Version classiqueVersion mobile

Australian Aboriginal Kinship

Laurent Dousset

Part three: Western Desert kinship ethnography

Texte intégral

Western Desert kinship terminology

1Western Desert terminology (also known as Aluridja terminology, since Elkin’s descriptions in the 1930s) has raised issues that are significant for anthropologists. Franklin Tjon Sie Fat notes that:

Some of the most intriguing descriptions of anomalous or inconsistent terminological systems combining a variety of Dravidian, Iroquois, and ‘Hawaiian’ or generational features with a range of affinal terms and extensive marriage prohibitions pertain to the Western Desert peoples of Australia (Tjon Sie Fat, 1998:78).

2Indeed, since the publication of A.P. Elkin’s Kinship in South Australia (1938-40), Aluridja systems have been characterised by their inability to distinguish siblings (brothers and sisters) from cross-cousins, which is very unusual for Australia, the allowing of marriage between parallel cousins, and the absence of sections, subsections and exogamous moieties (see Part 4 below). To take the most debated example, Lévi-Strauss (1967 [1947]:231, 251 and Figure 56, page 249) had obvious problems with the incompatibility between the Aluridja and other Australian systems, so he labelled this system ‘aberrant’. Another example is Scheffler’s (1978) work, also an important contribution to our understanding of Australian kinship systems and semantics. Scheffler discusses the Pitjantjatjara terminology — considered an Aluridja type — as consisting of characteristics similar to those mentioned by Elkin. Indeed, while he confirms that a MB marries a FZ and that the cross-parallel distinction is therefore introduced in Ego’s parents’ generation, he nevertheless gives as the English gloss for the term watjirra (a cross-cousin) the expression ‘distant sibling’.

3Testart (1996), a French scholar, believes that the Aluridja system is guided only by an elementary incest prohibition, thus lacking all the systematics that theorists writing on Aboriginal Australia have glorified. The absence of what he calls objective or totemic patrilineal or matrilineal groups (1996:300) that could exchange spouses in accordance with some principle of exogamy is interpreted as the reason for the total absence of marriage rules.

4Recently, both Tjon Sie Fat (1998) and Trautmann and Barnes (1998) have briefly tackled Aluridja kinship. While Tjon Sie Fat recalls the problems that this system poses for formal analysis, Trautmann and Barnes categorise it as a variation of what they call the type B crossness, while we would generally expect for Australia what they call a type A crossness (Dravidian).

5These problems have found their way into the literature because Elkin’s original ethnography was flawed and because he did not recognise some important differences that exist in Western Desert culture: notably, that which exists between sociological and egocentric contexts for the use of kinship terms (see Dousset 2003 and 2005).

6Rather than reproduce Elkin’s errors, I shall now describe how Western Desert people use kinship words in particular contexts. The first, which I call ‘egological’, is the classic approach to terminologies. This terminology is based on an Ego who addresses or refers to kin people around him or her. The second context, which I will describe below, is one in which general social distinctions are the focus of discussion and practice.

  • i A situation in which personal interrelationships determine which terminology or behaviour to adopt (...)

7In the egological contexti, Western Desert people have a terminology that distinguishes cross from parallel cousins, but they also have specific affinal terms for potential or actual in-laws. The following figure present these terms. Note that these are the terms used in the central parts of the Western Desert, among Ngaatjatjarra, Ngaanyatjarra, Pintupi and to some extent Pitjantjatjara people. For the eastern terms, you may want to check Goddard (1985 and 1992) for example, or for the western terms, Tonkinson (1991).

Figure 20: Western Desert terminology: Ego’s and Ego’s parents’ generations

Figure 20: Western Desert terminology: Ego’s and Ego’s parents’ generations

8As expected, the terminological system is of the Dravidian type. Unlike the Eskimo (European) system, MB is distinguished from FB, and FZ from MZ. Bifurcation takes place in G+1. Moreover, FZ’s and MB’s children are distinguished from brothers and sisters, who are in the same category as parallel cousins. The figure is not complete since there is a further distinction between brothers and sisters. Older brothers are called kurta and older sisters are called tjurtu, but both younger brothers and sisters are all called marlany or marlanypa, irrespective of their gender. Additionally, the word nyarrumpa designates all siblings and parallel cousins of opposite gender to the speaker, irrespective of his or her own gender.

9Another interesting distinction is birth order between siblings. To indicate birth order, people use the following expression, in which A is the oldest and D the youngest sibling:

A-lu pakarnu pirnirnu B-nya.

A has grown with B on his/her shoulders.

B-lu pakarnu pirnirnu C-nya.

B has grown with C on his/her shoulders.

C-lulta pakarnu marlankatjanya pirnirnu.

Finally, C has grown with D on his/her shoulders

(marlankatjanya is the last one, and thus stands for D).

10Moreover, among parallel cousins, birth order does not follow chronological age but the age of their parents. For example, all children of ones mother’s younger sister are called marlanypa (younger sibling), irrespective of whether they are older or younger, and all children of ones mother’s elder sister are called kurta and tjurtu, irrespective of whether they are older or younger. It is the relative age of the parents that is taken into account. Similarly, a younger brother of ones father may be called mama marlankatja, a ‘father who comes behind’. This expression is also used for the second husband of ones mother. Ngunytju marlankatja is the equivalent for mother’s younger sisters or for the father’s second wife.

11As in all Dravidian systems, the mother’s brother (kamuru) is in the same category as ones father-in-law (waputju, see below when I discuss the affinal terminology), and the father’s sister (kurntili) is in the same category as ones mother-in-law (yumari, see below). However, ones actual MB or FZ are never potential in-laws. To distinguish those kamuru that can become waputju from those that cannot because they are genealogically too close and to distinguish those kurntili that can become yumari from those that cannot because they are also genealogically too close, the adjective tjiwangkatja is added. Kamuru tjiwangkatja and kurntili tjiwangkatja mean a ‘distant uncle’ and a ‘distant aunt’ respectively.

