Version classiqueVersion mobile
OpenEdition Books

Australian Aboriginal Kinship

 | 
Laurent Dousset

Part two: Some basic concepts of kinship

Texte intégral

1This chapter will present and discuss the basic concepts and tools used and needed to understand kinship in any culture, not just in Australian Aboriginal societies. However, we will be careful to regularly use Australian examples and progressively lead the reader to Chapter Three, which will present in a more particularistic way the system in use in the Western Desert and elsewhere in Australia.

What is kinship?

Kinship and marriage are about the basic facts of life. They are about ‘birth, and conception, and death’, the eternal round that seemed to depress the poet but which excites, among others, the anthropologist. […]

Man is an animal, but he puts the basic facts of life to work for himself in ways that no other animal does or can (Fox, 1996 [1967]: 27).

2Everybody has some understanding of what kinship is. The most common conception, at least among European people, is that of a person who is born into a family, with parents who are married or live together as a couple, grows up and marries a person and will have children and grandchildren. The situation may, however, become blurred and complicated when scholars attempt to define the notion of kinship and to describe its characteristics and implications for other social domains.

3We can approach kinship by dividing this large concept into several sub-domains (terminology, descent, marriage and relatedness), which themselves reflect the evolution of the discipline of anthropology and the ways in which social scientists have attempted to analyse elements of kinship. Each academic school tends to stress different ethnographic realities and theoretical conclusions, but over time all have in some way helped us to understand better what kinship is and does.

Figure 7: Schematic view of the domains of kinship and the evolution of anthropology

Figure 7: Schematic view of the domains of kinship and the evolution of anthropology
  • i A set of words used to address or refer to people with whom one is related through a biological idi (...)

4The first domain studied was kinship terminologyi which became prominent after Morgan had put forward his evolutionary theory. Every language has a set of specific words that belong to the domain of kinship. In English, these are ‘mother’, ‘mum’, ‘father’, ‘dad’, ‘son’ and so on. Each term describes and reflects a particular relationship: a mother is usually the person who gave birth to the speaker; the father is typically the husband of the speaker’s mother and so on. The words themselves, the persons referred to via these words and the underlying or overarching definitions that accompany these words constitute what is called a ‘kinship terminology’.

Unilineal evolution

Lewis Henry Morgan published two major books. The first was Systems of Consanguinity and Affinity (1871), to which we will return later since he defined in this work an important typology of terminological systems. The second was Ancient Society (1877). It is in this second book that he expounded his theory of unilineal evolution. While he was not the only intellectual to put forward such a theory (Karl Marx, Friedrich Engels, Edward B. Tylor, Auguste Compte or Herbert Spencer are some of the other scholars who advocated such a theory), he was one of the major representatives of this movement. In this work, he attempted to link kinship systems with economic, technological and political structure and endeavoured to demonstrate that the various forms of societies, for which we have records from the past and which we observe in the present, each represent a particular step in the evolution of social forms. According to Morgan, human society evolves through three such stages: savagery, barbarism and civilisation. He associated the invention of the bow and of pottery with the first stage, agriculture with the second stage, and writing with the third. Democracy as a political system is yet another characteristic of the third stage. The kinship systems of societies similarly evolve from what he called ‘group marriage’ (a group of women marrying a group of men, with children belonging to the entire group) through various other forms to reach the nuclear family of Euro-American societies. According to Morgan, Australian Aborigines were in the lowest stage of social evolution while European society was in the most advanced state of civilisation. Evolutionary theory has been highly criticised and is today completely abandoned. It reflects a Eurocentric and ethnocentric point of view that does not take into account the particularities of human adaptations and the diversity of value systems. Some researchers, such as Julian Steward (1972), have suggested the concept of multilinear evolution which sees each culture evolving through its own adaptations to ecological and economic environments.

  • ii A group of people, or a number of lineages, whose members claim to be descendants of a distant ance (...)
  • iii In this process of socialisation, the individual comes to regard what has been inculcated as normal (...)
  • iv The systemisation of certain principles of filiation (see filiation).
  • v The particular relationship between children and parents who may share particular substances, right (...)
  • vi The union of two individuals, which brings together two families or groups. Marriage is always subj (...)

5Following this phase of kinship studies came a group of British anthropologists, whose theoretical approach to the study of human societies became known as ‘structural-functionalism’. Their attention was centred on kinship groups, such as clansii and lineagesiii, and on the processes of descentiv and filiationv that constitute such groups. Around the same time, some French scholars developed a structuralist approach which focused mainly on marriagevi and exchange. They attempted to discover and formalise universal rules defining marriage patterns around the world.

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

6Finally, these rather formal approaches were overtaken, beginning around the 1970s, by new ones that added an emphasis on the dynamics of kinship and on its negotiability; for convenience, I label these ‘relatednessvii’ studies. Later, we will return to each of these domains and discuss the particular concepts that scholars employed. It is most important that these schools and approaches should not be interpreted as paradigms, as theories that replaced each other, but as innovative additional elements, all of which should be taken into account when attempting to understand local kinship structures. First though, we need to examine some general conventions.

Symbols, abbreviations and conventions

7Anthropologists and other genealogy specialists usually use a simple set of symbols to represent persons and connections. These symbols were largely inherited from the International Federation of Eugenic Organizations in 1932 by the Sociological Research Committee of the Royal Anthropological Society of Great Britain (see the journal Man 1932, vol. 32: 120-121) in a paper that was termed The Standardization of Pedigree Charts. Since then, symbols and rules have in some respects changed, and individual researchers use their own particular set of symbols appropriate to the conditions in which they work. However, the symbols discussed below are widely accepted. Since we will be making recurrent use of these in later chapters, some detailed discussion is necessary.

8Two types of symbols are distinguished: the first type represents individuals and the second indicates the relationship between individuals (these are called ‘connections’). There are three basic individuals: a female, a male, and a person whose gender is unknown or irrelevant. This last symbol is rarely used but can be applied to represent, for example, the English word cousin since this word applies both to a male and a female. A male is represented by a triangle, a female by a circle and the unknown or irrelevant gender by a square.

Figure 8: The basic symbols representing individuals in genealogies

Figure 8: The basic symbols representing individuals in genealogies

9Additionally, it may be necessary to indicate if a particular person is living or deceased. The above symbols represent living persons. In the case of deceased persons, the figures are either coloured in with black or crossed out with a slanting line.

10These symbols representing individuals do not convey much information, so we need to add some more showing how these individuals are connected to one another. Here again, three basic conventions are used: one (actually, there are two from which to choose) indicates marriage, one indicates siblings (brotherhood and sisterhood) and a third represents filiation or descent.

Figure 9: Basic conventions representing relationships.

Figure 9: Basic conventions representing relationships.

11These connections sound complicated as described in the paragraph above, but do not worry about this because they are very easily understood when depicted on a chart (see below). A marriage connection, also called an alliance in certain cases, is represented as a line that goes from below one person to below another person. Some scholars use an ‘equals’ sign (=) to represent the marriage relationship. A sibling connection is represented as a horizontal line above the children and from it vertical lines drop to the siblings depicted. A filiation (parent-children relationship) is represented as a line that goes from below a person down vertically to indicate a single person or to a horizontal bar from which two or more siblings will be depicted. Additionally, you may want to represent adoptions in genealogies. This connection is usually represented as a filiation, using either a dotted or broken line.

  • viii In many domains, in particular language, this rule reflects the fact that social processes and stru (...)

12These symbols and connection types are based on a first, important rule to remember in kinship studies and genealogical representations: the rule of economyviii. Graphic or logical representations, as well as verbal descriptions of kin relationships, should always use the shortest and most efficient link and symbol to describe a relationship, unless other ways and symbols add information that substantially change the understanding of the relationship. This rule will become self-explanatory once you see how the symbols are applied to represent real-world kinship patterns.

  • ix In popular terms, a ‘family tree’. It is a chart that reconstructs an individual’s total known kin (...)

13These connections and symbols are then combined into what is called a genealogyix or a genealogical grid, in which every person represented is linked to one or several other persons through the connection types. In Western societies, these are commonly called ‘family trees’. The figure below shows a nuclear family: mother, father, their daughter and son.

Figure 10: Graphic representation of a nuclear family

Figure 10: Graphic representation of a nuclear family
  • x In order to be meaningful in anthropological terms, every kin category or every person in a genealo (...)

14This leads us to the second rule, that of multiple connectednessx. To make genealogies informative, the aim is to connect each person to at least one other person and, if possible, to two or more other persons using distinct types of connections. In the above example, the daughter is connected to her brother, as well as to her parents. In fact, she is connected through one line of filiation to her mother and through another one to her father, but these two connections are represented following the same path and thus form one line only. The daughter is thus multiply connected, in this case in three different ways.

