'I can't follow you on this horde-clan business at all': Donald Thomson, Radcliffe-Brown and a final note on the horde.

Peterson, Nicolas

Until Les Hiatt published his critical assessment of Radcliffe-Brown's classic 1930-1 formulation of Australian territorial organization in 1962, the model had remained virtually unchallenged. (1) It is true that W.E.H. Stanner (1933) in the Daly River area, Lauriston Sharp (1934) and Donald Thomson (1939: 211-12) working in Cape York, and Phyllis Kaberry (1939:30-1, 136) and Theodore Hernandez (1941) working in the Kimberley, had all indicated that some aspect of the model did not apply to groups of bush dwelling Aboriginal people they worked with, prior to 1962, but it was only with Hiatt's critique that the evidence was marshalled, and a compelling and coherent case made for the incoherence of Radcliffe-Brown's formulation of the concept of the horde. (2) In the terminology of today, the horde is referred to as the band or land-using group, as opposed to the clan or land-owning group. Although Radcliffe-Brown tried to maintain this distinction, he used the term horde inconsistently, sometimes for the land-using and sometimes for the land-owning group, and asserted that the land-using group was made up of males only from the landowning group (clan). (3) Hiatt clearly demonstrated the falsity of this assertion by showing that no such land-using group had been reported in the Australian literature.

Outside native title applications and land claims, and ethnographic reconstruction, this debate is only of historical importance today, but in the middle 1960s it was heated and significant. If Hiatt was correct, it meant the elegant and aesthetically appealing fit between Aboriginal social and local organization, was disrupted. The purest expression of this fit was Claude Levi-Strauss's (1949) model of Aboriginal society, in which the males of each patrilineal clan lived on their own clan's estate and were linked to surrounding clans by the exchange of sisters. Hiatt compounded his 1962 attack on this orderly system at the Man the Hunter symposium in 1966 (Hiatt 1968) by telling Levi-Strauss that Gidjingali women, and by implication many other Aboriginal women, were not exchanged between patrilineal clans, but were bestowed by individuals outside their own clan, namely their mothers and mother's brothers. After the symposium the editors of the Man the Hunter volume invited Levi-Strauss to comment on Hiatt's work, which he did dismissing it on the grounds of the recency of his fieldwork (1958-1960), which meant that he was dealing with 'what is left of a collapsing tribe' (1968:210-211). These fighting words are a good measure of the significance the debate had at that time. Joseph Birdsell provided his own version of Levi-Strauss's argument in 1970, in what was the last defensive sally in this debate.

The correspondence between Donald Thomson and Radcliffe-Brown considered here helps explain why the evidence against the orderly model of local organization was not clearly formulated until 1962, even though there was some strong evidence against it. (4) Ironically, Donald Thomson was writing to Radcliffe-Brown on this topic in the very year Levi-Strauss published his magnum opus on kinship (1949) based on the orderly model.

Donald Thomson was in the first small cohort of Radcliffe-Brown's students in Australia, and the first person to graduate with the Diploma in Anthropology in 1927. Subsequently he carried out fieldwork both in Cape York (1928-1933) and in Arnhem Land (1935-1943) with people still leading self-supporting lives in the bush. That is, he saw a number of bands and spent short periods living with them.

In 1935, on the basis of his time in Cape York he published a paper on 'The joking relationship and organized obscenity in north Queensland' in which he provides his first statement on territorial organization in a long footnote (1935:462-3). (5) .... Thus a horde consists of all the male members of the clan whose territory it inhabits, with their wives, who, though they are members of the horde are not members of the clan (since entry to a clan is by birth alone), (6) and less those women of the clan who have married into other hordes. But while they may change their hordes by marriage, they can never change their clans. It is clear, therefore, that although the horde is the war-making group, the clan, and not the horde, is the land-owning group; a clan is a stable, permanent, structural unit of society; but the horde is unstable; it is a sociological entity the membership of which is constantly changing. I do not propose to go more fully into this subject here, but brief mention must be made of the bond that unites members of the two groups. It may be noted that solidarity within the clan is maintained by the bond furnished by (1) common descent, (2) the possession of common totems, (3) the possession of a common territory. Solidarity within the horde rests upon none of these permanent foundations: it depends solely upon the cohesive force supplied by such social institutions as marriage and the bond set up between a man and a woman (who are members of different clans) by the family, centered in their children, and by the sharing of normal activities of everyday life, by fighting with other hordes--in all of which the bond of solidarity within the horde is affirmed and strengthened by collective ceremonies such as dancing, especially war, funeral, and vengeance dances.