12If they are close, and even more so if they are Ego’s actual MB and FZ, kamuru and kurntili have considerable responsibility for Ego’s education. Along with the parents, an aunt and an uncle will have a say in the ways a child should be educated, in the choice of an adoptive family should this happen, in the planning of a boy’s initiation and a child’s marriage in general. In many cases, I have observed that the relationship between a child and their uncle and aunt is closer than that with their parents, towards whom the child must show respect while maintaining a measure of physical and emotional distance.

13Watjirra, cross-cousin, also requires further discussion since its use is different from one dialectal group to another, even within the central parts of the desert, and because there are other words to denote persons of the same category (some of which are further explained below). The following table summarises the words and usages for cross-cousins:






Used in the northern part of the central Western Desert


Like watjirra

Alternative word used in the northern part of the Western Desert


Like watjirra

Word used in the western part of the Western Desert


Like watjirra

Word used around the Rawlinson Ranges


eB, eFZS, eMBS

In the southern part of the Western Desert, people call cross-cousins by the word for ‘brother’ before marriage. This changes after marriage when the in-law (affinal) terminology is used.


eZ, eFZD, eMBD

In the southern part of the Western Desert, people call cross-cousins by the word for ‘sister’ before marriage. This changes after marriage when the in-law (affinal) terminology is used.


yB, yFZS, yMBS, yZ,yFZD, yMBD

In the southern part of the Western Desert, people call cross-cousins by the word for ‘younger sister/brother’ before marriage. This changes after marriage when the in-law (affinal) terminology is used.



Descriptive term, actual brother-in-law


Like marutju

Descriptive term, actual brother-in-law


Like marutju

Descriptive term, actual brother-in-law



Brother’s wife, used only once brothers are married.


W, H, mFZD, mMBD,wFZS, wMBS, mFM,wMF, mDD, wDS

Husband and wife in general terms. Can be used for all opposite sex cross-cousins and for some persons of the grandparent and grandchildren generations as well. If the person is not ones actual spouse, it considered somewhat insulting if not used in a joking relationship.


Like kurri

Generally meaning companion, lover (see kurri)


Like kurri

Respectful towards a person who is a potential spouse but who is already married.

14Conversely to what Elkin and his successors claimed, the Aluridja system distinguishes cross-cousins from brothers, sisters and parallel cousins. In the southern and eastern parts of the central desert, cross-cousins are in some instances addressed as siblings, but this is only the case so long as marriage is not envisaged; that is, as long as these cross-cousins are not of marriageable age. The underlying structure, terminology and way of classifying people clearly follow the bifurcate merging pattern of Dravidian terminologies. We now examine the terminology used for the remaining generations.

Figure 21: Western Desert terminology:Ego’s grandparents’, children’s and grandchildren’sgenerations

Figure 21: Western Desert terminology:Ego’s grandparents’, children’s and grandchildren’sgenerations

15Most notable here is that in the generations of grandparents and grandchildren, only gender is distinguished: kaparli for females and tjamu for males. Note also that these terms are self-reciprocal if the interlocutors are of the same gender. A female Ego will call her grandmothers and her granddaughters kaparli, and they will call her kaparli in return. The same is true for tjamu. Not all groups have self-reciprocal terms in these generations, however. The Pitjantjatjara and Yankunytjatjara people, for example, call their grandfathers tjamu while they call their grandsons pakali. The grandmothers are called kami and the granddaughters puliri (Goddard 1985: 153).

16The generations G+2 (grandparents) and G-2 (grandchildren) are thus not Dravidian but Hawaiian or generational. They do not introduce any distinctions between cross and parallel, and the terminology does not reflect any particular lines of descent. Irrespective of whether they are on the father’s or on the mother’s side, the relationship between grandparents and grandchildren is one of equity and equality, of reciprocity and closeness.

17The grandparent and grandchildren terminologies are also the vertical limits of the terminology. There is no equivalent to something like ‘great-grandmother’. In fact, the terminology of the Western Desert is cyclical. The son of a son’s son (SSS) is again classified as a father and the father of a father’s father (FFF) is classified as a son. I have encountered a case where an old man called a two-year old boy kamuru (MB). The boy was his actual WZSDS (wife’s sister’s son’s daughter’s son) and thus a classificatory SDS (son’s daughter’s son). A son’s daughter’s son, for whom there is no term, is a MB according to a cyclical Dravidian type of calculation. It was with some amusement that the old man claimed that, according to the rules, this boy will have to give him his daughter as a wife.

Figure 22: Western Desert terminology: affinal terms (left: male Ego; right : female Ego)

Figure 22: Western Desert terminology: affinal terms (left: male Ego; right : female Ego)

18Affinal terms are used from the time that Ego’s marriage is envisaged, discussed and negotiated or has actually occurred. Men are able to envisage marriage once they have completed the first few stages of initiation, when they become wati (adult men as opposed to boys). Women are marriageable after they have had their first menstruation and their breasts have started to grow. Girls will also have to proceed through various steps of ceremonial initiations before becoming senior women.

19During and after male initiation, the main initiator has a very important role. I return to marriage later, but note for now that the word used to designate this initiator is waputju, the same as the word for father-in-law, and in fact, the initiator typically promises one of his daughters, actual or classificatory, to the young man as a spouse in response to his having ritually ‘killed’ the boy, who is later symbolically ‘reborn’ as a man. Waputju is married to yumari, the mother-in-law. Both are cross relatives: a waputju is a MB (kamuru), and a yumari is a FZ (kurntili) who are of some genealogical and spatial distance apart. Before marriage was envisaged, the boy and the girl would call their ‘future’ parents-inlaw ‘aunt’ and ‘uncle’.

20Waputju is again self-reciprocal, since it is applicable to the father-in-law and the son-inlaw. However yumari, who is a person Ego needs to avoid completely and who can only refer to her and never address her, is only self-reciprocal for a male Ego. A female Ego will refer to her daughter-in-law as mingkayi.

21Other interesting terms are marutju and tjuwari. The former designates a man’s sister’s husband, as well as a man’s wife’s brother, and corresponds well to the English understanding of brother-in-law. The same is true for tjuwari: it designates a woman’s brother’s wife and a woman’s husband’s sister. All cross-cousins are allowed to have a relaxed and close relationship, joking and sharing without restraint, even when they are of different gender. This is even more so the case between men who call each other marutju and females who call each other tjuwari, in particular if they are actual brothers and sisters-in-law.

