1. Some Theoretical Considerations

In the last article, ‘The Lineaments of Sacrifice,’ I showed that the highest rite of the Murinbata has a marked resemblance to sacrifice, and is therefore a good deal more than we ordinarily mean by initiation. How much more, and how best to interpret the facts, remain central difficulties.

It was apparent also that the rite exhibits other interesting similarities. For example, some of the symbolisms are reminiscent of rebirth and baptism. Observers of many other Aboriginal initiations have noted the same fact.

I concentrated in the article on the structural resemblance to sacrifice since apparently it had not been noticed or stressed by others. What further constructions of the rite may be made?

The first duty of anthropology in dealing with Aboriginal religion is to try to elicit the kind of reality the facts of study have for the people responsible for them. The data might be described as natural facts of human conviction about the ultimates of life. In other words, they are products of human passion, aspiration and imagination. They are what they are because Aboriginal mentality is what it is. In this sense they have their being, as realities studied by anthropology, within what could be called an ontology of life. It is that reality which anthropology must set up for study as best it may (Fig. 2).

In the course of doing so the typical difficulties become rather more marked than when other topics are under examination. The necessary means—abstraction, empirical methods, and an indicative-analytical language—are at all points being transcended by the facts. One neglects 96some features of reality in order to attend to others; empiricism does not exhaust even what is selected; the chosen language may dry out and at times dessicate what is under study. And in an attempt to outwit the difficulties the analyst can scarcely avoid what Professor Firth has called ‘the personal equation,’ an implicit valuation of facts under the guise of objectivity. In dealing with the more mundane topics one hardly notices the problems. But they make themselves felt in an acute way when one engages in study the ultimate facts of religious form and concern.

In these papers one of my primary aims is to make as explicit as I can aspects of Aboriginal religion which tend to remain tacit in more conventional approaches. To try, as a conscious aim of study, to bring out the onotology of a type of thought and life must involve one, I think unavoidably at present, in some degree of implicit valuation. On the other hand, the distortions made by the spurious ideal of complete objectivity can have even worse consequences.

The tradition of anthropology has always gone beyond the study of man simply as Homo sapiens and Homo faber. Beyond, too, the rather inelegant conception of Homo socius. The analytical construct ‘man-acting’ has never been more than a convenient travesty. The true subject of study is really Homo convictus, to use the term suggested by Zuurdeeg.1 That is, man to whom it is natural to act socially within a system of life depending on overwhelming convictions about ultimate values.

A certain aridity has lain over anthropological studies of religion for some time. That self-styled ‘encyclopæædic inventory’ Anthropology Today was perhaps a straightforward case of neglect. It did not contain a single paper on the topic and in its index listed only 11 of 929 pages as dealing with central questions. But many studies which are now remedying the neglect are still affected by a kind of theoretic blight. I refer in particular to the presupposition that the social order is primary

99and in some sense causal, and the religious order secondary and in some sense consequential. Thus, studies may issue in general propositions to the effect that religion ‘reflects’ or ‘expresses’ the social structure. It is quite difficult to see why such statements seem important or even interesting. They are not clear even as metaphor, since a reflection would be a reversed image, which the theorists do not of course mean, and an expression which is not a concentration—and, in this sense, an elucidation—does not advance analysis. I do not adopt any such formulation, and prefer rather to regard facts of the social order, if they are disconnected from religion, as providing one of the symbolical languages through which ontological reality is conveniently expressed. I have allowed questions of social structure to remain tacit largely for that reason. A few further words may be in place.

Conceptions of social structure seem usually to issue in models about human relations. They must be distinguished from models of or after the observable reality of those relations. The relations of social structure seem usually to be visualised as connections or interdependences between points of force which are characterised as role-positions on a kind of network. The conception is wholly unlike that which I use. One may actually see the constituent operations of Punj. One has to infer the role-positions and enactments required to clothe the structural models. I would argue that the operational model is perceptual of or after the matter and constitution of Punj, whereas the structural models would be about form and function. They would also be more preceptive than perceptual.

I have sought to avoid using any formulation which would simply be a reversal of the terms of the rejected presupposition. That is, to avoid setting up the religious facts as primary and social facts as secondary, which I do not think is the case. It is thus necessary to examine the relationships between at least three sets of phenomena: the ontological system, the conjoint phenomena of myth and rite, and the main social relationships as continuously interdependent over time.100

In this article I shall follow Robertson Smith2, Roheim3 and Elkin in making use, though in my own way, of the general idea of ‘sacramentalism.’ By such means a certain order can be found in the ontology, more particularly with respect to the motivation of rites and the institutions of totemism.

The main rite is marked by the use of external and visible signs betokening men’s dependency on otherworldly powers for an endowment and flow of life-benefits. It is the set of relations which obtain between these elements which constitutes sacramentalism. Men act through signs towards the ground of dependency; the flow is accompanied—or is held to be—by external signs signifying that a solidary relation holds between that ground and men; and in this way the acts, signs and flow not only interpenetrate each other but in a long established and involuted religious system compenetrate each other, that is, pervade each other in every part. In such a sense Murinbata religion is sacramentalist through and through. It could be described as a totemic sacramentalism. In order to understand it one has to examine the nature and principle of the endowment (that is, the totemic foundation), the exchange of signs (the rite), and the plan of dispensing or distributing the flow among men (the social institutions). One is dealing with what might be called a religious ‘economy.’ It is inherently possible to set up comparatively a counterpart of the formulation which, in a particular tradition, deals with ‘an exchange of prayers for grace’ as a system of determinations within an ‘economy of salvation.’ Such an economy is not the whole of a religion but is an important part of it.

The relation between rite and myth is one that seems to me very perplexing. Anthropological thought on the matter is quite unsatisfactory. In the main I depend on Robertson Smith’s formulations, which seem to me still the most cogent. The myth of Mutjiŋga or The Old Woman is now coordinate with the rite of Punj, and functions correlatively with it. But the conjunction is a fact of history and, although positive evidence

101is lacking, I do not doubt that the myth derives historically from the rite. I take the view that the rite is obligatory and constant, whereas the myth is discretionary and variable. The study is confined to the coordinate and correlative aspects in view of the facts that only speculation would be possible about the historical question, and that there are three other complicating problems: the rite is heavily adorned æsthetically, many systems of symbolism are in use, and the cultus is infiltrated by the political power of men. The historical aspect is certainly important both in relation to the particular facts of Murinbata religion and to the general theory of mythologised rite, but I do not feel able to do more than refer incidentally to it.

An objection has been made (in a private communication) that in the first article I did not specify what kind of sacrifice Punj resembles. I consider that the criticism is based on a misunderstanding. I used the words ‘Lineaments of Sacrifice’ precisely because I had elicited an implicit general form, which might be considered a logical and conceptual possibility of the religious culture. The known species or varieties of sacrifice are phenomena of particular traditions, which are much less than fully known. It seems to me a waste of time to try to infer from half-known histories to an eventuality which might have emerged in a history that did not happen.

2. Relations between Punj and Circumcision

The Murinbata distinguish in the most explicit way between Punj and its prepubic counterpart (Djaban) on the one hand and the circumcision of youths at puberty on the other. The two classes or varieties are—and are seen as—quite distinct in kind, purpose and importance.

In the major rite the veritable signs of The Mother—blood and the bullroarer—are used to transform young men who, if they are to carry life on, must learn the signs which alone have efficacy in drawing the flow from the endowment. The signs are reminders of the past; they have a transformative efficacy in the present; and, somehow, they 102guarantee what is to come. In such respects they are of the essence of all sacramental observance. The rite itself is episodic and transitory, but it has a permanent effect. It fixes an altogether indelible character on those who are submitted to it. A Murinbata man who becomes Kadu Punj becomes one who, in their own language, ‘holds something in the ear,’ i.e. ‘understands.’ No one ever undergoes the rite a second time as principal. He attains, irrespective of subsequent conduct, an irreversible state and condition. The distinction between those who have submitted, and those who have yet to submit, is always absolutely clear. The effect of the rite is recognised universally within its region of force and meaning.