15Before we move on to the second set of conventions used in kinship studies, it is crucially important to understand a distinction that has not been sufficiently marked, even by the most prominent scholars such as Lévi-Strauss. This is the distinction between a genealogical system and a terminological system. The symbols presented above are sometimes confusingly used to represent two quite different facts. First, they can be used to represent relationships between actual people, between David, Margaret, their daughter Annie and their son Henry. This is a genealogical usage of the symbols within a genealogical grid, whereas the second use is part of the domain of terminology and is general (not specific) and abstract (not concrete). They can be used to represent kinship words, rather than individuals, and stand for ‘father’, ‘mother’, ‘son’ and ‘daughter’. In the case of a nuclear family, confusing David with ‘father’ is not a big problem. But, as we shall see in the case of complex kinship systems such as those of Australian Aboriginal people, these two domains need to be clearly separated since in many cases the word ‘father’ is used for someone who is not necessarily a genitor, a (actual or biological) father. In this book, we shall seldom encounter genealogical usages of the symbols mentioned above, so you need to think of these as abstract and generic individuals rather than as real individuals.

16Moving on to the second set of conventions, we note that, besides the graphical representation of terminologies, anthropologists and linguists also use abbreviations to describe persons and relationships. I avoid making intensive use of these abbreviations in this book because, although straightforward, they seem to confuse some beginners rather than help them. You will see that most of the books indicated in the further reading sections use these abbreviations, so it will be useful for you to familiarise yourselves with them.

17The driving idea behind these conventions is the fact that some kinship terms cannot easily be translated from one society to another; for example, the English word uncle does not have any exact equivalent in Australian Aboriginal languages. In English, an uncle is your mother’s, as well as your father’s, brother. In the Western Desert, however, these two persons are identified using different words and are quite different types of relatives. Anthropologists therefore use abbreviations that are descriptive, that is they are not a translation of a specific kin term (uncle) but are based on primary or immediate relationships (such as mother or brother, for example) in order to reduce ambiguities. The table below summarises these conventions and the corresponding relationship.

Abbreviation

Specific

General

F

Father

any person’s father

M

Mother

any person’s mother

B

Brother

any person’s brother

Z

Sister

any person’s sister ('Z' is used for sister in orderto avoid confusion with Son)

S

Son

any person’s son

D

Daughter

any person’s daughter

H

Husband

any person’s husband (the person speaking canobviously only be a female, with the exception

of cases of homosexual marriages)

W

Wife

any person’s wife (the person speaking can obviously only be a male, with the exception of cases of homosexual marriages)

Sp

Spouse

husband and/or wife

Additional abbreviations

e

elder

e.g., eB signifies an elder Brother

y

younger

e.g., yB signifies a younger Brother

f

female speaking

e.g., fZ signifies a woman’s sister

Ego

The person speaking

Ego (not an abbreviation) comes from Latin, meaning ‘I’ (the first person singular). Ego simply denotes from whose point of view we are speaking.

18As was the case with the graphic symbols, these abbreviations are combined so as to describe every possible individual or position in the grid. Hence, an English uncle is an MB (mother’s brother) or an FB (father’s brother). The combination of these abbreviations can sometimes become rather complex. For example, one of your second cousins would be labelled MMBDS: your mother’s mother’s brother’s daughter’s son, where your mother’s mother is obviously one of your grandmothers.

19As already noted, important differences exist between kinship systems and terminologies across different languages and cultures: what an English-speaking person calls an uncle is not simply an uncle in the same way everywhere. Exactly how important and significant these differences are is an interesting question for those with an interest in kinship studies. We now examine the domain that was the first to be investigated by anthropologists and has remained significant for understanding cultural differences: kinship terminology.

Kinship terminology

20When we speak of kinship, we are dealing with cultural features, rules and words pertaining to a number of important domains: kinship terms, kinship groups, descent and filiation, marriage, and relatedness. Yet when anthropologists speak of a ‘kinship system’ they very often limit the field of investigation to kinship terminology alone. Terminology has indeed remained one of the central aspects of kinship studies. Words connected to kinship are part of language, which is taken to represent cultural differences and similarities and to provide a first approach to how people think and feel in different cultural contexts. For instance, the way colours are conceived of, differentiated, expressed and talked about in different societies and cultural contexts is a good illustration of the power of language to articulate differences.

21Harold Conklin (1955) is a well known anthropologists interested in studying how people perceive the world around them. He is particularly interested in the way languages classify and name colours and has investigated these questions among the Hanunóo people of the Philippines. Conklin concluded that the Hanunóo way of naming and classifying colours or seeing colour was completely different from the way Americans differentiate the colour spectrum. They describe and name hundreds of different colours, many more than the English language does. Additionally, factors such as texture, moisture and shine provide different names that distinguish colours. These colours are grouped and classified according to specific criteria, which Conklin summarises as lightness, darkness, wetness or freshness and dryness. His study is an excellent example of the necessity of breaking down cultural features into locally meaningful elements by investigating the differences that make sense in a particular cultural context.

Componential Analysis and Markedness

An important study in colour was carried out by Berlin and Kay (1970). The approach Conklin and others used is called componential analysis, a method that was primarily developed to investigate kinship and its terminologies (Goodenough 1956). The basic idea is that words of kinship are organised. The elements or components that remain once descriptions have been broken down into their constituting elements relate to each other in a systemic way. Just as alphabets can be combined in certain ways to produce written words, the components of kinship terminologies produce a kinship system since they are systemic and produce patterns of cultural significance. The above-mentioned abbreviations used in kinship studies are the first vital steps in componential analysis as they do not involve culturally predefined values. Componential analysis does, however, add other criteria in order to distinguish relationships. We shall return to these when discussing terminologies and investigating Australian social organisation. For now, we note that in Aboriginal Australia the most important components are generation, ‘crossness’ and gender. In some cases, relative age also plays a role. Componential analysis was inspired by the work of people such as Kroeber (1909), who suggested analysing kinship terminology with reference to eight categories: generation, lineal versus collateral, age difference in one generation, gender of the relative and connecting relative, sex of the speaker, consanguinal versus affinal, and the condition of the connecting relative, all of which later became the bases for componential analysis. Greenberg (1990 [1980], 1966:72-87) later added the criterion of ‘markedness’ to the analysis of universal aspects of terminologies. As noted by Scheffler (1987:203), this term was first applied by Trubetzkoy and R. Jakobson in the 1930s to ‘the asymmetrical and hierarchical relationship between the two poles of any opposition’ (Waugh 1982:229). Scheffler is a kinship studies specialist who used componential analysis. As an example, the English word man can designate a male but can also designate human beings in general, whereas woman designates only a female human being. In a common context, the two words man and woman are in an opposition which, however, is not equivalent but hierarchical since man may also mean human being in general, ignoring the gender attribute. According to the markedness theory, the word woman is considered marked against the word man as the former is also in some contexts included in the latter.

  • 1 An ‘idiom’ is a special way of speaking that is considered ‘natural’ to a speaker of a particular l (...)

22Kinship, as Tonkinson (1991) puts it, is a mass of networks of relatedness which radiate from each individual, and this network is expressed in a biological idiom1. This idiom is a set of words or expressions — a terminology — that is attributed via what Fox (1996 [1967]) terms the ‘basic facts of life: conception and birth’. Whatever the local term that stands for ‘mother’ may be and whatever other relationships or things may be expressed by this term, at its very basis it describes the unique relationship between a person and a woman from whose womb he or she was born. Every language and, in some cases, even every dialect has its own set of such words that distribute the network of relationships around an Ego, the male or female we take as our starting point.

23The number of elements in this network of words is culturally ascribed and unique, yet also follows certain universal rules. Not in all languages is each possible position of a relative distinguished from the others, as we have seen for the term uncle in English where it designates both the father’s and the mother’s brother. Those positions that are distinguished from the others are meaningful and labelled ‘categories’ or ‘classes’. In English, the distinction between mother’s brother and father’s brother is not meaningful but that between aunt and uncle, on the other hand, is meaningful and is based on the distinction of gender. All such kinship categories of a culture or language constitute a terminological map, or kinship system. These terminological maps can be compared from one society to another and have been typified. What is of particular interest is that human societies draw on only a limited number of such terminological types to construct their local kinship system. Indeed, identical or near identical terminological systems are found in societies as far apart as lowland Amazonia, India, China, North America and Australia. The existence of a limited number of types of terminological systems makes some prediction possible and renders systems, and thus cultures, comparable in the domain of kinship, although this always needs to be carefully confirmed.

24As already explained, terminological systems were the first elements of kinship systems to attract anthropologists’ attention, starting with Morgan’s book Systems of Affinity and Consanguinity of the Human Family (1997 [1871]). Morgan collected terminological systems through corresponding with people from different parts of the world, such as with A.W. Howitt in Australia, and concluded that, despite the diversity in the ways cultures and languages describe a person’s genealogical environment, there are important similarities that seem to be systemic. Morgan and other early social scientists were not only interested in terminologies but also studied marriage patterns and filiation. However, their legacy is principally crystallised in their systemic analysis of kinship terminologies from different societies. Hence, Morgan produced a first typology of these systems that has since been amended many times by various anthropologists (for example Murdock 1949), yet is still widely used today.

25We shall now examine some of these system types since understanding their mechanisms is important for understanding Australian Aboriginal kinship systems. Our particular focus is on what is called the Eskimo system, in use among English speakers, and the Dravidian system, which is predominantly used in Aboriginal Australia. I will however also rapidly mention other systems such as the Crow-Omaha, which is in use in some Australian societies, or the Sudanese system, which is not used in Australia but which can be considered as standing opposite the Australian-Dravidian system. I refer the reader interested in systemic discussions and definitions of Dravidian and Crow-Omaha to a paper written by Trautmann and Barnes (1998).

