Previous Next

The Aranda’s Pepa

VI. The ‘Marriage Order’ and Social Classification

Strehlow’s editor remarked in 1906 that ‘The views of Spencer and Gillen, as well as of other Australian researchers, on the meaning of kinship terms, as well as of the marriage classes, seem still hypothetical.’1 At the turn of the century the inclination towards evolutionistic theory was prevalent in Australian kinship studies. Reflected in the work of Fison and Howitt, Roth, and Spencer and Gillen, it led to a focus on ‘marriage order’ and kinship terminology. Questions about group marriage, primitive promiscuity, the transition from a four (section) to an eight (subsection) class system, and the origin of human society, were central in anthropological debate.

The Lutheran missionary Louis Schulze, who had arrived in Hermannsburg in the late 1870s, appears to be the first to report on the subsection system in this region (Schulze 1891: 223–227). However, it was through Spencer and Gillen that their forms of social classification became a seminal case. In particular, the eight-class system, today called the ‘subsection-system’, was much discussed. Radcliffe-Brown even named the system and its attendant kinship ‘Arandic’ after them. Frazer understood these aspects of indigenous Australian culture as survivals of a past stage in human social development, from a distant past, like other facets of Aboriginal life.

L.H. Morgan had put the classificatory kinship systems of indigenous societies on to the anthropological agenda, but his aim was to fit the kinship systems of the world into an evolutionary chain. Fox remarks:

At least half the anthropological literature on kinship has been largely concerned with the terms various systems employed in addressing and referring to kinsfolk and affines. Morgan saw in the study of terminology the royal road to the understanding of kinship systems. He was the first to see that the terminology was a method of classification, and that what it told us was how various systems classified ‘kin’. If we could understand this, we could understand the system. ‘Understanding’ for Morgan, however, meant understanding the evolution of kinship systems, and what the terminology held for him was the clue to the past state of the system. (Fox 1967a: 240)

The study of kinship as social organisation in indigenous society was only slowly emerging at the turn of the century. W.H.R. Rivers had given it an impetus with his genealogical method which proceeded from the ‘concrete to the abstract’ (Langham 1981; Stocking 1983: 85–89). Rivers’ method involved collecting genealogies – a genealogical grid – and on it imposing the particular terms, or social classifications, of a particular people. The grid was constructed by requesting the personal names of a person’s ‘mother’, ‘father’, ‘children’ and the like and then the ‘native’ terms for these relatives were listed. Rivers recommended multiple sources as a methodological check. The same set of relatives with their personal names and kin term could be elicited from a range of linked individuals. Through this method, Rivers ‘rediscovered’ the phenomenon of kinship ‘classification’ common in Australia whereby parallel cousins, for instance, are designated by the same term, ‘sister’ and ‘brother’, and by their children as ‘mother’ and ‘father’, just as reciprocally these children refer to each other and are referred to as ‘sister’ and ‘brother’. Rivers, however, took a further step of seeing in the genealogical method a means for studying ‘society’. The codes for conduct (see Schneider 1968: 29) or social rules attached to these terms provided a portrait of social order, or ‘social structure’ as Radcliffe-Brown would term it. According to Fox, ‘Radcliffe-Brown – also turning his back on evolution, but retaining the interest in terminology, produced a new and elegant comparative approach to kinship which sought to make generalizations about kinship systems, comparable to the “laws” of natural science’ (Fox 1967a: 21).

Carl Strehlow’s research lacked a framework that would have led him towards such a study of social structure. He did not integrate his data into a theory of how a society ‘functions’ as, afterall, the study of kinship as social organisation was just emerging. His collection provides, however, a starting point for the analysis of indigenous kinship systems, because it shows how people name their kinship universe and the manner in which they use kin terms as terms of address. What he did not do, in the fashion of Rivers, was superimpose his recording of kin terms on the genealogies that he collected. Therefore his grasp of a kinship terminology as classificatory, and its implications for a marriage rule for instance, remained somewhat tenuous. Neither was he able to superimpose the Western Aranda’s subsection system over the kinship system in its entirety.

What he did do was to use his genealogical material, or family trees, as frameworks on which to record data concerning personal attributes of individuals – their ‘totems’ and their skin or subsection names. He also looked at family trees in tandem, interpreting the multiple relations between affines across a number of generations. He thereby gave a sense of what it was to address a small-scale society through kinship and in this task von Leonhardi posed a series of scholarly questions. I will discuss three different aspects of Strehlow’s data on Aranda and Loritja kinship and individuals: the subsection system, kin terminology and genealogies. Although he did not employ his genealogies as Rivers and his followers would, his use of them had far-reaching implications for Aranda and Loritja people at Hermannsburg, and for subsequent anthropologists.

Carl Strehlow’s data on social classification

Without doubt Carl Strehlow’s main contribution to Australian anthropology was his myth collections and his language studies. Yet his ethnographic work includes some useful and important data on social classification. He collected a vast number of kinship terms that are still in use today. His collection provides a starting point for the analysis of indigenous kinship systems, because it shows how people classify their kinship universe and ‘when we want to understand the kinship rules and behaviour of any people we must ask how they classify kin and on what basis they make distinctions’ (Fox 1967a: 262).

Through his long residence at Hermannsburg, Carl Strehlow developed a sense of specifying different forms of Western Aranda and Loritja social relations. He saw that they had a particular form of social life and moral arrangements. He documented this in regard to the class system (section and subsection systems) and by compiling impressive lists of kin terms that showed how Aranda and Loritja people classified their kin in a kin universe, as well as how these systems connected with each other.

His kinship data were based on research he had conducted since 1892 with Diyari, Aranda and Loritja people, and by living and participating in everyday life of Aboriginal people at Hermannsburg and Bethesda. Letters by Gillen (1896),2 Siebert (1899)3 and Mathews (1906/7)4 indicate that Carl Strehlow had systematically collected kinship terms and data on the subsection systems of the Aranda and Loritja at least since 1896. Strehlow approached social classification initially through collecting section, subsection and kin terms, and seems only to have started to collect family trees in 1907 or 1908.5 Von Leonhardi sent some samples of family trees from Germany showing how they were best recorded from an individual’s point of view.6 His editor may have known the ‘genealogical method’, as he read Rivers, and this may have prompted his request. Carl Strehlow used these genealogies to illustrate how subsection systems categorised people into groups whose members could or could not be marriage partners.

