Routledge is an imprint of the Taylor & Francis Group
To purchase your own copy of this or any of Taylor & Francis or Routledge's collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk
Typeset in Joanna by RefineCatch Limited, Bungay, Suffolk
Printed and bound in Great Britain by
TJ International Ltd, Padstow, Cornwall
All rights reserved. No part of this book may be reprinted
or reproduced or utilised in any form or by any electronic,
mechanical, or other means, now known or hereafter
invented, including photocopying and recording, or in
any information storage or retrieval system, without
permission in writing from the publishers.
British Library Cataloguing in Publication Data
A catalogue record for this book is available from the British Library
Library of Congress Cataloging in Publication Data
A catalog record for this book has been applied for
ISBN 0-415-25550-3 (hbk)
ISBN 0-415-25396-9 (pbk)
Foreword by Dr D. F. Pocock 1
1 Sources and Historical Background 14
2 A Definition of Magic 22
3 The Elements of Magic 31
1 The Magician 31
2 The Actions 55
3 Representations 75
4 General Observations 106
4 An Analysis and Explanation of Magic 112
1 Belief 113
2 An Analysis of Ideological Explanations Concerning the Effectiveness of Ritual 120
3 Mana 133
4 Collective States and Collective Forces 150
5 Conclusion 174
by David Pocock School of African and Asian Studies, University of Sussex
In introducing Dr Brain's excellent translation I am acutely conscious that Professor Lévi-Strauss has not only written about Théorie de la magie at some length but also made it the occasion for a major, if early, statement of position. 1 Because this important document has not yet been translated I believe that I shall do the reader of the present work more service by drawing his attention to it rather than attempting something in the nature of an original essay. I shall write of Mauss's Theory of Magic in the perspective of Lévi-Strauss's own achievement which, in my opinion, gives it retrospectively its significance for the modern reader.
It may seem paradoxical to say that the importance of the present work is that it contributes to the dissolution of 'magic' as a category, nevertheless this is the claim made for it. The need for such an act of dissolution is to be found in the earlier history of ethnology.
Criticisms of social Darwinism in the nineteenth century are easily made and, in the process, many of the still relevant achievements of the period are neglected. One of the most important of these, at a time when notions of 'enlightenment' and 'progress' threatened to divide humanity along what today we would call 'racialist' lines according to innate capacity, was that the application to primitive societies of the theory of evolution re-established the fundamental human unity. If it now seems to us absurd that certain societies should have been thought of as representing stages in the evolution of the human species, we should remember that this was the price paid for the renewed belief in the unity of that species and the potential for change in the societies concerned.
Few theories are pure in their application and it should not surprise us if writers of that time occasionally sinned against the conception of unity by imputing, particularly in the area which they thought of as the supernatural, modes of thought which, if true, would as effectively have cut the primitive off from communication with the modern as a genetic difference would preclude interbreeding.
The vice is not dead: some modern accounts of non-European societies, again especially in the area of 'religion', 'magic' and the like, seem to rest on assumptions about human nature which would not stand the test of application to ourselves. Indeed sometimes it seems that the primitive is to be defined as that about which any nonsense can be believed. Modern writers have less excuse than their forebears: they, for the most part, shared beliefs in, or derived from, a revealed religion which, but for the labours of their missionary brethren, was indeed closed to the majority of mankind.
'Magic' was perhaps even more prone to this treatment than 'religion'. The ethnologists of the nineteenth century knew what they, at least, meant by the latter; magic on the other hand was a peculiar and alien phenomenon and its persistence in sections of European society only heightened the scholar's sense of estrangement. Much of the theoretical discussion which preceded Mauss's work had for its effect not so much that of overcoming the apparent division between those who believed and those who did not (i.e. the ethnologists themselves) as of reinforcing it. Thus it comes about, for example, that we learn more about Sir James Frazer's beliefs about 'magic' than we understand about the examples which he cites.
This intrusion of the subjective is not bad in that it is inevitable; it is common to all the social sciences. However, consciousness of it imposes upon them all the perpetual task of re-examining in relation to the facts the most tried and accepted categories of their apparatus. If categorical distinctions of the Western mind are found upon examination to impose distinctions upon (and so falsify) the intellectual universes of other cultures then they must be discarded, or, as I have put it, dissolved. I believe 'magic' to be one such category and need only cite here by way of evidence the fact that it is perpetually opposed to 'religion' and 'science' in our literature.
Marcel Mauss certainly had no such work of demolition in mind, although I seem to see in his two concluding paragraphs some hint of an awareness that his researches had led him further than his original intent. Certainly the modern reader can derive from Mauss's wide-ranging survey of the facts and his many profound insights the materials for a further advance.