The Aborigines are a contemporary people, and the ideas and forms of conduct I am dealing with are also contemporary or recent. It is impossible to know or say anything about the remote history of the forms and ideas. It would therefore be a wrongful procedure to make, or to try to make, a comparison between Murinbata and Christian sacramentalism such that the former is made to appear primitively in a uniform, linear, serial sequence with the latter. To do so would be to narrow what has to be widened, that is, generalised. The viewpoint I adopt has been well stated by Atkins in the observation that ‘the deeper fault of all Churchly sacramentarianism is our undue limitation of its meaning’ and that the idea of sacrament is ‘capable of a vast and helpful extension.’4 At the same time it is at least of interest to point out how fitting to the Murinbata evidence are some early conceptions of Christian sacramentology. The rite may well be compared to a seal (sphragis) which puts an impress on those who submit to and receive it, so that it signs them in a character (signaculum) which is indelible. In Murinbata eyes the character or property is one of ‘understanding,’ and it will be recalled that one of the early comparisons of the effect of baptism was with ‘enlightenment.’ Indeed, very much as sacrifice may be thought of as a logical possibility of Murinbata religious culture, though a possibility which did not eventuate, so one may say that many conceptual possibilities exist within their ruling ideas. But I feel that

103it is not for anthropology to force the emergence in a speculatively comparative way. We have to study the religion as a sacramentalist complex with differentiations of its own. The pattern of differentiation is distinctive.

A characteristic of the system is that neither birth nor marriage attracts rites and ceremonies of a sacramental kind. The tradition does not select these occasions of life as religiously significant.

The circumcision is not Punj, in other words is not a secret, obscure and dangerous thing. Its purpose is to make a youth into a man, whereas the purpose of the rite of the bullroarer is to make a man into a man of mystical understanding. He is a man transformed, one who knows the truth—or as much of the truth as anyone in the tradition can tell him—about a cardinal mystery of human affairs. And nothing like the importance which attaches to the bullroarer rite is attached to circumcision. In this system it would be entirely meaningless to fix on Punj any such label as ‘spiritual circumcision’ in the manner in which a link was made, rather putatively, between baptism and the complex of initiation rites existing before the Christian period. However, nothing I have said implies that circumcision is unimportant among the Murinbata. Far from it. It has high importance, sufficient to override every unconnected activity for days at a time, and to leave lifelong marks on personality, repute and status. It too has the aspects of sphragis and signaculum. But it is not nandji ŋala ŋala (lit. ‘thing,’ ‘big,’ ‘big,’ the duplication of ŋala being an intensification), and it in no way ranks in value, or competes in purpose and function, with the major rite. The two are of distinct orders, and are on different planes in that at circumcision mysticity recedes to the far background, and the considerations are dominantly temporal and secular. The fact that it is not intended to make youths ‘understand’ shows that it is not a spiritual transaction, and cannot be raised to the dignity of a sacrament.

The only other sacramental rite of importance in the regional culture is that attending death. Those which have been called ‘increase’ 104rites, intended to regenerate the vital principle of the totems, were not developed to any extent in the region. Nor were the first menses of females marked by elaborate ceremonial.

3. The Sacramental Plan

Aboriginal social life was a volitional and purposive system deferring to traditional data or givens. One may speak of it as tradition-directed. Any order the actualities of life exhibited came from a consistent observance of the givens. That system is now changing fairly rapidly. There are new objects of life to pursue through activities which have no, or very unclear, data to defer to as a means of rationalisation. The younger and very activist generation has yet to find by what to systematise itself. In these articles I am concerned with the traditional system.

The essential point of approach to the religious life of tradition is the phenomenon of totemism. It appears to be the main link between cosmology and ontology on the one hand and the social order on the other. Certainly, it forms the best approach an external observer can make to Aboriginal conceptions of the reality they experience.

An understanding of Aboriginal totemism also allows one to proceed to the stage of theoretical formulation, which must necessarily be stated in terms of a foreign intellectual tradition. I am taking such a step when I refer to a religious ‘economy,’ with the meaning that the facts are being phrased as those of an ordained (and, therefore, fixed) endowment which is dispensed on a certain plan of entitlement, so that its flow is utilised under conservational forms. As I phrase the formulation, the religious system is sacramental because it is totemic, and it is economic because it is conservational. Totemism is the foundation as well as the frame of the sacramental plan. There are four classical types of totemism in the region, a fifth having been imitated from tribes to the south over the last three generations. The function of Murinbata totemism is to mediate the first ordained order to living people. It may be said to dispense the ancestors’ endowment, and to 105provide channels for the flow. The types will be listed with a minimum of explanation.

1. Personal or Conceptional Totemism. Each person has a private totem (mir) that links him as an individual with a non-human entity either made by the ancestors or given its first formed relation to social life by them. The totem is identified by means of some notable incident associated with a child’s conception. The ŋaritŋarit or child-spirit is held to use an incident to draw a husband’s notice to the fact that a wife is conceiving a child. The totem is the corporeal agent of the spirit-child. Virtually any entity in the universe—fire, lightning, smoke, the sea, sickness—can be mir.

2. Sex Totemism. All men as a set are linked jointly with a class of non-human entities, as are all women. Kinship terms are extended to the totems (unnamed except by the names of the species). Thus, all men hold the woollybutt tree as being in a relation of kaka, mother’s brother, to them. The stringybark stands in the same relation to all women.

3. Clan Totemism. All members of each exogamous, patrilineal clan are linked with a large class of animal or natural entities which are thought of as being connected, as species, with particular places within the clan territories. The connection is intrinsic and perennial. The species, if vital, reproduce themselves through the sites, and can also be reproduced there by the totemites through magico-religious procedures.

The nature of the vital principle is obscure. Evidently it lies ‘in’ the species and not ‘in’ the association of a species with a place, for it can be stolen and reproduced elsewhere. Units or even fragments of the species (but not the place) are often stolen. The vital principle then works elsewhere. Such thefts are mortal offences. The clan members have an unquestioned right to possession of the totems (ŋakumal), the sites (ŋoigumiŋgi), the region (da), its resources, and any incorporeal correlatives such as songs, dances and emblematic designs. The system may be called ‘cult totemism’ since the designs, which have public and secret forms, are worn by males at ceremonies.106

4. Moiety Totemism. All persons, whether male or female, in each patrilineal moiety jointly possess a totem (unnamed except by the names of the species) distinguishing them from all members of the other moiety. Each totem is pule (old, senior, authoritative, friend, and in English ‘boss’) over all totems of the clans constituting its moiety, but the relation is somewhat vague and lacks precise system. Mythology represents the moieties, which are named after Tiwuŋgu, the eagle-hawk, and Kartjin, the kite-hawk, as composed originally of different orders of people in continuous conflict. Each possessed a vital resource—fire or water—without which the other could not live. They were saved from an animal-like existence and possible death only by co-operation. The moieties compete with each other in games, may abuse each other in conventional terms, and assist in the regulation of marriage. But they are not stressed as such in ceremonies or in camping patterns. The type is logically derivable from (3).

5. Sub-section Totemism. All members of each sub-section or ŋinipun (in English, ‘skin’), of which there are eight, jointly possess a class of non-human entities (named ŋulu) which are not connected locally with places. The system was copied from tribes to the south in historic times, and is neither well understood nor assiduously maintained except among the oldest men, who tend to regard it now as one of their secrets.

The effect of the totemic complex as a whole is to parcel out, on a kind of distributive plan, all the non-human entities made or recognised by the ancestors, and given relevance one to another, that is, set up in a moral system. The outcome is that, with the exception of (I),5 the totemic sets are sacramental corporations of a perennial order.

107It is the associating of a totem with a set of people which makes them into a group, that can operate, or be operated on, in virtue of the totemic sign. Members of the group have a ground of right over the totem and anything it stands for. They are also held answerable for all it imports. The rights are not always clear. In (2) the relation of totem to group does not seem to be much more than emblematic.

The nature of the bond between totem and totemite is best set out in relation to (4). The ŋakumal totem is commonly referred to as ŋatan, ‘brother,’ and in English as ‘my dreaming.’ When the site or ŋoigumiŋgi is referred to a possessive adjective is also used. I have not succeeded in finding out the literal meanings of the terms or in separating the morphemes distinctly. In English, the totem place is called ‘my dreaming’ or ‘my dreaming place.’ Persons of the same clan are ŋakumal numi, in English ‘one dreaming,’ often intensified to ‘one dreaming, one country.’ The bond, primarily mystical, can also be said to be ‘essential’ in that there is an identification of the intrinsic self with the site (‘myself there,’ ‘just like myself,’ ‘my ŋjapan (spirit) there’), and ‘substantial’ in that a corporeal connection is also asserted. There is the idea of a transmissible biological type, but the what and how of the connection are obscure.6 In a limited sense the bond is historical, for common knowledge of actual persons who were associated with the place and totem may go back for three generations, rarely more. It is historical too in the longer measure if we accept the pseudo-historicity of myth. But it could not accurately be called ‘genetic,’ or even ‘pseudo-genetic,’ for the following reasons.