The Eskimo system

  • xi The range of people towards whom kinship terms can or must be applied, which in some classificatory (...)
  • xii The kinship relationships that exist for a person before marriage (see affinal). In English, mother(...)
  • xiii Kinship relationships that come into being when marriage is envis-aged or has occurred. Affinal rel (...)
  • xiv A particularity of kinship terminologies that denotes the differentiation between maternal and pate (...)
  • xv Those relatives or principles that are on ones mother’s side.
  • xvi Relatives on ones father’s side, and principles relating to descent in which males are the central (...)

26As the name indicates, this system is found among the populations of the northern polar regions, particularly in Canada. Note however that, while the name has been retained to denote the kinship system, these people are called Inuit today; they consider the word ‘Eskimo’ to be a negative or pejorative term. The Eskimo system is also that found among the !Kung people in Africa and is predominant in Europe. Major characteristics include its emphasis on the nuclear family and its very limited extension of rangexi, meaning that the use of kinship terms is limited to genealogically very close persons and that others are grouped into large categories. For example, the English word cousin which includes all of one’s parents’ siblings’ children. Also, in the Eskimo system there is no connection between consanguinalxii and affinalxiii terms. Affinal terms are those that name relatives by marriage (‘in-laws’), whereas consanguinal terms denote all other relatives. Indeed, in this system there is no kinship term for a spouse before marriage and marriage takes place between persons not ‘biologically’ related. Another characteristic is that bifurcationxiv does not occur. We will revisit this term when considering the Dravidian system, but for now simply note that there is no terminological difference between mother’s brother and father’s brother or between mother’s sister and father’s sister. This can also be expressed in a more general way by stating that there is no structural difference between matrilinealxv or matrilateral kin (kin on the mother’s side) and patrilinealxvi or patrilateral kin (kin on the father’s side).

27The figure below depicts the generic Eskimo system, using the English terminology. Note that only two generations are shown: Ego’s generation (called by convention G-0) and Ego’s parents’ generation (called by convention G+1). The terminology shown below is thus not complete. You should note, however, that these two generations are the most important ones when distinguishing systems of terminology. With other generations, such as those of the grandparents, there are so many variations and so much overlapping that systematising these is impossible and, in fact, not relevant. We will return to some of these issues when we encounter concrete Australian examples.

Figure 11: The Eskimo system (example of the English terminology)

Figure 11: The Eskimo system (example of the English terminology)

28Several identical categories, or what are called equivalences, are apparent. In Ego’s parents’ generation, uncles and aunts constitute two such generic groupings and are distinguished from mother and father. The equivalences are expressed using the equal sign, while differences are expressed through the unequal sign.

29G+1: M ≠ MZ = FZ and F ≠ FB = MB

30The above line is necessary and sufficient to fully define an Eskimo system at the G+1 generational level: mother is not equivalent to mother’s sister but mother’s sister is equivalent to father’s sister; father is not equivalent to father’s brother but father’s brother is equivalent to mother’s brother. We will see that the equivalences in other systems are different. The line reads as follows: in the generation above Ego (Ego’s parents’ generation), a mother is not the same as a mother’s sister, but the latter is equal to a father’s sister (aunt); a father is not the same as a father’s brother, but the latter is equal to a mother’s brother (uncle). The same exercise can be done for Ego’s generation:

31G-0: B ≠ MBS = MZS = FBS = FZS and Z ≠ MBD = MZD = FBD = FZD

32In both these lines, we have assumed that gender is always distinguished, which is not the case in every system. If we wanted to be as complete as possible, we would write the first line as follows:

33G+1: M ≠ MZ = FZ ≠ F ≠ FB = MB

The Dravidian system

34Named after a people in India, this is a very widespread system, found on all continents and among the most diverse cultures, though it is typically associated with small-scale societies. At some level or other, this system is identifiable in all Australian Aboriginal kinship terminologies. Variations and adaptations occur, but they are overlays of a general schema that is Dravidian.

  • 2 ‘Canonical’, in mathematics, means something that is reduced to its most basic and simplest possibl (...)

35First identified by Floyd Lounsbury (1964), who distinguished it from the Iroquois system, it was later followed by descriptions of many variants and has been the object of numerous discussions about its canonical2 definition. We shall not go into these details since those features that interest us here are generally accepted.

  • xvii One of the particularities of Dravidian terminologies (see bifurcation). Bifurcate merging creates (...)

36The main feature of this system is what is called bifurcate mergingxvii, which means that categories are separated one generation above Ego (at G+1) according to gender: mother and mother’s sister are merged, father and father’s brother are merged; but mother is distinguished from father’s sister and father is distinguished from mother’s brother. Their children are merged back again in Ego’s generation (at G-0) in particular ways, as we shall see. ‘Bifurcation’ means change of gender with respect to Ego’s parents. Bifurcate merging is a straightforward procedure that distinguishes fathers, mothers, uncles and aunts but according to a different principle from that underlying the Eskimo system.

Figure 12: The Dravidian system, using English words : only very close kin are charted above, but it must be remembered that Dravidian terminologies are usually extended(see below). Also, this figure depicts only the generations of Ego (the speaker) and of his or her parents.

Figure 12: The Dravidian system, using English words : only very close kin are charted above, but it must be remembered that Dravidian terminologies are usually extended(see below). Also, this figure depicts only the generations of Ego (the speaker) and of his or her parents.

37Terms distinguish Ego’s ‘father’ from his ‘mother’. The father’s brother, being the same gender as the father, is also called ‘father’. Similarly, Ego’s mother’s sister, being female, is also called ‘mother’. The father’s sister, on the other hand, is of a different gender from Ego’s father and is called ‘aunt’ (or FZ, father’s sister). Similarly, the mother’s brother is called ‘uncle’ (or MB, mother’s brother). In other words, only the mother’s brother is an ‘uncle’ and only the father’s sister is an ‘aunt’ if we use the English words. This is the basic feature from which all others are derived in Dravidian systems. The Dravidian equations at Ego’s parents’ generation can be written as follows (the Eskimo equations are also added to underline the differences):

38Dravidian: G+1: M = MZ ≠ FZ and F = FB ≠ MB

39Eskimo: G+1: M ≠ MZ = FZ and F ≠ FB = MB

40The equations and distinctions produced in G+1 have direct consequences on the way people are classified in Ego’s generation. Since ones mother’s sister is called mother and since a mother’s daughter is a sister, ones mother’s sister’s daughter is a sister as well. Or, to put it (wrongly) in English: if you call your aunt mother, then you will call your aunt’s daughter a sister, since your ‘mother’ can only give birth to girls who are your sisters. This leads us to the second line of equations in the Dravidian system:

41G-0: B = FBS = MZS ≠ MBS = FZS and Z = MZD = FBD ≠ FZD = MBD

  • xviii Ones father’s brother’s children and ones mother’s sister’s children.
  • xix Ones father’s sister’s children and ones mother’s brother’s children. Cross-cousins are relevant in (...)

42Anthropologists therefore use specific terms to identify and label different types of cousins. Thus, children of FB (father’s brother) and MZ (mother’s sister) are called parallel cousinsxviii. In a Dravidian system, they will be called brothers and sisters. Bifurcation, meaning change of gender with respect to Ego’s parents, has not occurred. On the other hand, when bifurcation does happen, that is your father’s sister’s children and your mother’s brother children are called by the same term which is different from that used for your parallel cousins, then they are called cross-cousinsxix. Thus, bifurcation simply means that the difference is created when gender changes in the generation above Ego. This merging (remember the ‘bifurcate merging’ concept mentioned earlier) means two things: first, the fact that children of both your MZ and your FB are merged into the parallel cousin category. More importantly, it also means that children of your MB and of your FZ are themselves brothers and sisters and that their parents are therefore married. But to fully understand this, we need first to explain a few other things.

43Dravidian terminologies are often — and this is certainly the case in Aboriginal Australia — extended in such a way that every person with whom one has a relationship of any kind needs to be addressed or referred to by a kinship term. This extension to people other than close genealogical relatives follows this precise way of calculation we call ‘bifurcate merging’. I provide some examples below. My mother’s mother’s sister is a mother’s mother, but my father’s father’s sister is in a different category. You may remember that one of the particularities of the Eskimo system is that it does not contain a kinship term for a spouse whom you have not yet married. This is not the case in other systems. In a Dravidian system, since everyone is or can be addressed by a kinship term, your future spouse must also be addressable.

