The value question in India

This work is licensed under the Creative Commons | © Chris Gregory. Attribution-NonCommercial-NoDerivs 3.0 Unported. ISSN 2049-1115 DOI: http://dx.doi.org/10.14318/hau3.1.007

The value question in India

Ethnographic reflections on an ongoing debate

Chris GREGORY, Australian National University and University of Manchester

The terms of the debate about anthropological approaches to the value question in India have been set by Dumont, whose theories were based on his ethnographic studies in North and South India, his knowledge of the Sanskrit literature, his synthesis of the comparative ethnography of India, and his studies on the history of European economic thought. His theory of affinity as a value, one element of this general theory, was based on a critique of Lévi-Strauss and was, in turn, critiqued by Trautmann, among others. On the basis of fieldwork done in Central India, I draw attention to an unexamined assumption that all three theorists share, and I also consider its consequences.

Keywords: value, Dumont, India, kinship, fraternity

The value question has been a preoccupation of Brahman scholars for millennia, and their theories of value as total social phenomena bring together family values (e.g., kanya daan, the virgin-bride gift), religious values (e.g., Brahmans as pure), and economic values (e.g., profit taking as a virtue for the merchant classes) as part of a structured hierarchical whole. Of course, the Brahmans did not speak with one voice, nor did all non-Brahmans accept the “impure” status foisted upon them, a fact that became apparent when the subaltern found her voice. Regional variation, too, has complicated the debate because cross-cousin marriage is regarded as a virtue in the south and a vice in the north.

Anthropologists have contributed to this debate with ethnographically informed critiques from different regions. Such was the approach adopted by Dumont, whose ethnographic fieldwork in villages in both North and South India has informed the development of his controversial theories which, for better or worse, have defined the terms of the debate today. The breadth and depth of Dumont’s scholarly work place him in a league of his own as a value theorist. His ethnographic studies in North and South India and his comparative anthropological work on the sociology, economy, and religion of India and Europe—combined with his study of the history of ideas about value in the Sanskrit, English, French, and German literature—cover a terrain that few of his many critics have even attempted to tread. On the other hand, the central theme that permeates all his work—hierarchy is to India as equality is to Europe—is so simple that he provides his critics with an easy target to shoot down.

Dismissing Dumont’s work, then, is easy, but critiquing it in the old-fashioned sense of criticism, modification, and transcendence is an altogether different matter. This is because fieldwork in the classic British empiricist tradition is the basis for all his abstract generalizations. He uses this method, along with the comparative ethnographic method, to construct arguments of ever-increasing generality and ever-widening geographical spread. Dumont greatly admired British ethnography (Galey 2000: 325), and he produced fine-grained studies that can be compared to the very best of that tradition. For example, his A South Indian subcaste (Dumont 1986b), first published in French in 1957, “remains,” as his editor and translator notes, “the most extensive and comprehensive description of an Indian caste in the ethnographic literature” (Moffatt 1986: iii). The fieldwork for this book was conducted in Tamil, a Dravidian language. His two years in Tamil Nadu starting in late 1948 was followed by fifteen months of fieldwork in North India in two stints over the period from 1954 to 1958. This latter work was carried out in Hindi, an Indo-Aryan language. It is the original data from this fieldwork that Dumont has mined for his theoretical insights; he is not a comparativist in the armchair tradition of Indian kinship theorists such Lévi-Strauss (1949) and Trautmann (1981). Both Dumont’s fieldtrips were preceded by studies of ancient Sanskrit texts, a language he picked up while a prisoner of war. His classic synthetic work on caste, Homo hierarchicus: The caste system and its implications (Dumont 1980), first published in French in 1966, followed. His ethnographic work on kinship led to comparative work on Australia in his Affinity as a value (Dumont 1983), while his comparative work on economic value and ideology led him to Europe in his work on Homo aequalis (Dumont 1977, 1986a, 1991), a study that involved reading the original French, English, and German texts.

It is much easier, then, to dismiss Dumont than it is to critique him. To dismiss is to ridicule with arrogance, but to critique is to dignify from a position of humility. The latter is the position I try to adopt here by focusing on one narrow aspect of his general theory of value: affinity. His argument about affinity as a value in South India is but part and parcel of his general comparative argument about equality and hierarchy in the world at large, but I restrict myself here to some ethnographically informed arguments about the nature of alliance in India.

My own fieldwork was done in Central India (hereafter CI) on the frontier zone between the Indo-Aryan north and the Dravidian south. I use this perspective, triangulated with some of Dumont’s ethnography, to draw attention to an unexamined assumption that informs the theories not only of Dumont (1966) but also of Lévi-Strauss (1949), whose theories he critiqued, and Trautmann (1981), who in turned critiqued Dumont. Differences aside, all three theorists equate alliance with marital alliance, thereby overlooking the significance of fraternal alliance. Lévi-Strauss’s great achievement was to raise the question of fraternal alliance, but that was only to dismiss its significance relative to marital alliance. Trautmann’s theory of Dravidian kinship in India is premised on the assumption that a cross-cousin marriage rule is the key to understanding the semantic value of reference terms not only in India but, pace Dumont, among the Australian Aborigines and elsewhere.

Dumont’s affinity as a value thesis, which Trautmann (1981: 174) finds to be a “bizarre way of looking at the matter,” also takes, at least at face value, a marriage-centric approach to alliance. Dumont’s theory is indeed bizarre, but this is because it contains a contradiction in that it embodies the seed of the contrary argument that fraternity is the key value. This contradiction is not fatal to his theory; to the contrary, it is a creative contradiction that opens up new questions for research.