Genealogical tables of up to six generations are not particularly difficult to make. They are purported to be true and may well be true in part, for a high proportion of entries withstand repeated cross-check between people not in collusion and with nothing obvious to gain by lying. But the tables are worthless as evidence of genetic fact. There is now, and presumably was in the past, a high rate of extra-marital sexuality;

108classificatory brothers have a right of sexual access to wives; multiple and short-term marriages by women have been a high proportion of all marriages; no importance whatever is attached to conception or birth outside the clan country; the period of gestation is not known with any accuracy; and there is the immense fact of the mystical theory of procreation. The men who help to construct the genealogies are not thinking genetically. What they do is to trace the issue of children from women who were then consorting with men in conditions in which the men had a right of claim to the issue. Murinbata patriliny is the exercise of a political and jural right arising from spiritual agnation.7⁸

The genitor property of paternity is a function which can be—and, on sufficient evidence, is—performed just as well by a man’s brother as by him. A man can institute the minga or ‘firestick’ relation with his brother’s child by ‘finding’ a child-spirit in a dream and directing it to his brother’s wife. The pater property is the exercise of jural and political right founded on the same mystical ground. The clan is a set or collection of people made into a group, with the unity we try to denote by the word ‘clan,’ by a sign affixed perennially to them. The sign is the ŋakumal. Its referent is the mystical complex ‘one dreaming, one country.’ Its connotation is a range of rights and duties. The ‘patrilineal clan’ is patrilineal because there are both a spiritual or essential link and, somehow, a corporeal or substantial link over time between fathers and sons; it is a clan because the sign makes a unified group from a set or collection; and it is perennially corporate in its estate, corporeal and incorporeal, because everything has been from the beginning the

109inalienable8 right and duty of those who have had the sign from the beginning.

Adoption of members of other clans—especially of the children of captured women—is common. It is fully jural in the sense that there are rules, but no adoptee ever loses his rights and duties in animum and in rem in the estate of the consort of his mother at or about the time of birth, or acquires those of his adoptor clan. A complete entry into the sacramental corporation by adoption is impossible.

A ground made in this way is a condition of the intelligibility of many events or institutions, for the necessity or requiredness is consequential on totemism. A bare summary may suffice.

The ku or gu in ŋakumal and ŋoigumiŋgi evidently means ‘flesh.’9¹ Birth is the event of arrival, on the plane of the visible, and in the world of the living, of flesh somehow shaped or constituted by a child-spirit (ŋaritŋarit) and a soul (ŋjapan). The doctrine of the commerce between child-spirit and soul is obscure. But at birth a spiritual duality shows itself in flesh affiliated mystically with a place and a species. The actual birthplace is simply where a baby ‘falls’ so as to ‘make a dent in the ground,’ and is otherwise not important. The word ‘totemism’ conveys very poorly the idea of a ground on which life forms and to which it returns. At death, the former practice coupled platform-exposure of the body with (in the case of men) an eventual rite to free the soul (nothing is said of the child-spirit) to go elsewhere to make a new entry to the

110visible world. The doctrine allows for the entry of a given spirit into persons of other clans and even of the opposite moiety. The last ashes of the burned remains of the body were then stamped into the ground of the ancestral clan by men of all clans. In this way a cycle was completed, but not closed, for the spirit retained its capacity to return to the visible world.

The preferred marriage—with a classificatory cross-cousin—takes that form evidently because at that point the totemic signs of father and mother have lost force sufficiently to be disregarded. The condemnation of marriages which are said to be ‘too close’ can be analysed to have that meaning.

Trade, which is carried on intertribally, is a transmission between clans of valuables which have been held or touched—‘owned’ is not the word—in a given sequence by every member of each clan,10 excluding young children, before being passed onto the next.

Outside the clan there is, with one exception, nothing heritable or inherited. At a man’s death his chattels are destroyed, his debts paid, and anything else stays with his clan. If he has the title and office of pule (which means in this case seniority in the clan) they too stay within the clan; likewise visionary skills (to see spirits, to dream songs, to divine malefactors) which gifted men like to bequeath to their sons. The exception is the rights of safety, hospitality and hunting-privilege which a man has in his mother’s clan-country (kaŋatji), which merely lapse. In every other place but his own and his mother’s clan-country a man is kamalik, a stranger.

I am trying here to give in brief compass an account, not of social structure as it is usually understood, but of the way in which groups are constituted and have a set of functions under what may be called on these evidences a totemic determination. The totemic system has a threefold efficacy. (a) It provides a wisdom or principle or logos by which the Aborigines conceptualise groups of people. (b) It unifies sets or collections of people as more than sets or collections, in short,

111gives them the ‘group’ property, so that they are supposed to exemplify physically that which characterises them socially in their own eyes and the eyes of others. (c) It points two ways through a special class of totems: back, to a ground and source which are self-authorised and self-authorising; and on, to powers and rights in various classes of functions for the members of groups which have their being—and a fortiori the powers and rights—only by virtue of the authority set up. The ‘necessity’ of the relations is a totemic necessity, and totemism is the language of the ontological system.

But there are many such clans, neither solitary nor self-sufficient, ranged in parallel (Kartjin clans tend to cluster along the coast) and in series (because the clustering is only a tendency). Their common life of interaction (through marriage, friendship, trade, war, hunting and other types of activity) requires a different method of analysis. The necessities behind these relations are only distantly, if at all, concerned with totemism, but arise from different conditions presupposed for each class. I do not propose to embark on them here. But the principles of such associations are obviously multiple. No one principle—certainly not that of totemism, or that great standby of Australian anthropologists, ‘kinship’—orders the totality of interaction. The fact is that the principles of social interaction are conjugate,11 by which I mean that each is limited in range, true only in given conditions, and to some extent inconsistent with every other. Thus, the spatial principle of ‘closeness’ or ‘distance’ modifies the totemic principle sufficiently to allow men to say cynically ‘distant sisters are good sweet-hearts,’ and to kill without compunction men who, under the terminology of kinship, are ‘brothers.’

It follows from the above that, except under hypothesis, Murinbata society and/or culture cannot be set up as ‘a unified whole.’ The hypothesis seems to me quite unpersuasive. The ‘principles’ of social interaction do not appear to have a ground of unity which can be stated. Indeed, the metaphysical conceptions which the Murinbata

112have developed about their own ontology of life are filled with what for the present may be called ‘dualism.’ It is thus impossible to deal, in any rigorous sense, with a social system which is supposed to exhibit a structure. The many classes of relations of association exhibit several distinct though connected structures. A brief discussion may help to remove ambiguity.

The relations of association between clans through their members are visible as conjoint acts. These when repeated exhibit form or pattern. One may characterise the form or pattern as a structure of operations. The ceremony of Punj is a type-instance. This structure (let us call it S) is a sequence or process of observable features, the acts or operations which are clearly named or identifiable. But a necessary condition of Punj is that its principals be guided and guarded throughout by one or more members of a class of naŋgun, or wives’ brothers. The quasi-genealogical relation is a necessary condition of S, and part of the ceremony may be studied as a relation between referents (naŋgun) and relata (initiands).12 In such a case we have also to study operations but, if desired, we can study them as functions of naŋgun towards the young principals; and thus, if one likes, as functions between ‘positions’ in a ‘network.’ The schema of description, abstraction and interpretation is then quite distinct. A structure of functions in this sense may be distinguished as S, and S, the operations, may be regarded as constituting the matter or content of which S is a form. But the persons present on the ground of Punj are there as members of families, clans, moieties, sub-sections and the like. That is, as members of groups which form a segmental structure. This structure, which we may call SS, is visible only from time to time and in masked forms. What one sees at Punj is the structure of operations (S); one has to infer or construct the structure of functions (S) and the segmental structure (SS). As Professor Firth and many others have shown, the ‘structure’ idea could be taken a good deal farther.