  • ii A situation in which the relationship between general social categories determines the terminology (...)

22Having considered the egological context, we now turn to the sociological contextii, in which the terminology is much more limited. Such sociological contexts are, for example, ceremonies where each generation has particular tasks or a context in which people discuss general rules and relationships rather than individual relationships. In these cases, it is possible for Western Desert people to temporarily ignore the underlying Dravidian system and to switch to a Hawaiian system. We have seen that the Hawaiian system is already in use in the grandparent and grandchildren generations. In these sociological contexts, this principle is extrapolated to all generations so that all females of the parents’ generation are called ngunytju (M), irrespective of whether they are FZ or M, and all males of the parents’ generation are called mama (F), irrespective of whether they are F or MB. The same is true for Ego’s generation, where all females are called sister (tjurtu) and all males are called brother (kurta). Temporarily calling a cross-cousin ‘brother’ does not however transform this man permanently into a brother or parallel cousin. It is usually only a temporary and contextual label. There are, though, particular situations in which this transformation is more permanent; for example, in the context of marriage, to which we now turn.

Finding a spouse

23As explained in Part 2, there are proscriptive and prescriptive marriage rules. Every society has proscriptive rules: you cannot marry so-and-so because it would be considered incestuous. Most people also know prescriptive marriage rules that flow from either obvious and explicit norms and discourses or from the terminology’s systemic properties. Calling parallel cousins ‘brother’ and ‘sister’ is such a systemic property, since it extends the proscriptive rule (the prohibition to marry actual brothers and sisters) to other, genealogically distant or even unrelated persons.

24In the Western Desert, the proscriptive rule is not limited to a genealogical factor alone but is extended to include a spatial factor as well. In other words, incestuous marriages are considered to be those involving a category of persons other than cross-cousins but also those including persons who are genealogically and geographically close. Very rarely are actual cross-cousins allowed to marry, and only very exceptionally may spouses be affiliated to identical territorial sites in the landscape. There is much variation in relation to these rules across the Western Desert, so we concentrate here on two similar cases: the Ngaatjatjarra and the Ngaanyatjarra dialectal groups. With both, there are explicit, somewhat strict rules that bring together a combination of several proscriptive and prescriptive elements.

  1. The first rule concerns the formal aspects of the kinship system. You may not marry someone who is a brother or sister, or who is in an adjacent generation (e.g., mother, daughter, uncle, son) to you. Your spouse must therefore be a cross-cousin or a grandchild or grandparent who is of the cross category. A FM or a DD for a male Ego, and a MF or a SS for a female Ego, can be considered acceptable as spouses on occasion, but such relatives would be reclassified as cross-cousins, which is perfectly compatible with the Dravidian system and does not alter other relationships.

  2. The second rule concerns genealogical distance. You may not marry someone genealogically close or anyone for whom you can identify known, named identical ancestors. Put in prescriptive terms, you should marry someone at least three generations removed (third cross-cousins). In other words, you are encouraged to conform to a strong genealogical exogamy.

  3. The third rule concerns spatial distance. You may not marry someone with whom you share too much spatial identity; that is those conceived or born on the same site, or who have lived for prolonged periods of time in the same area or community as you. Put in prescriptive terms, you should conform to this strong spatial exogamy by finding your spouse in a distant and different place. The rule of spatial and social distance is so important that, in cases recorded by Tonkinson (1975) among the Mardu and by me among the Ngaatjatjarra, some persons who potentially fit into the spouse category, being cross-cousins, are renamed brothers and sisters. This can happen with people who are either geographically close or who should not become affines for a range of other emotional, social, political or economic reasons. Through the ‘reclassification’ of these persons as siblings, the possibility of a marriage is annulled since it would now be considered incestuous.

25A marriage that ignores these rules is yinyurrpa, which is considered incestuous. Interestingly, this word also means ‘the other generational moiety’, a social grouping discussed below (Part 4) which includes people of adjacent generations (those of parents and children), with whom marriage and sexual relationships are forbidden. Applying all of these rules rigorously has nowadays become difficult. While the formal rule and the rule of genealogical distance are frequently adhered to, that of geographic exogamy has become increasingly difficult to maintain in large, settled communities, where intracommunity marriages are now common and accepted. The ideal of an extreme geographical exogamy, also called ‘local exogamy’ in anthropology, is nevertheless still strong in people’s discourses and alliance strategies. The various reasons for this include the influence of exogamy on the creation of wide-ranging networks of solidarity and exchange.

26How then does a person find a spouse? Here again there are several mechanisms or institutions that work in parallel (Dousset 1999): the ceremonial promise (pikarta), the non-ceremonial promise (karlkurnu) and the elopement (warnngirnu). Before going into the details of each of these possibilities, it is important to note that actual marriage patterns do not reflect what alliance theorists would expect to be the case. You may recall that in Part 2 there was a discussion of one of the marriage systems theorised by Lévi-Strauss (1967 [1947]), which he called ‘direct’ or ‘symmetrical’ exchange. This type of marriage was allegedly found in Dravidian terminological systems where a mother’s brother is also the father-in-law and where the father’s sister is a mother-in-law. I have repeatedly underlined the necessity to distinguish genealogical from terminological aspects: it is not because one may call a mother-in-law father’s sister that this mother-in-law is ones actual father’s sister. She may well be someone who stands in the same category but who is genealogically, socially and spatially removed. The Ngaatjatjarra marriage patterns confirm this hypothesis. Among the 282 marriages I analysed with respect to this question, only eight can be considered direct exchanges and only four of these reflect the ideal form of sister exchange between two future brothers-in-law (the others being a different form of exchange involving close parallel cousins). If direct exchange is rarely followed in practice, even though people may explain marriage as a direct exchange between two brothers, this is because other values are followed. I have already, but indirectly, alluded to these several times; for example, the necessity for people to increase their capacity to rely on the resources of others in difficult times. The archaeologist Richard Gould (1969: 273) suggests that

Compared to the Kalahari people, the Western Desert people must subsist on significantly less water, game, and plant foods. [...] Western Desert Aborigines have managed to survive in the harshest physical environment on earth ever inhabited by man before the Industrial Revolution.

