Navigation – Plan du site
Kinship, gender, and social relatedness

From structure to substance and back: materialities in Australian Aboriginal kinship

Laurent Dousset

Texte intégral

1In Race et Histoire, Lévi-Strauss wrote a few very interesting sentences, which I must quote here to some extend as a start:

“For all that touches on the organization of the family and the harmonization of the relationship between family group and social group, the Australians, although backwards with regard to economy, occupy such an advanced position in comparison to the rest of humanity that it is, to understand the systems of rules that they have elaborated in a reflected and conscious manner, necessary to appeal to the most refined forms of modern mathematics…. They have thus surpassed the level of empirical observation to elevate themselves onto the level of knowledge on mathematical laws that govern the system. They go so far that it is not exaggerated to consider them not only to be the founders of general sociology, but even more, to consider them as the true introducers of all measures in the social sciences” (1987:48-49; my translation).

2One notes of course that, rather infrequent for Lévi-Strauss in such explicit words, he places social forms in some kind of an evolutionary comparison: Australians are backwards in economy, but the most advanced in the social mathematics. What is even more interesting however in this quote, is the following short but strong and very unexpected bit: “they have elaborated in a reflected and conscious manner” a system of rules. And as such, they become the architects of their own social system, but more, the founders of general sociology.

3This clearly testifies of the admiration he had for Australian kinship and social category systems, which in other places he has depicted as the crystalline beauty of Australian classes and kinship (1996:41-42). But it is particularly unexpected because, as we know, he was first of all interested in the deep structures of things and relationships, those that happen unconsciously, those that are part of the human mind; and he was not interested in practice, as his famous response to Less Hiatt at the Man the Hunter conference testifies. He said: “my work on kinship has been concerned with a different problem: to ascertain what was the meaning of rules, whether they are applied or not” (1968).

4But here, elaboration in a reflected and conscious manner, as he says, of what he elsewhere calls deep structures, must necessarily be part of practice. How could one consciously elaborate unconscious rules and systems?

5I will not further engage with Lévi-Strauss, just to mention that the move from a conceptualization of kinship as abstract or unconscious systems underlying thought and practice, to an understanding of kinship as providing conscious means to engage in other domains, has been I guess a general tendency amongst most anthropologists working on questions of social organization in Aboriginal Australia; and has to some extend, along with the context and requirements of Native Title, been the reason for a resurrection the last 15 years and more of interest in these questions. – At least, this was and is my own approach to things that pertain to the domain: people know of course their system, and they know it so well, that they do bend, adapt, change what seem to be undisputable rules; and they can do so while still engaging with a meta-langue about and beyond these rules and their own practice. In other words, to make a long story short, when Lévi-Strauss said that people had conscious and reflected thoughts about kinship rules and their fabrication, he was right. But what he missed, is that these conscious fabrications, ultimately expressed in nearly mathematical language, have as object and subject other things than those these rules seem to be regulating in the first place. What he missed, was the materiality of kinship.

6One other quote is close to being enough to make my point. And it comes from our friend Bob Tonkinson. In an unpublished seminar given at ANU in 1975, of which he has provided me with his preparatory notes, he explains the practice of ngaranmaridi among the Martu of the Australian Western Desert:

…at the time of ritual introductions of strangers from different areas … when the particular kinship links are being determined, an element of choice exists as to whether to designate “FZ” as umari [WM] or gundili [FZ] and thereby differentiate their children accordingly. ... Discussions are held by Ego … and others to decide which if any of the stranger women who are initially all related as 'spouse' will be 'cut out' and thus become ... Z.

7Here again, let us underline a few words in the quote: people will decide, he writes, whether a woman is a sister or cousin / wife, and thus reclassify their mothers as either mothers-in-law or aunts. Those of you who know what a so-called classificatory system of relationship is, and in particular one that is fundamentally based on a distinction of the world into affines (parallel kin) and consanguines (cross-kin), will understand that this “cutting out” of wives, as Tonkinson calls it, is not a simple operation flowing from rules alone. It is a decision; a decision involving the capacity to adapt realities to ontologies and vice versa; a decision which involves some of the most fundamental aspects of cultural values, such as incest and its prohibition, or marriage and reproduction.