The following discussion is based on a chapter called The Marriage Order in volume five (Strehlow 1913: 62–89) which explores not only ‘marriage rules’, but also the moiety division, section and subsection systems, classificatory kinship and family trees. At the time it was believed that at the core of indigenous kinship lay the function of marriage regulation and indeed the section system ‘used to be called [a] “marriage-class system” and was believed to regulate marriage’ (Fox 1967: 188). However, the section and subsection systems are not the basis of marriage rules (Dousset 2005: 15) and marriage calculations are not their only function. These systems are mainly intra- and inter-language group devices to facilitate interaction and communication – often at ceremonial events. Nor is a kinship system a marriage system. Rather, such a system contains a marriage rule. Carl Strehlow did not distinguish clearly between a kinship system, a marriage rule that complements the kinship system, and a subsection system which classifies people according to kinship categories but is not a kinship system or a marriage rule in itself.

The subsection system

Carl Strehlow’s account of the ‘marriage order’ starts with the basic division that organises Aranda society into two groups: exogamous moieties. These patrimoieties were called ‘Nákarakia’ (our kindred or people)7 and ‘Etnákarakia’ (those people or that kindred) or ‘Maljanuka’ (my friends) by the Aranda. These terms were not names for one or the other moiety but were reciprocally used by both groups (Strehlow 1913: 62). Today Arandic people still refer – from an egocentric point of view – to such groupings that are maljanuka or malyenweke and nákarakia or ilakekeye. Malyenweke means ‘them’ or ‘our in-laws’ while ilakekeye means ‘us’. In one’s own patrimoiety, that is ilakekeye, are one’s actual and classificatory fathers and their siblings, father’s fathers and son’s children, and also one’s mother’s mother’s patriline which is part of ilakekeye, ‘us’. In the opposite moiety, malyenweke, in addition to one’s spouse and brothers-in-law there are one’s actual and classificatory mothers, mother’s brothers and mother’s fathers and also one’s father’s mother. Like the Aranda, the Kukatja-Loritja used particular terms for the members of these groups reciprocally. All relatives of an ego’s group (own patrimoiety) were called ‘Ngananukarpitina’ meaning ‘all of us’; and all relatives of the opposite moiety were called ‘Tananukarpitina’ meaning ‘all of them’ (Strehlow 1913: 79). In societies organised in this way one should always marry someone from the opposite moiety.

Strong kinship ties exist between these two social groups. Some people in the opposite moiety, for example, play a crucial role in ceremonial matters relating to land. They create relationships which serve to articulate ownership of land such that the ‘patrimoiety division broadly correlates with complementary roles associated with rights and responsibilities associated with country, sites and ceremonies’ (Green 1998: 11). The most important partners in matters of land management and ceremony, are usually recruited from the opposite moiety, and preferably, also from a particular patricouple.

The two exogamous groups are further divided into two or four classes, called sections and subsections today. Carl Strehlow recorded that the Southern Aranda had a ‘4-class system’ and the Aranda ‘living north of latitude 24 degrees possess 4 marriage classes in each moiety, they have thus, a 8-class system’ (Strehlow 1913: 62). He wrote that according to Aranda tradition these divisions were established in a mythological past:

This division of the people into different marriage-classes is regarded as being of very ancient origin and is already hinted at in the legends concerning the people of primordial times. Even before Mangarkunjerkunja had formed the people, the undeveloped rella manerinja were divided into two strictly separated groups. While the members of one group lived on dry land and were therefore known as alarinja, the members of the other group, having long hair and feeding on raw meat, lived in water and were therefore called kwatjarinja. (Strehlow 1913: 62)

According to Aranda mythology the moiety called ‘alarinja’ was divided into Purula (Pwerrerle), Kamara (Kemarre), Ngala (Ngale) and Mbitjana (Mpetyane); and the other moiety ‘kwatjarinja’ into Pananka (Penangke), Paltara (Peltharre), Knuraia (Kngwarreye) and Bangata (Pengarte). In the Southern Aranda myth on the section-system, the alarinja group was composed of Purula (Pwerrerle) and Kamara (Kemarre), while the kwatjarinja group was comprised of Pananka (Penangke) and Paltara (Peltharre).

Northern/Eastern/(Western) Aranda

Southern Aranda




Purula-Kamara, Ngale-Mbitjana



Pananka-Bangata, Paltara-Knuraia


16. Alarinja and Kwatjarinja of the Aranda.

Source: Strehlow 1913.

Each moiety includes two generational pairs in father-child relationships. These pairs are called patricouples and in Aranda ‘njinaŋa’ (nyenhenge) (T.G.H. Strehlow 1947, 1965). The patricouples Kamara-Purula and Ngala-Mbitjana form one moiety and Paltara-Knuraia and Bangata-Pananka the other. Aranda marriage rules prescribe that Kamara marries Paltara, Purula marries Pananka, Ngala marries Knuraia and Mbitjana marries Bangata. Carl Strehlow (1913: 63) shows this pattern in the following way, A and B are parents and C their children:




Purula m.


Pananka f. :


Kamara m.


Paltara f. :


Ngala m.


Knuraia f. :


Mbitjana m.


Bangata f. :





Pananka m.


Purula f. :


Paltara m.


Kamara f. :


Knuraia m.


Ngala f. :


Bangata m.


Mbitjana f. :


17. Aranda ‘marriage rules’.

Source: Strehlow 1913.

This system, called by Aboriginal people in central Australia ‘skin’, has been chartered by the Institute of Aboriginal Development in the following way:

18. Arandic skin chart.

Source: Henderson and Dobson 1994.

Strehlow wrote that the Western Loritja, i.e. the Kukatja-Loritja, had a ‘marriage order’ identical to the one of the Aranda. They too divided their society into two exogamous groups and into subsections. He made a brief remark on the Southern Loritja, observing that they did not have a section or subsection system. Nevertheless, they did use the reciprocal terms ‘Ngananu-karpitina and Tananukarpitina’ for patrimoieties, Western Loritja kin terms and the same basic marriage regulation of the Aranda and Loritja, i.e. the grandchildren of different sex siblings, or the children of cross-cousins were preferred marriage partners (Strehlow 1913: 87).