27And Tonkinson (1991: 15) noted that, ‘Coping with the desert is hard work, especially during the seven or eight months of the hot season. You soon develop immense admiration and respect for the Aborigines who have conquered it and made it their own’.

28With peaks as high as 45 degrees Celsius during the day in summer, the temperature can drop at night to below freezing in winter. Rainfall is scarce throughout the year, with a slight increase during the winter months. Riverbeds are dry most of the time and are filled with water for only a few days after heavy rain since it soon disappears into the sand. The demographic density is extremely low but, as noted earlier, a key factor is not simply the scarcity of rainfall but its unpredictability. It may thus rain abundantly in one localised area while just a few kilometres away no rain has fallen for many years. Managing this uncertainty has led Western Desert people to develop a dynamic and flexible system of landownership. However, it was also vital to create and reproduce exchanges, partnerships, a sense of closeness and obligations of reciprocity with other families and groups living in neighbouring areas. Being in a close relationship with other people, in particular through the institution of marriage, provides a sense of security in the event of hardship and drought. In times of drought, it allowed people to visit and exploit resources located outside their traditional foraging area. There is no other way human society could have survived and thrived in this environment so long ago.

29Although direct exchange can be considered a formal prescription and is explained by people (particularly men) as the most appropriate way to marry, as an adaptive strategy it does not conform to ecological and economic imperatives. A far more adaptive system of marriage would be one in which each individual, family and group diversifies its network of relationships. Keen (2002) calls these multiple links ‘shifting webs’. Indeed, if two men ‘exchange’ sisters, they have only each other to rely on during times of hardship. However, if one of them gives his sister in marriage to another man and then marries the sister of a third man, he enlarges his network. The strategic goal here is the diversification of alliance.

30There is also a parallel here with the process of obtaining a first wife for a young man during initiation. When a boy is old enough to be initiated, any time after the age of 15, his immediate family, and in particular his actual aunt (FZ) and uncle (MB), discuss and negotiate the place and time of the ceremony amongst themselves and with potential host families and communities. Initiation rarely occurs in the boy’s own residential community but ideally takes place among other, distant families. The central role is that of waputju, his initiator and future, or at least potential, father-in-law. His waputju will be chosen from a distant family, if possible one with whom the boy’s family has yet few or no sustained relationships. Indeed, as we have seen, the waputju must promise the boy his daughter, or a close ‘daughter’, in marriage. This girl will be called pikarta, ‘the one obtained through pain’. At the same time, however, a potential mother-in-law, called pampurlpa during the ceremony, will promise her daughter or one of her close ‘daughters’. Pampurlpa is usually not waputju’s wife, but another classificatory aunt (FZ). After the initiation ceremony, the boy will thus already have two potential wives who are usually from distinct geographical (and sometimes cultural) communities. His network has already started to expand.

31Here, we need to go further: the boy will only rarely marry one of these two girls but will become involved with yet another woman and her family through other social processes. Even though he does not marry the promised girls, the promise itself has already established an affinal relationship on which he can count. Additionally, all the boys who are initiated simultaneously constitute a group called ngalunku and henceforth become close friends and allies, bound by mutual friendship and obligations of care and reciprocity.

32Most marriages are indeed consequences of either karlkurnu or warnngirnu. The first kind is a promise between a potential son-in-law and his potential parents-in-law. In former times, when people where still nomadic hunter-gatherers, the man would follow his family-in-law and hunt and provide food and gifts for them until the parents-in-law agreed to give him their daughter. Today, offerings of money, a car or other material goods are expected by the girl’s parents.

33The second form, warnngirnu, is elopement. If two young people fall in love but both have already been promised to others in marriage or if parents reject the idea of marriage for other reasons, the young couple may simply run away. The elopement is sometimes ritualised. The family and community pretend to be following the couple in order to bring them home and separate them, but the pursuers make sure that the couple escape, thus acknowledging the marriage. In cases where there is anger and the marriage is vehemently opposed, the hunting party will do its utmost to stop the elopement. To avoid this, the young couple will travel to a distant area and return only after they have a child, which in most cases will ratify the marriage. Whichever situation, karlkurnu or warnngirnu, has led to the marriage, it is evident that each man has at least three potential or actual wives. He has established relationships with at least three families: two promises during initiation and one through actual marriage outside the ceremonial context. In addition, while still relatively young, he has enlarged his network through strong relationships with his co-initiates.

On the family and related issues

34As we know, marriage is also about establishing a family, not just a network. In the Western Desert, the words walytja or yungarra marri are those that seem to equate with the English word ‘family’. However, we need to be cautious because ‘family’ has a culturally specific meaning and is the result of hundreds of years of social evolution and change, as Goody (1999) has shown.

35The historic and social background of this word also implies a kinship system, a way of recognising ‘who or what’ is part of this unit. The fact that by the term ‘family’ Europeans understand a limited group of people is directly linked to the ‘Eskimo’ kinship terminology and to the limitation of the extension of range in such systems. It may therefore be appropriate to translate the English ‘family’ into the French ‘famille’, the Italian ‘familia’, or the German ‘Familie’, because these terms have a similar historic origin and mean something comparable, if not identical, in all these languages: a wife, a husband, their children; included are sometimes grandparents and grandchildren, brothers and sisters of the parents and their children, even though they may not in all cases, or even in many cases, actually live together. In the English meaning of ‘family’, the idea of ‘common blood’ or consanguinity is implied: children are of the same blood as or share some genetic material with their parents.

  • iii A residential unit. People who live together in the same abode or camp.
  • iv An economic unit. The smallest unit of people that produce and consume material goods in common.