8What this seems to be all about is the capacity to adapt form to content, rather than the other way round. The form is well known: Australian classificatory systems of relationship. The content, however, is of a different matter altogether, and it is changing depending on context.

9Let me illustrate this through a few examples taken from my own experience... and I must apologize to those who have already heard them in other places. With respect to the Ngaatjatjarra-speaking people of the Western Desert, terminology and practice, or form and content, have indeed caused a few serious problems to the mathematics. Elkin and others, staying often only a few weeks or even days with Western Desert groups, have claimed to have observed a system in which people do not distinguish sisters from wives. And since the distinction is not made, or in fact since wives are supposedly called sisters, they thought that Western Desert, or Aluridja as it was called, was a culture that did not even embrace the most elementary and nearly-universal aspects of incest prohibition. Difficulties increase obliviously when one observes, and Elkin did so already, that these people distinguish mothers from father's sisters, and thus fit well into the general cross-parallel distinction. Theorists, such as Lévi-Strauss or Tjon Sie Fat, have struggled with that problem. The former ended up calling it an “aberrant system” (1967), while the latter recalled that it was described as one of the most “intriguing” system with “anomalous or inconsistent terminologies” (1998:78).

10The inconsistency is structural, it somehow does not reflect the exact mathematics. In fact, it reflects a different mathematics, and the solution is long-term fieldwork; when one is capable of observing the availability of different modes of classification within one and the same language; depending on context, and thus politics. Indeed, similarly to what Alan Rumsey (1981) said about Ngarinyin people, or David Kronenfeld (1973) about Fanti, skewing (also see Dousset 2012) — conflating cousins with a person of the upper-level generation — does most often not reflect actual systems from the classic anthropological perspective, but reflects terminological usages pertaining to particular contexts in which discourse and usage takes place. The basic calculation, however, remains that of a cross-parallel distinction.

11What happens is that Ngaatjatjarra-speaking people, and other groups of the Western Desert, apply three terminological sets, or in fact three modes of classifying differently the same people; without, let me reassure you, creating a situation of confusion or aberration. The first terminological usage is one that testifies of closeness, if not of identity. It flows from an idiosyncratic expression: Kungkankatja, Minalinkatja (Dousset 2002), freely explained as “children of a brother and of a sister are identical”. What should be cross-cousins thus are first of all considered as brothers and sisters themselves. The cross-parallel is extinguished. They are made of the same substance, the same heritage, have often grown up close to each other, and thus become siblings. Cross-cousins are people that, potentially at least, could have sexual relationships and marry. Children of actual siblings of opposite sex, even though mathematical cross-cousins, would not dare to engage in such bodily relationships. Similarly, never would I call my actual uncle “wife's father”, or my actual aunt “wife's mother”, even though, again, mathematically speaking, they sit in the same category. To summarize; closeness is here expressed through a mathematically-speaking unexpected usage of the terminology, one that transgresses a structural rule. At the same time, this systematic transgression seems to point to another level of cultural prescription; about which I will come back below.

12What is significant in here, is that the usage of the sibling terminology on people that are or should be cross-cousins is not limited to this apparent deviation to a structural rule, but is the result of other processes and situations that have not much to do with what we traditionally understand as being kinship. Whether children of actual siblings or not, people who share experiences or substances in a prolonged and repeated manner fall into the same appreciation: they become simply too close, too identical to being cross-cousins and thus potential spouses. Having grown up together or close to each other, having eaten food cooked on the same fire, for example, are fundamental elements for defining relatedness. A good example is nyuyurlpa (Dousset 2003), adoption, a very frequent practice in the Western Desert. The idiosyncratic expression used to explain adoption is:






The adoptive mother

from another one



[and] raised

13Adoptive mother is signified through the word nyuyurlpa, and a nyuyurlpa becomes then an actual ngunytju, mother. But the word nyuyurlpa denotes in fact “the one that adds timber to the fire”, that is, the one that creates the conditions for eating together food cooked on the same fireplace, and warming oneself at night around the same heat. Adoption is a process where difference becomes identity and sameness. But more, the value of ngunytju, mother, is one that describes the conditions for becoming same; and these conditions have not much to do with kinship as such.