113If S, S and SS are confounded, as they tend to be in many discussions of ‘social structure,’ an ambiguity which is quite intolerable results.13 My analysis here is intended to avoid this kind of ambiguity. The objects are limited. I wish to distinguish a class of groups relevant to the religious life (the totemic class); to sketch the types of functions which fall to them by virtue of their constitutive sign; and to show that the acts of sociality taking place between them, and thus forming the ‘relations,’ may be studied as operations with a distinguishable structure. I believe that ‘a structure of enduring relations’ must be taken in this sense, as a structure of acts of sociality, when the ontology of the Aboriginal life is under study.

Although the sketch is brief, I hope that it brings out a main point: the totality of Murinbata life is one of multiple principles. Because the principles are conjugate they affect different regions of life which overlap and are in conflict. The sacramental or totemic principle sets an ideal which covers only part of the total field of necessary interaction. If the principles have a unity among themselves I have not been able to find what it is, and doubt if it exists. Certain aspects of Murinbata tradition suggest a working towards a unified system or unified whole. But to utilise a theory in which the fact of a unified whole is a postulate, or even a hypothesis, seems to me to be without warrant.

If one were to try to make a picture of the structure of sociality, it would have nothing in common with the ‘network’ imagery of structuralist anthropology. The Murinbata themselves make a kind of picture of the articulation of the segmental groups. They use sticks or stones in such a way that what emerges looks a little like a branching tree or a flung fish-net. But it is not a picture of sociality. That picture exists in the dramatisation given by Punj in complex symbolisms of mime, song, dance and rite. The ontological reality stated there is not

114reducible to points of force on a network. A ‘theory’ of that reality would have to be a rationalisation of a reality which, if my account is correct, the Murinbata put to themselves as a joyous thing with maggots at the centre. It takes considerable temerity to try to improve on this imagery.

I shall therefore not extend the sketch, as would be necessary if my purpose were different, to the complex structures and segments of Murinbata organisation, or to the functional classes. Instead, I shall pursue the argument that the religion may best be studied for what it shows itself to be: a celebration of values and at the same time a dramatisation of the moral imperfection of social being. The thesis is much that stated by Höffding14 many years ago: that religion is determined by ‘the fate of values in the struggle for existence.’ He was speaking of the precious social values and of religion as their conservation. The view has been described as ‘a bad and grovelling philosophy of religion’ by a writer15 from within a particular tradition which itself appears to illustrate the thesis. But Höffding added cautiously: ‘if this is so, we must not ignore the possibility that this underlying element of religion may exist and operate without expressing itself either in myth, dogma or cult.’ In the traditional Murinbata religion that element finds an expression.

The material which follows lends itself to many uses and interpretations. It can easily be wrought into fancied shapes by those of particular convictions. I would regret such uses though I cannot guard against them. I do not myself regard the facts as giving any weight to general theories of natural religion or to the conceptions of any particular religion.

4. The Search for a Paradigm

The attempt which I made in the first article to draw a picture of the rite of Punj resulted in a spiral path or course of connected acts. If the same approach were made to the facts of the two earlier initiations—Djaban

115and circumcision—and if the three were connected together we should have a whorl-like path or course which gives a kind of total picture of the ritual pattern to be observed in the development of male social personality from tender youth to manhood. As a picture it would have only the worth of its logic and assembled facts. But the totality of a community’s life is, as Radcliffe Brown observed, ‘not any sort of entity.’ The only hope of attaining some understanding within that totality is by the study of the identifiable processes which are entities. In the circumstances even a poor picture is better than none. The initiations are separable and identifiable entities; they are processes or tasks taken on by men of authority, carried through, and ended with well-stated tasks attained. The observation and analysis can be kept at all stages under inductive control. The factual descriptions I have given, while by no means perfect, are not, I believe, greatly in error. The expository device used brings out a property or set of properties of the processes which seems to me to be paradigmatic of Murinbata religious culture. Another approach of the same kind would no doubt produce an equally interesting shape. The task of theory is to enquire into the properties of processes which take on such shapes when pictures are drawn of them. An equal simplicity and clarity are not attainable at present by the use of words and sentences.

Each process is a serial sequence of connected acts. The three processes lie end-to-end in a temporal order. The principle of each is the same: a setting aside from normality, a kind of destruction, a kind of transformation, and a return to a new normality of the same order as, but qualitatively distinct from, the original. A simple word to describe the movement would be ‘zigzag,’ a more ponderous one ‘dialectical.’ The facts are quite observable, and if I use ‘dialectical’ I do so without any reference whatever to the philosophical and argumentative meanings of the word. Nor do I conjure up any images of T. S. Eliot’s ‘vast impersonal forces.’ I am using for things which I have seen, and could have been seen by anyone, a word that describes perhaps better than any other word the form of process in which the things occur. I propose for the moment to leave the usage abstract and allusive, but to try to connect 116it with a generalisation made by a different approach to the religious culture.

On the facts put forward in these papers, Murinbata religion might well be described as the celebration of a dependent life which is conceived as having taken a wrongful turn at the beginning, a turn such that the good of life is now inseparably connected with suffering. The terms of that statement may now be considered in a little more detail.

(a) The Celebration of Dependence

All the peoples of the region live in what is one of the most favourable environments in Australia. They look on it as abundantly stocked with the means of life. The traditions have little to say of hunger, nothing of famine. Certainly, it is no Polynesia; scarcity makes itself felt; conflict takes place over unlawful use of clan preserves and over the theft of species for magical propagation in places where they are dwindling or have disappeared. But the well-watered countryside has been able to sustain a relatively dense population without difficulty, more easily, for example, than the fertile Murray River Valley. The coastal clans are better off than those inland because of tidal flats and mangrove forests, which are rich in food species. A widely-distributed plant, the zamia palm, provides green or dry nuts as a staple food all the year round. But the clans of the plains, swamps, savannah and hills do not consider their lands less fruitful. Many do not relish seafoods even when available, since the taste does not please them. Every clan tends to boast of the plenty with which the ancestors endowed it.

The idea that living men are lesser beings than the ancestors, and dependent on them, is strongly held. It is justified by a mythology which uses a simple but vivid imagery to show how great the powers which men have lost were. The ancestors stocked the land with rivers, springs, food, weapons and other means of life, raised up hills and mountains, put spirit-children into the waters, used the wind and songs as agencies of will, went up into the sky, provided dreams as a means of communicating with the living, and performed a host of similar marvels. Perhaps the 117greatest were those by which the ancestors transformed themselves into animals,16 thus instituting the relationships of totemism, and left life-giving principles in the estates of the patrilineal clans. The nature of the marvels is the measure of men’s dependency. But the Murinbata attain a buoyant and even high-spirited attitude to life in spite of its contingency. There is nothing even approximately equivalent to the ancient Nile-dwellers’ brooding on ‘the carnage of the year.’

In writing of them one is always tempted to use the word ‘celebration’ instead of ‘ceremony.’ It would put no strain on the meaning of words to write of Punj and Djaban in such a way. From all accounts the same spirit ruled the Karamala, the first-stage initiation last performed about fifty years ago, and another rite—the Tjimburki, on which I have only a little information—which ceased about the same time. The word would even fit, though not quite as well, the expressive quality of the circumcisions, the formal fights between clans, and the games between moieties. I do not of course suggest that ‘ceremony’ could be dropped. Far from it, since ceremoniousness is the most visible property of the several classes of formal conduct. But ‘celebration’ peculiarly fits the quality and function of the class of formal conduct which is also religious.

All trustworthy accounts of Aboriginal life have brought out the facts of the vitality and enthusiasm that accompany the religious rites. The account given in the last article was not overdrawn in its suggestion of a rapt celebration of whatever is central to Punj.

(b) The Wrongful Turning of Life

It is impossible to say to what extent the Murinbata of the past believed, or those of the present believe, what are put forward as their beliefs.

118The older men and women certainly appear to, and scepticism is by no means complete among those who are younger. The mythology, which is well formed, is still widely cited. The element relevant to the inquiry is the persistent suggestion of many myths that there has been some kind of ‘immemorial misdirection’ in human affairs, and that living men are committed to its consequences.

The myth which is central to Punj is not the only one which could be referred to,17 but is particularly valuable in the present context. It is given here in a contracted form.

(The myth then relates, in much the same phraseology, how eight more children, making ten in all, were swallowed).

Even allowing for differences of idiom, the myth and the rite do not match each other—at least on the analysis so far put forward—in all particulars. There is a good organic connection but, since we are dealing with the myth of the rite, we have a natural expectation that the two will fit together. But the respective broad patterns can be narrowed to virtual identity of theme and form: a setting apart of life from normality, an act of destruction, a transformation, and a return to normality in a new

121status and a new locus. Just as the study of the pattern of the rite led in a surprising direction, with its suggestion of depth beyond depth, so does a study of the pattern in the myth.