Further, Strehlow described how the sections of the Aranda-Lada and Aranda-Tanka in the south and the sub-section system of Western Aranda could interlock, and that Loritja subsections are compatible with the ones of the Aranda. To him the Loritja subsection terms seemed to have been originally based on the Aranda terms, with the addition of the prefix ‘Ta’8 to indicate a male subsection name and ‘Na’ a female subsection name. Thus, the Loritja have differentiating subsection terms for their male and female members, which the Aranda do not have.

Eastern/Central Arrerntei

Western Arandaii


























i. Reproduced from Henderson and Dobson (1994: 42).

ii. Carl Strehlow’s Aranda spelling of subsections.

iii. Carl Strehlow’s Loritja spelling of subsections. For a modern spelling of the Luritja terms see the Pintupi-Luritja kinship learning material (Institute of Linguistics 1979).

These observations on how the section and subsection systems interlocked, how people were shifted into different categories and from which areas they came as well as the time frame when this data was collected (1896–1909), are particularly interesting when placed in the context of comments on the diffusion of section and subsection systems and names across Australia. Among others Spencer and Gillen (1899, 1927), T.G.H. Strehlow (1947, 1971, 1999), McConvell (1985) and Dousset (2005), have observed that these systems diffused into the desert areas of Australia. McConvell (1985) suggests that the systems of the Aranda had come from the Pilbara; it had diffused fanlike to the east and south-east as far as Southern Aranda territory. Dousset (2005: 40) maintains that the section names Kemarre and Penangke of the Southern Aranda had come from the Pilbara. As the western section system met in the north-east with a section system in the Victoria River Downs (VRD) area, it created a subsection system that facilitated marriage arrangements and probably ritual and social interaction. This subsection system then made its way south towards Aranda country (McConvell 1985).

In 1896 Spencer and Gillen (1899: 72; 1927: 42; T.G.H. Strehlow 1947: 72) recorded that the Central Aranda had originally only a section system and that the additional terms for a subsection system had been a recent borrowing:

This division into eight has been adopted (or rather the names for the four new divisions have been), in recent times by the Arunta tribe from the Ilpirra tribe which adjoins the former on the north, and the use of them is, at the present time spreading southwards. At the Engwura ceremony which we witnessed men of the Ilpirra tribe were present, as well as a large number of others from the southern part of the Arunta amongst whom the four new names are not yet in use. (Spencer and Gillen 1899: 72)

It is believed that the subsection system is a relatively recent borrowing or innovation in Arandic cultures. The cosmologies of the Western Aranda and Loritja, however, may indicate that the subsection system is an institution of ‘ancient origins’ (relatively speaking) in Carl Strehlow’s study area (Strehlow 1907, 1908, 1913). His data support Spencer and Gillen’s and McConvell’s hypothesis of the southwards movement of the subsection system, in so far as he described in detail how one system locks into the other and that it had been spreading southwards. At the same time the narratives about Mangarkunjerkunja ancestors (Strehlow 1907: 6–8), suggest that the subsection system had been for quite some time in use on Western Aranda and Kukatja-Loritja territory when Spencer and Gillen were studying the ‘Arunta’ at Alice Springs in 1896.

Strehlow’s myth data may indicate that the systems had possibly fallen into disrepair and had been ‘re-established’ (Strehlow 1907: 6–9). At different times, ‘Mangarkunjerkunja’ ancestors came from the north teaching the subsection system and the ‘marriage-rule’, and even later on a third ancestor called Katukunkara had to reinforce the system that had been abandoned. What this really means is impossible to know. It may indicate that at different times in the past, regional meetings of people occurred that introduced new concepts or reinforced communication modes that had not been used for a while. The subsection system is very likely to have been one of them that cross-cut linguistic and social boundaries. He remarks that it is noteworthy that in Aranda traditions all good laws come from the north and the bad spread from the south (Strehlow 1907: 8).

On a more practical note, T.G.H. Strehlow’s material, recorded in the 1930s, suggests that the diffusion was not a simple process. Difficulties were encountered to fit one system into the other. He recorded a scathing remark about the subsection system from the north that the Southern Aranda felt was being forced upon them:

The four-class system is the better of the two for us Southerners; we cannot understand the eight-class system. It is mad and purposeless, and only fit for such crazy men as the Northern Aranda are; we did not inherit such a stupid tradition from our fathers. (Strehlow 1947: 72)

Still decades later in the early 1980s Ray Wood, a consultant anthropologist, recalls his senior Pertame (Southern Aranda) informants complaining about the subsection system:

The older Pertame generation I worked with in the 1980s told me that there was still only a 4-section system when they were young, and the 8-subsection system has been coming in since then. They said its introduction made for all sorts of complications, even splitting descent groups and sometimes siblings into different patricouples, due to e.g. different mothers, marriages, and/or its differing introduction at different places in the Pertame region, like Horseshoe Bend versus Orange Creek.

I often noticed that they themselves still struggled with the 8 system quite a bit, and sometimes told me a given apical figure was of this subsection, only to later revise it to another, told me it would be x if you figured it out through Maryvale or through certain of their kin, but y if you figured it out another way etc.9

Today most Pertame use the subsection system in inter-group dealings and will readily supply the information of the existence of eight skin names. However, it seems that intra-group dealings may not rely entirely on the subsection system and may be the reason for the oversight of two subsection names that did not appear in a recent Pertame wordlist (Swan and Cousen 1993). T.G.H. Strehlow made another interesting observation on the transition from one into the other system in his family tree FT I. 28, which is based on his father’s family tree F.T. XXVIII (see also T.G.H. Strehlow 1999). He remarked that ‘C.S.’s class-names have been preserved throughout, so as to show the continual wavering and hesitation of his informants when assigning class names to the people in this F.T’.10

Today, it is clear that the function of the subsection system is to facilitate group interaction, ritual-exchange and marriage (Elkin 1932; Myers [1986] 1991; T.G.H. Strehlow 1999; Dousset 2005: 78–80). They are convenient social labels and propose global categories for ranges of behaviour that are especially useful in inter-group gatherings and communication. Dousset writes that the section system is ‘convenient in the context of contact’ and ‘that contact is indeed their vehicle for diffusion’ (Dousset 2005: 82). What Dousset says about the section system may also be said about the application of the subsection system of the Aranda and Loritja today:

Such contacts were either traditional – based on networks linking neighbouring groups for ceremonial, economic and marital exchanges and relations – or they were “new”, resulting from colonisation’s and settlement’s increasing effect on inter-group relations and modes of communication. In every case, sections are a lingua franca of kinship, which in turn propose a formal framework for interaction among humans. (Dousset 2005: 82–83)

At Hermannsburg where people were forced together, the compatibility of Aranda and Loritja subsections would have been of invaluable use, because many people were concentrated at the mission who under other circumstances would not have had to interact with the same intensity. It is likely that in this period the compatibility of Aranda and Loritja ‘skins’ became firmly established, as they had to accommodate and reconcile the new living conditions at the Lutheran mission settlement.