My paper is divided into three sections. The first section critically examines Dumont’s affinity as a value argument in light of the theoretical arguments of Lévi-Strauss and Trautmann, as well as some ethnographic evidence on kinship reference and address in Central India. The second section develops the idea that a consideration of fraternity as a value raises the question of the pragmatics of address in addition to the classic question of the semantics of reference. The third section, on fraternity as a value, develops the implications of this idea for understanding the cultural specificity of Dravidian and non-Dravidian kinship in India.

Dumont on affinity as a value

If some of Dumont’s main thesis about equality and hierarchy is expressed in clear, comprehensible language, then some of his minor theses are expressed in opaque language that is often visually illustrated with equally incomprehensible diagrams. His affinity as a value thesis is, alas, a proposition of this kind. However, if we see his general theses as second- and third-order abstractions from his ethnographic data and follow him on his journey from concretion to abstraction, then his arguments not only become clearer, but they also become persuasive. Here is his summary formulation of his thesis:

With us modern Westerners, affinity is subordinated to consanguinity, for my brother-in-law, an affine, becomes an uncle, a consanguineal relative, for my children. In other words, affinity is ephemeral; it merges into consanguinity in the next generation. As values are by definition conceived of as permanent, durable, I may say that affinity is inferior to consanguinity, or undervalued in relation to it. Now my thesis is that the specificity of the South Indian kinship system lies in the fact that affinity there is transmitted from generation to generation, is thus permanent and durable, and so has equal status with consanguinity, or a value equal to it. (Dumont 1983: vii)

This statement poses two unsubstantiated assertions and a number of questions. The unsubstantiated assertions are that in South India (hereafter SI) consanguinity is equal to affinity, whereas it is greater than affinity in Europe. These assertions can be formulated as follows:

SI:        consanguinity = affinity        (1)

Europe:        consanguinity > affinity        (2)

These formulations pose the question of the situation in North India (hereafter NI). Is consanguinity greater than, less than, or equal to affinity? Dumont does not address this question explicitly in his Affinity book, but he alludes to it in another article where he critiques the cross-cousin marriage criterion as a basis for contrasting NI kinship with SI kinship. Such criterion, he notes, negatively characterizes North Indian kinship and “does not take into account the numerous and important features common to both regions, which can be roughly summed up as a considerable stress on marriage” (Dumont 1966: 90). The logic of his argument leaves only one possible interpretation of this statement: that consanguinity is less than affinity in NI (i.e., that it is different from the situation found in Europe and similar but different from that found in SI). We can write this as follows:

NI:        consanguinity < affinity        (3)

These three propositions constitute the essence of Dumont’s affinity as a value thesis and do so by posing a key question: What is the nature of the value that informs the greater than, less than, and equals signs? This is the question that I shall try to address in this essay.

Propositions (1) and (2) were elaborated on in his infamous 1953 Man article (reprinted in Dumont 1983)—infamous because it was greeted with derision. The offending argument was his assertion that “the basic meaning of the terms for the ‘cross’ category is affinal—my mother’s brother is essentially my father’s affine” (Dumont 1983: 12). This can be formally expressed as

SI:        MB = FWB        (4)

and visually as in the following diagram:

This proposition was greeted with complete incomprehension: “I cannot claim that I understand the article of Dravidian kinship terminology by Mr. Dumont,” Radcliffe-Brown admitted, “but I can assure Mr. Dumont that amongst the Australian natives the maternal uncle is thought of as the brother of Ego’s mother and not as the brother-in-law of the father” (reproduced in Dumont 1983: 20). Trautmann (1981: 174) called it a “bizarre way of looking at the matter.” His powerful critique of Dumont’s theory affirms the conventional wisdom that the key defining characteristic of Dravidian kinship is cross-cousin marriage (i.e., that consanguinity, not affinity, defines it). The proof of this is that one term, mamam, refers to both MB and FZH.

Dumont’s 1953 article also contains a number of other unusual propositions that need to be introduced before I move on. The first is that “cross cousins are essentially ego’s affines” (Dumont 1983: 14), which can, for a male ego, be written as

SI:        MBS = WB (ms) = FWBS        (5)

and illustrated visually as follows:

Dumont uses a diagram of this kind, which embodies propositions (4) and (5), to illustrate his argument that affinity as a value is inherited from father to son. Few have been persuaded because Dumont’s choice of language does not make sense. Why insist on calling a consanguineal relationship an affinal relationship? This argument has the appearance of a decree of the kind that declares that, henceforth, black is white.

We can begin to make sense of it when we get to the final proposition, the one Dumont rightly considered most important. This is the idea that, in South India,

the two categories of kin and affines comprehend all relatives without any third category. This may be understood without resorting to dual organization; the opposition between kin and affines constitutes the whole—the affine of my affine is my brother; marriage is in a sense the whole society, which it unites, and at the same time separates in two from the point of view of one Ego. (Dumont 1983: 17, emphasis added)

Two key, but related, issues are at stake here, as indicated by the italicized phrases. The proposition “the affine of my affine is my brother” is equivalent to saying that my wife’s brother’s wife’s brother (WBWB) is my brother (B), which can be formulated as follows:

SI:        B = WBWB        (6)

This is visually captured in the following diagram, which also embodies another version of proposition (5):

From Alter-ego’s perspective, we get two equivalent propositions: ZH = MBS and B = ZHZH, which are visually illustrated as follows:

Dumont’s argument that affinity “has equal status with consanguinity, or a value equal to it” (1983: vii) is tantamount to the claim that affinity can create consanguinity. It is one thing to look at an actual relationship of consanguinity and call it a relationship of affinity (as illustrated in (4) and (5) above) but quite another to look at an actual relationship of affinity and call it a relationship of consanguinity as (6) illustrates. Of course, it is not Dumont who has made up this proposition; rather, the Tamil speakers he worked with did. It is their mode of valuing certain kin. Dumont’s “bizarre” propositions (4) and (5) simply take the idea to its logical conclusion. But is his use of theoretical language appropriate? If Tamil speakers say that Ego’s wife’s brother is deemed to be a cousin, while his double affine, his wife’s brother’s wife’s brother, is deemed to be a brother, then this formulation throws into question the very meaning of the words “affine” and “consanguine.”