36Another idea usually associated with the notion of family is the householdiii. A family usually lives as a united group in a distinguishable shelter: they are co-residents. In anthropology, however, the (nuclear) ‘family’, as a group constituted by marriage and offspring, is sharply distinguished from the notion of household — those who live together — as well as from the domestic groupiv. The latter defines the smallest economic unit of production and consumption. These distinctions are necessary because not all societies equate family, household and domestic group. In the Euro-American families of today, grandparents do not often live in the same household as the parents and their children and do not always take part in the household’s production and consumption activities but are nevertheless considered to be part of the family. Another example is the cohabitation of students in one apartment or house. They constitute a household and may even be a domestic group because they perhaps share their earnings and goods, but they are not considered a family in the traditional European sense.

37Before illustrating in more detail why the word ‘family’ cannot be directly applied in the Western Desert or in Aboriginal Australia in general, we must first take a small historical detour to see how the notion has been applied and understood by anthropologists.

38The first definition of ‘family’ in a systematic and serious anthropological work appeared in one of the first books ever published in ‘modern’ anthropology and already proposed a dynamic and cross-culturally informed understanding of the ‘family’. You may recall my earlier mention of Morgan’s Systems of Consanguinity and Affinity of the Human Family (1871) when discussing kinship terminologies, as well as of his Ancient Society (1877). Many anthropologists consider these two books to mark the birth of anthropology as a science in its own right, as well as the starting point of systematic kinship studies in human cultures. Morgan compared what he called ‘systems of relationship’ — what anthropologists call ‘kinship systems’ — and investigated what he termed the ‘forms of consanguinity and affinity’ of various cultures. From his own journeys among American Indian people and from correspondents all over the world, he obtained data on kinship terms, marriage rules and other elements of social organisation.

39Comparing these various cultural features, he elaborated a picture of the evolution of the human family. We know today that his evolutionary ideas, which were a prevalent ideological background of academic discourse in his period, are largely erroneous. Nevertheless, we owe to Morgan many of the most important concepts used in kinship studies and, more importantly, the ability to conceive of the ‘family’ as a concept that could change meaning in history and in cultural space. He also understood, as the title of his first book clearly indicates, that words such as ‘consanguinity’ and ‘affinity’ do not have absolute meanings but that their use and importance depend on the culture studied. That the family comprises people who share a common blood is more of a ‘cultural’ or ‘social’ statement than a truly biological one. Proof of Morgan’s understanding of ‘family’ as a dynamic feature of human societies lies in some of the titles and phrasing used in his book. The third part of his Ancient Society, for example, has as its title The development of the idea of family.

  • v A kind of marriage involving one woman or man and her or his spouse. A monogamous family comprises (...)
  • vi A marriage between a group of brothers and a group of sisters. This mistaken notion was used by evo (...)
  • vii Having several simultaneous spouses.
  • viii A form of polygamy. A woman who is married to more than one husband at the same time.

40Nevertheless, at times Morgan could not detach himself from what he believed to be the ‘end product’ of human social evolution: the Europeantype monogamousv family comprising a man, a woman and their children. He was heavily influenced by his own ethnocentrism, that is an inability to accept cultural constructions other than his own as ‘natural’. He understood the various appearances of the family in different human cultures to represent different stages in a unique social evolution, one that he described as starting from a culture where promiscuous intercourse or group marriagevi was allowed and developing towards the monogamous family of European societies through various stages such as the Hawaiian customs, barbarian phases, polygamyvii and polyandryviii, the patriarchal family and finally the ‘civilised’ family.

  • ix A long-running argument over how to explain human behaviour, customs, etc. Are they caused by genet (...)

41His mistaken idea that the monogamous European family was the form towards which human societies ‘naturally’ tend leads us to the second step in the analysis of the utility of the notion of family. This second step occurred decades later, during what would become known as the nature-nurture debateix. In 1913, Bronislaw Malinowski, an anthropologist who subsequently became renowned for his research in and writings about the Trobriand Islanders, wrote The Family among the Australian Aborigines, a book that challenged some of the conceptions of the cultural evolutionists. In Malinowski’s opinion, the nuclear family is a human universal, whatever the kinship system, because it answers a human being’s primary needs: to be nurtured and reared. Much later, in a discussion with the historian Robert Briffault, broadcast on the BBC and published by Ashley Montagu (1956), he went even further and explained the following:

The science teaches us that marriage and the family are rooted in the deepest needs of human nature and society; that they are associated with progress, spiritual and material. The real task of anthropology consists in giving us insight into the essentials of marriage and the family, as well as the understanding of their value for society.

[...] the most fundamental point of the debate [is] that marriage and family are based on the need of the male to face his responsibility and to take his share in the process of reproduction and of the continuity of culture.

42What did Malinowski mean? First, that various types of kinship systems and social units in which people live and work exist but that the family is the universal minimal social unit because it is rooted in the deepest needs of human nature. Marriage and the family are institutions that make possible such elementary things as being fed, looked after and cherished, feeling secure and so on. This view of the family is deeply rooted in Malinowski’s general conception of culture. In his so-called functionalist approach, every social institution and every cultural feature satisfies particular functions that contribute to the integration of society as a whole. Social institutions and structures are like organs in a living body, each having a particular task. The nuclear family is one of these ‘organs’.

43Malinowski adds another point that reflects even more deeply his view of the family as a human fundamental. One could imagine ‘families’ in which the father takes no part in the bringing up of children and may not even live with them. Such families have since been observed in China, among the Na people, where the mother lives with her own mother, her sisters and her brothers and where the father is, according to Hua (2000), absent and often even unknown. These are ‘families’ where males would only be anonymous donors of semen. Similar features are nowadays observed in Western families where monoparental families, in which one parent alone educates his or her children, are increasingly frequent. Why then is the (nuclear) family universal according to Malinowski? He explains it as a consequence of power relationships: men (or culture) ‘force’ woman and children into families in order to make sure they are included in the reproduction of society and to ensure that males are not solely anonymous semen-givers but actively involved in decisions and action involving children’s education.