The myth—like the rite—is obviously dense with import, but not immediately understandable import. There is rich material for analysis within several disciplinary approaches. The immediate task, as I see it, is to try to resolve from the complex what is primary to our purpose, and to leave what—in that sense—is secondary to different studies. The remainder, of course, includes many matters of intrinsic interest to anthropology, e.g. the illumination of relationships of social structure; the adduction to the myth of common-life situations so that they have symbolical meanings; and the syntax and idiom of symbolical conceptualisation. The psychoanalytical schemata are not within my province and, in any case, not every anthropologist is persuaded that truly universal symbolisms have been demonstrated beyond question.

Our first question must surely be: what construction does the Murinbata themselves put on the myth?

With many myths it is often a task of great difficulty to answer such a question. But there is little difficulty in finding out the mentality in this case, since the mature men freely discuss the significance of Mutjiŋga’s death.

From many conversations with the men the following points emerged quite strongly. (I) Mutjiŋga was once Kadu, that is a truly human person, not one of the self-subsistent spirits recognised within the theogony. (2) She had primal authority: what she did determined for men much of the subsequent shape of reality. (3) Her death was, and still is, a matter of sorrow: ‘she should have lasted a long time’; ‘the people did not want to kill her’; ‘they wanted to keep her alive for Punj.’ (4) Her death was the consequence of her own mysteriously motivated act, and was inevitable: ‘she went wrong herself’; that is (as Left Hand reminded her) the fault was hers, not the people’s; the act was maŋe nigunu, ‘her own hand.’ (5) The loss to man was irreparable: the sentiment is usually expressed by a phrase20 which is translated into

122English as ‘bad luck,’ but is possibly better expressed as ‘sad finality,’ with an overtone of something like ‘loss’ or ‘waste.’ (6) Mutjiŋga’s act was wrong, but apparently only in that it was premature: she should have waited until the children were grown and ready to become Kadu Punj; swallowing them would then have been right. (7) Because she died, men now have only the bullroarer, which was made in order to take her place i.e. stand for her and (as I said in the first article) to be her emblem, symbol and sign.21

The problem of the ‘meaning’ of the myth is one to which I have no satisfactory answer. Whatever may be the situation in other societies, in Aboriginal Australia it is impossible to ask questions bearing directly

123on the matter. There is no way of forming the questions unambiguously in the vernacular, and Aborigines who speak English well not only have nothing relevant to say themselves but cannot phrase the rationalistic-type questions in their own tongue.22 The usefulness of both direct and indirect questions falls off sharply. Even old men of intelligence and stamina who survive many inquisitions are apt to shrug, and say: ‘it is a thing we do not understand.’ This always struck me as an excellent definition of mystery.

In the most general sense we are in touch with a world-and-life view beginning to take what would eventually be a credal form. But it is the kind of embodiment which is innocent of detached intellectualism. With such materials Robertson Smith’s trenchantly-expressed cautions need to be kept very much in mind. ‘The myth apart from the ritual affords only a doubtful and slippery kind of evidence’ as to the fixed and statutory elements of the religion. While ‘men would not be men if they agreed to do certain things without having a reason for their action’ we cannot look to the myths for a statement of those reasons. Nevertheless, accepting the cautions, we may still find at least indirect evidence that to Aboriginal minds the nodal experiences of life—and thus, the primordial formula stated in the myths—have a certain tragic quality. The evidence is rather more extensive than I am able to indicate conveniently here. But one may say that it depicts man-in-the-world as exposed to untimely death, treachery, violence, warlockry and harmful influences of many kinds; as given to malice, bad faith, egoism and jealousy; and in a sense as bereft of justice.

The actualities of life undoubtedly cohere with such an image to a large extent. There is a stratum of living on which much disappointment and bitterness accumulate. Marital infidelity, back-biting, mischief-making, violence, treachery and delict are very common indeed. But the

124mythological conception is on a grander scale than that of the domestic scene. The scale is what I refer to as macro-experience (Fig.1).

To argue that the Murinbata act as they do because they have the conception of life sketched in the myths would be absurd. To argue that the facts of life are the reason why they have the conception would be an undue simplification. It seems to be nearer the truth to say that actuality and conception are variables which have developed together with others. The two cohere, but to say anything of an evidential kind about the development seems impossible.

Ordinarily, Aboriginal religion is not represented as concerned with metaphysical problems. I am not able to share such an opinion, and think that the impression of a rather mindless participation in rites created by some works reflects the outlook of the analyst rather than that of the Aborigines. A prolonged exposure to the rites comes to suggest both a depth and a dignity of outlook which may lack formulation but not reality. The implicitness—by which I mean wordlessness—of the conceptions does not detract from the kind of reality they have. The celebrants of Punj cluster in a shallow, circular hole in the sacrosanct place where the real presence of the Mother is supposed to manifest itself. The hole is conceptualised either as a ‘nest’ or a ‘wallow.’ The first is a fairly clear symbol of family and sociality. The second is in some sense its reverse. The buffalo, in this region usually a solitary wanderer, makes or seeks a wallow against the heat of the day or to free itself from irritating pests and parasites. The symbol thus denotes what seem like positive and negative statements of the same truth about life: at the centre of things social, refuge and rottenness are found together. In other words, there is an intuition of an integral moral flaw in human association.

Few facts or institutions of living are left without some sort of mythological warrant for being what they are. A supposed past is described in ways which are held to account for the present. The standards of sequence, completeness and logical relatedness attained in the myths are not impressive if the stories are taken in a literal sense. The gaps, variations and contradictions do not allow one to say that, 125in any rigorous sense, the myths constitute a ‘system’ of belief. There is also a certain amount of evidence that we are in touch with an historical composite. At the same time we are clearly dealing with a world-and-life view expressing a metaphysic of life which can and should be elicited. But the myths do not allow anything whatsoever to be inferred concerning the remote past with which they purport to deal. What we encounter is a contemporary form of thought about the recent past and present.

(c) The Connection of Suffering and Good

There is nothing like the idea of a Golden Age in the mythology. Some of the evidence that the time of heroes and marvels is not idealised was given in (a) and (b). The absence of such a point of view should mean that on the moral side of mythology there are no entelechies, and this is the case with the Murinbata myths. The ideas of perfection are simply not formed. Indeed, the narration of the myths is sometimes accompanied by laughter among the listeners, and I have sometimes thought it had a sardonic tone.

It might be said that the more important myths are unanalytical essays on acts of will, the motives of the will remaining unstated and rather mysterious. We thus learn of the will of Mutjiŋga to put an end to burgeoning life, and of her act and its consequences, but nothing whatsoever of her motive. The same is true of Tjinimin’s will to kill his father Kunmaŋgur, and of the latter’s will at death to remove all the fire from the world. Then too there is the will of Waak, Old Crow, to die when confronted by the demonstration of Bali, Old Crab, that life could continue by casting off its restricting shell. The myths remain tantalisingly silent about the why of the acts which had such vast consequences. The Old Woman was killed: and men have to maintain her emblem, sign and symbol as the means of immolating youth. The Father was murdered: and his death at the hands of his son gave men the means of the perennial life which The Mother of All must touch. Waak put men under necessity of death: but the vital will of men to persist 126is evident in all they do; the greatest single cause of human conflict is the attribution of death to warlockry. The facts create a first impression of dualism. To support a thesis of dualism would be easy. There is no suggestion of a first cause, or a spiritual personage who is all-good, or one who is so all-powerful that all are subject to him in all things. There is no independent entity, and none of a wholly unitary nature. Mutjiŋga killed, but is mourned. Kunmaŋgur was killed but in death gave benison, only to try then to deprive the world of fire so that men would live like animals, eating raw flesh. One could say that dualism in this sense is the norm of the mythology, but it is only an apparent dualism. It is rather a kind of counterpoise, a unity of opposites.