Aranda and Loritja kin terminology

Carl Strehlow made a significant contribution in the area of kinship studies by collecting a vast number of kinship terms (Strehlow 1913: 66–69, 81–85). He described the kinship system as encompassing the whole society. Aranda and Loritja kinship terms could be used for all members of the society without taking ‘blood’ ties into consideration, although there were ways to describe closeness of relatedness. He observed that every child is born into a particular subsection and thus enters into a certain kin relationship to all other subsections, regardless of whether or not blood ties exist. On the Loritja terminology he remarked that ‘The relationships between a member of a certain subsection and the members of all the other subsections is expressed in kinship terms in Loritja society, regardless of existing blood relationship or the lack of it’ (Strehlow 1913: 79). In daily life, the presence or absence of known genealogical connections are not distinguished (Strehlow 1913: 63), and each person stands in a set of relations to others described in kin terms. On both these counts this system differs from a European one.

In a classificatory system, certain kin terms are used to cover a wide range of relatives who are regarded as equivalents of one’s father, mother, brother, sister and so forth. For example, in Western Aranda, kata (karte), the term for father, covers all father’s brothers, who in English would be called uncles. The term for mother, maia (meye), is used for all mother’s sisters, who in English we would call aunties. Maia also includes daughter-in-law, from a man’s point of view. Wanna (wenhe), the term for aunt, is applied to father’s sisters as well as mother’s brother’s wife, and from a woman’s point of view, it can also include her mother-in-law. The term for uncle, kamuna (kamerne), includes mother’s brothers and father’s sister’s husband. From this it can be seen that the classificatory kinship system not only incorporates consanguineal, but also affinal relatives. Loritja classify their kinsfolk, including affines, in a similar way, although there are differences.

Where some relatives whom Europeans distinguish are classified together in the Aranda system, others who bear the same European term, are distinguished by Aranda speakers. For example, Western Aranda used and still use four different terms for one’s grandparents. The terms are aranga (arrenge) for father’s father, tjimia (tyemeye) for mother’s father, palla (perle) for father’s mother and ebmanna (ipmenhe) for mother’s mother. The grandparental terms are also used to cover one’s grandparents’ siblings as well as one’s grandchildren on a reciprocal basis. The term for one’s father’s father, aranga, for example, includes father’s father’s brothers and sisters and son’s sons and daughters. Relationships with grandparents are of particular importance to the question of land-ownership. Carl Strehlow’s Loritja kin data records also four grandparental terms (Strehlow 1913: 81–82), but today only two seem to be in use: tjamu for grandfathers (father’s father and mother’s father) and kami for grandmothers (father’s mother and mother’s mother) (Sackett 1994: 31; Vaarzon-Morel and Sackett 1997: 36). However, there are ways to express which grandfather is spoken of, in particular when reference is made to land-ownership (Sackett 1994: 31–32).

Strehlow explains the kinship classification of the Aranda and Loritja via the subsection system which locks into the kinship system. He uses as an example of how a Purula man not only calls his natural brothers kalia (older brother) or itia (younger brother) but also calls all other Purula belonging to the same generation as himself brother or sister. At the same time he calls all Purula in the generation above or below him aranga, which is the term for his father’s father as well as his natural (and classificatory) son’s son, who are both Purula.

As an Aranda person, who has been born into a subsection, can be placed into three connections to the other subsections – on an equal, higher or lower generational level – it follows that just 24 classificatory kin-terms would be required. However, gender and age (whether older or younger than the person speaking), also bear on the kin terms used making for a larger number of terms (Strehlow 1913: 63). To illustrate this point, Strehlow compilled an extensive list of the terms used for classificatory and ‘blood’ relatives which show how close and distant kin are labelled (Strehlow 1913: 66–70).

Typical for his time Carl Strehlow presented kinship terms at a distance from social life. He did not indicate that they may imply ‘codes for conduct’ which include avoidance and respect rules, obligations and rights, but for a brief remark in a section called Marriage Customs (Strehlow 1913: 89–94) on obligations and behaviour of spouses towards their in-laws:

The husband is obliged to continue to furnish his father-in-law, whom he calls antara tualtja, with food, particularly with meat. Should he kill a kangaroo, for example, then he has to give a large piece of it to his father-in-law. … He is further required to give his shorn-off hair to his father-in-law, who will make strings etc. out of it. At the death of his father-in-law he will let his shoulder be scratched with a stone knife (unangarala kalama, from unangara = shoulder, and kalama = to cut oneself) until the blood flows, as a sign of sorrow. Were he to omit this, he might conceivably be clubbed to death by his own relatives. Following the death of his father-in-law, he gives his own shorn-off hair to a brother of the latter.

The husband is not allowed to speak to his mother-in-law marra tualtja while she resides in the camp. Indeed, he may not even approach her. Should he encounter her outside the camp, he may communicate with her from a distance by means of the common secret language ankatja kerintja, or in the sign language to be described at a later stage. … The mother-in-law on her part must avoid the hut of her son-in-law and is obliged to give him the hair shorn off her head, so that he can make himself a belt or other strings from it. At his death, the mother-in-law punctures her head with a stone so that blood gushes out of it. (Strehlow 1913: 90–91)

Later, the study of kin terms developed into a study of social terms of address and inter-relations. Green’s account of the use of kinship terminology in Arandic languages, for example, demonstrates how the terms work in their social context and how kinship relationships contain behavioural patterns (Green 1998; see also Institute of Linguistics 1979; Centre for Indigenous Development Education and Research 1996). While Carl Strehlow described the regular use of kin terms, Green (1998) explores their actual and pragmatic application taking social context into account which determines their use and may appear as an irregular use of terminology. Also T.G.H. Strehlow (1999) shows that in reality irregularities were not out of the order. There was and is flexibility in a classificatory kin universe that allows variations.