Dumont’s decision to continue using these terms as the labels for his theoretical categories was an unhappy choice. Two factors were no doubt at play here as he struggled to express this idea the he gained from his informants in the general language of theory. First, his Tamil informants, unlike my Halbi informants, had no names for the general theoretical category he identified in their kinship system; secondly, Lévi-Strauss’s magnum opus, The elementary structures of kinship (1949), had just been published, and his language of “marital alliance” was in the ascendancy.

The Halbi word that best captures Dumont’s notion of affinity as a value is dadabhai, a compound word that translates as “brotherhood.” It consists of the word dada, an address term for elder brother and father’s father, and bhai, a reference term for “brother.” The word dadabhai, along with its complement saga, refers to an “opposition between kin and affines” that “constitutes the whole,” to use Dumont’s language. Brothers, bhai, fall into two classes. The first is maina bhai, which can be glossed as “cross brother” (XB). Among the referents within the category XB are genealogical referents such as WB, MBS, FZS, and the like. Dumont’s proposition (5) in his theory then becomes, in the language of Halbi speakers:

maina bhai = XB = MBS = WB = ZH

The second is an unmarked category of bhai (B) that includes referents such as B, FF, and WBWB, among many others. Thus, Dumont’s equation (6) becomes

bhai = B = WBWB = ZHZH

These are the Halbi words for the two types of fraternal alliance that have the appearance of marital alliance. The following diagram captures the Halbi usages for Ego as EGO and Alter-ego as (ego).

Halbi kinship is an example of a Central Indian system that has an Indo-Aryan lexicon with a Dravidian semantics (Trautmann 1981). Lexicons are culturally specific, but their semantic structures are general; the sub-categories (unmarked) bhai (B) and maina bhai (XB) capture the super-category bhai (B) that Dumont identified but failed to name. In other words, the word bhai is a polyvalent category that has many levels of meaning that equations of the type given above can merely hint at.

Following Lévi-Strauss, we can call these affinally created alliances fraternal alliances. Lévi-Strauss’s great achievement was to distinguish between marital alliances and fraternal alliances, but, wedded as he was to the idea that moieties (dual organization) rather than brotherhoods were the elementary structures of kinship, he raised the concept of brotherhood only to dismiss its significance: “It is far from our mind to claim that the exchange or gift of women is the only way to establish an alliance in primitive societies,” Lévi-Strauss (1947: 483) notes. The establishment of blood brotherhood, he adds, creates a bond of alliance, which entails the prohibition of a marriage with the sister. But this, he argues, is a form of mechanical solidarity, which “adds nothing and unities nothing,” unlike marital alliances, which are a form of organic solidarity and bring about “an integration of the group on a new plane” (1947: 484).

While Lévi-Strauss concedes that marital alliance has its limits, he is unable to go the full distance and allow for the fact that fraternal alliance, too, may be consequential. To admit this much would be to question the fundamental assumption upon which his theory depends: the incest taboo. The relationship between incest and exogamy was a much debated subject when Lévi-Strauss was writing. Indeed, the Seligman referred to in his opening sentence is Brenda Seligman, who argued that it “it is rash to assume the universality of the idea that the supreme gift is that of a woman and that marital relations cements social solidarity” (1935: 91). She also quotes the case of Tikopia, “where the whole population is in reality related, [and] marriage with the classificatory brother or sister, though said not to be correct, is inevitable and is tolerated” (1935: 86). In India, sibling marriage is the norm, but not all siblings are marriageable. In Bastar, for example, marriageable siblings are marked by the adjective maina. Thus, a woman marries her maina bhai, her cross-brother (XB), as the diagram above illustrates. To call this cross-cousin marriage as anthropologists always do is to deny oneself access to the native point of view, which Malinowski (1922: 25) rightly stressed was the goal of the ethnographer.

Lévi-Strauss’s (1949: 394) marital-alliance-centric approach is apparent in his analysis of data from Central India, where he correctly notes that dadabhai is not a moiety but sees it as an “exact equivalent” of the Kachin wife-giving group mayu ni. However, he has to change the data to make it fit his model: in Central India, patrilateral cross-cousin marriage, not matrilateral cross-cousin marriage, is the moral ideal, as his figure 75 depicts. However, he does notice that “it is curious that in these conditions the most frequent marriage is not with the mother’s brother’s daughter but with the father’s sister’s daughter” (1949: 395). If one takes a fraternal-alliance-centric approach—the insider’s perspective—then the data ceases to be curious.

Dumont, dismissive as he was about so-called tribal India—an administrative category devoid of explanatory adequacy that he uncritically accepted—nevertheless realized that Australian models of dual organization were not applicable to India:

While Australia has global models of the society, we are left, in India and elsewhere, with representations attached to an individual ego, representations that are in that sense irremediably local, and there is no way to infer the global from the local.