Although this account starts out being consistent with Radcliffe-Brown's model, two points can be made. First, the reference to the horde being an unstable unit, the membership of which is constantly changing, is consistent with Radcliffe-Brown's model, if one assumes that the constant changing only refers to the coming and going of women at marriage and does not refer to males. Second, the remarks on solidarity in the horde are curious because if all the males were of a single clan then one would expect common descent, common totems and common territory to be a basis of solidarity between the males in the horde, as they are in the clan. However, he is emphatic that solidarity in the horde rests on none of these but on marriage etc, which might be taken to indicate that the males in the horde do not all come from the same clan.

In a footnote in his now very well known paper on the influence of the seasonal factor on the life of the Wik-Mungkan in western Cape York (1939), which it is possible that Radcliffe-Brown did not see as it was published in an archaeological journal, Thomson speaks of the multi-clan composition of what we would now call the band, but which was then called camp or horde. From the perspective of today his statement is very slightly ambiguous, although at the time it can safely be assumed that it was seen as a statement of the Radcliffe-Brown orthodoxy: ...the group of people popularly spoken of as a 'camp' [i.e. band], that is found at any time within a clan territory, really consists of members of many clans, and for this group the term horde will be used. It is in reality an aggregation of families of the male members of the clan (1939:211).

The phrase, 'really consists of members of many clans', unqualified, would imply that there were both men and women of many different clans in a band, however, the final sentence pulls the statement back into line by indicating that the men are of one clan, so it must be the women who are from a range of different clans. Although the correspondence below relates specifically to Arnhem Land, Thomson does indicate, at one point, that his views apply to his entire experience of bush living groups (see letter of 2/11/1949).

THE CORRESPONDENCE

In 1949 Thomson published two works both of which deal with territorial organization and the horde (1949, 1949a). It seems highly likely that it was the preparation of these two items that prompted him to write to Radcliffe-Brown in July 1948 to try to get him to clarify his use of horde: ... there is one point on Australian social organisation that I want particularly, to clear up. On pape [sic] 35 of your Oceania publication (Vol I, No I, [1930]) you speak of the horde as "a small group of persons owning (italics mine) a certain area of territory ... I would be very glad if you would clarify this point for me. The group you are speaking about there is what I have been calling the clan--that is, the permanent, stable unit of social life, consisting of a number of families related or believing that they are related, by common descent, and possessing common totems. The occupational group, the loosely-organised group which at any time occupies that clan territory, I have spoken of as the horde. According to my idea the clan is exogamous but not the horde. I regard the horde merely as the nomadic group found in a territory at any time. The horde would then be the law-making, and war-making, but not the land-owning group, since in a patrilineal clan, its female members would be living with other hordes occupying territories of the clans of the men they married. This would mean that a woman left the {territory of the} (7) clan to which she belonged to join the horde of her husband, into which she was adopted. Would you please refer to my definition of horde and clan in the "Joking Relationship" paper [see above], ... and tell me if you disagree with that statement. I thought that I was following your teaching in this, and was rather surprised when, recently, I referred to your Oceania definition and found that there was some difference in terms. As you will see, I had regarded ... the horde as a loosely organised occupational group only--certainly not stable, but because autonomous, having a kind of functional solidarity--based that is, on (possibly only temporary) common interests, not on a solid foundation like the solidarity of the permanent stable group, the clan. In Arnhem Land a woman's solidarity with her clan--though she is living in another horde--seems to be kept alive during the clan initiation ceremonies, mardai'in, by her obligation to assist in the making of food for ceremonial presentation--as well as by the permanent ties of kinship.

I would very much like to hear from you on that subject (26 th July 1948). On the 27th October 1948, Radcliffe-Brown replied in a handwritten letter from the then Farouk University in Alexandria: I am sorry that my statements about clan and horde were not sufficiently clear. There are undoubtedly some variations in different parts of Australia. But the same general type of organization is found in very many parts. I was thinking of three kinds of groups. The first is the clan and consists of all persons born as children of a man of the clan. What I called the horde consists of all the men of a clan together with their wives, their sons and their unmarried daughters. This group is composed of families. The third group is a changing one consisting of persons camping together for a shorter or longer time. I think of this as a camp. In Western Australia and in some other parts if you examine a camp you find that the families of one horde camp separately from the others. A large camp for religious ceremonies may consist of five or six separate camps each containing families of a single horde. The country belongs to one of the hordes, who are the hosts, the other hordes being visitors. (8) Certainly in some parts I have found it easy and useful to distinguish the horde as a group of families of which the men constitute (with their sisters married elsewhere) the patrilineal clan. Is this clear? If not I will write more at length.