44What is important to know at this stage is that all the terms available to name all these people are listed in Figure 12 above. In the generation of Ego’s parents, these terms are mother, mother’s brother, father and father’s sister. All persons of G+1 must be called or referred to using one of these four basic categories. Similarly, all the persons belonging to Ego’s generation must be called or referred to using one of the available terms, which are sister, brother, female cross-cousin (MBD, FZD) or male cross-cousin (MBS, FZS). We conclude from this that ones future spouse must be a cross-cousin since you are not permitted to marry someone you call brother or sister. In principle then, your cross-cousins’ parents are also your parents-in-law and they themselves are married. Since the only terms available at G+1 are M, F, MB and FZ and since you cannot marry someone who is the offspring of a F or a M, it means that your parents-in-law must be a MB and a FZ; it also means that MB is the husband of FZ. It does not mean that a person’s actual mother’s brother is married to a person’s actual father’s sister, but it means that someone you call father’s sister is the spouse of someone you call mother’s brother. Remember that, as I have mentioned earlier and as I will show below, in Dravidian systems, kinship terms can be extended to include people who are genealogically very distant and even unrelated. There are, thus, actual father’s sisters and distant father’s sisters. This is the second part of the merging: your FZ is split away from your F and FB but reunited on the other side with your MB. We have the following schematic representation of these relationships: a FZ is married to a MB, and their daughter is Ego’s spouse.

Figure 13: Bifurcate merging in a Dravidian system (example of marriages)

Figure 13: Bifurcate merging in a Dravidian system (example of marriages)

45You should now understand more clearly why I stressed earlier the importance of distinguishing the use of these symbols and their analysis in a terminological context from their use in a genealogical context. In this chapter, we are talking about a terminological context: it is the terms people use for each other that lead us to establish the particular links displayed in Figure 13. It is rarely the case that an actual mother’s brother marries an actual father’s sister; it is also rare for a person to marry an actual first cross-cousin. This, however, is how people should address each other, whether they are closely related or not. You may be reassured to learn that even some of the most prominent scholars, among them LéviStrauss and Dumont, have consistently confused these two levels of social reality.

46Now that you know there is no limitation to the extension of range, that is all members of a group or society must be addressable through a kinship term and that the attribution of a kinship term to a person is not random but follows formalised rules, it is necessary to explain how people who are very distant relatives, and from a European point of view are not related, still nonetheless know how they should call or refer to each other. The underlying principle is that bifurcate merging is in a sense inherited over generations and generations. Let us take an example of two sisters.

Figure 14: Transmission of the bifurcate merging calculation without bifurcation

Figure 14: Transmission of the bifurcate merging calculation without bifurcation

47Each of the original two sisters depicted above has a daughter (their husbands are not represented). Because these offspring are children of sisters (same gender, no bifurcation), they call each other’s mother mother, and they therefore call each other sister. In the next generation, these sisters have boys. The boys call each other brother because they call each other’s mother mother. These brothers have children as well. Their daughters call each other sister because their fathers call each other brother. This may go on over hundreds of generations and although in time people would no longer know how they are genealogically linked, since, in Australia in particular, genealogical memory is usually quite shallow, they know how they are linked through the kinship terminology. The character of their relationship is passed on from generation to generation. Australian Aboriginal and other Dravidian-like kinship systems are structured in such a way that, from a formal point of view, society in theory descends from an original group of two brothers and two sisters in which each brother married the other’s sister. Thereafter the system ‘remembers’ how all descendants are linked to these original two couples. The same principle described for brothers and sisters is valid for all types of relationships and kin categories. The following figure adds some complexity to the picture by adding cross-cousins to the descendants.

Figure 15: Transmission of the bifurcate merging calculation with bifurcation (cc: cross-cousins; siblings : brother-sister)

Figure 15: Transmission of the bifurcate merging calculation with bifurcation (cc: cross-cousins; siblings : brother-sister)

48Columns are numbered from f1 to f4 (f stands for filiation; remember that a filiation is a parent-child relation). Rows are numbered from g1 to g4 (g stands for generation). This will make for easier reference. For example, the first woman top left is in filiation 1 and in generation 1. We call her f1-g1. The last woman bottom right is f4-g4.

49The figure shows that, from generation 2 onwards, the relationships between filiation 2 and filiation 3 are of the cross-cousin type. This is because f3-g1 is the brother of f2-g1 (change of gender) and the children of each become cross-cousins (marked cc in the figure). The father of f3-g2 is the mother’s brother of f2-g2. We have seen that mother’s brothers and father’s sisters are the parents of cross-cousins and not of siblings (brothers and sisters). This is where bifurcation takes place: lines of filiation bifurcate, or divide, because there is a change of gender.

50Even if you could follow up this genealogy for hundreds of generations, f2 and f3 would always stay in a cross-cousin relationship unless there was a change of gender.

51There is no need to know how these people are connected or how distant their genealogical links to each other are: the children of every person I call father’s sister (FZ) I will call cross-cousin.

52The woman f3-g2 had a boy and a girl (f3-g3 and f4-g3). The girl (f4-g3) has a girl herself (f4-g4). Because f3-g3 and f4-g3 are of different gender (they are siblings of opposite sex), bifurcation takes place again. The man f3-g3 is a mother’s brother for f4-g4, and the former’s children are therefore the latter’s cross-cousin.

53Another interesting feature is that the relation between f2-g4 and f4-g4 is not a cross-cousin relation, but a sibling (in this case sister) relation, because two bifurcations took place, one in filiation f2, the other in filiation f4. One happened in generation 1, the other in generation 3. A double bifurcation cancels itself out: the cross-cousin of my cross-cousin is my sister or brother. Again, you could go on and on: a triple bifurcation changes the relationship back into cross-cousins, a quadruple bifurcation into siblings, and so on.

  • xx Kinship labels that can be used for a whole category of persons (see descriptive). The English word (...)

54After hundreds of generations, people become very distant ‘relatives’, in theory just as distant as two European persons who are unknown to each other. Nevertheless, through the transmission of the bifurcate-merging principle in this classificatory systemxx, Aborigines know that they are related; they know in what type of relationship they stand to everyone else. A classificatory system with no limitation to the extension of the range of persons to whom kinship terms can be applied means that the terms are not limited to close kin but can be extended to all known persons and, theoretically, to humankind.

55The examples discussed above are formal illustrations of the working principles that lie beneath the Dravidian system. Reality is more complicated in some cases and more straightforward in others. Indeed, in real life, people usually have more children than in our illustration and they also have children of both genders, so that bifurcation takes place in every generation and filiation. The patterns of real genealogies considerably increase the complexities illustrated in our formal model, not forgetting that we should add marriage to our figures, a problem to which we shall return shortly.

56In real life, one does not have to memorise every person’s position within the bifurcate merging model or remember every person’s relation and position in the system. A complete overview of the genealogical grid, including generations of genealogical memory, is not necessary for the system to work and for people to live with it. In fact, what I have called the relational triangle (Dousset 2008) is sufficient for every person to know how he or she stands with regard to every other person. Let us take a look at this relational triangle, and show how Aboriginal people actually determine their relationships.

Figure 16: The relational triangle

Figure 16: The relational triangle

57The figure above illustrates two examples of relational triangles. In the first example, let us suppose two women meet. The first woman is Ego (you), the second is X. They need to find out their relationship to each other but because they are genealogically very distant there is no way they can trace it. What they will do therefore is find a third person both of them know their relationship to. This third person in our example is Y, a man. Let us suppose Ego’s relationship to Y is MB (a mother’s brother; not an actual MB, as you understand now, but a distant or classificatory one). The woman X’s relationship to this same man Y is of the daughter-type, that is X is a classificatory daughter of Y. Let us recapitulate: Ego calls Y mother’s brother, and Y calls X daughter. Ego and X therefore now know that they stand in a cross-cousin relationship to each other because a mother’s brother’s daughter is a cross-cousin, and the problem is solved.

58The second example is a little more complicated. Ego calls Y mother and this mother calls the man X cross-cousin. Because a person can only marry cross-cousins and not siblings, X’s cross-cousin Y is also a potential spouse and ones mother’s spouse is obviously ones father. Ego would therefore call the man X father.

59We will often return to features that have been explained in this chapter, in particular when discussing marriage and describing Aboriginal Australian kinship systems. Before then, some further comments are warranted on the other major kinship types recognised by kinship scholars.

Crow-Omaha, Sudanese and Hawaiian systems

60It should now be obvious why I have focused most closely on the Eskimo and Dravidian systems. The Eskimo system is understood by people of Euro-American background working in Aboriginal communities; and the Dravidian system is heavily represented in Aboriginal Australia. Since the most likely readers of this book are Aboriginal people and others interested in Aboriginal culture, the choice was obvious. However, some of the other systems are found in one form or another in Australia, so it is necessary to briefly refer to them here.

61The Sudanese system is the most descriptive. As far as I know, this system is not represented in Australia. It is however, in some respects, nearly an opposite example of the Dravidian system. Mentioning it briefly here will contribute to our understanding of the distinction between descriptive and classificatory systems or terms. Anthropologists, starting way back with one of their ancestors, Lewis Henry Morgan, make this distinction between classificatory and descriptive terminologies. Classificatory terminologies denote with a single term several categories or classes of people (such as uncle in the English terminology) while descriptive systems have the characteristic of naming every category of kin differently. The term for cross-cousin in a Dravidian system is typically a classificatory term, while that for MB is a descriptive term. In fact, it is not possible to distinguish entire systems as being either descriptive or classificatory since all have some elements of description and some elements of classification. The Eskimo system discussed above, for example, was long considered to be a descriptive system, but in fact it has classificatory terms since the word cousin, for example, covers four categories of people: matrilateral cross-cousins (MBD and MBS), matrilateral parallel cousins (MZD and MZS), patrilateral cross-cousins (FZD and FZS), and patrilateral parallel cousins (FBD and FBS). The Sudanese system though is one of the most descriptive systems, if not a totally descriptive one. Every single kin category is named using a distinct term. If this were to be applied to the English terminology, then the uncle on the mother’s side and the uncle on the father’s side would be distinguished.