14Similarly, as Fred Myers (1988:17; see also Hansen & Hansen 1992) has explained, the word walytja, among Pintupi people, just north of the Ngaatjatjarra, which means vaguely family, or those with whom one identifies a connection and reference of similarity and inclusion, also means “my belongings”, my luggage, so to say, what I own and take with me. Thus, to own something is not that different from having relationships. Belonging to, and ownership of, are of the same culturally identified semantic entity.

15And this goes all hand in hand also with the humanization of space, where people can be called and referred to using the geographical and mythological site to which they most closely affiliate to; and, vice versa, where these sites can be referred to using the person's name. But more, where the tracks between these sites in space have the same connotation as the lines that draw relationships between people in our genealogical charts.

16Hence, yes, kinship terminology is associated to generic and typified behavioral expectations and norms depending on which category people sit in. However, the opposite is true as well: experiences, life experiences, can produce the conditions in which the usage of terminology is defined by other means than the categorical position in which the relationship stands. Experience creates kinship; a kinship that looks as formal as the one we deduce from category and terminology alone.

17The second terminological usage, which I refer to as “sociological”, goes even further. In certain contexts, when generational opposition is the core of the discussion, actual uncles and aunts can be called father and mother without hesitation or confusion. In certain ritual contexts, where generational moieties play an important role and distribute people in space and function, all people of the opposite generation, and not just uncles and aunts, may be called father and mother, which work as cover terms. The sociological context is one in which generation and gender are the principal, if not sole factors creating distinctions. In these ritual conditions, the conflictual opposition but ontological complementarity of the parent-child couple is extended to all participants. They are the contexts in which people place themselves as actors within the reproduction of the cosmological order. Actual genealogical relationships here lose their significance.

18The third terminological set sheds some light on the previously mentioned usages; at least from a materialist point of view. The other usages and their contexts underline the question of similarity and identity, of commonality. The third context, which I call “egological”, on the other hand, redistributes people in accordance to the mathematics Lévi-Strauss and others expected. It is at stake when interlocutors are discussing actual marriage of an actual person. It is about setting up alliance strategies between individuals and families. Here, distant aunts are aunts, not mothers; and distant uncles are uncles, not fathers: who knows, they could become actual in-laws. It is the potentiality of future affinal relationships that places people in the category in which they structurally already sit; they are not “cut out”, to use Bob's expression again. Obviously, placing people in the affinal category defines who can become a spouse or a parent-in-law.

19The reasons for cutting out ? or not cutting out ? for using rather one or the other terminological set? Obvious, at least for those that have worked in the Western Desert. It is about network, about inclusion, about extending and creating new nodes of connectedness; as in opposition to the idea of bounded society and exclusion, of course. It is about what Ian Keen calls “shifting webs” (2002).

20Marriage, so closely tied to terminology in this context, is the main element under discussion. Marriage and all the obligations and rights that come with it, with all the networking that is involved. And Ngaatjatjarra people are clear about how this should be done: the initiator, who must be someone from a distant community, promises his or a close daughter; and in return the initiator's son will marry the initiate's sister. Direct exchange as in the textbook. But Ngaatjatjarra people also have another set of rules and discourses about these rules: you should not marry someone you are already linked with, but you should marry somewhere else. No need to marry a man who is already your actual brother-in-law… it wouldn't be the right way, a woman once told me. Thus, hardly anyone marries his promised spouse, but rather finds yet another partner, from another distant location, while both, his promised and his actual spouse being, with their families, tied through comparable obligations and rights. Diversification is the motto: diversification of people and of access to resources.

21We are in front of what seems to be one of those usual contradictions between discourse and practice; — between a structural repetition of identical marriages through direct exchange, which never happens and remains in the domain of discourse and meta-language; and actual practice where people tend to diversify as much as they possibly can the nodes of their embodiment in the geographical and social landscape. The opposition between discourse and practice, at least in this context but possibly beyond, is however not a fortunate representation of what is happening.