This statement of his position differs from the definitive 1930-1 statement by the addition of a third concept, that of the camp. He had originally used this term fleetingly in his first paper on Aboriginal territorial organization in 1913 (1913:147). (9) The usage here is slightly more elaborate and somewhat confusing since he is using camp in three ways. First it is: a changing group consisting of persons camping together for a shorter or longer time. Second it refers to a large residential gathering for a ceremony. Third, it is a subdivision of the large ceremonial gathering that is identical with the horde. It seems that Thomson's query kept him thinking because on the 16th November 1948 he wrote to Thomson from Egypt about the concept: Camp. I would prefer, on the whole, to use this for the body of persons who are camped together at a particular time, and I have been in a "camp" in the desert of Western Australia that stretched nearly two miles in length and contained nine hordes. The statement [in a draft text Thomson sent to RB] that a horde is nomadic 'only within the boundaries of the clan territory' (page 51) is, I think misleading. In times of peace members of a horde, or even the horde as a whole, may go to visit another horde and hunt over their territory for days or weeks. There are obligations of hospitality, not only to the sister's sons, but also to certain other clans, which sometimes look as if the other clan had rights over a clan territory at certain seasons when a particular food supply is abundant. An example is the harvest of bunya nuts in the Bunya Bunya Mountains of Queensland.

Besides dealing with the issue of the word camp, he added an addendum to this letter which reveals another weakness of Radcliffe-Brown's analysis, the absence of an adequate consideration of women: In dealing with the clan-horde structure I think it might be useful to point out that in Australia it is only a man (after initiation) who occupies an independent position in society. Women and children are in a dependent position as attached to some man, father or husband. So far as a clan and its associated horde are concerned the real group of independent persons (initiated men) is the same. This may help the reader who may find difficulty over what is really a simple matter. In much anthropological literature the word clan is used both for the persons connected by birth and also for the group that includes their wives. This is confusing and unscientific, and I used the terms horde and clan in order to get rid of the ambiguities. (10)

In 1948 -1949 Thomson (1948; 1949) published two accounts of territorial organization in Arnhem Land where he had spent time living with people who were truly independent. In his book on the Economic Structure and the Ceremonial Exchange Cycle in Arnhem Land (1949:11-12), he writes: The clan, or in practice its occupational group, the horde ... is the most important unit of social life in Arnhem Land. The clan is the territorial or land-owning group, but the horde is the war-making group, and each horde is to a large extent independent and self-governing.... There is no individual ownership of land. Land ownership is vested in the clan as the hereditary group, and as this is a patrilineal society, a man is born into his father's clan and has hunting rights over the territory of his own clan and generally over that of several others, particularly that of this mother ... the group which actually occupies the clan territory is made up of the male members of the clan and their children, plus their wives (who remain members of other clans--the clans into which they were born), and minus those female members of the clan who have married into other clans. To this occupational group the name "horde" is applied. The clan is an hereditary group, entry into which is by birth alone, but a person may enter a horde by marriage or adoption. The clan is a permanent and stable unit of social organization; the horde rests on no such solid foundation and is unstable.

This statement is pure Radcliffe-Brown until the very last clause. The earlier statement in 1948 (1948:154) is slightly different because it is stronger on mixed composition and clearly states that the land-using group is made up of members of several clans: The clan is the land-owning group. Each group is nomadic within the confines of its own territory, and sometimes in the territory of certain other clans, over which it exercises special hunting rights. As these people are patrilineal, and as the clans are exogamous (the women of the clan marrying into other clans and the men marrying women from other clans), the group which occupies a given clan territory actually consists of members of several clans. For this group the name "horde" is used. As will be seen, the horde in Arnhem Land is an independent, self-governing unit. It is also the war-making group, and the fighting which rages constantly is between hordes, either singly or in groups united chiefly by affinity within the kinship organization.

The first two self-supporting groups Thomson met in Arnhem Land at Bennet Bay and on the Koolatong River, were large ceremonial gatherings so it is unsurprising that they had men from a number of different clans. Subsequently, however, he visited several day-to-day living groups or hordes (i.e. bands), and where he provides details of group composition, they involved men from several different clans. (11) In the second account, the phrase, 'the group which occupies a given clan territory actually consists of members of several clans' would probably have been read by most people at the time of writing as meaning that the diversity of clan membership was produced through the in marrying of women, although it does not say this, and in the light of the correspondence with Radcliffe-Brown and of his field experience he, in fact, probably means men from different clans as well. This is further suggested by the fact that he refers to the land-owning group sometimes hunting in the territory of certain other clans.