The myths are a sort of statement about whole reality, a declaration about the penalties of private will, and by implication a thesis on the spoiling of possible unity. They also come very close to the spirit of certain insights within other cultures, e.g. The Buddha’s observation that suffering is a product of the striving for being. But we simply do not have the evidence from which to infer clearly what Mutjiŋga strove for. Why should she have wished of her own volition—maŋe nigunu—to end life? This, to the Aborigines, is a dismaying mystery. The case of Kunmaŋgur is not easier. Tjinimin, The Bat, emerges as a figure again acting maŋe nukunu, in self-will for gratuitous motive: He seduced his sisters, but there is no suggestion in the myth that the father knew, or was angry, or reproachful. The Murinbata, thinking it astonishing that anyone, of volition, should have preferred death to life, hold the crow in opprobrium, and hold in contempt peoples who have been known to eat it. We can perhaps say that the evidence shows a dispensation of duality that the formation of familiar being and its constituent entities made for the cohesion of good to bad. But the cohesion is not truly or clearly moralised. The idea of an entelechy has simply not differentiated.

Taken as a whole the myths deal with cosmology rather than cosmogony. That is, they deal less with origins as such than with the instituting of relevances—the beginnings of a moral system—in a life which already was. The tacit assumption invariably is that something existed before the marvels. No imagination is exercised about that 127aspect. The myths rationalise and justify familiar entities, forms and relations. In that sense one may say that they deal with being rather than with existence, or with existence become intelligible by having taken on familiar forms. The enduring nature of the forms is a subject of strong Aboriginal sentiment. The authority of the marvels by which the forms came about or were associated with weighs very heavily. Experience with the traditional Murinbata confirms in every way the impression made on other Australian anthropologists: the Aborigines seem to feel bound by some kind of necessity to what was instituted anciently.

5. Relations between Myth and Rite

The myth of The Old Woman is not told as a feature of the rite. It is not, like the rite, a secret. Women know at least something of it. Some men know it better than others. But for all the initiated men it has an esoteric significance which is a male secret. My earlier statement about the myth of the rite thus requires qualification. Empirically, we can speak only of an attachment. The anthropological interest lies in that fact, and how it is to be characterised under a theory.

The type of analysis required is one that will link the two within the ontological system as I have tried to sketch it. All the difficulties referred to earlier—the limits of the indicative-analytical language, empirical methods, and abstractive selection—are here at their worst.

An analysis through linguistic categories is not only outside my scope but, in my opinion, is also unsuitable. To be sure, the myth is a special kind of language but I reject as inappropriate an analysis as a special kind of language. My concern is with the use of language or speech-forms as myth in a situation of rite. The first aim of study from such a viewpoint is to find similarities between rite as a structure of operations and myth as a structure of comparable or analagous elements. Unless such elements can be brought into an approximation, however rough at first, then there is nothing sufficiently isolate and concrete for concepts to refer to, and in consequence there cannot be any theory of myth attached to rite.128

129It is useful at this stage to bring together in a table of condensation the main facts of the rite and myth, and to put alongside them those of Djaban and circumcision. Descriptions of the last two are probably unnecessary, since Djaban closely resembles Punj in all essentials, and the circumcision is of classical type.

Table 1 is an arrangement of the main events or part-events of the four entities set out in columnar sequences which are intended to reproduce the actual processes acted out or told. That is, the order in which the events are set out, from top to bottom of the columns, is the actual order (with certain omissions of detail) in which they occur in the originals.

The tabular presentation makes drastic selection and concentration unavoidable. Each entry is as far as possible a datum, something I have seen, heard or been told. The summary accounts of the myth and the rite already given will enable others to check for distortion.

My aim has been to abstract similarities, and I have adapted the language of description to that purpose.

The table shows three things plainly: an affinity of constituents between columns; a similarity of sequences from the onset to the completion; and a good measure of total congruence. It also shows that the distinction made by the Murinbata between I and II is not true of the structural forms. And it reveals the essential problem: a cryptic similarity between (III) and (IV) across a frontier set up by facts belonging to two distinct order. The similarity is not an effect of the schematic ordering. It is ‘in’ the facts themselves.

The contents of Table 1 are somewhat more intelligible if considered against certain other relevant facts (Table 2). The first table is an empirical, the second a rational, ordering of facts.

Each of the first three columns of Table 1 denotes an entity—the ‘ceremony’—which is also a system of functions and a process. The descriptions are in no case complete. Column (IV) is an imaginative system within a wider system of conduct composed—in the eyes of the Murinbata—of (II), (III) and (IV). The wider system defers to data in part set out by Table 2. It includes invisible as well as visible reality. The 130functions, if fully analysed, would be a description of the state of the system. The process is the entity seen in time and space.

As I have said, the Murinbata classify (I) separately from (II) and (III) by saying that it is not intended to make young men ‘understand’ and is not as ‘big’ or as ‘heavy,’ thus providing a simple linear scale of their own. They insist that one must know (IV) in order to understand (III), and as best they can expound what they mean by The Dreaming as the condition of understanding (I)—(IV).

It is likely to be fatal to an understanding ‘from within’ to divide the entity, as by dissociating the use of the language from the acts, or the ideas from the acts, or by setting up beliefs in a causal connection with the acts. Likewise, to impose on the functions a set of organicist or mechanistic metaphors, or to suppose that the process is illumined by treating it as a variable of kin-relations.

Vertically, Table 1 exhibits the activity-aspect of entities and systems of functions as processes over time, each process being distinct and separable from onset to termination. Horizontally, it exhibits what seem to be equivalences. Both axes are fully meaningful only through things which are either not shown, e.g. the symbolic culture, or shown in part only in Table 2, e.g. the data of which each vertical axis of Table 1 is a system. The tabular arrangement has obvious limitations, for at every item one has to consult, as it were silently and by imagination, things which cannot be depicted in this way.131

133There is certainly sufficient natural metaphor on both axes to impel caution in the use of interpretative metaphor suggested by the organicist and mechanical-structural hypotheses. And while the real-life connections between the columns are made by actual persons at the business of life, to select the classes of kin-relations as having some kind of primary explanatory power is arbitrary and unwarranted.

6. Analogy; Allegory and Explanation

The art and method of explanation among the Aborigines are to find apt likenesses between the familiar and the unfamiliar. To hear or see this done is one of the most commonplace experiences of fieldwork among them. They take delight in repeating the mistakes their fathers made: how sugar was called ‘sand,’ flour ‘paint,’ tobacco ‘fæces’ and so on. In more complex matters they use the word ŋinipun which in this context signifies shape, form, aspect, contour or outwardness: ‘general appearance’ would be a good translation. I conceive that this process of mind occurs in the making of myth. Things of the social order provide them with ŋinipun or shapes or images aptly resembling those which their intuitive minds discern in rite. It is an extension of the process by which one subject, an unknown, is likened to another, a known, the likening constituting a type of explanation. If the main functions of myth are not cognitive and communicative then I doubt very much if we can understand them at all.

In a striking passage23 Robertson Smith observed: ‘As a rule the myth is no explanation of the origin of a ritual to anyone who does not believe it to be a narration of real occurrences, and the boldest mythologist will not believe that. But, if it be not true, the myth itself requires to be explained, and every principle of philosophy and commonsense demands that the explanation be sought, not in arbitrary allegorical theories, but in the actual facts of religious custom to which the myth attaches.’

134In other words, if there is no theory of rite then there can be no theory of myth. It is a position with which, on the whole, I agree. I do not consider that we can speak seriously of ‘a theory of rite’ as yet existing in anthropology, and for this reason interpretations of myth remain conditional. But, within this limitation—and it is one which I regard as almost crippling—certain things seem reasonably clear. If the rite of Punj has an ontological significance then so has the myth. If the rite expresses something about the macro-experience of living then so does the myth. If the rite has to do with a mystery—and the Aborigines say so in words which are quite distinct—then our concern lies not with the literal but with the figurative meaning of the myth. The literal language is simply a case of the non-mysterious being used to give shapes or ŋinipun to the mysterious. For my own part I see neither true interest nor significance in dissecting the figurative aspect of the myth for ‘reflections’ or ‘expressions’ of the social structure. That there are such I do not doubt. What I would think remarkable would be to find myths which did not contain such reflections or expressions. For what other image or idea-language could be used? One must keep sharply in mind that one is dealing with the symbolical constructions of a people amongst whom there is no class of scholars or detached intellectuals, who know nothing of writing, who rarely if ever ask the how or why-questions from philosophical motive, and who face a vast undifferentiation of entities and relations which, in such circumstances, can only be intuited. For the same reasons, understanding can founder before it begins if the method of inquiry is controlled by a rational logic which suppresses considerations of an ontological reality that—as is clearly the case among the Aborigines—has begun to excite feeling. Both the myth and the rite of Punj are evidences of a something differentiating, a something the Murinbata say they do not ‘hold in the ear,’ that is, a mystery.