22There are two series of ideals and discourses that cohabit, in fact. One on initiation and direct exchange where closure and confinement, repetition and confirmation are central values. Mythology and ritual are fundamental here. The second series, on the other hand, is about diversification, openness and network. Here we talk about politics and economy. The distinction between these rather arbitrarily separated domains becomes apparent when observing people's actual decision-taking processes. Shall one call a man father or close uncle, or shall one call him distant uncle and father-in-law. The decision has its consequences, as we now know. And it is elaborated, discussed and finally promulgated after taking into account the normative capacity of a man or a woman to become an actual in-law, of course, but more interestingly an explicit discussion about the intention and the possibility of getting involved in an affinal relationship with enumerated benefits: with cash economy, crystallizing people's attentions to particular “wealthy” families and communities has an impact on which terminological set should be applied. There is however no reason to believe that these principles were not at work in former days. There are indeed a few examples of religiously and politically important men who, before contact with the Western world in the late 1950s, were highly polygynous (more than two wives in an area were the incidence of polygyny is very low) and who were referred to by numerous persons as waputju, father-in-law, rather than kamuru, uncle.

23I would like to recall what is my strong conviction. Opposing or differentiating these phenomena as a contradiction between discourse and practice is not the right approach. Western Desert groups are well known for their inclusivist ethos, as Tonkinson (2003) has remarked; that is, their capacity to underline connections with outsiders rather than stressing what differentiates them. Kinship terminologies or emic explanations of rules about marriage are not discursive devices meant to implement particular practices, but are devices that speak about values: in our case, these values reflect an accomplished inclusivist ethos in which the repetition of relationships stands for the exclusion of others; in which the reproduction of sameness is formulated through the repetition of relationships as forms. The existence of a parallel discourse on the prohibition of incest through diversification (as opposed to repetition), reflects these same values, this time however not as accomplished, but as a processual inclusivist ethos.

24One could go further, and ask, how much of this is again a particularity of the Western Desert? Is this again an aberration? In the AustKin project in which I am involved with linguists from the ANU and in which Ian Keen is involved as well, this is the kinds of questions we can ask. Where and how much are there discursive devices (that is terminological elements, for example), that point to openness and closeness at the same time. That testify of reproduction of sameness and shifting webs simultaneously. While this project is yet not at its end, there is already evidence testifying that the coexistence of these two value-systems is far from being limited to the Western Desert alone, including extensive regions in Queensland and south-east Arnhem Land.

25One could postulate the continuous tension between inclusion and exclusion, and between openness and boundedness, as signifying the notion of society (the thing that attracts an intention of belonging), which, as Godelier (2009) explains, is not coextensive with that of culture. There are at least two approaches depicting the idea of society based on these tensions. One is to reflect emic categories in etic typologies; an approach not that distant from the usual anthropological modus of investigation. The other has been designed as an actor-network theory that does not assume ontological entities.

26Following the first approach a society is something that emically needs to erect three fundamental pillars of value-systems in order to reproduce itself in time and space: place, memory, and modes of recruitment and belonging. According to the second approach, on other hand, society is not a thing, but is some vague result of the extension and accumulation of relationships in which values, and thus discourses themselves, are simultaneously means of action and actors: cultural objects (Latour 1987).

27Western Desert ethnography seems to be best depicted as a subtle combination of these two approaches, as I have tried to illustrate. In the light of current Native Title exigencies, however, where the continuous reproduction of bounded entities is expected, this poses obviously significant problems. How establish boundaries, be they social or spatial, when one important value defines the exact opposite: extending these boundaries, or in fact not being limited by them? More: the procedure of going through the Native Title process and other types of negotiations with the State, as I have observed it among Ngaatjatjarra people, has had significant impacts, shifting the local value-system from an open processual inclusivist ethos progressively to an accomplished and thus closed inclusivist ethos. People think of themselves increasingly as being “tribal”: with boundaries and enumerated membership. To take just one example. Ngaatjatjarra, the name of the group itself, was formerly designating a geographical area in which a particular mode of speech was put onto country by mythological beings. People that lived in this area thus were Ngaatjatjarra people because they lived in a country to which the Ngaatjatjarra dialect was associated. A person that would move out of this country would adopt the local mode of speech, and would not unambiguously be called a Ngaatjatjarra person, because the language stayed behind, in its original country. Nowadays, however, people use increasingly the word Ngaatjatjarra as a nearly genealogical marker. People take their dialectal name with them, whether they live in Ngaatjatjarra country, Alice Springs or elsewhere. The distinction between belonging and otherness, between insiders and outsiders, has in the past 18 years during which I have visited Ngaatjatjarra people, increasingly become an explicated and explicit means of talking about oneself.