Well after these two publications were in press, (12) Thomson returned to the local organization issue, writing to Radcliffe-Brown on 2nd November 1949 as follows: I have re-read your excellent and simple notes on Social Organisation and these will help a lot. But I can't follow you on this horde-clan business at all. In a letter you wrote to me about my book you advised me to treat the horde first. (13) As I see it, however, the clan is the vital group--the group which, in the territories I know, owns the land ... The women marry out; others come in. The resultant group is the horde. Yet on reading your notes in Oceania it seems to me that you attribute to the horde many of the features which seem to me to be essential characteristics of the clan. But there is this--what you wrote to me more recently seems to conflict with this. Or rather, I was myself rather misled by much that you have written until I read your footnote on page 59 of the Oceania account (Vol. I part I [1930]). Where, for example you speak of a horde as 'owning' a totem, I can't follow you because you define a horde as consisting only of the male members (of a patrilineal clan) and their descendants and less the women who marry out and go to other clans. But these women, though members of other clans retain their membership of the clan into which they were born and its totems are their totems--not the totems of the horde into which they marry. (14) And again, you speak of a horde as oten [sic] or sometimes, being exogamous. But under the levirate, apart from any other consideration--does that not rule out exogamy of the horde--because a woman must then necessarily marry a man (her deceased husband's brother) who is a member of the clan to which she belongs--and which she entered at marriage. Is it not the clan that is exogamous--never the horde? On page 35 [1930] again you speak of the horde as a group of persons owning a certain area of territory and the horde as the primary land-owning and land-holding group. But in the areas I know I feel that that group is the clan and not the horde. Where I may be wrong is in this: I have been regarding the clan as the stable, permanent hereditary group--the horde merely as the group of people who at any give time occup [sic] its (the clan's) territory. And both in Cape York and Arnhem Land that group does vary and it consists of men of many clans who for kinship reasons have some right in the territory and adopt it. There may on this basis, be men of several clans and even of both moieties, in occupation of any given territory--and living in it for years. That was the position for example, at Caledon Bay. I called that occupational group the horde. You would perhaps call it just a "camp". But if so, what then is the horde in that area. The clan is definite and well-defined, and this composite functional group also. But if it is not to be called a horde, then I don't think there is a horde there. Later: I am afraid that this is getting diffuse and that I am just talking around the subject...... My point is this. Much that you say of the horde seems to me to be a function of the clan. In Arnhem Land in Cape York you get a large group of people living (say) in the territory of the Arrawiya clan in the Caledon Bay district. It has a name--the name of the clan territory, which is definite and well-defined and well understood by the people. But that territory is now occupied by a lot of people besides the male members of the clan and their wives--the people we have been calling the horde. These are often members of other clans--even of other tribes or of several tribes because there is no tribal organisation at Caledon Bay. This group is the occupational group in that area. But it changes. I would like to call that the horde, it is unstable but then, I do not think, on mature reflection, there is any occupational group (except in theory) that really does correspond without [sic] definition of the horde. The occupational group does alter in composition more often than you seem willing to admit. It is unstable. But if you boject [sic] that this group isn't the horde, what is it--and what, and where, is the horde as a functional group? I am going to say now that in the territories I know the horde does not exist in the form you describe--except ideally, or in theory. The group I have just described is the one I know everywhere. It is made up predominantly of the male members of the clan which owns the territory--and their wives, who are members of this horde--but invariably there are others of such long standing that they can't be dismissed as visitors. They hunt and fight with the group with which they live, but they are not really linked with it by any bond that I can find except that of kinship and custom. They have lived with the group, which contains kin--and they throw in their lot with it. Please let me have your comments and your final word on this because I can't make my group fit yours as defined in Oceania and I am going to suggest with very real deference, that it needs to be modified.

In this letter Thomson lays out his problems with the concept of horde absolutely explicitly and they are the core of the objections that others have raised since the 1930s. The letter appears to have caused an hiatus in the correspondence for just over a year. When Radcliffe-Brown did write to Thomson on the 29th November 1950 he made no mention of the issue of the horde. It seems it was only in October 1954, that Radcliffe-Brown mentioned local organization again in correspondence with Thomson, although he published twice on it subsequently (see 1954, 1956), in a postscript to a letter: What is most needed in the book [Thomson was preparing on Arnhem Land] is a really satisfactory account of the local organization. You will find a very confused account by Berndt in the next number of the American Anthropologists and, I hope, a brief comment of mine (posted 20th Oct 1954).