What one is thus studying is a moment in the development of cult. The historical aspects, for reasons I have already stated, are too nebulous to deal with here, since their significance requires careful, extended reasoning. I will simply interpolate the observation that 135a large inter-regional study is now required into the wide diffusion of the bullroarer cult, and into the fall or stasis of the significance of circumcision, or its failure to take on the factitious significance of cult in the northern region.

Regarded from the above point of view, the myth of Punj appears as an elementary attempt to make an identity between a social reality and a new intuition of a suprasocial reality. The known and non-mysterious—the social order—provides both a literal and a figurative language of shapes by which to interpret the unknown and mysterious. Literally, the myth is a story about people. Figuratively, it is an allegory made up of extended metaphor formed from analogies of resemblance. It is not a fully expressed or explicitly formulated analogy, although it approaches that stage. For the most part it is a mystical and figurative statement; not mystical in the sense of a dreamy confusion of thought, but in the sense of having to do with mystery. Metaphor resting on analogy of resemblance is its necessary means.

On the assumptions that the myth of The Old Woman is a story about one thing under the guise of another, and that the story is thus an allegory, there may be some usefulness in an attempt to extract the allegory in a form which does not simply repeat the story. The procedure can only be arbitrary and rationalistic, and the product is unverifiable, but provided it is done with an informed knowledge of Aboriginal symbolism the result may be not without value.

An allegorical interpretation might run approximately as follows:

(1) Innocence or new life (childhood) in mortal peril (death) from private motive (the act maŋe nigunu). (2) Mysterious power (femaleness) using responsibility (seniority) to spoil necessary trust (the people’s unavoidable request). (3) The flow of life (the stream) being used as concealment (submersion) after wrongdoing. (4) Life becoming tortuous and secretive (the winding, deepening water) after being simple (shallowness). (5) Wrongdoing, at first untraceable (absence of signs in the clear water) finally becoming identifiable (the murk in the water). (6) Opposed but complementary elements of life (left hand, right hand) by agreement (the plan of search and ambush) and diverse means (the 136different routes) which are nevertheless complementary (the successive meetings) attaining their object (the entrapment). (7) The assertion of joint male authority (two men send by all the people) to do justice (the rhetorical attribution of guilt) and retribution (the justified slaying). (8) The persistence of life (the moving stomach of The Old Woman), but its powerlessness to save itself (the disembowelling). (9) The avoidance of contamination (by excrement) as a condition of renewed life. (10) The cleansing of a new life (washing and drying) and bright adornment (by painting) before restoration (escort to camp) in a different status (initiation). (11) The restoration (welcome) to the loving care (weeping) of those with mysterious power over life (the mothers).

The interpretation will seem an exploitation of symbolical obscurities only to those who do not have a developed sense of the Aboriginal power of imagination. A fair degree of plausibility can be given to the general-specific meanings. I do not propose to do so here, or to pursue the approach further since it lies across the main object, but I might point out that the construction is not very much more tenuous than are many statements about the social structure of peoples who—as one distinguished anthropologist has said—may be only dimly aware, or not even aware at all, that they live under such structures. The interest of the construction is that, in conjunction with Table 1, it reveals more vividly, by a change of language, that both the myth and the rite exemplify processes which we can characterise as having a dialectical form. I regard this fact as one of high importance.

The processes are not merely changes or movements over time but developments from one state or situation or condition to another, such that new and old belong to the same order but are qualitatively distinct, the old not quite annulled and the new not quite unfamiliar. The developments are also attained through a train or sequence of opposed acts or operations. We thus seem required to characterise the sequence as ‘dialectical.’ It is certainly temporal; it is certainly also connected or systematic; and it certainly has the appearance—though an obscure appearance, coming through very diverse and complex elements—of motion and direction through opposites which become resolved. 137Exactly what a dialectical process amounts to in a society which anthropology tends to regard as more or less stationary is a matter for investigation. The analysis of processes has not been distinguished by clarity or precision, unlike the studies of the morphology of kinship. But developmental processes with such characteristics occur so commonly among the Aborigines—for example, the supposedly static structure of the sub-section system is a type of dialectical form—that once the form is extracted the natural corrective of a static structuralism of approach is provided. However, I do not propose to divert to such matters here.

In Table 3 I set out the materials of Columns III and IV of Table 1 in an effort to clarify some of the details of the form. The sequences from A to F are listed vertically, like the vertical arrangement in Table 1. The table is divided by a central line. To the left of it are placed acts of a given class defined as positive (+) for a quality, and to the right of it acts which are the negative (–) of the class. The general conception is that of signed conduct referred to earlier. On the far right of the table I have included major entries made in Fig.1 in the first article. The table aims at setting up broad correspondences only. An effort to force the comparison at such points would detract from whatever value the arrangement has. As stated earlier, it must be remembered that we are dealing with a moment in the development of a cult.

Any errors made in the placements in Table 3 are of course transferred to Fig.3. The classes of acts here regarded as (+) and (–) require analysis in respect of their logical characters, a task I have been unable to attempt. Both table and figure are thus to be regarded as provisional. The congruence of the profiles, and a study of the detail of the movements, show that the visual representation of Punj given in the first article is at best very notional. The differences of the profiles probably, but not necessarily, indicate incomplete observation or errors of fact or classification. I wish to make clear that I do not regard the facts as perfectly established even though I have taken pains over them.138

7. Myth, Rite and Reality

One of the things that distinguish the myth of The Old Woman from every other is that it is told as a sorrowful story. This and a number of other features put one in mind of the madre dolorosa and the Greek cults of mystery. But it is nearer to my purpose here to ask: can reasons be found for the growth of the whole cultus? And, if the process of the rite has been described correctly as dialectical, has it any necessary relation to the cult?

(a) The bullroarer existed in the region before the rise of the present cult. It was smaller, with a narrower range of types, of different colour (black rather than red-brown), and was decorated externally (by white feathers or down) in a style wholly distinct from the incised patterns which came later. (b) There is a dim memory of a great rite—the Tjimburki—extending over many months, which celebrated the coming of the dry season, and in all probability was connected with the black bullroarer. The rite was a male secret, and centred around deep, circular excavations in the earth. Traces of the excavations may still be seen. (c) Yet another rite, of a distinct kind, was connected with the use of great ovoid or circular stone structures still to be seen in the countryside. Nothing is known of the rite except that it was connected with dancing.24 Nowadays the stone rings are usually connected—probably in a fictitious way—with Kunmaŋgur, The Rainbow Serpent, who may have been the subject of an earlier rite. Some Aborigines still connect him rather than Mutjiŋga with the bullroarer, and I found this to be so among the Nangiomeri twenty-five years ago. (d) There is evidence that songs, dances and decorative designs formerly connected with a senior rite were transferred in the past to the junior rite of Djaban, and a probability that those associated with Djaban were in turn transferred to the Karamala initiation of still younger boys.

142

Facts of this kind strongly suggest an historical succession of cults over a period of perhaps a century. If one adds the indirect evidence of the sacred rock-art the historical perspective is deepened. Some rock-paintings are connected with Kunmaŋgur through a myth which still survives strongly, but without a connected rite. Other paintings, which have a strong affinity with the Wandjina figures of the Kimberleys, cannot be connected with any rite of which a memory survives.

It seems to me probable that the cultus of the bullroarer, now linked with The Old Woman, is but the last phase of a complex religious development expressing itself through successive transformations of a rite of initiation of constant form. The fact that four ceremonies with an initiatory purpose can be identified gives a peg on which to hang a reconstruction.

On the other hand, the type of process depicted in Fig.2 is thoroughly characteristic of at least three of the initiations, on which I have good information, and probably also of the fourth, the Karamala. There is some slight evidence too connecting the Karamala with the mime of the blowfly. The facts suggest a crescive and involuted development of theme and form over a fair period of time. The Murinbata live close to the northern boundary of the region where the practice of circumcision ceases, and just beyond the northern boundary of the region where sub-incision ceases. My hypothesis is that the diffusion of circumcision to the region supervened on a system of initiation of which Karamala and Tjimburki were characteristic, replacing them as the highest cult 143of the time, only to be superseded in turn by the cult of Mutjiŋga. The tradition is that the present cult came from the south through the Victoria River tribes. I propose to examine the evidence for the view in a separate study. A comparison with the cult of Kunapipi studied by Dr. Berndt thus becomes necessary.