In his Introduction Professor Lévi-Strauss reminds us that if we are to do justice to Mauss we must remember the date at which the Theory of Magic was published. 2
It was at a time when comparative ethnology had not yet been abandoned, largely at the instigation of Mauss himself, and as he was to write in the Essay on the Gift: 'That constant comparison in which everything is mixed and where institutions lose all local colour and documents their savour.' Only later did he devote himself to drawing attention to societies 'which truly represent the maxima, the excesses, which better allow the facts to be seen than those in which, although no less essential, they remain small and undeveloped.'
This is to compare Mauss with himself and certainly no one would seek to displace the Essay on the Gift as his masterpiece. Nevertheless I, having given the Theory of Magic a role in retrospect, wonder whether its contribution to the work of dissolution does not lie in the fact that it does cover so wide a range of material. Professor Lévi-Strauss himself appears to be partly of this mind when he defends Mauss, and Durkheim also, from the common criticism alleging that they 'were wrong…to bring together notions borrowed from widely separated regions of the world and to constitute them as a category.' The same author continues: 'Despite all the local differences it seems certain that mana, wakan, orenda represent explanations of the same type; it is therefore legitimate to constitute the type, to attempt to classify and to analyse it.'
It is upon this assimilation of geographically distinct notions that Professor Lévi-Strauss is able to advance the proposition: 3 conceptions of the mana type are so frequent and so widespread that we should ask ourselves if we are not confronted with a permanent and universal form of thought which… being a function of a certain situation of the mind in the face of things, must appear each time that this situation is given.
Lévi-Strauss then cites both the example of the Nambikwara who, on being introduced to cattle, designated them by a term very close in its connotation to manitou and the example of French words used for essentially mysterious objects. From these he passes to the observation that in our own society such terms are fluid and spontaneous whereas elsewhere they constitute the base for considered and official systems of explanation, a role which we reserve for science. 4
Lévi-Strauss's argument leads him finally to see the mana type notion as pure symbol or as having zero symbolic value: a formulation analogous to the linguistic zero phoneme. This analysis, which carries us beyond the category 'magic', is explicitly related to the contribution of Mauss who 'was one of the very first to denounce the insufficiency of psychology and traditional logic and to disrupt their rigid frames by revealing other forms of thought, apparently "alien to our adult European understanding."' 5 Lévi-Strauss's first and simplest formulation runs as follows: 6
Always and everywhere notions of this (mana) type intervene, somewhat as algebraic symbols, to represent a value of indeterminate meaning (signification), which being itself empty of meaning (sens) is therefore susceptible to the reception of any meaning (sens) whatsoever. Its unique function is to make good a discrepancy between signifier and signified, or, more exactly, to draw attention to the fact that in certain circumstances, on a certain occasion or in certain of their manifestations, a relation of inadequacy exists between signified and signifier to the detriment of the anterior relation of complementarity.
This is an important step: it does not dissolve the concept 'magic' so much as, so to speak, cut the ground from beneath it. A field of explanation is opened in which what we call 'magic'-pre-eminently an activity-is only one of, and of the same order as, many symbolic actions which overcome the discrepancies of thought. Rituals do what words cannot say: in act black and white can be mixed; the young man is made an adult; spirit and man can be combined or separated at will. Indeed actions speak louder than words.
Let me give an example of a very simple magical act which I have observed in Gujarat, in western India. A Hindu by accident brushes against, touches an Untouchable. To free himself from the consequent pollution he then touches a Muslim. The structure of the situation is as follows. The Muslim is not a Hindu, he is outside the caste system and the world of purity and impurity as are all non-Hindus. Nevertheless, he is given a place in the Hindu caste hierarchy in the village. The Untouchable, on the other hand, is of his nature an essential element in the system; he personifies the negative pole of impurity, an impurity, let us note, which at the level of thought, embraces Muslims, Christians, etc. The Untouchable's function in the system is to be excluded from its activities; to be treated as an outsider. There is then, so to speak, an absurdity in the system: the outsider is thought of as an insider, and the insider is treated as an outsider.
It is against this absurdity that we now observe the Hindu who has by accident touched the Untouchable. It is like a game of tag-he has equated himself with the Untouchable. He frees himself from this association by deliberately touching the Muslim whose contradictory nature-insider/outsider, pure/ impure-provides the conduit for a restoration of normality: the Hindu and Untouchable are again separate. It is important that I stress that it is only on this occasion and for this purpose that the contradiction of the Muslim's position is used. In certain circumstances he may be recognized as a Muslim, in the majority of circumstances he is assimilated in the Hindu caste system, one or the other. Only in the circumstances that I have described is his double and contradictory value recognized.