While, from the present point of view, Australia is characterized by a global model of society as made up of intermarrying groups (or marriage sections as they are sometimes called), India has only categories, which regulate the repetition of intermarriage—or, as I call it, the marriage alliance—but which operate in each case by reference to a particular locus and thus exclusively take the form of what is called in anthropological jargon “cross-cousin marriage.” (Dumont 1983: viii, italic emphasis in original, underlined emphasis added)

Thus, Australia is to South India as global is to local, and as groups are to categories. But what is “ego-centric” opposed to? The zero sign is obviously “socio-centric,” from which it follows that moieties are opposed to brotherhood as dual organization is to kindred groupings. These contrasts can be arrayed in a table as follows:

Australia India
Global Local
Groups Categories
Socio-centric Ego-centric
Moieties Brotherhoods
Dual Organization Kindred Groups

In sum, Indian kinship is not an “elementary structure” in Lévi-Strauss’s (1949) terms. Rather, “elementary form”—an idea that evokes the eye of the beholder—is more like it. The local ego-centric kindred groups called brotherhoods make no sense when looked at from the perspective of global socio-centric moieties of the dual organization type found in Oceania and elsewhere.

The pseudo-mathematical language of “equations” that kinship theorists use is a very imperfect way to talk about kinship equivalences of the type I have discussed above. This language covers a semantic field that includes absolute identity at one extreme, via various notions of similitude, to generalized agreement on the other; but, as a word that stands opposed to inequality in some form or another, it is useful in that it poses the question of value. What is the process of valuation by which diverse entities are brought into some sort of agreement, be it absolute identity or generalized sameness on the one hand or inequality of some kind on the other? Such is the question that value theorists have always pondered, for it is a question that has concerned people of all ages. For example, the principle that informs relations of equality and inequality within and between castes is not just an academic question; in fact, people in India live the issue on an everyday basis as they are confronted with the problem of whether to accept food or water from the hand of someone they consider inferior. Dumont’s answer to the abstract theoretical question this poses is well known, but when it comes to the valuation of kin, his theories are pregnant with ideas that remain to be developed—and his notion of affinity as a value is one such idea. Dumont has perceived the problem but has not found the language to pose it. As we have seen, the problem is not one of affinity as a value but of fraternity as a value: By what process are some non-kin in India reckoned to be siblings? Who is the valuer? What are the values he or she uses in the valuation process?

A short theoretical interlude is necessary before looking at the data.

Theoretical interlude on the notion of value

Marx and the exchange-value of commodities

The preceding analysis has added four value equations to the initial two with which we began, but the question of the theory of value that informs the equations remains. Such was the question Marx (1867: 70) asked when confronted with the problem posed by commodity exchange: What, for example, is the common substance that equates twenty yards of linen with one coat to form a commodity equation of the following kind?

Commodities:        20 yards of linen = 1 coat

The equation between B and BWBW in (6) poses the analogous question for the kinship theorist: Wherein lies the fraternal oneness between a male ego and the affine of his affine whom he refers to and/or addresses as brother? Or, in other words, by what thought processes do SI people establish that BWBW is a brother? The theoretical language in which this question is expressed is crucial, because if we ask the wrong question, we will get the wrong answer. We need a language that is transcultural—one that does not bring with it implicit assumptions from another culture.

Marxian value theory, concerned as it is with commodities and the labor involved in their production, is a quantitative value theory because the problem of value presents itself as one of price as expressed in units of a currency. For example, if a coat costs $15, the problem presents itself as the following:

Money-price: $15 = 20 yards of linen = 1 coat

In contrast, the valuation of kin presents itself first as a problem of language in the form of the lexical items used to refer to people, and second as a problem of face-to-face address, which may involve non-verbal as well as verbal signs. The first problem raises the question of semantics and the second of pragmatics. Saussure had some interesting things to say about linguistic values of the former kind, and Jakobson about the very different value question that face-to-face address poses.

Saussure and the exchange-value of reference terms

The notion of linguistic value comes from Saussure, who argued that the move from lexicology to semantic structure is a move from signs to values. Linguistic value, Saussure (1915: 114–17) argues, is concerned with the exchange-value of the signs in much the same way that economic value is concerned with the exchange-value of commodities. Furthermore, just as Marx argues that the values that inform exchange-values come from outside the sphere of exchange, so Saussure argues that linguistic value is determined by the greater system of signs of which it is a part.

Sheep,” he argued, “has the same meaning as the French word mouton, but not the same value, for mouton can also be used to mean the lamb meat, whereas sheep cannot, because it has been delimited by mutton” (Saussure 1915: 115). Insofar as kinship terms are concerned, we can see that the reference term mama in Hindi has the same meaning as the word mamam in Tamil, but not the same value, for mamam in Tamil can also be used to mean the FZH, whereas mama cannot because it has been delimited by the referent MB. Speakers of any given language are formulating exchange-values all the time without necessarily being aware of it. If you ask a Hindi speaker, “What is the relationship between FZ and MB?,” they will instantly reply, “A sister/brother relationship.” In other words, the exchange-value of FZ/MB is Z/B, which can be written as the following value equation:

NI:        bubu/mama = FZ/MB = bahin/bhai = Z/B

This exchange-value is culturally specific to NI Hindi speakers; SI Tamil speakers have a different answer, as we shall see. In Saussure’s terms, they are in a different “environment,” which raises the very difficult extra-linguistic question of the identity of the valuers in those environments.

It also should be noted that the formal linguistic-value of kinship terms is in the form of an exchange ratio that abstracts from the ego. This is because reciprocal reference cancels out the ego, as it were. Equations of the kind above can be deduced from sentences such as “The bubu/mama relation is a bahin/bhai relation” or “My bubu and my mama refer to each other as bhai and bahin.” The speaker in the latter sentence is the ego for two of the terms, bubu and mama, but not for the terms bhai and bahin; the ego in this example is a classic example of a “shifter” (Silverstein 1976).

We are now in a position to grasp the linguistic values that inform proposition (6) B = WBWB. From male Ego’s perspective, when he refers to his WBWB, the valuation is a semantic valuation of a linguistic kind. In other words, it is a purely linguistic phenomenon—an exchange-value between reference terms. When male Ego refers to WBWB as B, the implication is that Alter-ego refers to ZHZH as B, too.