What emerges immediately is a correction of older views of Aboriginal life as static and unchanging. In terms of the evidence offered here an analytic problem can be rephrased. Instead of the rather bare conception of ‘rite,’ or of ‘rite of initiation,’ we may use the conception of a ritual dialectic of initiation. A necessary step is the study of the logical properties of processes of this type within an ontology of life which predicates many relational properties about life in a given world. The symbolical expressions of the metaphysical conceptions of the Aborigines are thus not an optional study for anthropology but a necessity. I regard the questions as absolutely central to theoretical anthropology, but their generality puts them beyond the scope of these articles. I shall attempt a sketch, however, at a later stage.

Each of the rites de passage initiated a male into a world transcending that in which he had lived hitherto according to the usages proper to his stage and locus of life. But not only were mind and personality widened and deepened by the successive rites. At each stage the subject discovered—or had unfolded to him—a reality of life transcending the former limited reality. On this view the rite of Punj widened experience to a terminal. It was, so to speak, the closest approach possible to the whole of life-reality. Hence perhaps the awe and mystification suffusing the rite. The viewpoint at least suggests a logical ground for the cognitive and affective states. One may thus construct the initiatory rites, on the metaphysical plane, as acts of thought and feeling towards the whole of reality, and the myths as expressive statements about it. But the history of rites in the region suggest that the conception of reality has been expanding, and with this no doubt there was too a deepening mystification. We may note at this point that there is also positive evidence of an increasing complexity of social organisation at least in part over the same period, e.g. the development of a relatively simple 144system of patrilineal moieties into the more complex system of subsections which, at least in theory, descend by indirect matrilineality. Many tensions were built up by that process. It is tempting to see in the conjugate principles of matrilineality and patrilineality a possible source of the displacement of Kunmaŋgur, the totemic ‘father’s father’ of one moiety and the ‘mother’s father’ of the other, by Mutjinga, who is called kale neki, ‘the mother of us all.’

At all events, the fullness of the rites de passage can be apprehended only if they are studied with the Murinbata ontology, and if related to the metaphysical conceptions of a reality being widened and deepened by history. It is in that context also that the ritual dialectic must be studied. The subject requires as its prior condition a careful study of the Murinbata symbolism, and this will form the subject of another article. But I should like to say with care that, since there are so many possibilities of misunderstanding, the argumentative or philosophical senses of ‘dialectical’ have nothing to do with my usage. Each step in the processes can be checked by observation; the oppositeness of the things related at each step is either implied or stated by the Aborigines themselves; and in the description of such processes from the ‘outside’—that is, from the analyst’s viewpoint—no better word than ‘dialectical’ seems available for processes so constituted.

8. The Mood of Assent

The genius of Murinbata religion may be said to lie in three things. It affirms reality as a necessary connection between life and suffering. It sees the relation as continuously incarnate and yet as needing reaffirmation. It celebrates the relation by a rite containing all the beauty of song, mime, dance and art of which men are capable.

That the rite of Punj is—whatever else it may be—a high, joyful celebration no one could doubt who has seen it; or that it is at one and the same time a fearful approach to mystery; or that it typifies what in native eyes is an ineluctable condition of men.145

The re-enactment of the primordial tragedy brings understanding—or so it is held—to the youths through whom life is to continue. The covenant of duality is thus endorsed by and on a new generation. Each young man is taken out of his empirical and social self, as though to meet his essential self, is touched by something transcendental—The Mother’s blood, which is a symbol both of life and of suffering—and is then returned bearing her sign—the bullroarer—to help perpetuate the relation, within a logos which gives life and suffering a common source and a joint imperium.

There does not seem much reason to doubt that Punj could have developed in the appropriate conditions into the species of sacrament known as sacrifice. On the analysis, sacrifice is a logical possibility of the religious culture. Theoretically, it could have differentiated itself. For such a consequence, however, a growth of moral imagination would have been necessary.

One cannot judge an unknown history by its outcome, but there is nothing in Murinbata tradition which suggests an insight that men might be either free or perfectible. They are a people to whom the invisible has been the test of the visible; the unknown has been mysterious and terrifying; the dead have been feared; and men have been bound to the past. One can guess only at the historical conditions, for them, which might in time have suggested that, as St. Paul would have it, the visible is the test of the invisible; or that the unknown is only the unknown, the dead are simply piteous, and living men are able to shape their future. It seems that such ideas have simply not occurred to the Murinbata. Their moral imagination has been stultified.

The logical possibility of sacrifice, inherent in what they do in rite, evidently has had neither occasion nor motive to emerge in events. The emergence has been denied, one would think, by the force of the tradition that first things are also last things.

The ordainment of a once-for-all life puts its terms beyond human initiative. An abandonment of self, as in sacrifice, could not be requited by anything, for the compassion of invisible powers was 146given once-for-all. Voluntary suffering could not increase merit because the necessity of suffering was part of the founding covenant. The act of sacrifice, to be intelligible at all, needs a ground of moral freedom, one which is understood as such. The motive can scarcely appear until men have seen with clarity that they have no remedy of their own for inherent weakness, and the insight is ineffective unless there is a strong moral conscience.

The Murinbata do not give any significant impression of having, or thinking they have, moral freedom. The sense of the corruption of things lacks sharpness. As far as one can tell from outward show the formations of conscience are not strong.

If rite and symbolism were the only data one might be tempted to see a society ‘trembling on the edge of tragedy.’ The trouble is that nearly as good a case could be made that it trembles on the edge of laughter. It depends whether one deals primarily with the religious or the mundane life. A sense of tragedy is coconscious with the religion, and we can link it—though vaguely—with certain facts of the mundane life, but the latter is far too concerned with the pursuit of food, valuables, leisure, enjoyment, safety, lovers, spouses and other goods of life to have any equivalent concern.

In the mundane life there is an emphasis difficult to state, but in some ways the most revealing of all Aboriginal attitudes. The difficulties are perhaps less in the facts than in our ability to handle them. The character of the emphasis is possibly best stated in the first instance by contrast.

One may say that people have set up a kind of quarrel with life, or with its terms, when they rail against the gods, spirits or fate; or when they threaten or make offertories to the invisible powers; or plead or pray; or shut themselves off from life in hope of inner consolation; or look for final justice to someone or something standing over human history. Such attitudes and types of conduct are alien to Aboriginal mentality. The natural range of human temperament is of course represented, but in the Murinbata ethos one does not find any evidence of a pessimistic, cynical, apathetic or even quietistic tenor. A high 147intensity of bitterness and violence accompany injury and loss. But the expression is not directed at the self (as in suicide, which to the best of my knowledge never occurs), or in a diffuse way at society, or directionlessly at life or fate. It is directed at people and, to a large extent, along institutionalised lines. A juridical system provides a patterned means of redress for the delicts of actual life. It provides categories of wrongdoing, formal means of accusation, ways of meeting evidence with evidence, a means of bringing wrongdoers to penalty, a code of equivalent injury, and methods of limiting the spread of trouble. It is an elementary and defective system which is deeply undermined by an egoistic insistence on the right and duty of personal retaliation, and by mystical divinations which are often used for other, cloaked enmities. I am unable to embark on these topics here, but will limit myself to the statement that an attitude of ‘assent’ shows out in the astonishing extent to which the Aborigines appear to forgive and forget after the juridical system has worked. Continuous reproach or obloquy are quite uncharacteristic.

If there are an intuition and a symbolism of tragedy, then they do not issue in a quarrel with the proximate terms of life. The motive seems to be the desire to observe the continuity with The Dreaming. This entails the endurance of the joint imperium of the good and the bad. One detects no nostalgia for the past or yearning after a perfected futurity. In the tradition no one offered himself to any gods—indeed, there were none—and did not want to be accepted by any. Evidently what was done was done, and much of it was good. The insight that the bad is the condition of the good is not contemptible. In this religion it is expressed with a certain nobility that transcends the strange symbolisms.

The relation between the religious and the mundane life is continuous. It is a functional relation of complex interdependence. A striking phenomenon is the way in which joyous sociality—one might almost say the perfect sociality of laughter and fun, in apparent mockery of almost every stable institution—opens and closes the highest rite. To say that Tjirmumuk alone makes possible the parallel celebration of tragedy would be too much. But it is psychologically appropriate 148that ‘good fun’ (for that is the way the Murinbata describe Tjirmumuk) should make a jointure with ultimate concern. It allows them to assent to life, as it is, without morbidity.