62The Hawaiian system, also called a generational system, is in some respects the simplest because it distinguishes very few categories of kin. As we shall see, it is present in Australia in certain contexts. However, it is predominantly used among Polynesian societies of the Pacific. In this system, in Ego’s parents’ generation all women are called mother and all men are called father. Thus, all people of ones own generation who are children of people one calls mother or father are brothers and sisters. It is important to note that Hawaiian systems can be further subdivided into two sub-system types. Because in a Hawaiian system Ego only finds brothers and sisters among persons of his own generation and because he or she is supposed to marry someone from this generation, possible spouses for Ego need to be distinguished other than by terminology alone. There are two solutions. The first is to limit the use of terminology to very close kin and to apply a strict rule of exogamy (the necessity to marry out into genealogically or spatially distant families). This solution limits the extension of range of the kinship terminology. The other solution, when the terminology is used even among genealogically distant kin, is to somehow differentiate the categories of cross-cousins and of siblings as in a Dravidian system, even though before marriage they are all designated using the sibling terminology. In the latter solution, the terminology is of the Hawaiian type but the underlying system and the marriage rules follow the Dravidian type where cross-cousins are potentially also spouses. The Hawaiian terminology should not be seen as an independent system but as an overlay of a system that is fundamentally Dravidian. This is the case in the Western Desert of Australia where in some situations people use a Hawaiian type of classification while in others they use a Dravidian type. The distinction between pre-marriage terminology and post-marriage terminology is relevant here. I shall return to this point when discussing Western Desert systems.

63Crow and Omaha are sometimes considered distinct systems. However, in Australia at least, in most cases they are more likely to be interpreted as variations of the Dravidian system (Godelier 2004; Needham 1971). As Kronenfeld (1973, 1991) shows for the Fanti people of Africa and Rumsey (1981) for the Ngarinyin of north-west Australia, Crow and Omaha features are particular terminologies used in particular contexts. In other, more ‘usual’ contexts, a Dravidian system is used. For Australia, McConvell (nd) identifies Crow-Omaha features among a few groups in the northern part of the Kimberley (Ngarinyin people), in south-east Arnhem Land (Nunggubuyu and others), in Central Australia (among the Gurindji and Warumungu), in the Gulf of Carpentaria among the McArthur River people (such as the Yanyuwa), and in the Cape York Peninsula (Gugu Yimidhirr and Ayabadhu). Refer to Figure 2 to locate these groups on a map.

  • xxi A particularity of Omaha and Crow systems of kinship reckoning. Persons from different generational (...)

64The particularities of the Crow and Omaha systems are that they operate what is called skewingxxi. In certain contexts, cross-cousins are called by the same term as a person from another generation, as if, for example, you called one of your cross-cousins ‘uncle’ in English. In other words, two persons of two different generations are skewed into one and the same kin category.

65The equations that can be constructed for these systems are no longer generational, but cross over several generations (note that not all equations are used systematically):

66Omaha: MZ=MBD=MBSD and/or MF=MB=MBS

67Crow: FM=FZ=FZD=FZDD and/or F=FMB=FZS=FZSS

Figure 17: Examples of possible Ohama and Crow skewing. The persons marked in yellow in each figure are called or referred to by Ego using the same word

Figure 17: Examples of possible Ohama and Crow skewing. The persons marked in yellow in each figure are called or referred to by Ego using the same word

68Note that in these systems, skewing is either on the father’s side (Crow) or on the mother’s (Omaha), never on both. Marriage still occurs with a cross-cousin, as in a Dravidian system. However, since one of the cross-cousins is skewed, they disappear as a potential partner. In other words, in Crow systems with patrilateral skewing, the potential spouse must be matrilateral; in the Omaha system with matrilateral skewing, the potential spouse must be patrilateral. This marriage pattern is called asymmetrical cross-cousin marriage, which is one of the features of Crow-Omaha systems. As we will see below when discussing marriage, these systems are also called semi-complex or indirect exchange systems.

Other conventions concerning kinship terms

  • xxii Words that people use when directly addressing other kin (e.g., ‘dad’) (see Terms of reference).
  • xxiii Words that people use to refer to kin (e.g., ‘father’) (see Terms of address).

69As we have seen, kinship always comes with a specific terminology. A terminology is the collection of words used to address or refer to persons considered to be kin. Each term typically refers to a specific type or category of kin. Anthropologists distinguish between terms of addressxxii and terms of referencexxiii. Father is a term of reference, dad a term of address. It is only in particular contexts that one addresses ones father as father; generally, ones father is addressed as dad, but referred to as father. The sentence ‘I call my father dad’ makes sense because, while the terms refer to the same person, they really mean something slightly different. Many kinship terms are address and reference terms at the same time, such as the English cousin or uncle. This is the case in the Australian Western Desert where most terms are both terms of address and reference.

  • xxiv Terms of address or reference between two persons who are structurally identical (e.g. ‘cousin’ in (...)

70Another qualifying distinction made for words of kinship is whether they are self-reciprocalxxiv or not. A self-reciprocal term is a term applied by both persons, the one who is being addressed and the one who does the addressing; for example, the word cousin, which does not distinguish gender, age, or type of cousin. The English brother is another such selfreciprocal term, but only in the case of two men so related who are talking to each other. The word does not, for example, distinguish if one brother is older than the other. On the other hand, the English mother is not self-reciprocal since the person who calls someone mother will be addressed or referred to by this ‘mother’ with another word: daughter.

71Finally, anthropologists distinguish between descriptive and classificatory terms (and even systems), as we have seen. When the use of kinship terms is limited to very few people, that is when the kinship terminology is not extended to distant or unrelated people, as in the Eskimo system and European languages, or when every possible kin category has its own specific term, such as in the Sudanese system, the terminology is called descriptive. Every word describes one very specific relationship and one person only: a mother can only be the person who gave birth to you. The word mother therefore defines a relationship as well as a specific person. Terminology and genealogy are here identical. Mother is therefore, in the Eskimo system, considered a descriptive term of reference. Mum is a descriptive term of address.

72This is not the case in Australian kinship systems nor in fact in most systems of the world. These systems are classificatory or have many classificatory terms because kinship terms are not limited to close kin but can be extended to all known persons and, theoretically to humankind. The word for ‘mother’ may be applied to persons other than ones birth mother. In a Dravidian system, as we have seen, your mother, mother’s sisters, mother’s parallel cousins and any other person your mother calls sister are in the ‘mother’ category and can or should be named mother. Hence, mother is a classificatory term because it places a certain number of persons in the same class or category of kin. The most important distinction between descriptive and classificatory terminologies and systems is that in a descriptive system terms usually describe a person, whereas in a classificatory system terms describe a type of relationship.

73Researchers did not always agree on the concepts of descriptive and classificatory types of kinship, particularly those who suggested that classificatory systems were older or ‘primitive’ while descriptive systems were thought to be more modern and ‘advanced’. These theoretical positions, held in the 19th century by scholars such as Lewis Henry Morgan, soon became untenable. I shall not discuss the moral and scientific grounds on which certain systems came to be considered ‘primitive’ or others ‘advanced’ since there is little evidence to support either category. What made the distinction between descriptive and classificatory systems shaky was that, as Alfred Kroeber (1909), for example, showed, all systems have both classificatory and descriptive features, as we have already seen.

74Australian kinship systems, which are supposed to be classificatory systems, have a significant number of descriptive features. In the Western Desert, for example, the word for ‘brother-in-law’ is usually applied to ones actual brother-in-law only and cannot be applied to another person. So although the notion of classificatory and descriptive terms is useful since it allows us to distinguish terms that are applicable to various persons from terms that are closely related to ones genealogy, we cannot talk of classificatory or descriptive systems since most if not all kinship terminologies have simultaneously descriptive and classificatory terms.

Filiation and descent

75After discussing some general aspects of kinship terminologies, we now move on to the second stage of acquiring the necessary concepts for understanding kinship: filiation and descent.

76While terminologies are among the most important elements one needs for comprehension in the field, they are not the only aspects that concern the network of relatedness. The emergence of the British structural-functionalist school early in the 20th century, with scholars such as Malinowski, Radcliffe-Brown, Fortes, Evans-Pritchard and others, resulted in an increasing concentration of research on the relationship between kinship and political and economic systems. How people were recruited into corporate groups and how these groups also become landowners or land users, for example, were recognised as important to an understanding of human societies.

  • xxv In this process of socialisation, the individual comes to regard what has been inculcated as normal (...)
  • xxvi A group of people, or a number of lineages, whose members claim to be descendants of a distant ance (...)
  • xxvii A group of people who consider themselves to be identical in some respects because they are descend (...)