Although Thomson lived until 1970, his last published statement on local organization appears to be in the 1951 edition of the Australian Junior Encyclopaedia (1951:80-1), where his account is classic Radcliffe-Brown.

WHY DID THOMSON PUBLICLY ACCEPT RADCLIFFE-BROWN'S MODEL?

How are we to understand Thomson's behaviour? It seems clear that Thomson did not feel that his experience was just an anomaly in the context of doing normal science, to use Kuhnian (1970) terminology.

There is no doubt that Thomson felt the situation in eastern Arnhem Land was starting to be disorganised by the impact of Europeans and more importantly by the hundreds of Japanese pearlers off the coast each year (see Peterson 2005:34). However, the bands he saw in Cape York also had males of several clans living in them, and even though the missions there were obviously having an impact on the life of people in the bush, he did not feel that the bush life there was being disorganized in the same way. Presumably, therefore, the mixed clan composition was not a product of disorganisation. I think the reason for the difference between the two areas in his mind was that he saw the sexual relations with the pearlers in Arnhem Land upsetting things, whereas in Cape York the absence of pearlers meant there were no such intensive relations.

Other people, as has been noted, had also been documenting mixed clan composition in the 1930s but playing it down. Theodore Hernandez's account of local organization among the Drysdale River Tribes of the Kimberley, is a case in point. Where for the most part he felt that the people in the 1930s were living like their ancestors (Rumsey and Redmond 1999: 44), when he reported that bands had little unity and cohesion, by which he meant they were not systematically made up of members of a single totemic group (i.e. patrician, 1941: 221), he felt obliged to qualify this and say it was different in the past. Alan Rumsey and Anthony Redmond suggest that this was because he was 'unduly influenced by the then widely accepted view of Radcliffe-Brown' and what he in fact reported is what is now understood as a textbook image of a flexible residential, land-using group (i.e. band 1999: 44).

But Thomson's case is different. He first voiced his concerns when he must have been drafting his Economic Structure book (1949) but his strong critique of Radcliffe-Brown's model came only in private correspondence after he had the galley proofs.

Other aspects of the correspondence between Radcliffe-Brown and Thomson provided some suggestions as to why this came about. In the period until 1950 the only Department of Anthropology was at Sydney under the direction of A.P. Elkin. Neither Radcliffe-Brown nor Thomson liked Elkin and Thomson had become alienated from the Sydney school. (15) Thomson felt isolated academically, and communicated that to Radcliffe-Brown on the 16th January 1948 when he wrote: I shall be most grateful if you will help me with criticism--however severe, for I would rather get that now than when it is published. Please remember that I was your first Diploma student in Sydney. I wish that I had known you better in those days, and in the years which followed. However the work has been done, and the grounding you gave me in Sydney has stood me in good stead I hope.... It is rather difficult to stand and to work quite alone here--and of course it means no recognition for one's work except overseas--and I have always known that it [is] that that matters anyway. I appreciate your offer of help just now when I am preparing to publish.

Thomson made good use of this offer. Between June 1947 and October 1955 Radcliffe-Brown and Thomson exchanged 65 letters: 22 being sent by Radcliffe-Brown and 43 by Thomson. In more than a dozen of these letters Thomson explicitly asked Radcliffe-Brown for advice and help. This ranged from how to deal with published criticism of his work from the Berndts and how to write about totemism in a general book he had been commissioned to prepare on Aboriginal culture, to assistance in the analysis of his Cape York kinship materials--particularly those from the Wik-Mungkan--and for comments on various manuscripts and papers. Mindful of Thomson's situation Radcliffe-Brown wrote on the 14th November 1949 that, 'Stanner has promised me that he will give you any help he can' when he returns to Australia. (16) Again a year later he comments: I feel that in Melbourne you are very isolated. Now that Stanner is returning to Australia you might be able to forget your military conflict with him and get together with him on Australian anthropology. He is capable of doing extremely good work if it is not too late as the result of his lack of contact for so many years during the war. If you can succeed in making friendly contact with him you would find him, I think a congenial and helpful colleague. He has one thing in common with you, his appreciation of the blackfellow.