Thus, proposition (6) has the following more precise exchange-value formulation:

SI:        WBWB/ZHZH = B/B        (6a)

The reality in SI is that relative age divides brothers so that they are never equal: Ego’s elder brother is referred to as akka in Tamil, the younger as tambi. Thus, the equality is expressed by the mutual use of these terms by Ego and Alter-ego. Assuming WBWB is older, we get the following proposition:

SI:        akka/tambi = eB/yB = WBWB/ZHZH = B/B

The diagrammatic form of this equation is as follows:

The terms akka and tambi are equal in the sense that they both belong to the same semantic field of “brother” (B), but they are unlike in that they belong to different sub-fields in this semantic space. As members of the category “brother” (B), which is unmarked in Tamil but not in Halbi, they are unlike cross-brothers (XB). But at the super-category level B, they are in the same field as XB but opposed to the super-category “sister” (Z). In other words, formal linguistic value is determined by the position of the term on the branches of a semantic tree. The equations above are useful as a shorthand mode of expression only because they flatten the different levels; linguistic value of a semantic kind must be understood as a tree or its equivalent matrix form.

This poses the comparative question of how many semantic trees of the sibling kind there are in India. Are these species of the same kind of generic tree or radically different generic types? For Dumont (1980), a lumper, India is one and his unfinished task was to demonstrate this fact. For his critics, such as Trautmann (1981), a splitter, Indian kinship consists of two incommensurable paradigms.

Jakobson and the moral value of face-to-face address

Using the exchange-value of reference terms is only one way to think about linguistic value; the other is address terminology. This introduces a completely new mode of valuation for reasons that Jakobson (1960) outlines. Language has many functions, and reference is only one of these. There is a world of difference between speaking about kin in the third person and speaking to them directly, face to face. The former situation raises questions about the referential function of language, the latter about the emotive and conative functions of language. These are moral sentiments of a socio-linguistic nature, not pure linguistic values. Reference terms raise the question of the abstract semantic relationship between ego and alter-ego; address terms, on the other hand, raise the altogether different question of the concrete, face-to-face relationship between the addressor and the addressee. The emotive or expressive function, notes Jakobson (1960: 6), is focused on the addresser and “aims a direct expression of the speaker’s attitude toward what he is speaking about. It tends to produce an impression of a certain emotion whether true or feigned.” The conative function, which finds its grammatical expression in the vocative and imperative, is oriented toward the addressee. Thus, it is one thing for Ego to say, “He [Alter-ego] is my father,” but quite another for a son to address his father as “Hi, Dad” and for the father to reply “Hello, Chris” in return. The relationship between Ego and Alter-ego is synchronic and abstract, and that between addressor and addressee is a concrete diachronic inter-temporal relationship whose emotive value varies according to the mood of the addressor and the choice of address term used, be it “daddy,” “father,” or “sir.”

Verbal address is just one mode of address among many that include non-verbal modes such as salutes. Valuation of this kind moves us from the purely linguistic world of the meaning of the exchange-value of reference terms to the meaning of reciprocal actions and reactions of people in face-to-face encounters. Here, kindred equality is expressed as mutual respect. In other words, a culturally specific notion of respect is the key value that informs equations such as (6).

Dumont, careful ethnographer that he was, paid meticulous attention to the way kinship terms were used, distinguishing not only between reference and address but also between generic categories and individualizing terms, and potentially also distinguishing marriageable kindred from specific categories of married kindred. Trautmann (1981), by contrast, is concerned only with reference terminology and does not even pose the question of address. However, being the library-based scholar that he was, he cannot be blamed for ignoring the special problems that address terms pose because good ethnographic data on modes of address in India are virtually non-existent.

To recapitulate my argument so far: Dumont’s thesis about affinity as a value embodies a number of minor theses, specifically that:

  1. it is culturally specific to India and cannot be understood from a dual organization perspective;
  2. it must be re-phrased as “fraternity as a value”; and
  3. value in this context means both the semantic value of reference terms and the moral value of address terms, the basis of the latter being familial love and respect.

Fraternity as a value

Fraternity as a value is of pragmatic importance to speakers of an Indian language because it distinguishes between marriageable and unmarriageable kin along an affinal chain of otherwise unrelated people. The following diagram captures the paradoxical fact that in Dravidian kinship systems, affines who are one marital step away are marriageable but those two steps away are not. A male ego places his BWZ in the marriageable cross-sister category XZ, but his BWBWZ is a sister, Z, and hence unmarriageable.

The shift in perspective that Dumont is urging us to take on SI Dravidian marriage is difficult to grasp because it is so simple and obvious. It forces us to problematize a fundamental assumption that has informed discussions about Dravidian kinship for over one hundred years—namely, that the reference terms imply a prescriptive cross-cousin marriage rule. That is one way of interpreting the terms, but it privileges consanguinity. But because consanguinity is equal to affinity, the other way to interpret the reference terms is via the affinal links. From this perspective, there is no rule to follow but a pool of marriageable siblings to choose from. This forces us to see things from the perspective of potentialities rather than prescriptions. The language of Halbi kinship, with its distinction between marriageable and unmarriageable siblings, enables us to grasp the native point of view. It also enables us to make sense of Dumont (1983: 22) when he says that “affinity in a way precedes the actual marriage, that an individual has potential affines before he acquires actual in-laws by marriage” and that “affinity in a wider sense is inherited just as our ‘blood’ relationships are.”