Finally, following the explanation of how section and subsection systems interlock, how they related to a kinship system, and how kin terms are used in relation to close and distant relatives, Carl Strehlow addressed the Aranda’s marriage rule. In their system it is the rule to marry one’s second cross cousin: a mother’s mother’s brother’s daughter’s daughter or MMBDD (Scheffler 1978: 42) who is also a father’s mother’s brother’s son’s daughter or FMBSD (Fox 1967: 196). Carl Strehlow’s investigation puts great emphasis on the fact of this preferential rule:

Its most important principle is found in the rule that the pallukua, the grand-children of brothers and sisters (it is immaterial whether they are real brothers and sisters or regarded as siblings according to their class), should marry each other, and that according to their class they are in a relationship of noa = spouse to each other already from birth. The following two tables should demonstrate that this will often lead to the marriage of the grand-children of two natural siblings, and many more examples could be given. (Strehlow 1913: 70)

He included here a discussion on patrilineal descent, although he considered that Spencer and Gillen’s work had sufficiently demonstrated that the Aranda and other peoples in central Australia traced descent through the patriline and that subsection names were inherited through fathers or more correctly from father’s fathers in alternating generations. The discussion of patrilineal descent and that the subsection is inherited from father’s father, one’s aranga (Strehlow 1913: 71–72), was motivated by his disagreement with R.H. Mathews who was using material supplied by him11 without quoting him and arranging it arbitrarily to support his theory that descent is traced matrilineally amongst Aranda people and their subsections are allocated through their mothers (Mathews 1908). Von Leonhardi assured Strehlow that ‘Mathews does not understand the marriage laws and classes, not that he would be the only one.’12

In Aranda society, Strehlow as well as his son maintained, that it is the father’s fathers who always give their grandsons their subsection, whether the mother is from the correct subsection or not, the children always belong to the subsection of their father’s father (Strehlow 1913: 71–72; Strehlow 1999: 23, 29). To this day, Western Aranda people generally allocate subsections according to father and father’s father’s subsections.

Carl Strehlow’s genealogies

Although Carl Strehlow had recorded all births, deaths and Christian marriages at Hermannsburg since 1894, he seems only to have started to collect family trees when researching social classification with von Leonhardi. He compiled 28 family trees13 of Aranda families incorporating Loritja people who were living at the Hermannsburg Mission and had Aranda spouses (Strehlow 1913: 85). Only a small portion of his genealogical material was published to illustrate the ‘marriage-order’ of the Aranda and one ‘imaginary family tree of a Loritja belonging to the Takamara class’, because he believed that ‘one would have had to live among the Loritja for several years and have gained sufficient knowledge of the individuals in order to draw up a really reliable family tree’ (Strehlow 1913: 85). His genealogies were supplemented in 1920 with a remarkable index of all indigenous names of the people appearing on them and what they mean (Strehlow 1920: 15–39).

At the beginning of the twentieth century it was not clear in anthropology what ‘kinship’ should really mean, hence, sociocultural and biological aspects were not carefully distinguished or recognised. European notions of what constituted a family and marriage, what descent and kin terms like father, mother, and so on meant, were taken for granted or at least not well defined (see Schneider 1984: 97–112). To make this point Schneider maintained that ‘anthropology’s whole enterprise of treating kinship as a genealogical grid laid over the assumed facts of biology was misguided; instead, it was the “core symbols” that defined what kinship was for a given culture’ (Silverman 2005: 289). During the 1970s he even rejected the anthropological concept of kinship itself, ‘claiming it was nothing more than anthropologists turning their own, Western symbolic system into a universal theory’ (Silverman 2005: 320).

Clearly, the European concept of ‘family tree’ was for Strehlow’s study of indigenous kinship inadequate. The notions of consanguinity, apical ancestors and bias towards patrilineal descent attached to a European family tree14 were not sufficient to describe the complexities of classificatory kinship systems, that, for instance, included affines to a much larger degree and categorised close and distant kin not by the sole criteria of ‘blood ties’. A European family tree represents a particular universe of biological facts (ideally) which includes the same range of relatives that can be theoretically traced in every society. Descent is commonly understood as consanguine and patrilineal, which suggests ancestry many generations deep with apical ancestors. Aranda and Loritja people did not think of themselves and their relatives, or relatedness, in this way. In Aboriginal societies these links are often assumed, putative and classificatory, and can include ‘consanguinity’ under certain (ethno-scientific) aspects. The kin universe of indigenous people was vertically shallow, but horizontally very differentiated and wide. The main direction of reckoning kinship was and still is within two descending and ascending generational levels. For instance, the descendants of older and younger brothers and sisters would have been of importance, not an apical ancestor who had lived before their time, and a large number of affines would be included.

The fact that Aboriginal people generally do not remember the names of their great-grandparents, although there are ways to address them if they are still alive (Green 1998: 29), and that the names of deceased people in central Australian cultures were (Strehlow 1915: 17; Meggitt 1966: 5) and are taboo (see also Sansom 2006: 156–157), indicates that a European mode of genealogy-taking could not and would not capture Aranda and Loritja kinship adequately. In addition, it was and is inappropriate in many situations to make inquires about the deceased.15

Aranda people today, for example, use the word Kwementyaye (Breen 2000: 27) in place of the name of a deceased person or try to replace the name of a living person altogether finding a synonym. Only after an adequate amount of time has elapsed, is a name put back into circulation and will be associated with a living person. Hamilton (1998: 102) remarked that ‘the taboo on the names of deceased persons, and the desire to erase their memory as soon as possible, ensures that no precise genealogical knowledge can be maintained.’ Also Meggitt ([1962] 1986: 194) observed that Warlpiri ‘men were rarely sure of details of genealogy in their grandparents’ generation-level’ and Peterson adds that ‘young children often do not remember their genitor and this, combined with the prohibition on the mentioning of the names of the dead and the dependence of children on their mothers who are therefore likely to be the main teacher of the terminology, emphasises the tracing of social links through women’ (Peterson 1969: 29). Sansom (2006: 153; 2007) makes a strong point that in Aboriginal cultures there are mechanisms specifically to support ‘forgetting’.

It is rather unusual to find an Aboriginal person even today who can reproduce their genealogical links beyond their grandparents without the help of archival records. In some cases the answer, when seeking names of great-grandparents, may be jukurrpa or tjukurrpa by Warlpiri and Luritja people – referring to the dreaming. I have, however, not heard this reply from Western Aranda people. This may be the result of their sustained exposure to Lutheran culture and a ‘family tree tradition’, as well as their relative early sedentarisation at Hermannsburg. Aboriginal people did not have a tradition like the Pashtuns of Afghanistan, or Hawaiians, who incorporated their genealogies into oral traditions (Sansom 2006: 153, 158).

