oxford University Press.

[250]Yanagisako, Sylvia Junko, and Jane Fishburne Collier. 1987. “Toward a unified analysis of gender and kinship”. In Gender and kinship: Essays toward a unified analysis, edited by Jane Fishburne Collier and Sylvia Junko Yanagisako, 14–50. Stanford: Stanford University Press.

Parenté, égalité, hiérarchie: distinction de sexe et valeurs dans une perspective comparative

Résumé : Cet article traite de la notion d’égalité dans l’oeuvre de Louis Dumont, plus particulièrement en relation avec la distinction de sexe et l’idéologie. Cet exercice de lecture des textes de Dumont sur la parenté revisite son analyse de la terminologie et du mariage en Inde du Sud, qui le conduit à discuter la valeur égale de l’affinité et de la consanguinité dans la société hiérarchique des castes, et l’égalité des sexes. Ses travaux sont ici utilisés pour ouvrir une perspective comparative sur la distinction de sexe. L’article propose d’étendre ce type d’analyse à d’autres sociétés et d’autres idéologies, pour tenter de mieux comprendre les changements contemporains dans les relations de parenté.

Cécile BARRAUD is a Senior Researcher Emeritus at the National Center for Scientific Research (CNRS, France). She has conducted extensive fieldwork in Southeast Moluccas, Eastern Indonesia, since 1971. Her research focused initially on understanding “houses” as social groups. She is the author of Tanebar-Evav: Une société de maisons tournée vers le large (Éditions de la Maison des Sciences de l’Homme, 1979), Sexe relatif ou sexe absolu (with Catherine Alès, Éditions de la Maison des Sciences de l’Homme, 2001), and of an essay on the notion of property in Austronesian languages and in the Indo-Pacific area (in André Iteanu, ed., La cohérence des sociétés: Mélanges en hommage à Daniel de Coppet, Maison des Sciences de l’Homme, 2001). She is currently engaged in a cooperative German–French program on the theme “Encountering religions in Southeast Asia and beyond.”

Cécile Barraud
Centre Asie du Sud-Est
Centre National de la Recherche Scientifique
École des Hautes Études en Sciences Sociales
Paris France


1. Dumont’s papers on kinship were often first published in English. Most of them have been translated, including the most contentious article (Dumont 2006). This text provided a comparison between English and French kinship studies. Translation had to face the different meanings attributed to key concepts in English and in French: the meaning of “kinship” itself, and others such as “descent,” “filiation,” and so on. Parkin’s “Introduction” (Parkin 2006) is very useful in this respect. See also my discussion at the end of this article. Dumont himself was long opposed to a translation because of these terminological issues.

2. I use Dumont’s definition of ideology as a configuration of ideas and values or a set of ideas and values common to a society.

3. This question is still at the heart of kinship studies (Sahlins 2013).

4. Dumont further develops this idea in his essay “On value” (Dumont 1980b: 238). A reprint of this article was published in HAU: Journal of Ethnographic Theory, Volume 3, Issue 1, pp. 287–315. Available here: http://www.haujournal.org/index.php/hau/article/view/hau3.1.028

5. Debates since the 1980s (see, e.g., Ortner and Whitehead 1981; Collier and Yanagisako 1987a; Moore 1988) have addressed social systems of inequality, treating “difference” as a social construction defined by each society. I have discussed these questions at length in Barraud (2001) and in Alès and Barraud (2001).

6. A French philosopher, Vincent Descombes (1999), analyzed and developed these aspects of equality in a comment on Dumont’s discourse pronounced when he was awarded the Tocqueville Prize in 1987.

7. Hierarchy and marriage alliance in South Indian kinship was published in French in 1975 (Dumont 1975a) and reprinted in Affinity as a value (Dumont 1983).

8. The French version adds: “of a certain type of affinity” (Dumont 1975a: 48).

9. Whenever I use the expressions “with us,” “our systems,” “we Westerners,” “our”/“their,”’ and so on, or whenever I quote Dumont using these expressions, it must be recalled that Dumont, since his earliest analyses, always advocated the analysis of the subject studied (caste, kinship, hierarchy, ideology, etc.) with the anthropologist’s own society, culture, ideology, and so on (see, e.g., his introduction to Homo hierarchicus). He asserts that anthropology has the aim of “putting modern society in perspective in relation to the societies which have preceded it or which co-exist with it, and of making in this way a direct and central contribution to our general education” (Dumont 1980a: 2).

10. The French version (Dumont 1975a: 50) adds, “in the vocabulary while marriage rules specifically emphasize its importance.”

11. In another passage, Dumont (1983: 169) expresses this idea in simple terms: “with regard to our problem of the relation between caste and kinship, or hierarchy and equality, in South India.”

12. I diverge from Dumont on this point, prefering the formulation “equality of the sexes” to “equality between the sexes.” The latter could be interpreted as equality between men and women, whereas the analysis at hand focuses solely on the terminological level.

13. The sex distinction, and particularly its relative sex modality in different societies, is central to Alès and Barraud’s Sexe relatif ou sexe absolu? (2001).

14. This feature is well known as being a characteristic of Eastern Indonesia societies.

15. “To understand the situation, we may perhaps recall the more extreme case of our own, Western, societies in which kinship does not, as in simpler societies, order directly or indirectly the whole of social life” (Dumont 1966: 110, emphasis in original).

16. An anonymous reviewer of this article rightly underlined the great influence of the church in defining Western kinship and affinity and property transmission. There is indeed a historical context to take into account to explain the contemporary devaluation of affinity, which I did not include in the comparison of English and French terminologies. I am grateful to this reviewer for this important observation.

17. See the recent French Report (Théry and Leroyer 2014) on the question of “access to origins, and the familial place of beaux-parents (stepparents); and a reflexion on filiation, to account for the evolutions of contemporary family.” If the discussion of marriage is central in this interesting report, nowhere are the statuses of affinity or the question of the affinal terms used in the context of filiation examined.

18. See Parkin’s (translation of Dumont’s Introduction to two theories, for his discussion of the translation of French filiation: “Dumont also addresses the double meaning of French filiation (meaning both descent and the parent–child links which descent consists of: though it is relatively rare in English, the same word is used there in the latter sense only” (Parkin 2006: xii). And again, in his discussion of Radcliffe-Brown’s definition of kin and of siblingship, I quote Parkin’s translation of Dumont: “Siblingship is left aside, consanguinity thus being linked only to the relation of filiation (elementary) or descent (multiple), a nuance to be noted in the way the French term filiation is used.” And in the translator’s note 1: “In other words, French filiation is used to translate both ‘filiation’ (ties between parents and children) and ‘descent’ (the extension of such ties backwards in time from the perspective of the present) from English” (Parkin in Dumont 2006: 6).