Fraternal alliances of this kind imply that people in India marry people they refer to and address as siblings. This, of course, does not imply incest, a morally loaded category if ever there were one. Rather, it implies what I will call sibling endogamy. SI kinship categories imply a brotherhood of intermarrying siblings, which is nothing more than a definition of caste from the local perspective of Ego. This ego can be male or female, but the native point of view obliges us to take the male perspective if we are to fully grasp Indian kin categories. This is because wives are incorporated into brotherhoods in India but not into moieties, as is the case, for example, in Aboriginal Australia.

We must now travel with Dumont and consider the values that inform fraternity as a value in North India, a region where affinity is greater than consanguinity, as equation (3) tries to capture.

The word bhai (“brother”), notes Dumont (1966: 99), takes on diverse meanings, “from proper brother, born of the same mother and father, to an acceptation such as ‘man approximately equal in age to Ego’, in his village or even beyond it. The latter extra-kinship meaning occurs in address only.” He adds that

nothing prevents its being extended indefinitely and its including all the members of the endogamous group or of the sub-caste (within certain age limits), as there is a vague notion of their being of the same blood. All men might even be included, insofar as they are all sons of Adam, or Manu. The transition is thus easy to the extra-kinship uses encountered in address. This way of looking at the relationship does away with some of our difficulties. For instance, it allows us to understand how an affine, say the wife’s brother sala, can be opposed to bhai in the stricter sense, and can also merge into the category bhai in the wider sense. For the most stringent interdictions of marriage prohibit it between persons who have a common ancestor within a certain range of generations (sapindya rule), or again, as long as memory of an intermarriage persists. Beyond that point, the opposition between consanguinity and affinity disappears, the affine becomes a “brother” either in the vague sense of member of the same endogamous group, or even in the sense of a sufficiently remote common ascendancy. (Dumont 1966: 100)

Dumont’s analysis of NI kinship terms was the beginning of a long and technically complex debate (Dumont 1975; Jamous 1992, 2003; Vatuk 1972), but at the center of it is the nature of the brotherly relationship between male Ego and his wife’s brother. This can be expressed as a linguistic value equation of the form eB/yB = ZH/WB = jija/sala. This is an asymmetrical respect relationship: sala defers to jija regardless of age. It is like an eB/yB relationship in that the respect the yB gives to the eB is transferred to the jija as wife taker. This implication of this species of fraternal alliance is that the unmarried sister of B and the unmarried WB are siblings and hence unmarriageable. This is reflected in the reference terminology, where we find that BWB/ZHZ = B/Z = bhaiya/bhaini in Fijian Hindi.

The following diagram summarizes the preceding argument:

The corollary of the unmarriageability of the BWB/ZHZ = B/Z = bhaiya/bhaini sibling relations is the marriageability of the WBW/HZH = saraj/nandoi relationship (including, most importantly, the unmarried same-sex siblings of men and women in this relationship, i.e., the WBWZ/HZHB relation).

This linguistic valuation presents a stark contrast to the SI situation, where the same relationship is unmarriageable. The terms una (B) and akka (Z) are used to refer to this relationship.

The ZHZ/BWB relation, by contrast, is marriageable. The South Indian migrants in Fiji, who use a mixture of Tamil and Hindi terms in their lexicon (but in an SI way), refer to this this relationship as mama/bhauji, as shown below.

The term mama as used by both male and female Ego has the sense of cross-brother (XB), but it also is used to refer to one’s mother’s brother and, as such, is a classic example of Dumont’s (1983: 14) argument that the affinal links are given a consanguineal expression. The emotive value of mama is that of extreme familiarity—joking kinship.

These stark contrasts might lead one to believe that the SI system is a mirror image of the NI system. This much is true of the elementary forms of fraternal alliance, but when these links are used to construct a chain of affinal links, the substantive difference can be perceived.

The following diagram depicts the SI situation, and what is striking is the symmetry. Affinal chains, from the perspective of Ego, move in opposite directions: via the sister’s husband (to the left in the diagrams I have used) or via the brother’s wife (to the right). In other words, in the direction of wife takers (left) and wife givers (right), the Ego-centric brotherhood—the point of reference—is in the center. The moral sentiment informing face-to-face relations between the brotherhood and its wife takers, on the one hand, and between the brotherhood and its wife givers, on the other, is marked by mutual respect in SI and CI. All the relatives on this chain are referred to as siblings, but distinctions are made between different categories of siblings. The first is between marriageable and unmarriage-able siblings. The latter are found in the center of the chain and at two steps removed, where we find, as Dumont noted, that the affines of affines are siblings, albeit siblings of a kind where familial moral sentiment is weaker.

The crucial defining characteristic of the Indian brotherhood is that it is defined by husband and wife sets, on the one hand, and brother and unmarried sister sets, on the other. The former can be called an affinal couple, the latter a consanguineal couple. The symmetry to be found in SI comes about because these two sets are equal in that sense Dumont suggests characterizes the relationship between consanguinity and affinity in SI. To express this formally, the equation

affinity = consanguinity

finds its expression in the relationship between

(H+W) = (B+Z).

The values operating here get us beyond Saussure’s linguistic values and Jakobson’s emotive values. These are the familial sentiments of respect and love. The equals sign here signifies mutual respect but one where the wife is always obliged to defer to her husband and where the sister has to defer to her brother but only if he is older. The elder sister commands respect and gets it.

Respect relations differ in NI, and the specificity of NI respect relations gives the linguistic values of the reference terms a sociological twist. In the north, it is not only the elder sister who gets respect but all sisters regardless of age, and especially those who are married. Her status is the reflected glory of that of her husband. He becomes male Ego’s new elder brother and commands respect. His sisters, too, become elder sisters and hence unmarriageable. This continues all along the affinal chain as it heads in the sister’s husband’s direction (to the left). Male Ego, therefore, must look in the other direction for a wife, a younger sister of his younger brothers in the wife-giving direction. This gives the chain a tilt, as shown in the diagram below.