Despite the limited notions of European family trees, Carl Strehlow’s genealogies contain valuable data. He did not present ego’s descent as strictly patrilineal in his published family trees. He included a number of ancestors whose descendants had intermarried and shows their relatedness, rather than unilinear descent from one apical ancestor. His published family trees illustrate multi-lateral descent of a particular individual and his spouse. The couples Ipitarintja and Laramananka (1,1a), Loatjira and Ilbaltalaka (2,2a), Nguaperaka and Lakarinja (3,3a) and Erenkeraka and Kaputatjalka (4,4a) were placed in the centre of his published genealogies (Strehlow 1913: Stammbäume). His unpublished family trees, in contrast, traced patrilineal descent from an apical ancestor, which were the model that his son and the Finke River Mission would adopt.

The obvious data on Carl Strehlow’s genealogies include personal Aboriginal and sometimes European names, if they had been baptised at birth or had converted,16 subsection affiliations and ‘consanguine’ relatedness; this data has assisted in land and native title claims to identify appropriate claimants. He also included the ‘ratapa’ and ‘altjira’ of most people appearing in his family trees. Both terms are polysemic expressions. In the context of his family trees, ‘ratapa’ means the conception dreaming of a person, which could be acquired in three ways. Von Leonhardi summarised how this dreaming association could be acquired:

Either an embryo (ratapa), living in the metamorphosed body of an altjirangamitjina, enters the body of a woman passing by, in which case the child would be born with a narrow face, or a “totem ancestor” emerges from the earth and throws a small bullroarer at a woman, in whose body the bullroarer turns into a child which would then be born with a broad face. Apart from these two methods of conceiving a child some of the blacks also report rare cases of an altjirangamitjina entering a woman and thus being reincarnated. The old men, too, eventually admitted this. Such a reincarnation is possible only once. (Strehlow1907: Preface)

19. Genealogy of Ipitarintja and Laramanaka.

Source: Strehlow 1913: I.

In Carl Strehlow’s work the word ‘ratapa’ is not only used as a synonym for ‘totem’, but also for ‘spirit child’ or ‘child-seeds (Kinderkeime)’. He wrote that this word derived from the verb ‘ratana’ meaning ‘coming from, originating’. These spirit children were said to be invisible, but fully developed children with reddish skin colour (Strehlow 1908: 52).17 He writes that as soon as a woman knows that she is pregnant, i.e. that a spirit child has entered her, the paternal or maternal grandfather carves a small tjurunga with the designs of the ancestor from whom it emerged and stores it in the rock cave where all the other objects are stored. When the baby is crying, it is said to be crying for the tjurunga that is lost when entering into the mother. The tjurunga is called in the presence of women and children ‘papa’. To calm the child the relevant tjurunga is taken from the cave, wrapped with strings, to prevent women from seeing it, and laid in the wooden baby carrying tray where it emanates secret powers into the child that makes it grow quickly (Strehlow 1908: 80).

The word ‘altjira’ in this context18 references yet again another spirit entity. It is used for mother’s conception dreaming. Carl Strehlow describes the relationship that a person has generally to mother’s conception dreaming as follows:

However, every person is also connected with another particular totem which is called altjira. This is the totem of his mother. Every native sees this as the animal or plant, whichever might be the case, that belongs to him, and therefore calls it his garra altjira or deba altjira. The Aranda permit the consumption of these maternal totem animals or totem plants respectively. Although all the children of one family, i.e. of one mother, may each belong to a different totem (ratapa), they nevertheless share another totem (altjira). (Strehlow 1908: 57)

There are a number of remarks which indicate that also other words could be used to denote personal and mother’s conception dreaming. He noted, for example, that ‘A person’s specific altjirangamitjina is called iningukua; the altjirangamitjina of one’s mother is simply called altjira’ (Strehlow 1907: 3). The word ‘iningukua’ means ‘spirit double’ and does not seem to be in use anymore. Western Aranda people call this type of spirit ‘pmere kwetethe’ (Kenny 2004a,b). Thus, altjira can also mean spirit double of one’s mother; and one’s own spirit double is called ‘iningukua’. However, in a more general context von Leonhardi remarked that ‘iningukua’ was an alternative name for ‘altjirangamitjina’ (Strehlow 1910: 7), which means dreaming ancestor. He explains that ‘the specific altjirangamitjina, from whose metamorphosed body the ratapa emerges, is described as the iningukua of the person concerned’ (Strehlow 1908: 53).

Carl Strehlow does not mention in his entire work that a dreaming could be patrilineally inherited. This is rather intriguing, in view of later emphasis on patrilineal connections to dreamings in Australian anthropological literature, including his son’s work and among the Western Aranda themselves. I will discuss this issue in the following chapter.

Carl and T.G.H. Strehlow’s family trees in the present

Carl Strehlow did not use family trees as an instrument of social analysis. He used them as a matrix to show classification of kin, and as a vehicle for collating data on individuals. Nevertheless, they have influenced the family documentation of Western Aranda people. His family trees have determined in two main ways the perception and development of family documentation in Arandic society. They provided a starting point for his son, T.G.H. Strehlow, and they seem to have initiated a practice of recording genealogical data at the Finke River Mission that went beyond the usual practice of the Lutheran church to record birth, marriage and death. It continued for the better part of a century and has to some degree impressed onto the Western Aranda themselves a concept of ‘families’ with apical ancestors and may have strengthened patrilineal emphasis.

T.G.H. Strehlow was in possession of his father’s family trees when he produced his own genealogies nearly 50 years later. During his own research he was often only able to gather data reaching back to the grandparental generation of his informants. He incorporated his father’s groundwork, only occasionally referencing it. He compiled 150 genealogies, which usually begin with an ‘apical ancestor’ and his wife. About 50 of these family trees are based on his father’s work and thus, can be dated back to circa 1800, and further, and ‘from which all authentic facts can be extracted to substantiate theories of aboriginal land rights and law’.19

The information on T.G.H. Strehlow’s family trees is very rich, but it can only be understood in its particular theoretical and ethnographic context, which is not immediately apparent. The genealogical data are not clearly defined, as he may have assumed that anyone interested in his family trees would have read his extensive oeuvre and would be able to contextualise them. Like his father’s work, his data presupposed an enormous amount of knowledge. Carl Strehlow, for example, included the ‘ratapa’ and ‘altjira’ of most people appearing in his family trees published in volume five (Strehlow 1913). It is not clear without having read volume two (Strehlow 1908) what the terms ‘ratapa’ or ‘altjira’ describe on these family trees.