44These two points addressed by Malinowski deserve our attention. First, he believes that the (nuclear) family is a necessity, a cultural or social universal response to natural needs: a human baby, unlike some animals, requires feeding and rearing. Simultaneously, the ‘family’ (and kinship in general) is where power relationships are expressed and applied. Malinowski’s second thesis, according to which the family is a consequence of males’ apprehension about being excluded from the reproduction of society, is difficult to defend. However, we can agree with him that the family, and kinship in general, is an important nexus of local politics. It is about who rightfully does what, to whom and why. This is particularly true in Australian kinship systems.

45While many anthropologists, such as Radcliffe-Brown (1941:2), have followed Malinowski in his definition of the minimal social unit as being formed by parents and their children, others have challenged this conception. Meyer Fortes (1978:21), for example, notes that the mother-child couple, excluding the father, is ‘the unique and irreducible source of all human existence’. Other researchers, such as Bohannan (1963:73) and Goodenough (1970:18), define the ‘nuclear family’ as universally comprising only a mother and her children.

  • 1 ‘Tautological’ is a definition that is a repetition of the thing that has to be defined. For exampl (...)

46The nuclear family, it is now acknowledged, is neither universal nor inevitable (see Holy 1996: 33). Indeed, having such a preconceived conception of the nuclear family as a building block could result in erroneous accounts of cultural systems. As Godelier (1977:138) explains, a family cannot survive on its own but is embedded in larger networks of exchanges and mutual support. We shall not venture further into the question of how and why anthropologists attempted to define the family (or the basic social unit) and its origins. Héritier summarises these endeavours and suggests that these authors try to answer ‘[...] what the family is there for [...] without being able to produce a definition which is universally correct, which is not somehow tautological1’ (Héritier 1991: 274).

Domestic group, household and adoption among the Ngaatjatjarra

47We now examine a local case. As I have noted, the English word family does not correspond exactly to any indigenous category in the Western Desert or in most of the rest of the continent. Rather than conveying that the relationship between children and their parents is not recognised or unimportant, I simply imply that the definition of the minimal social unit rests on criteria other than those usually associated with the notion of family.

48There are two ways of approaching the question in the Western Desert. The first is to consider the parent-child relationship as a starting point and to describe the types of social mechanisms that relate to or emerge from this relationship. This approach takes as its starting point a biological or genealogical fact (being the child of) and endeavours to explain social phenomena that may either diverge from expected principles or reflect them. Another approach is to define how the minimal social unit is conceived emically (that is, from people’s own point of view) and to extrapolate from this understanding by considering other observable social units. This second approach is a sociological one — rather than the genealogical approach mentioned before — as it first describes social mechanisms and, in a further step, correlates these to possible biological relationships. The notion of ‘meaning’, that is the culturally specific interpretation of genealogical or other links, is of primary interest in this second approach.

  • x The feeling and practice of having or establishing connections with other people through various me (...)

49These two approaches reflect the two principal (and to some extent conflicting) approaches adopted by anthropologists in their study of human kinship. The genealogical approach is also the method used by former functionalist and structuralist anthropologists who, comparing cultural specificities construed upon the genealogical grid, elaborated generalities and typologies of kinship systems. The second approach is more like that of ‘cultural studies’. Many anthropologists were dissatisfied with former methods, contending that these were based on an Euro-American conception of kinship, considered to be based on cultural interpretations of the natural facts of procreation. The ‘cultural studies’ approach they proposed, which is interested in relatednessx in general terms, seeks to understand local cultural processes and meanings in relation to a relationship’s constituents. I now endeavour to combine these two approaches, since both produce insights and are complementary even if distinctive in their interpretative nature. I intend to use structural notions and concepts but shall attempt to demonstrate that these concepts have cultural equivalences that are meaningful in the Western Desert.

50A first step in this endeavour is to clearly differentiate the family as a unit of kinship from both the household as a residential unit and the domestic group as an economic unit. The aim is to elaborate a definition, even if narrative, of the minimal recognised social unit in Western Desert culture: the smallest unit that ‘makes sense’ in Aboriginal Australia. This is an important preliminary question in the study of a kinship system since it is the building block on which social structure is based.

51During anthropological fieldwork in a Western Desert community, the researcher soon develops an intuition that the minimal or elementary social unit is something similar to the nuclear family, is somehow linked to a household or camp and also reflects an economic entity. However, intuitions are not always right and scientific rigour demands that we construct the argument necessary to discover and demonstrate the local substance or essence that justifies this unit and organises its character and meaning.

52The starting point must be one in which the nuclear family based on genealogical (and affinal or marital) ties, the domestic group and the household, a notion that is widely used especially in the anthropology of Eastern Europe, are independent. What we then need to consider are the multidimensional interrelations between social units constituted by genealogical links, production and consumption activities and residence. This particular study is based on a specific community in the Western Desert, about which I first need to present some general historical and social facts.

53The community in question was established in 1976 as an outstation or country camp of a larger community. Outstations are small settlements which emerge around a larger community that have a few important facilities, such as a shop and a school. These outstations were established in response to the desire of desert peoples to live closer to or on their traditional lands yet still benefit from the advantages of a settlement with its infrastructural resources, such as a water bore, an electricity generator and a store. Outstations sometimes grow and become ‘independent’ communities themselves. They are thus incorporated under Australian federal law and usually obtain the right to run a school and obtain governmental financial assistance. In this particular case, people gave two main reasons for developing an independent community: being closer to their old homelands (having been nomadic hunter-gatherers until the 1950s) and, secondly, having had a family member die in the community where they had settled. Many Aboriginal cultures avoid the place where a person dies, just as they do not pronounce their name for a year or more. People whose names are identical to that of a deceased person are, after the death, called Kumarnarra, no-name. Because this person died in the settlement, they felt that the time had come to relocate.

54The newly established community initially had only a bore and a couple of corrugated iron sheds. Other relatives soon moved in and the community later incorporated, built more houses and sank a second water bore. A shop, clinic, school and community office were erected, and by 1995 the community was a well organised and independent settlement comprising some 120 people. In 1996, of these 120 people, 93 could be considered permanent Aboriginal residents, living in 20 out of the 28 available houses in the community, with an average of five persons per dwelling. Not all houses were or are in good condition (see Tonkinson and Tonkinson 2010 on desert values and attitudes to property). Many are built of corrugated iron and become infernal ovens in summer or lack the most basic facilities such as water and electricity. In many cases, people prefer to live outside, choosing to camp between the houses where fires burn and where swags are used as beds. There are various reasons why people in many cases prefer camping to living in houses, among them the fact that houses considerably diminish the capacity to adapt the residential structure to social relationships. It is important to be able to alter the distance between families as well as the orientation of their camps with respect to affinities and expected behavioural norms (see White 1977 and Dousset nd.b).