However it was from Radcliffe-Brown, not Stanner that Thomson wanted advice, assistance and mentoring. Initially, it is clear that Radcliffe-Brown was keen to take up this role as it involved the Cape York kinship systems which interested him: indeed 11 of Radcliffe-Brown's 22 letters involved this topic. Towards the end of their correspondence Radcliffe-Brown commented, 'I am quite aware that you are one of the best observers amongst the students we had at Sydney and that you find difficulties about the analytical description of a kinship system and the production of charts or tables.... I shall be very glad to help you to the best of my ability' (received 16/11/1953).

It seems that as Radcliffe-Brown became sick in 1954 the relations with Thomson became closer, although a key letter is missing. However from Thomson's reply to the missing letter on the 28th September 1954, the import is clear. He writes, '... I was very deeply moved by the affection I felt in your letter and by your asking me to talk to you about personal things', which he proceeds to do, summarizing much of his fieldwork career. After emphasizing the many hundreds of pages of notes and the more than 2000 photographs he has from Arnhem Land, most of which were still unpublished, he writes: Will you help me and advise me as a matter or urgence [sic] to vindicate your teaching and to get this out in a series of major works? It is done in the most meticulous detail. But I need advice and criticism, and the guidance of a master.

Because of publications on Arnhem social organization by Elkin (1953) and Ronald Berndt (1955), which inter alia criticized Radcliffe-Brown, and also because of Peter Murdock's muddled views on the Murngin, Radcliffe-Brown was keen to get Thomson to work on his rich material on Arnhem Land social organization and clan groups. Thomson felt that he needed to return to Arnhem Land to clarify things before he could publish, although he kept promising to send material to Radcliffe-Brown, who went so far as to write to Sir John Latham, a former politician and recently retired Chief Justice, urging him to persuade the government to financially support the return visit. The guidance Thomson sought in dealing with his Arnhem Land materials was not forthcoming, however, as Radcliffe-Brown became increasingly unwell and then died towards the end of 1955. This left Thomson without a mentor and completely isolated from academic anthropology in Australia, since all of his energy had gone into the relationship with Radcliffe-Brown.

CONCLUSION

Radcliffe-Brown was, of course, fully aware that no other anthropologist had such intensive contact with Aboriginal people leading self-supporting lives that were, in effect, on or beyond the frontier, as Thomson. Yet even knowing this, and given the explicitness of Thomson's views about the lack of fit between Radcliffe-Brown's own view of the horde and the actual situation in the bush, he ignored them. The correspondence clearly illustrates the limits to Radcliffe-Brown's thinking and imagination.

Although Radcliffe-Brown was not really interested in what went on on the ground, the irony is that it was his ecological explanation for the horde (see 1931:107) that was at the theoretical core of Julian Steward's (1936) influential cultural ecological theory of the patrilineal band. Thomson, the ecologist par excellence, had no impact, theoretically, on the issue of land use despite his fieldwork experience and evidence. His ecological contribution that has commanded attention was to use the concept of seasonality as a descriptive device for the integration of annual variation in the food quest with variation in material culture in both Cape York and in Arnhem Land (1939; 1949a: opposite page 8).

It is clear that Thomson had no doubts about the evidence from his fieldwork. While he was clearly fearful of losing Radcliffe-Brown's friendship and guidance, I do not think that his failure to continue to make an issue of the difference between his views and those of Radcliffe-Brown can simply be put down to the desire not to jeopardize this relationship. As the correspondence indicates, Thomson was having difficulty knowing how to organize and analyse the data he had from Arnhem Land on social organization and religious life, which at least had its own internal organization in the context of Radcliffe-Brown model. So I think it is justified to assume that he would have been even less clear what to make of the information on the mixed clan composition of bands: other than to say Radcliffe-Brown was wrong, there was nothing he could do with the information. Further it only muddled what was a seductively totalizing picture of Aboriginal life that had such great appeal to his mentor, and to Claude Levi-Strauss. It is not known whether Thomson ever read Hiatt's paper, and if he did, what he thought of it.

ACKNOWLEDGEMENTS

I would like to thank Dorita Thomson for generously providing me with access to the original correspondence, Rosalind Peterson, and the anonymous referees for their most helpful comments. The paper was prepared in association with an ARC Linkage Project (LP0347221), Anthropological and Aboriginal perspectives on the Donald Thomson Collection: material culture, collecting and identity.

REFERENCES

BERNDT, R. 1959. The Concept of the 'Tribe' in the Western Desert of Australia. Oceania 30: 81-107.