Obvious data on T.G.H. Strehlow’s family trees include personal names, subsections and relatedness through apical ancestors, which evokes patrilineal descent and physical kinship. He also included labels such as ‘half caste’ (H.C.) or ‘full-blood’, which he colour coded, and sometimes supplemented with fractions, i.e. 7/8. This coding made the notions of descent and blood ties unmistakably clear. Nearly every person on these family trees has a footnote that is often cross-referenced to his diaries or to other family trees. These footnotes contain an immense variety of historical, cultural, social, geographical (location of sites, sometimes the description is in Aranda) and additional kin information as well as gossip.

The conception sites, called ‘pmara kŋanintja’ in T.G.H. Strehlow’s work (1971: 596) and the conception ‘totem’, ‘kŋanintja’,20 of most people can be found on his family trees. However, he does not explain these terms and abbreviations or include an explicit key or legend. The reader is left to his or her own devices to interpret, for example, ‘from Emalkna; imora kŋ.’ Thus, it is not surprising that many (mis)-interpretations and -understandings occur. The information on a personal conception site is often interpreted by descendants of a particular individual that this place is also associated with them and that they have traditional rights to own it. Sometimes it is even understood as the name of a traditional country or estate. The abbreviation ‘kŋ.’ stands for kŋanintja and ‘from’ refers to the place where the spirit-child entered the mother to be, i.e. the conception site of an individual, it is not a patrilineally owned place as wrongly assumed by many Aboriginal people. The example above therefore shows only that a particular person’s conception site is ‘Emalkna’ (Mt Heuglin in the Western MacDonnell Ranges) associated with ‘possum dreaming’. See also Morgan and Wilmot (2010) who have made similar observations on these matters.

According to T.G.H. Strehlow, a conception site was the place where a pregnant woman felt for the first time her baby move in her womb, and theoretically could have been on any estate where a woman had the right to forage. At these places the spirit part of a person – left behind in the landscape by the ancestral dreaming beings – entered the mother (Carl Strehlow 1908: 53, 56; T.G.H. Strehlow 1947: 87). The human soul begins its existence when the ‘spirit-child’ enters a pregnant woman giving the embryo a soul (Strehlow 1908: 52–56; Strehlow 1971, [1964] 1978). T.G.H. Strehlow wrote that these spirits were part of the trails of ‘life’, left behind throughout the landscape by the ancestral dreaming beings ([1964] 1978: 20, 22). Human children could come into being at all places situated along these trails. In Aranda Traditions (1947: 88), he used the word ‘ŋantja’ for spirit child. He also called them ‘life cell’ or ‘life giving property’, which entered a woman and developed into a human being. Pink (1936: 288–290) wrote that baby-spirits gave the spirit part to human beings when entering the mother. Her Northern Aranda informants maintained that baby-spirits were left behind by a dreamtime ancestor who had left some tjurunga in the landscape.

While, according to T.G.H. Strehlow, the conception site of an individual was of great importance and prominent in an individual’s life, it did not confer automatically landholding rights to any of his or her descendants, but they had the right to learn about it, if they were prepared to do so, which required engagement and effort. A conception site was associated with a particular person, unlike the dreaming places or country claimed through father and father’s father which would belong to a well defined group of persons.

T.G.H. Strehlow’s genealogies usually represent njinaŋa (patricouple) groups with a male apical ancestor; to a degree these genealogies were understood by him as one of the instruments for the analysis of land ownership. To most users of T.G.H. Strehlow’s family trees it is not clear that they are dealing with what he called a patrilineal ‘totemic clan’, i.e. a njinaŋa section. He wrote in Aranda Traditions that he had ‘attempted to introduce the term njinaŋa section to denote a group of men forming a local totemic clan’ (Strehlow 1947: 143). Only the people of this group, who are patrilineally affiliated, belong to the main dreaming associated with a particular place (pmara kutata21 in T.G.H. Strehlow’s terminology) of the male apical ancestors on that particular family tree which may or may not be his conception site.

The lack of an explicit key to the Strehlow genealogies has caused much confusion and misunderstanding of what they represent. In particular when Aboriginal people access T.G.H. Strehlow’s genealogical material at the Strehlow Research Centre in Alice Springs and mistake the conception sites of their ancestors with a place they may claim as their own, believing that it confers primary rights to a place or country. Also the memory of apical ancestors was not preserved in Aboriginal societies. It is only with genealogical records like the ones produced by the Strehlows, Tindale and the Finke River Mission that Aboriginal people today are able to reproduce such ‘deep’ genealogies. As already mentioned this is not likely to have been the way Aranda people perceived their relatedness. The reality of desert life with its particular social circumstances and traditions determined who was emphasised in a person’s kinship net. The cultural significance and the appropriate interpretation of the information on these family trees can only be understood through close reading of the ethnographic works of both Strehlows. This means they have to be set in their theoretical and historical context with consideration of contemporary indigenous community politics.

20. Hesekiel Malbunka’s family tree drawn by T.G.H. Strehlow. Exhibited in 2007 with the consent of the Malbunka family in the tourist facility of the Museum of Central Australia in Alice Springs.

Source: Strehlow Research Centre, Alice Springs.

The concept of Western Aranda families, resulting from a long influence of constant reinforcement of family trees produced by the Finke River Mission, T.G.H. Strehlow and the Central Land Council, may be seen as a paradigm for Sutton’s (2003: 206–231) conceptualising his ‘Families of Polity’ as a modern kinship form among indigenous Australians that are not residential or local groups (Austin-Broos 2004: 61). Generally these ‘families’ are today characterised by a patronymic identity and cognatic descent many generations deep. A different treatment of this phenomenon takes a view of kinship application in the contemporary social context of day-to-day living in which elaborate and complex networking by individuals takes place (Austin-Broos 2003a, 2006, 2009). Carl Strehlow’s treatment of kinship was still at a considerable distance from such a description and analysis of social organisation.