This tilt comes about because affinity is greater than consaguinity in the north, as Dumont might say, because the husband/wife couple is more respected than the brother/sister couple. Formally:

NI:        consanguinity < affinity        (3)


(B+Z) < (H+W).

This tilting divides siblings into respected siblings, who are treated like elder siblings, and others, who are treated like younger siblings. This obliterates the cross/parallel distinction and severs the joint brother/sister ego, creating the elder sister ego and a younger brother ego who endow the reference terminology with new linguistic values. In short, the semantic structure of NI kinship acquires a husband-centric bias. The following figure captures the essence of this transformation in the form of the semantic tree for siblings as one moves from SI to NI.

We are now in a position to consider the question of who are the valuers. For Marx (1867: 4) the relationship between value and valuers was that of exchange to production: the values of equality and freedom in the marketplace were informed by the unequal relationship between capital and labor on the factory floor, a relationship that was masked by commodity fetishism. In the domestic domain, the relationship between value and valuers is not veiled in this way; it is there for all to see. Family relationships are tussles for power and status between men and women on the one hand and within a group of men and women on the other. The key players in the affinal chains found in India are obviously women as wives and sisters and men as husbands and brothers. Insofar as the NI case is concerned, the husband-centric bias of the linguistic values and their complementary moral values betray the role of the husband as the central valuer: jija (ZH), as wife taker, is a term of respect; sala (WB) as wife giver, is a term of abuse. In SI and CI, the relative strength of the brother and his values has left its mark on the form of the chain. The history of this struggle between husbands and brothers is impossible to reconstruct, but it is one that persists to this day and is fought out in the life cycle rituals of the people of India, especially the wedding rituals and the tussles over dowry, which have had many tragic consequences in North India. In Fiji, where relationships are more egalitarian, the Fijian kin term tavale has entered the Fiji Hindi lexicon as a loanword in place of sala. In Fijian, tavale refers to both WB and ZH and is a relationship of mutual respect and familiarity.

The critic of the preceding argument can rightly point out that my discussion has been restricted to Ego’s generation to the exclusion of all others. Such a discussion is beyond the scope of this paper, but one of Dumont’s (1966: 111) examples of an actual marriage he recorded while doing fieldwork in NI gives some indication of where a consideration of other generations might take us. According to conventional wisdom, NI is characterized by the absence of cross-cousin marriage, but Dumont found that Sarjupari Brahman males predominantly marry their classificatory MBD, who are referred to as bahin, or sisters. However if we conceive of the MB as an affine, as Dumont says, then the favored marriage turns out to be with a man’s actual FBWBWD. The following diagram illustrates what is at stake here. As we ponder the difference between the classificatory form of the marriage and the actual form, it becomes obvious that affinity is a good way of thinking about consanguinity of the classificatory type.

The term kaki, the father’s younger brother’s wife, can be glossed as classificatory “junior mother.” Thus, a man’s FyBWB is a mama, a classificatory “mother’s brother.” His daughter is a sister bahin but a classificatory marriageable one. Her mother and her mother’s sister were potentially marriageable affinal sisters for Ego’s father, a potentiality that is inherited by the son, as Dumont might say.


Trautmann’s critique of Lévi-Strauss and Dumont now defines the terms of the debate about Dravidian kinship. The value question is posed in terms of the semantics of culturally specific reference terms in cultures where a rule of cross-cousin marriage of some kind or other prevails. Dravidian kinship is thus no longer a problem of Indian sociology but an abstract, transcultural theoretical one of typology based on different notions of crossness. Trautmann and his collaborators have been very successful in solving these abstract problems—so successful, in fact, that few questions remain to be resolved, and those that do probably have no solution.

The forgoing analysis suggests that we need to go back to Lévi-Strauss and Dumont to now go forward. Lévi-Strauss’s assumption that fraternal alliance is of secondary importance needs to be questioned. Dumont’s argument that Dravidian kinship has culturally specific characteristics that it shares with Indo-Aryan kinship needs to be revisited. The fundamental contradiction in his thesis of affinity as a value—and I claim that this theory only makes sense as a theory of fraternity as a value—provides a way forward. Such an approach moves us from a focus on reference terms to address, from semantic value to emotive and conative values, and from exogamous clans to kindred groupings. Such an approach requires a historical geography of India of the kind that Trautmann pioneered in his definition of an “Indian culture of kinship” (1981: 315). Indian Dravidian kinship is very much a part of this culture, not apart from it as he argued, but one needs to focus on the pragmatics of address to appreciate this. A historical geography of fraternal alliances in Indian kinship takes us beyond the shores of India to the lands to which the diaspora has spread. In Fiji, for example, this Indian culture of kinship lives on in the life cycle rituals and language of the fourth- and fifth-generation descendants of indentured laborers who went to Fiji to work on sugarcane plantations. However, the brotherhoods they have formed today are very different from the ship brotherhoods their great-grandparents formed on their way to Fiji. Fraternity as a value, then, poses many problems that have yet to be asked, let alone answered.

Some of these problems concern unexamined Eurocentric assumptions about the overvaluation of consanguinity relative to affinity and contiguity, a theme that Viveiros de Castro (2009) develops in the context of Amazonian kinship: the standard Western model, he argues, has consanguinity in “the province of the given: it is an innate, passive property of the human relational matrix, its essential bodily substrate. Affinity is active construction: it is differentiating choice, affective or political, and inventive freedom.” The Amazon model, he asserts, is the converse of the Western model: “Here we find affinity as a given, internal and constitutive relation, and consanguinity as constructed, external and regulative” (2009: 259).