77Anthropologists became particularly interested in notions such as lineagexxv or clanxxvi, which are both so-called descent groupsxxvii. Early on, researchers recognised that, in both kinds of descent groups, the main mode of recruiting people (that is of selecting who can or must be part of a lineage or clan), was the control of descent and the transmission of membership from one generation to the next. As with terminological systems, anthropologists identified several kinds of such modes of transmission. I quickly explain each of these here, but first we need to recall what we understand by ‘filiation’ and ‘descent’. Filiation is the relationship between a parent and his or her child. However, it does not mean this relationship alone but includes the values associated with the relationship. For example, in some societies a child’s blood is believed to be inherited from the mother while the bones are considered to be inherited from the father. We thus have a double filiation in this case: blood from the mother, or matrifiliation; and bones from the father, or patrifiliation. Descent is when these principles of filiation are extended over generations and when, to take the same example, a man considers himself to have the same bones as his father, his father’s father and so on. In many instances, this group of men, including all of their actual brothers, are considered to be a descent group and thus constitute a corporation, in the form of either a lineage or a clan. Obviously, the same would be the case for female members. Of course, descent ramifies so that brothers or sisters constitute their own descent group, each with their separate descent line.

The British structural-functionalist school

This school of thought emerged in the 1930s in opposition to the theory of unilineal evolution advocated by Henry L. Morgan and others. One of the major changes it brought to anthropology was the insistence on the necessity of long term fieldwork. This was particularly advocated by Bronislaw Maloniwski who forged the expression ‘participant observation‘ and who had done fieldwork among the Trobriand Islanders in the Pacific (Malinowski 1932, for example). Besides Malinowski, one of the most prevalent structural-functionalist scholars was Alfred Radcliffe- Brown who had worked in the Andaman Islands and Australia. He was particularly interested in kinship and its relationship with social organisation and land tenure. Meyer Fortes (1945 for example), who had worked among the Tallensi and Ashanti in Ghana, Africa, was an important figure in defining the methodologies for comparative anthropology. He also coined the expression ‘complementary filiation‘ to indicate that even in unlineal systems (see below) there are characteristics which complement this unilineality. Another prominent figure was Edward E. Evans-Pritchard. His work on the Nuer (Evans-Pritchard 1937 and 1940, for example) is still today considered among the most important monographs in anthropology.Structural functionalists approached society ina holistic way and thus followed the French scholar Emile Durkheim’s approach in this respect: society is an integrated and more or less self-contained system which reproduces itself in time and space. Social institutions, such as kinship, technology, religion etc. constitute structural elements that function together to build a coherent ensemble: society.

  • xxviii A descent principle or group (see descent group) in which descent is in one line only, through the (...)
  • xxix A descent group (also see descent group, the opposite is ‘unilineal’) that draws its origins from b (...)

78Three general types of descent systems have been identified: unilinealxxviii, non-unilineal and mixed descentxxix systems. As the name indicates, unilineal systems trace the process of incorporation, inheritance and transmission through a single line: the father’s or the mother’s line, but not both. In a patrilineal descent system, the Baruya of Papua New Guinea for example, membership of a group is inherited through the male line. Children belong to the same group as their father but to a different group from their mother. In a matrilineal descent system, the Navajo of North America for example, membership of a group is inherited through the female line and children belong to the same group as their mother but to a different group from their father. In the latter, the matrilateral uncle (the mother’s brother) usually plays a very important educational role for his sister’s children and marriage bonds are usually quite weak. Unilineal systems are most often associated with clans and lineages. A lineage is a group of people who descend from a single known human ancestor while a clan is a group of people, or a group of lineages, who claim descent from one mythical ancestor. Lineages and clans are often landowning units. As I have already stressed, there are no clans or lineages in the Western Desert and no landowning kin groups as such.

  • xxx In cognatic descent, people consider they are descended from both their mother’s and their father’s (...)
  • xxxi A kind of unilineal system (see below) in which the individual can choose which side (that of the m (...)

79Among non-unilineal systems, the most representative is the cognatic descent systemxxx. Some anthropologists interpret all non-unilineal systems as being cognatic and differentiate within cognatic systems the bilateral and the ambilineal systemxxxi. Others, such as myself, talk of bilateral systems and of cognatic systems as being two words for the same thing but differentiate the ambilineal system (see below). We do not need to go into these discussions here, simply to note that non-unilineal systems are generally considered unable to sustain enduring corporate groups such as lineages or clans. Indeed, in a bilateral system, children belong indistinguishably to their mother’s group or family as well as to their father’s group or family and inherit rights and property indistinguishably from their father’s and their mother’s side. In practice, societies with cognatic descent systems can constitute enduring corporate groups through other mechanisms. In certain parts of eastern Europe, where the bilateral system is the rule, the norm of primogeniture also existed, according to which the first-born male would inherit landownership and constitute, with his close and enlarged family, a land-based enduring group. He in turn would transmit landownership to his eldest son and so on. Younger brothers had to leave the family group or work under the authority of the older brother without ever becoming an owner. Thus, while in theory and in discourses children inherit goods and rights from both their parents’ sides, in practice one side may be considered by people as being central and the other peripheral, as we shall soon see with the concept of complementary filiation and double descent.

80Among the mixed systems, the ambilineal system is probably the most difficult to identify, but it is also very interesting since it is open to political and economic contextualisation and strategies. In an ambilineal system, a person may choose to follow the mother’s or the father’s line, depending on opportunities or social pressure. However, once he or she chooses to link up to one or the other side, this decision may not be changed later in life; a person is expected to stand by the group or family from which she or he claims inheritance. In the Australian Western Desert it is possible to speak of an ambilineal system (Layton 1983) but, unlike that in Polynesia where ambilineal groups constitute more or less enduring corporations called ‘ramages’ (Firth 1957), this is of little consequence for defining landownership or membership of distinguishable units.

  • xxxii In double descent systems, people inherit their belonging or substances from both the mother’s and (...)
  • xxxiii In unilineal systems, people inherit one major characteristic from either their mother’s or their f (...)

81Another mixed system is double descentxxxii. People may trace their descent in different ways, depending on what is transmitted, or different people may trace descent in different ways. Such is the case among the Mundugumor of Papua New Guinea where girls and boys do not follow the same rule. Girls belong to their father’s group and boys to their mother’s. In other societies, such as the Apinayé of Brazil, descent lines are gender specific: women trace descent matrilineally, men patrilineally (see Holy 1996: 121 and following pages for a discussion of double descent systems). Descent describes the principles inherent in the transmission of material and non-material things, rights and duties over generations. Filiation, on the other hand, is the principle that links a person to his or her immediate parents only. On many occasions, filiation and descent produce similar effects, such as in the case of the transmission of landownership through the patriline. In many other cases though, there are additional criteria that establish a relationship to a parent in a different way, without constituting a corporate group. Fortes (1959) has coined the expression complementary filiationxxxiii to describe these additional factors and mechanisms. All of these concepts will be revisited when we discuss actual examples of Aboriginal Australian kinship pattern and behaviour.

Marriage and marriage alliances

82We now turn to the last formal, theoretical domain that requires elaboration before moving on to more ethnographic data: marriage. The central concept when discussing marriage is that of exogamy: the obligation to marry outside of a family, a group, a clan, a tribe and so on. The importance of marriage and exogamy was recognised by anthropologists even before the word ‘exogamy’ itself had been proposed by McLennan (1970 [1865]). Indeed, Tylor (1889: 267), recognising the political aspects of marriage, summarised them in a wellknown formula according to which ‘tribes must have had before them the simple practical alternative between marrying out or being killed-out’. Establishing links with other groups, in particular through marriage, and thus introducing the crucial notion of the exchange of human beings between groups was seen by Tylor and many of his successors as a vitally important means of reproducing society as a whole in time and space and not just human beings. The exogamy of one group supposes the exogamy of another and thus links these two groups in terms not only of political alliance but also of shared genealogical and thus ancestral history.

  • xxxiv The interdiction to marry a certain category of persons (see Prescriptive rule).
  • xxxv The obligation to marry certain categories of persons (see Proscriptive rule).

83Marriage is intimately related in several ways to the two domains discussed above: terminologies and principles of descent. In particular, we must assume that everywhere, in each society or cultural background, marriage follows certain rules and norms. Put negatively, nowhere can people marry just anyone they like. There are always restrictions on marriage with certain persons; these we call proscriptive marriage rulesxxxiv. There are also in many cases preferences, and even obligations, to marry particular categories of persons; these we call prescriptive marriage rulesxxxv, which can be jural (Leach 1965) or structural (Needham 1973).

84By ‘jural’ we mean that there are explicit rules and norms that define obligations of marriage. ‘Structural’ means that these rules flow from the kinship terminology itself; for example, in cases where particular persons are called spouse in the terminology and where it is obviously not possible to marry someone from another category. In most cases, jural and structural rules overlap since the structural obligation to marry into a particular kinship category is usually also associated with explicit discourses and explanations. This is the case in Aboriginal Australia.

85Proscriptive and, to some extent, prescriptive marriage rules are related to the prohibition of incest and to the notion of exogamy. We now discuss these concepts and their relationship to terminologies and descent.