T.G.H. Strehlow’s genealogies, by extension also his father’s and other family trees of the Finke River Mission, and more recently family trees generated by the Central Land Council during the land rights era, have assumed new meanings in the context of land-ownership. Genealogies are increasingly the way to claim affiliation or connection to Aranda country rather than through esoteric knowledge of myth which is in decline (Oberscheidt 2005). Sometimes T.G.H. Strehlow’s genealogical information is almost treated as secret-sacred material. There is a perception that once on a family tree one is ‘in’, even if the connection is marginal or affinal.22 T.G.H. Strehlow’s genealogies are often perceived as the last word on traditional membership of a landholding group by some indigenous people, and used by them as evidence for membership. However, they do not define how kinship confers particular rights and obligations in land or how kin networks function. In addition, placing weight on these genealogical records can obscure other processes including fission, fusion, the end of patrilines and political alliances of particular families.

Is there a responsible use of family trees?

In November 2007, Central Land Council anthropologists, Helen Wilmot and Rebecca Koser, presented at the conference of the Australian Anthropological Society a paper in which they addressed problems in the context of Aboriginal land claims, mining and royalty distribution processes created by genealogical materials obtained from archives. They raised the question: how are anthropologists to deal with problems, generated by archival materials, in particular by family trees, amongst Aboriginal people in central Australia. People are gaining positions of power and influence in Aboriginal decision making with these documents that under traditional laws and customs they would not be able to gain, and are rejecting others. They discussed how some facets of identification are based on these documents that are perceived as ‘quasi-traditional authority’ and how this information is reified.

Land Council anthropologists are witnessing that written documents are used as ‘proof’ of ownership and connections to land. They have observed that parts of genealogies, such as a footnote, are internalised even by senior people. These snippets of genealogical information, which may or may not be wrong, are on occasions recited as if they were traditional knowledge and misconstrue traditional ownership. Morgan and Wilmot (2010: 9) give an example of such a situation. As an explanation for how certain families were related to each other, various senior members of a particular group had repeatedly told a Land Council anthropologist that ‘All our mothers were sisters from Bambi Springs (location a pseudonym)’. Some time later it was discovered that this sentence had been plucked from a footnote of a Strehlow genealogy (the traditional owners had in the meantime lost their copy of this family tree), and after careful analysis of the genealogy it emerged that the connections claimed had been based on a misunderstanding of the document with significant implications for claims to land.

Ethically it is difficult for anthropologists and institutions, who hold this type of genealogical material, to address these issues (Morgan and Wilmot 2010: 3–4). Once such material enters the public domain it is not possible to control or guide what people do with this material, how they interpret it or base their identity on it. The use of written material as proof of identity is not common to all Arandic regions; degrees of urbanisation and westernisation in central Australia differ. There are Arandic people who barely speak English, and are embedded in their traditional laws and customs. These are often the people who suffer when the written artefact takes on a new life in the hands of relatives who are proficent English speakers and familiar with modern mainstream life and administration. Morgan and Wilmot remark that one of the issues of the rise of genealogical documentation as a new form of authority, is that it is used by Arandic people, who are print-literate and adept in processes of negotiation within the wider society, to successfully demand recognition as traditional owners or native title holders from recognised senior, knowledgeable people based on these genealogies. They write:

In these scenarios, it is often the case that such a heavy reliance on genealogical documentation is the result of limited knowledge about kinship rules, how country is inherited or knowledge of the country purported to be owned and even, in some cases, where that country is located. It is not uncommon to receive requests from senior traditional owners to hold workshops about some of these issues in order to pass on cultural information. There is clearly recognition that knowledge and traditional hierarchies are diminishing, which is related to socio-cultural change between generations, in a context of reinvented post-traditional forms of knowledge. (Morgan and Wilmot 2010: 9)

On the other hand Carl and T.G.H. Strehlow’s family records are highly valued today among many Aboriginal people in central Australia. For some descendants, to possess excerpts of T.G.H. Strehlow’s genealogies showing one’s ancestors is one of the most precious and cherished possessions, giving some indigenous people a kind of sense of belonging, and the feeling of empowerment knowing ‘who’ they are and from ‘where’ they come.

1 Von Leonhardi to Carl Strehlow, 2.6.1906.

2 F. Gillen to B. Spencer, 14.7.1896 (Mulvaney, Morphy and Petch 2001: 130).

3 O. Siebert to A. Howitt, 22.8.1899 (Melbourne Museum).

4 Carl Strehlow to R.H. Mathews, 1906–1907 (NLA 8006/2/4).

5 On the 18.8.1909 von Leonhardi confirmed in a letter to Carl Strehlow the receipt of 20 family trees.

6 Von Leonhardi to Carl Strehlow, 8.12.1907.

7 See Strehlow (1913: 62, fn. 5) for an elaborate attempt on the possible etymology of these reciprocal terms.

8 Today ‘Tja’ is used.

9 Ray Wood, email, 14.8.2006.

10 T.G.H. Strehlow’s FT I. 28.

11 Carl Strehlow to R.H. Mathews, 1906–1907 (NLA 8006/2/4).

12 Von Leonhardi to Carl Strehlow, 2.6.1907 and 26.2.1909.

13 T.G.H. Strehlow’s FT I. 28 and Book of XVII, p. 118a bottom.

14 European traditions tend to have a patrilineal bias for inheritance of surnames and property or accession to a throne, for example. It is nevertheless bilateral and/or cognatic in general character. These sorts of differences also apply in Aboriginal Australia, not least in relation to distinctions between ‘actual’ relatives as opposed to more nominal ones.

15 Often researchers have to describe ‘in a round about way’ a person who has a tabooed name and has to ask if it is allowed to say that name. In June 2006, for example, I was asked not to use the Luritja word tjala (honeyant), it had been temporarily taken out of circulation in this particular family due to a recent death. On this occasion I was also informed that the word apme (snake) had been replaced by arnerenye (belonging to the earth/ground or living in the earth/ground) in the Hermannsburg area.

16 At the mission generally only the Christians had European names.

17 T.G.H. Strehlow maintained that only the spirit children of Ntaria were called ratapa (Strehlow [1964] 1978).

18 See Chapter VI for discussion on altjira’s semantic field.

19 T.G.H. Strehlow F.T. I. 6.

20 Unpublished dictionary K: 92; see also Strehlow 1947.

21 See Kenny (2004a,b) for more on this term.

22 Affines can have contingent rights under certain conditions (see Sutton 2003: 12–13).

Previous Next