This argument echoes the argument I have been developing above about affinity being the cause of consanguinity in the Dravidian CI and SI systems. It is interesting, therefore, to see it being advanced in the context of an Amazonian Dravidian system and in the theoretical context of Lévi-Strauss’ (1949) discussion in The elementary structures of kinship about the distinction between fraternal and marital alliance discussed above.

The value question, as Ton Otto remarks in the introduction to this special issue, traps us all. As Dumont and Viveiros de Castro remind us, we cannot examine the values of “them” without interrogating the values of “us”; furthermore, we are all agents and victims of valuation processes involving market values, familial values, and religious values where language is used as both means and end, and it has been this way since the emergence of Homo sapiens as a moral being. Morality, as Mauss notes (2007: 156), is “the art of living together, and it can be recognised by the presence of good.” Disputes about the definition of the good have a long history and no simple answer. The question of value must be continually re-posed. Familial values are only part of the story but are a domain where anthropology has made a significant contribution. To continue to re-pose these questions, we must go back to our intellectual ancestors in order to move forward in light of new ethnographic evidence.


This essay was first presented at the Anthropology of Value workshop in Cairns in December 2011. I am grateful to the participants for their comments and especially to the editors for thoughtful suggestions on how to clarify my argument. The UK ESRC has provided support for some of the research on which this paper is based.


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

———. 1975. “Terminology and prestations revisited.” Contributions to Indian Sociology (NS) 9 (2): 197–215.

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

———. 1980. Homo hierarchicus: The caste system and its implications. Chicago: University of Chicago Press.

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

———. 1986a. Essays on individualism: Modern ideology in anthropological perspective. Chicago: University of Chicago Press.

———. 1986b. A South Indian subcaste: Social organisation and religion of the Pramalai Kallar. Translated and edited by Michael Moffatt and Lewis Morgan. Oxford: Oxford University Press.

———. 1991. German ideology: From France to Germany and back. Chicago: University of Chicago Press.

Galey, Jean-Claude. 2000. “Louis Dumont (1911–1998): A committed distancing.” American Anthropologist 102 (2): 324–28.

Jakobson, Roman. 1960. “Closing comments: Linguistics and poetics.” In Style in language, edited by Thomas A. Sebeok, 350–77. Cambridge, MA: MIT Press.

Jamous, Raymond. 1992. “The brother-married-sister relationship and marriage ceremonies as sacrificial rites: A case study from Northern India.” In Understanding rituals, edited by D. de Coppet, 52–73. London: Routledge.

———. 2003. Kinship and rituals among the Meo of Northern India: Locating sibling relationship. Oxford: Oxford University Press.

Lévi-Strauss, Claude. (1949) 1969. The elementary structures of kinship. London: Eyre & Spottiswoode.

Malinowski, Bronisław. (1922) 1961. Argonauts of the Western Pacific. New York: Dutton.

Marx, Karl. 1867. Capital. Vol. 1, A critical analysis of capitalist production. Moscow: Progress.

Mauss, Marcel. 2007. Manual of ethnography. London: Berghahn Books.

Moffatt, Michael, 1986. Introduction to A South Indian subcaste: Social organisation and religion of the Pramalai Kallar, by Louis Dumont. Translated and edited by Michael Moffatt and Lewis Morgan, 1–6. Oxford: Oxford University Press.

Saussure, Ferdinand de. (1915) 1959. Course in general linguistics. London: Philosophical Library.

Seligman, Brenda Z. 1935. “The incest taboo as a social regulation.” Sociological Review 27 (1): 75–93.

Silverstein, Michael. 1976. “Shifters, linguistic categories, and cultural description.” In Meaning in anthropology, edited by Keith Basso and Henry A. Selby, 11–55. Albuquerque: University of New Mexico Press.

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

Vatuk, Sylvia. 1972. “Kinship terminology in Northern India.” American Anthropologist 74 (3): 791–93.

Viveiros de Castro, Eduardo. 2009. “The gift and the given: Three nano-essays on kinship and magic.” In Kinship and beyond: The genealogical model reconsidered, edited by Sandra Bamford and James Leach, 235–68. New York: Berghahn Books.

La question de la valeur en Inde : : réflexions ethnographiques sur un débat en cours

Résumé : Les termes du débat relatif aux approches anthropologiques de la valeur en Inde ont été fixés par Louis Dumont dont les théories avaient pour fondement ses études ethnographiques au Nord et au Sud de l’Inde, sa connaissance de la littérature sanskrite, sa synthèse de l'ethnographie comparée de l’Inde, et ses études sur l'histoire de la pensée économique européenne. Sa théorie de l'affinité comme valeur (un élément de sa théorie générale) était fondée sur une critique de Claude Lévi-Strauss et a été, à son tour, critiquée entre autres par Thomas Trautmann. À partir d'un travail de terrain effectué en Inde centrale, j'attire l'attention sur une hypothèse que les trois théoriciens partagent, selon laquelle l'alliance est assimilée à l'alliance matrimoniale et néglige l'importance de l'alliance fraternelle, et j'examine ses implications théoriques.

Chris GREGORY has appointments in Anthropology at the Australian National University and the University of Manchester. He has conducted fieldwork in India, Papua New Guinea, and Fiji, and his research has focused on the relationship between kinship, economy, and religion. He is the author of Gifts and commodities (Academic Press, 1982), Observing the economy (with Jon Altman, Routledge, 1989), Savage money (Routledge, 1997), and Lachmi jagar: Gurumai Sukdai’s story of the Bastar rice goddess (with Harihar Vaishnav, Kaksad Publications, 2003).

School of Archaeology and Anthorpology
College of Arts and Social Sciences
Australian National University
Canberra ACT