1955. Murngin (Wulamba) Social Organization. American Anthropologist 57 (1): 84-106.

BIRDSELL, J. 1970. Group Composition among the Australian Aborigines: a Critique of the Evidence from Fieldwork Conducted since 1930. Current Anthropology 11 : 115-42.

ELKIN, A. P. 1953. Murngin Kinship Re-examined, and Remarks on Some Generalization. American Anthropologist 55: 412-419.

GRAY, G. 2005. A Deep-seated Aversion or Prudish Disapproval: Relations with Elkin. In B. Rigsby and N. Peterson (eds), Donald Thomson the Man and Scholar, pp. 83-100. Canberra: Academy of the Social Sciences in Australia.

HERNANDEZ, T. 1941. Social Organization of the Drysdale River Tribes, North-west Australia. Oceania 11(3): 211-232.

HIATT, L. 1996. Arguments about Aborigines: Australia and the Evolution of Social Anthropology. Cambridge: Cambridge University Press.

1962. Local Organization Among the Australian Aborigines. Oceania 32: 267-86.

1968. Gidjingali Marriage Arrangements. In R. Lee and I. DeVore (eds), Man the Hunter, pp.165-175. Chicago: Aldine.

KABERRY, P. 1939. Aboriginal Woman: Sacred and Profane. London: Routledge.

KUHN, T. 1970. The Structure of Scientific Revolutions. Chicago: Chicago University Press.

LEVI-STRAUSS, C. 1949. Les Structures Elementaires de la Parente. Paris: P.U.F.

1968. Gidjinglai Marriage Arrangements: Comments and Rejoinder. In R. Lee and I. DeVote (eds), Man the Hunter, pp. 210-212. Chicago: Aldine.

PETERSON, N. 2005. Thomson's Place in Australian Anthropology. In B. Rigsby and N. Peterson (eds), Donald Thomson the Man and Scholar, pp. 29-44. Canberra: Academy of the Social Sciences in Australia.

PETERSON, N. AND LONG, J. 1986. Aboriginal Territorial Organization: a Band Perspective. Sydney: Oceania Monograph No 30.

RADCLIFFE-BROWN, A. 1913. Three Tribes of Western Australia. Journal of the Royal Anthropological Institute 43:143-94.

1930-1931. The Social Organization of Australian Tribes. Oceania 1930, 1(1): 34-63; 1930, 1(2): 206-246; 1930, 1(3):322-341; 1931, 1(4):426-456.

1931. The Social Organization of Australian Tribes. Melbourne: Macmillan. The Oceania Monographs, No. 1.

1954. Australian Local Organization (a letter to the Editor). American Anthropologist 56(1): 105-106.

1956. On Australian Local Organization (a letter to the Editor). American Anthropologist 58 (2): 363-367.

RUMSEY, A. and REDMOND, A. 1999. Final Anthropological Report--Wanjina-Wunggurr-Wilinggin Native Title Claim. Report prepared for the Federal Court. Derby, Western Australia: Kimberley and Kamali Land Councils.

SHARE L. 1937. The Social Anthropology of a Totemic System in North Queensland. Unpublished PhD thesis, Harvard University.

1934. Ritual Life and Economics of the Yir-Yoront of Cape York Peninsula. Oceania 5: 19-42.

STANNER, W. 1933. The Daly River Tribes: a Report on Field Work in North Australia. Oceania 3: 377-405.

STEWARD, J. 1936. The Economic and Social Basis of Primitive Bands. In Essays in Honor of A.L. Kroeber, pp. 331-350. Berkeley: University of California Press.

SUTTON, P. 2003. Native Title in Australia: an Ethnographic Perspective. Cambridge: Cambridge University Press.

THOMSON, D. 1933. The Hero Cult, Initiation and Totemism on Cape York. Journal of the Royal Anthropological Institute 63:453-537.

1935. The Joking Relationship and Organized Obscenity in North Queensland. American Anthropologist, 37:460-490.

1939. The Seasonal Factor in Human Culture: Illustrated from the Life of a Contemporary Nomadic Group. Proceedings of the Prehistoric Society 5:209-221.

1948. Arnhem Land: Explorations Among an Unknown People, Part I. The Journey to Bennet Bay. The Geographical Journal, 112:146-164.

1949. Economic Structure and the Ceremonial Exchange Cycle in Arnhem Land. Melbourne: Macmillan and Co.

1949a. Arnhem Land: Explorations Among an Unknown People, Part 2. The People of Blue Mud Bay. The Geographical Journal, 113:1-8.