28My paper is entitled from structure to substance and back, and my intention was to illustrate modes and means through which kinship is the condition, but also the product of the material aspects of these mathematics Lévi-Strauss has evacuated, or in fact could not have taken into account because of his preoccupation with the deep structures. He was correct assuming that Australian kinship is the product of conscious reflections. In my experience, the capacity of Aboriginal people to enumerate, discuss and contextualize rules and systems is indeed phenomenal and testifies of a relativity, a cultural relativity, towards their own systems of norms and values. The application of kinship terminology does not flow from rules alone, but is conditioned by the context in which the terminology produces material and substantial effects on relationships. Kinship is not an independent social domain, and as such, cannot be investigated without placing it back into the context in which it is continually produced. We knew this. What appears interesting, at least to me, is that it takes the form of reflected and consciously elaborated systems.


DOUSSET, Laurent 2002. Accounting for context and substance: the Australian Western Desert kinship system, Anthropological Forum, 12(2): 193-204.

DOUSSET, Laurent 2003. Indigenous modes of representing social relationships: A short critique of the “genealogical concept”, Aboriginal Studies, 2003/1: 19-29.

DOUSSET, Laurent 2012. “Horizontal” and “vertical” skewing: similar objectives, two solutions?. In Thomas R. Trautmann and Peter M. Whiteley (eds), Crow-Omaha: New Light on a Classic Problem of Kinship analysis. Tucson: Arizona University Press, p. 261-277.

GODELIER, Maurice 2009. Communauté, Société, Culture: Trois clefs pour comprendre les identités en conflits (Huxley Memorial Lecture 2008). Paris: CNRS Editions.

HANSEN K.C. & HANSEN L.E., 1992. Pintupi/Luritja Dictionary (3rd Edition). Alice Springs: IAD (Institute for Aboriginal Development) [SIL, Summer Institute of Linguistics [1974].

KEEN, Ian 2002. Seven Aboriginal marriage systems and their correlates, Anthropological Forum, 12(2): 145-157.

KRONENFELD, David 1973. Fanty Kinship. The Structure of Terminology and Behavior, American Anthropologist, 75(5): 1577-1595.

LATOUR, Bruno 1987. Science in Action: How to Follow Scientists and Engineers through Society. Cambridge: Harvard University Press.

LEVI-STRAUSS, Claude 1967. Les structures élémentaires de la parenté. Paris: Mouton [1947].

LEVI-STRAUSS, Claude 1968. Gidjingali marriage arrangements: comments and rejoinder. Discussions Part IV, 22b. In R.B. Lee & I. De Vore (eds), Man the Hunter. Chicago: Aldine, p. 210-211.

LEVI-STRAUSS, Claude 1987. Race et histoire. Paris: Denoël Folios essais [1952].

LEVI-STRAUSS, Claude 1996. Anthropologie Structurale II. Paris: Plon [1973].

MYERS, Fred 1988. Burning the truck and holding the country: property, time and the negotiation identity among Pintupi Aborigines. In T. Ingold, D. Riches & J. Woodburn (eds), Hunters & Gatherers 2: Property, power and ideology. New York: St Martin's Press, p. 15-42.

RUMSEY, Alan 1981. Kinship and context among the Ngarinyin, Oceania, 51(3): 181-192.

TJON SIE FAT, Franklin 1998. On the Formal Analysis of “Dravidian,” “Iroquois,” and “Generational” Varieties as Nearly Associative Combinations. In M. Godelier, T.R. Trautmann & F.E. Tjon Sie Fat (eds), Transformations of kinship. Washington & London: Smithsonian Institution, p. 59-93.

TONKINSON, Robert 1975. The Riddle of the Non-Marriageable Cross-Cousin, : ms: Séminaire donné à l'ANU (Australian National University), Canberra.

TONKINSON, Robert 2003. Ambrymese dreams and the Mardu Dreaming. In R. Lohmann (ed.), Dream Travelers: sleep experiences and culture in the Western Pacific. New York: Palgrave Macmillan, p. 87-105.

Pour citer cet article

Référence électronique

Laurent Dousset, « From structure to substance and back: materialities in Australian Aboriginal kinship », in Australian Aboriginal Anthropology Today: Critical Perspectives from Europe (« Les actes »), 2014, [En ligne], mis en ligne le 13 juin 2014, Consulté le 24 mars 2017. URL :

Droits d’auteur

© Tous droits réservés