  • xi A residential unit. People who live together in the same abode or camp.

55The composition of these householdsxi is diverse; we consider here a household as people usually living together in or around one identifiably separate dwelling. Eight of the occupied dwellings hosted nuclear families (parents and their children); another eight were occupied by nuclear families, plus at least one other person, such as a married child with his or her children or another close relative. Two sheds were occupied by widowers and widows with their unmarried children. Widows’ camps are called yarlukuru, widowers’ camps tawarra. Two other households consisted of persons not closely related to one another. It is clear that, although the nuclear family plays a certain role in many if not most households, there is certainly no strict overlap between these two notions in a demographic sense.

  • xii An economic unit. The smallest unit of people that produce and consume material goods in common.
  • xiii Occurs when a person demands something from another on the basis of their kinship link and the obli (...)

56The domestic groupxii is often difficult to delimit. You will remember that a domestic group is the smallest economic unit that collectively consumes and produces material goods. The most elementary economic relationships among the Ngaatjatjarra are those between a woman, her husband and her brother. Indeed, if we measure the domestic group as the result of economic cooperation, then this group must be considered as the most basic unit. Information on wealth and availability of resources flows freely among these people and reciprocity is not a matter of demanding or demand-sharingxiii (see Peterson 1993). An adult daughter no longer living in the family household would ask before helping herself to food from a meal prepared by her mother. A husband and his wife’s brother, on the other hand, would not hesitate to just help themselves. Similarly, a woman would not hesitate to help herself from food prepared by her brother or her tjuwari (brother’s wife). As another example, a brother visiting his sister would not need to ask permission when entering the private area of her camp (the area reserved for sleeping and eating), even when no one was present. This is not the case with other relatives or community members. They would not enter the camparea if the owners were absent, and when it was occupied they would not enter it without manifesting their presence and receiving some words or gestures of invitation.

57This domestic unit hardly ever constitutes a residential entity. Every man is in principle involved in at least two domestic units: one with his wife, and another with his sister and her husband. Similarly, every woman is involved in two domestic groups: those of her husband and her brother.

Figure 23: Overlap of domestic units among the Ngaatjatjarra

Figure 23: Overlap of domestic units among the Ngaatjatjarra

58As the above figure shows, domestic groups are interconnected in such a way that each person has a double belonging, both through brother/sisterhood (consanguinity) as a brother or sister on one side, and through marriage (affinity) as a husband or wife on the other. Domestic groups are hence not bounded entities but constitute a chain of people linked through their positions in each of the two domestic groups. If this is to be considered the minimal economic unit, it calls into question the notion of the ‘domestic unit’ itself since its extent is in fact endless.

59I have not chosen a female Ego to be the centre of the domestic group, and thus the axis in the figure, but wife and sister occupy an important place in these relationships. The ‘nuclear family’ and the ‘domestic group’ never completely overlap, but these relationships are articulated around the woman. To investigate this articulation and analyse the meaning constructed around a genealogical connection, we can draw information from an interesting test case: that of adoption.

60Adoption is indeed the most ‘cultural’ or ‘social’ event of what could be considered a ‘natural’ process, which we can call ‘engendering’. It is the most ‘cultural’ aspect of a ‘natural’ process, and the most ‘natural’ aspect of a ‘cultural’ process. Adoption creates links that look or are made to look ‘biological’ yet are social. It is in the local understanding and explanation of adoption that we can discern whether there is an interpretation of genealogical connectedness and what form it will take.

61Adoption is very common among Ngaatjatjarra people and fosterage, the temporary caring for children, even more so. There are numerous reasons why children are claimed or given away for adoption. In most cases, however, it is the result of deceased parents or when an unmarried mother is considered unable to care adequately for her offspring. The children leave their natal familycamp or house to settle in the new household. The expression used to describe adoption is particularly interesting:






The adoptive mother

from another one



[and] raised

62According to the expression, the position of being an adopter can only be occupied by a woman, which is an interesting fact in itself. As far as I know, it is impossible to express adoption in masculine terms. The second interesting point is that nyuyurlpa (the adoptive women, or the verb ‘to adopt’) also means ‘to add timber to the fire’. This is not a linguistic reconstruction or a hypothetical reconciliation between the meanings of two words. The identity between these two words is explicit and conscious. We may thus rewrite the expression as follows: ‘the one (feminine) who keeps the fire alive took the child from another one, kept it and raised it’. It is also useful to understand what else this fire is about, in addition to it being something in the centre of a residential community that is personalised through a woman.

63Fire has a strong metaphorical and symbolic value in the Western Desert. Parents who have lost a child are called purrkutjarra (literally ‘with coals’), because they coat their faces with charcoal from an extinguished (dead) fire. Another example is the ritual that follows a child’s birth: women hold the baby in the smoke of a burning (living) fire and give it its first (secret) name. Also, after initiation, the youth is massaged by a classificatory motherin-law in the smoke of a fire, confirming his rebirth as a man. A more common example is the following: the smoke of a distant fire signals the presence of other human beings in an immense desert. Traditionally groups would announce their planned arrival by lighting smoke fires as they were approaching. Fire is also used to burn out weeds and shrubs so that the grass can regenerate and attract animals.

64The importance of fire as a metaphor for life, birth and rebirth is unquestionable. Moreover, the fire of a camp is associated with a woman, who is keeping it alive, perhaps in a camp where an adopted child is settling. It is difficult to define the underpinning substance or essence of the basic social unit among the Ngaatjatjarra, but the study of adoption has produced some new understandings: the central genealogical connection of the household is the woman, her material counterpart, the fire (household), and her relationships with her brother and husband (domestic unit).