1951. The Aborigines of Australia. In, C. Barrett (ed), The Australian Junior Encyclopaedia, vol. 1:70-97. Sydney: The Australian Educational Foundation.

WARNER, L. 1958. A Black Civilization. New York: Harper Brothers.

Nicolas Peterson

Australian National University

NOTES

(1.) References to Radcliffe-Brown's (RB in the footnotes) statements can be confusing. His original paper was published across four issues of Oceania, three of which are dated 1930 and the fourth 1931. He wrote about the horde specifically in the first and last issues. However, all four articles were republished as the first Oceania Monograph in 1931 with its own sequential page numbering.

(2.) See Hiatt (1996:20-25) for an historical discussion of the horde concept. Elkin (1953: 417) also commented that the band was not always the patrilineal horde as conceived by Radcliffe-Brown but it seems this was more on the basis of field reports he was getting from the people listed, than from his own fieldwork. Berndt (1959:95-6) was quite clear about Radcliffe-Brown's muddle of clan and band but not so clear on band composition, although footnote 64 is interesting.

(3.) RB recognized that some hordes (ie bands in this context) did have males of more than one clan, but this was because of irregular marriages (1931:106).

(4.) The strongest evidence was in Lauriston Sharp's PhD thesis (1937).

(5.) In a 1933 paper Thomson (DFT in the footnotes) mentions the word horde (1933:459) in the following context, 'the small nomadic groups popularly known as "camps," for which the name "horde" has been proposed, being independent, self-governing units'. He has nothing further so say about the horde in this paper.

(6.) I have adopted the following conventions in respect of brackets. Rounded brackets i.e. (....) are DFT's own; Square brackets i.e.[....] are introduced by me. At one point I have used {....} where Thomson has written on a copy of the letter. It should be pointed out that while most of the letters from RB are originals--one or two are copies typed up by DFT's secretary, of a handwritten original, presumably for easier reading although RB's handwriting is highly legible. DFT's letters are carbon copies of the letters he sent. There are, some drafts of letters for which there is no carbon copy of the original, making it unclear whether the letter was sent or not. None of the latter 'letters' are used here. Where a word is underlined in a letter, this reflects the original/carbon copy.

(7.) DFT added in the words in brackets here, by hand on the typed carbon copy he kept of this letter. It is not clear whether the handwritten words were on the original sent to RB. All of DFT's letters to RB were typed but several of RB's letters were hand written.

(8.) Here we see the classic confusion of horde, normally used for the land-using group, being referred to by RB as the land-owning group because, as emerges in the subsequent sentence he is really only concerned with males. Elsewhere in this same letter RB says, 'he is inclined to speak of it [the horde] as a quasi-domestic group'.

(9.) 'A native camp is composed of two parts, the married people's camp and the bachelors' camp ... If a visitor comes to the camp and brings his wife with him, he puts his fire and shelter near the married peeple [sic], on the same side as his own country lies. If he is unmarried, or if he has not brought his wife with him, he goes to the bachelors' camp' (1913:147).

(10.) See Hiatt (1996:22) for a discussion of this neglect of women in his analysis.

(11.) See Peterson and Long (1986) and Sutton (2003:70) for other records confirming this kind of composition.

(12.) On 21st April 1949 DFF writes to RB, 'I am taking advantage of the time it takes to bring this type [i.e. the tailed n] from U.S.A., to send you the second galley proof by air mail' of the Economic Structure book.

(13.) This provides some explicit evidence for the hypothesis advanced by Sutton (2003:48) that the horde precedes the clan in RB's understanding of the process of the clan's social reproduction.

(14.) As Hiatt points out (1996:23) Kaberry made this point in 1939.

(15.) RB to DFT 5th July 1949: 'Elkin has been a disaster both for the Department in Sydney and for 'Oceania'. He is a mean cuss and makes a special point of ignoring my work and what I did for him. Raymond Firth protested to him over this'. DFT to RB 4th October 1951, 'I think that Elkin is a very little and very jealous man. This is shown, of course, by his failure to give due acknowledgement to your own pioneering work in founding the Sydney School and setting the house in order in Australia in preparation for extensive field work'. See Gray (2005) for an account of how the alienation came about.

(16.) There is no real evidence that this came to anything except that Stanner, who organized the foundational conference of the Australian Institute of Aboriginal Studies in 1961, did include Thomson, but not in a significant role, nor did he include a paper from Thomson in the published volume