86Anthropologists have always been very interested in marriage because it is about bonding individuals, families and groups and reflects the dynamics of social strategy, interaction and structure. The birth of structuralism in anthropology under Claude Lévi-Strauss led to it being theorised as one of the leverages of a new theoretical proposition. According to this structuralist theory, society is not (or not solely) made up of principles of belonging (descent and groups) but primarily of principles of exchange. Lévi-Strauss (1967 [1947]) claimed that three types of exchange characterise the human social realm: exchange of words (language), exchange of goods (the economic domain) and exchange of human beings (marriage and kinship). Only when these three principles of exchange are systematised and functional in a group of human beings does the latter constitute a society with shared cultural codes. Alliance theorists, as the name indicates, together with Lévi-Strauss and many other anthropologists, concentrated their research efforts on marriage, the domain in which the exchange of human beings occurs. Marriage becomes a system of exchange if it is associated with exogamy (the obligation to marry someone from ‘outside’ your family or group). In these terms, brothers and sisters who, over generations, marry each other will not constitute a society. To do so, Lévi-Strauss argues, we need an incest prohibition, which compels people to obtain their spouses from other families. In his view, this prohibition allows the passage of a group of people from a state of nature to a state of culture.

  • xxxvi The reproduction over time and generations of identical marriages; that is, for example, brothers e (...)

87What is the difference between ‘alliance’ and ‘marriage’? Dumont (1968 [1957]), another alliance theorist, who in some respects disagreed with Lévi-Strauss, has conceptualised the difference in the most systematic way. He speaks of alliancexxxvi (or marriage alliance) when he observes and analyses the repetition of identical marriage types over generations or among people of the same generation level. In theory, this repetition occurs in Dravidian systems. Thus, marriage is the individual event occurring in a particular place with particular people in a particular context; it brings together two people (and families) with the aim of uniting them as spouses and future parents. Alliance as a system reveals patterns of regularity in the choice of suitable spouses and describes repetitions of identical marriage types over generations.

  • xxxvii Two men or women marrying each other’s sisters or brothers. This exchange is linked to Dravidian-ty (...)
  • xxxviii A complex version of direct exchange (see direct exchange) where at least three groups need to exch (...)
  • xxxix A marriage system in which there are no prescriptive marriage rules (see prescriptive rule). This i (...)

88Lévi-Strauss (1967 [1947]) and other alliance theorists distinguish three basic types of marriages: direct exchangexxxvii, the most basic or elementary system; indirect exchangexxxviii; and complexxxxix systems. In the direct exchange system, marriages place people and groups in a symmetrical relationship. Females of group A, for example a clan, marry males of group B and females of group B marry males of group A. Group A thus exchanges women or men with group B in a direct and reciprocal way. This exchange system can be linked to the terminological systems mentioned earlier. Most often, direct exchange systems are found where the terminology is of the Dravidian type. As we have seen, a mother’s brother (MB) marries a father’s sister (FZ) and their offspring are cross-cousins who are also potential spouses. Thus, if we combine this with patrilineal descent groups as mentioned above, we see that your father and your mother’s brother are in different groups that exchange their sisters or daughters. Let us illustrate this over two generations.

Figure 18: Direct exchange model

Figure 18: Direct exchange model

89For example, let us say that Henry and his sister Jane are members of the black crow clan (orange in the figure) and that James and his sister Deborah are in the white cockatoo clan (blue in the figure). People of the black crow cannot marry other people of the black crow (proscriptive rule) since they are brothers and sisters, so they need to marry people from another clan (exogamy), in this case the white cockatoos (prescriptive rule). The same is true for people of the white cockatoo clan. They marry people from the black crow clan. Thus, Jane marries James and Deborah marries Henry. In the eyes of alliance theorists, Henry and James have exchanged sisters.

90Henry and Deborah have a son Patrick and a daughter Dora. Both are of the black crow clan because in our case descent is patrilineal so children inherit their clan membership from their father. James and Jane also have two children: Arnold and Sarah. Both are of the white cockatoo clan, the same as that of their father, James. Patrick and Dora must marry people from another clan (exogamy), in this case the white cockatoos. The same is true of Arnold and Sarah. Thus, Dora marries Arnold and Sarah marries Patrick. Again, we see that Patrick and Arnold have exchanged sisters (or equally, that Sarah and Dora have exchanged brothers). We could follow such exchanges over generations, but the principle that operates always remains the same: in a direct exchange system, people marry a cross-cousin. When crosscousins consistently marry, whether they are actual cross-cousins or genealogically distant cousins, in this repetition of structurally similar marriages we see marriage alliance working as defined above. Thus, alliance does not just operate between two groups but is iterated.

91The second marriage type is ‘indirect exchange’, also called a semi-complex system, which is found in Crow and Omaha terminological systems (see above). Marriage in these cases is either patrilateral or matrilateral, meaning that a person can marry a cross-cousin from either the father’s or the mother’s side, but not from both. This is so because the cross-cousin on one side is related to you as your mother’s brother or father’s sister and you would not marry someone you call aunt or uncle. This means that you always marry ‘on the same side’, and, in terms of clans or descent groups, it also means that there are at least three exchanging units. The marriage pattern occurring in these systems is also called asymmetrical cross-cousin marriage. Figure 19 illustrates a simplified version of indirect exchange: women of the white cockatoo clan (orange) marry men of the black crows (blue); women of the black crows marry men of the brown eagles (yellow); and women of the brown eagles (yellow) marry men of the white cockatoos (orange).

Figure 19: Indirect exchange system

Figure 19: Indirect exchange system

92The least systemic exchange system is called the ‘complex system’. It is usually associated with Eskimo kinship terminologies. Marriage is here no longer a system of exchange on its own but merges into other types of exchanges, social structures and ideologies. The class of marriageable persons is not defined in complex systems, except as non-kin (with the exception of some cousins being marriageable in certain complex systems). This is why it is called ‘complex’: kinship alone is not sufficient to understand the rules of marriage. An example of such a system is contemporary European marriage, where alliances (repetitions of identical marriages) are very rare since children of a single family may marry in different ways and into the most diverse family backgrounds. Here, there is only a proscriptive marriage rule, which is concerned with the incest prohibition, but there are no formalised prescriptive rules. Statistically, of course, there are some regularities even among complex systems, but they are relevant for social factors other than rules of marriage alone. For example, people of a certain social class tend to marry into the same social class and people in rural areas tend to marry following strategies that retain the coherence of land ownership over time.

Summary and Conclusion of Part Two

93Rather than defining kinship in absolute terms, which would be a complex enterprise as anthropologists may have a wide variety of definitions, I have presented and discussed the principal domains of investigation in kinship studies, gradually defining what kinship is all about. After the presentation of the usual abbreviations, symbols and conventions used, three major domains have been illustrated: terminology, descent and marriage. These three aspects of kinship have continued to attract considerable attention throughout the history of anthropology.

94Every language and culture has a kinship terminology, that is words that belong to the particular domain of addressing or referring to relatives using a biological idiom. However, while there is considerable diversity in terminologies, they can also be categorised in ways that offer significant insights into how people interact in particular societies or groups. The Eskimo and the Dravidian systems have been discussed with some depth since the former is the European type of system while the latter is that generally found in Aboriginal Australia. There are clear differences between these two system types with regard to how kin relationships are calculated and to the extension of the applicability of kinship terms. In Aboriginal Australian Dravidian systems, there is no limitation to the extension of range and there are systematic ways of calculating everyone’s relationship to everyone else. What is crucial to remember for Dravidian systems is that they define potential spouses as being in the cross-cousin category while the Eskimo system does not define any category of persons who can become a spouse. However, while the Dravidian terminology defines these spouses, it is important to remember that a cross-cousin is not necessarily a close relative but can be genealogically distant. The cross-cousin position can be determined through the relational triangle. Terminology and genealogy are two different domains that need to be clearly distinguished.

95We then briefly discussed the notions of filiation and descent. Filiation is the relationship between parents and their children while descent is the systematisation of principles of filiation over generations. In many cases, rules of descent create groups of people who share certain characteristics, such as common substance or history. These principles can be of various kinds: patrilineal (or patrifilial), matrilineal (or matrifilial), ambilineal, cognatic and double descent. In Aboriginal Australia, patrilineal and matrilineal descent systems are the most common although the Western Desert system is best seen as an ambilineal system.

96Terminology and descent are closely linked to marriage. Despite issues of personal affinities and choices, all societies have marriage conventions or rules. The basic proscriptive marriage rule is the incest prohibition, forcing people to marry out of their own family or group. In most societies there are also prescriptive marriage rules: Ego is supposed to marry into a certain category of persons. In the Australian Dravidian system, these prescriptive marriage rules are usually jural and structural: they are what people expect you to do, but they can also be deduced from the kinship terminology itself, where only certain people can be labelled spouses.

97Marriage is also understandable as a form of exchange of people between groups. Three basic types of exchange have been discussed: direct exchange, that which in most cases is applicable in Australia; indirect exchange, which is applicable for example in CrowOmaha (or skewed) terminologies; and complex exchanges where marriage is tied to other social domains such as politics, economical conditions and social class.

Bibliographie

Further Reading

To learn more on the history and particularity of kinship studies

FOX, Robin 1996. Kinship and Marriage. Cambridge: Cambridge University Press [1967].