Summary and Conclusion of Part Three

65This section is clearly the most ethnographic one because in it I attempt to apply the concepts developed earlier to a specific region, the Western Desert. I began by summarising previous analyses of this region’s kinship system in order to highlight a few errors and misconceptions. The major one was to consider that people use only a single type of terminology, applicable in all situations and by all persons in the same way. The ethnography shows that, within a single language area and group, several terminological systems can coexist and be deployed. I have called these contexts sociological and egological. In the latter, a Dravidian-type terminology is used, one in which cross-cousins are clearly distinguished from parallel cousins and individual relationships are relevant. In the sociological context, however, the general social structure is relevant; a social structure in which generational moieties, which we address below, are relevant. Here, cross-cousins can be called by the terms for brothers and sisters and a Hawaiian-like terminology is applied.

66We then considered affinal terminology: the ways of addressing and referring to in-laws, whether actual or potential. This led us to consider how people find spouses and marry, how marriage rules are applied and how marriage, real or potential, leads to the establishment of social networks. I have stressed that every person has several alliance relationships regardless of whether marriage actually occurred. In an ecological situation in which networking is an important adaptive strategy, marriage is a means to assure reciprocity and solidarity between families that are often separated by long distances.

67Marriage typically creates a family, but the notion of ‘family’ is not straightforward so three types of entities were enumerated and discussed: the family as a genealogical unit, the domestic group as an economic unit and the household as a residential unit. Particular attention must be paid to the ways they interact. I showed that studying processes which lie on the border between a ‘naturalistic’ and a ‘cultural’ local view of kinship can prove useful. Adoption and the way it is handled and interpreted constitute one of these processes.


Further Reading

Concept of the family, household and domestic group

GOODY, Jack (ed.) 1971. The Developmental Cycle in Domestic Groups. Cambridge: Cambridge University Press [1958].

GOODY, Jack 1988. Production and Reproduction. A Comparative Study of the Domestic Domain. Cambridge:

Cambridge University Press, Series Cambridge Studies in Social Anthropology, 17 [1976].

GOODY, Jack 1999. The European Family: An Historico-Anthropological Essay. Malden (MA): Blackwell Publishers.

JOLLY, Margaret and Martha McINTYRE (eds) 1989. Family and Gender in the Pacific: Domestic Contradictions and the Colonial Impact. Cambridge: Cambridge Univ. Press.

LASLETT, Peter and Richard WALL 1972. Household and Family in Past Time. Cambridge: Cambridge University Press.

Australian family and community

DOOHAN, Kim 1992. One family, different country : the development and persistence of an Aboriginal community at Finke, Northern Territory. Sydney: University of Sydney, Oceania monograph 42.

ELKIN, Adolphus P. (ed.) 1957. Marriage and the Family in Australia. Sydney: Angus & Robertson.

MADDOCK, Kenneth 1972. The Australian Aborigines. A Portrait of their Society. Ringwood - Victoria: Penguin Books.

MALINOWSKI, Bronislaw 1963. The Family among the Australian Aborigines. New York (London): Shocken Book (Univ. of London Press) [1913].

MUSHARBASH, Yasmine 2009. Yuendumu everyday: Contemporary Life in Remote Aboriginal Australia. Canberra: Aboriginal Studies Press.

ROSE, Frederick G.G., 1960. The Australian Aboriginal Family : Some Theoretical Considerations. Berlin: Forschen und Wirken, III.

SMITH, Diane and Evelyn DALY 1996. The economic status of indigenous Australian households: a statistical and ethnographic analysis. Canberra: Australian National University, Centre for Aboriginal Economic Policy Research.

SMITH, Diane 1990. Looking for your mob: A guide to tracing Aboriginal family trees. Canberra: Aboriginal Studies Press.

TURNER, David H. 1980. Australian Aboriginal Social Organisation. Canberra, Atlantic Highlands: AIAS, Humanities Press.


1 ‘Tautological’ is a definition that is a repetition of the thing that has to be defined. For example to say a circle is round is tautological because the word ‘round’ does not explain what a circle is.

Notes de fin

i A situation in which personal interrelationships determine which terminology or behaviour to adopt (see the sociological context).

ii A situation in which the relationship between general social categories determines the terminology and behaviour to adopt (see egological context).

iii A residential unit. People who live together in the same abode or camp.

iv An economic unit. The smallest unit of people that produce and consume material goods in common.

v A kind of marriage involving one woman or man and her or his spouse. A monogamous family comprises a man, a woman and their children (also see Polygamy, Polygyny and Polyandry).

vi A marriage between a group of brothers and a group of sisters. This mistaken notion was used by evolutionists to imply that in ‘primitive’ societies mating was indiscriminate among groups of people, thus children would not know who their biological parents were. No such society exists.

vii Having several simultaneous spouses.

viii A form of polygamy. A woman who is married to more than one husband at the same time.

ix A long-running argument over how to explain human behaviour, customs, etc. Are they caused by genetics, ‘instinct’, etc. or by being raised in a society? In other words, how much in a human being’s behaviour, thinking and practice is culturally determined and how much is biologically determined? The two influences are so inextricably linked that attempts to separate them out and apply some kind of quotient would be impossible and pointless.

x The feeling and practice of having or establishing connections with other people through various means; it produces relationships similar to kinship.

xi A residential unit. People who live together in the same abode or camp.

xii An economic unit. The smallest unit of people that produce and consume material goods in common.

xiii Occurs when a person demands something from another on the basis of their kinship link and the obligations and responsibilities that this relationship entails.

Table des illustrations

Titre Figure 20: Western Desert terminology: Ego’s and Ego’s parents’ generations
Fichier image/png, 23k
Titre Figure 21: Western Desert terminology:Ego’s grandparents’, children’s and grandchildren’sgenerations
Fichier image/jpeg, 172k
Titre Figure 22: Western Desert terminology: affinal terms (left: male Ego; right : female Ego)
Fichier image/png, 70k
Titre Figure 23: Overlap of domestic units among the Ngaatjatjarra
Fichier image/png, 17k

© pacific-credo Publications, 2011

Conditions d’utilisation :


Volume papier
Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search