HOLY, Ladislav 1996. Anthropological Perspectives on Kinship. London, Chicago: Pluto Press. STONE, Linda 1997. Kinship and Gender:An Introduction. Boulder: Westview Press.

Contemporary studies of kinship

CARSTEN, Janet (ed.) 2000. Cultures of Relatedness: New Approaches to the Study of Kinship. Cambridge: Cambridge University Press.

McCONVELL, Patrick, Laurent DOUSSET and Fiona POWELL (eds) 2002. Kinship and Change in Aboriginal Australia. Anthropological Forum 12(2).

SCHWEITZER, Peter P. (ed.) 2000. Dividends of Kinship. London and New York: Routledge, European Association of Social Anthropologists.

STONE, Linda (ed.) 2001. New Directions in Anthropological Kinship. Oxford: Rowman & Littlefield.

Notes

1 An ‘idiom’ is a special way of speaking that is considered ‘natural’ to a speaker of a particular language.

2 ‘Canonical’, in mathematics, means something that is reduced to its most basic and simplest possible expression but remains valid. A ‘canonical’ expression is a rule or a definition that is considered valid wherever it is applicable.

Notes de fin

i A set of words used to address or refer to people with whom one is related through a biological idiom.

ii A group of people, or a number of lineages, whose members claim to be descendants of a distant ancestor, who is usually mythical (also see lineage).

iii In this process of socialisation, the individual comes to regard what has been inculcated as normal and natural.

iv The systemisation of certain principles of filiation (see filiation).

v The particular relationship between children and parents who may share particular substances, rights or duties.

vi The union of two individuals, which brings together two families or groups. Marriage is always subject to rules and has to be distinguished from sexuality (also see Prescriptive and Proscriptive rules).

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

viii In many domains, in particular language, this rule reflects the fact that social processes and structures tend to strive towards the elimination of redundancy.

ix In popular terms, a ‘family tree’. It is a chart that reconstructs an individual’s total known kin relationships to ancestors, contemporaries and descendants.

x In order to be meaningful in anthropological terms, every kin category or every person in a genealogy needs to be linked to at least two other persons.

xi The range of people towards whom kinship terms can or must be applied, which in some classificatory systems can be extended indefinitely.

xii The kinship relationships that exist for a person before marriage (see affinal). In English, mother, father, brother etc. are consanguinal terms while wife, brother-in-law, mother-in-law etc. are affinal terms.

xiii Kinship relationships that come into being when marriage is envis-aged or has occurred. Affinal relationships and terms include all inlaws. Affinal terms and consanguinal terms together constitute the entire terminological system.

xiv A particularity of kinship terminologies that denotes the differentiation between maternal and paternal relatives and which takes gender difference into account.

xv Those relatives or principles that are on ones mother’s side.

xvi Relatives on ones father’s side, and principles relating to descent in which males are the central point of reference.

xvii One of the particularities of Dravidian terminologies (see bifurcation). Bifurcate merging creates the distinction between crosscousins and parallel cousins (see cross and parallel cousins). A mother’s brother is distinguished from the father and from the father’s brother, and a father’s sister is distinguished from the mother and mother’s sister. A mother’s brother and a father’s sister are spouses and their offspring are affines.

xviii Ones father’s brother’s children and ones mother’s sister’s children.

xix Ones father’s sister’s children and ones mother’s brother’s children. Cross-cousins are relevant in systems with bifurcate merging, such as in Dravidian systems. In the later systems, cross-cousins are potential spouses. Cross-cousins are distinguished from parallel cousins.

xx Kinship labels that can be used for a whole category of persons (see descriptive). The English word cousin is a classificatory term since it covers all mother’s and father’s siblings’ children, irrespective of whether they are male or female or of whether they are on ones father’s or mother’s side. Most systems have both classificatory and descriptive terms.

xxi A particularity of Omaha and Crow systems of kinship reckoning. Persons from different generational levels are conflated (skewed) into one category only.

xxii Words that people use when directly addressing other kin (e.g., ‘dad’) (see Terms of reference).

xxiii Words that people use to refer to kin (e.g., ‘father’) (see Terms of address).

xxiv Terms of address or reference between two persons who are structurally identical (e.g. ‘cousin’ in English).

xxv In this process of socialisation, the individual comes to regard what has been inculcated as normal and natural.

xxvi A group of people, or a number of lineages, whose members claim to be descendants of a distant ancestor, who is usually mythical (also see lineage).

xxvii A group of people who consider themselves to be identical in some respects because they are descendants from one real or mythical individual (also see lineage and clan).

xxviii A descent principle or group (see descent group) in which descent is in one line only, through the mother or the father. Also see matrilineal and patrilineal.

xxix A descent group (also see descent group, the opposite is ‘unilineal’) that draws its origins from both the mother’s and the father’s side.

xxx In cognatic descent, people consider they are descended from both their mother’s and their father’s side (also see mixed descent).

xxxi A kind of unilineal system (see below) in which the individual can choose which side (that of the mother or the father) he or she wants to belong to. Once the choice has been made (either by the persons themselves or imposed by others), it cannot be changed.

xxxii In double descent systems, people inherit their belonging or substances from both the mother’s and the father’s side, but they do not inherit the same things from each.

xxxiii In unilineal systems, people inherit one major characteristic from either their mother’s or their father’s side, but may inherit complementary elements from the other side. For example, among the North African Touareg belonging to lineages as well as the house or tent is inherited through the female’s line, but the animal herd is inherited from the father.

xxxiv The interdiction to marry a certain category of persons (see Prescriptive rule).

xxxv The obligation to marry certain categories of persons (see Proscriptive rule).

xxxvi The reproduction over time and generations of identical marriages; that is, for example, brothers exchanging their sisters and their sons again exchanging their sisters etc. Alliance of marriage or marriage alliance occurs in theory in Dravidian systems. In practice, repetitions of identical marriages only rarely occur and what is called ‘shifting webs’, that is the diversification of marriages, is much more frequent.

xxxvii Two men or women marrying each other’s sisters or brothers. This exchange is linked to Dravidian-type terminologies and is reproduced over time, resulting in an alliance of marriage.

xxxviii A complex version of direct exchange (see direct exchange) where at least three groups need to exchange women or men for marriage. This system is found in Crow-Omaha terminologies.

xxxix A marriage system in which there are no prescriptive marriage rules (see prescriptive rule). This is the marriage system at work in Euro-American societies. In these societies, it is not the kinship system itself that determines potential partners but rather other characteristics such as belonging to the same social class.

Table des illustrations

Titre Figure 7: Schematic view of the domains of kinship and the evolution of anthropology
URL http://books.openedition.org/pacific/docannexe/image/562/img-1.png
Fichier image/png, 80k
Titre Figure 8: The basic symbols representing individuals in genealogies
URL http://books.openedition.org/pacific/docannexe/image/562/img-2.png
Fichier image/png, 3,7k
Titre Figure 9: Basic conventions representing relationships.
URL http://books.openedition.org/pacific/docannexe/image/562/img-3.png
Fichier image/png, 2,4k
Titre Figure 10: Graphic representation of a nuclear family
URL http://books.openedition.org/pacific/docannexe/image/562/img-4.png
Fichier image/png, 3,0k
Titre Figure 11: The Eskimo system (example of the English terminology)
URL http://books.openedition.org/pacific/docannexe/image/562/img-5.png
Fichier image/png, 38k
Titre Figure 12: The Dravidian system, using English words : only very close kin are charted above, but it must be remembered that Dravidian terminologies are usually extended(see below). Also, this figure depicts only the generations of Ego (the speaker) and of his or her parents.
URL http://books.openedition.org/pacific/docannexe/image/562/img-6.png
Fichier image/png, 64k
Titre Figure 13: Bifurcate merging in a Dravidian system (example of marriages)
URL http://books.openedition.org/pacific/docannexe/image/562/img-7.png
Fichier image/png, 15k
Titre Figure 14: Transmission of the bifurcate merging calculation without bifurcation
URL http://books.openedition.org/pacific/docannexe/image/562/img-8.png
Fichier image/png, 15k
Titre Figure 15: Transmission of the bifurcate merging calculation with bifurcation (cc: cross-cousins; siblings : brother-sister)
URL http://books.openedition.org/pacific/docannexe/image/562/img-9.jpg
Fichier image/jpeg, 124k
Titre Figure 16: The relational triangle
URL http://books.openedition.org/pacific/docannexe/image/562/img-10.png
Fichier image/png, 28k
Titre Figure 17: Examples of possible Ohama and Crow skewing. The persons marked in yellow in each figure are called or referred to by Ego using the same word
URL http://books.openedition.org/pacific/docannexe/image/562/img-11.png
Fichier image/png, 3,1k
Titre Figure 18: Direct exchange model
URL http://books.openedition.org/pacific/docannexe/image/562/img-12.png
Fichier image/png, 32k
Titre Figure 19: Indirect exchange system
URL http://books.openedition.org/pacific/docannexe/image/562/img-13.png
Fichier image/png, 21k

© pacific-credo Publications, 2011

Conditions d’utilisation : http://www.openedition.org/6540

Acheter

Volume papier

amazon.fr