One major criticism of Mauss's work is to be found in Lévi-Strauss's discussion. It is worth reporting at length not only because of its relevance to the present text, but also because it touches upon a continuing tendency in modern anthropology.
The anthropologist inevitably works with the categories of his own culture and consciously refines them through the experience of others. He may, and sometimes does, imagine that his categories are perfectly matched in the cultures which he observes; thus they are believed to practise 'magic' as he supposes it to be. From this it is an easy and dangerous step to imagine that the entire phenomenon is now accessible to his empathic understanding. Ironically an unreflective empiricism is thus transformed into simple-minded subjectivism. Because of this tendency I take the opportunity of translating Lévi-Strauss's critique at length. 7
We refuse to accompany Mauss when he looks for the origin of the notion of mana in an order of realities other than the relations which it helps to construct: the order of sentiments, voli-tions and beliefs which are, from the point of view of a sociological explanation either epiphenomena or mysteries, in any case extrinsic to the field of investigation. This pursuit is, to our mind, the reason why an enquiry in itself so rich and penetrating, so full of illuminations, falls short and ends deceptively. In the final account mana is but 'the expression of social sentiments which have been formed, sometimes fatefully and universally, sometimes fortuitously in relation to certain things, chosen for the greater part in an arbitrary manner….' But notions of sentiment, fatality, fortuity and arbitrary are not scientific notions. They do not throw light upon the phenomena which they claim to explain, they participate in them. We can see that in one case at least, the notion of mana does present those characteristics of mysterious power and secret force which Durkheim and Mauss attribute to it: it plays just such a role in their own system. There truly, mana is mana. At the same time one wants to know whether their theory of mana is anything other than an imputation to native thought of properties which were implied by the very particular role that the idea of mana was called upon to play in their own.
It would be peculiarly inauspicious to close on a negative note a foreword to a work of this nature. That the leading ethnologist of our time should lean with such weight upon a fifty-year-old argument is as good a testimony as one could wish to its vitality. One can on occasion become irritated with Mauss as with a contemporary and it is Lévi-Strauss himself who has insisted in the same Introduction upon the astonishing modernity of the mind of Marcel Mauss.
1 'Introduction à l'œuvre Marcel Mauss' in Marcel Mauss, Sociologie et anthropologie, Presses Universitaires de France, Paris, 1950.
2 Op. cit., p. xli.
3 Op. cit., pp. xlii-xliii.
4 Op. cit., pp. xliii-xliv.
5 Op. cit., p. li.
6 Op. cit., p. xliv.
7 Op. cit., p. xlv.
Up to now, the history of religions has consisted of a blurred bundle of ideas. We already have a wealth of authentic and instructive facts which will one day furnish a science of religions with abundant material. Unfortunately these facts are haphazardly classified, under vague headings; sometimes their presentation is spoiled by slipshod language. Words such as religion and magic, prayer and incantation, sacrifice and offering, myth and legend, god and spirit are interchanged indiscriminately. The science of religion has no scientific terminology. Nothing but gain would result from establishing one. However, our aim is not only to define words, but to set up natural classes of facts and, once we have established them, to attempt an analysis which will be as explanatory as possible. These definitions and explanations will provide us with scientific notions-that is, clear ideas about things and their inter-relation.
We have proceeded along these lines in our study of sacrifice. We chose it as a subject of study because it seemed to us to be one of the most typical of all religious actions. We decided to explain the mechanism and also the apparent multiplicity of functions which the rite had to serve, once it had been set up; in fact we tried to justify the importance of its position in the whole religious system.
The first problem led to others, including those we are dealing with now. We came to realize, during our study of sacrifice, the real nature of the rite. Its universality, its constancy, the logic of its development-all this gave it, in our eyes, a kind of inevitability, far superior to the authority of legal convention that seemed sufficient to impose observance of it. Because of this, sacrifice and, as an extension, rites in general, appeared to us deeply rooted in social life. On the other hand, we considered that the mechanism of sacrifice would not be explained except through a logical application of the idea of the sacred; we assumed this to be so and made it the starting point of our studies. We held, furthermore, in our conclusions, that sacred things, involved in sacrifice, did not constitute a system of propagated illusions, but were social, consequently real. We found, finally, that sacred things were considered to provide an inexhaustible source of power, capable of producing effects which were infinitely special and infinitely varied. In so far as we could consider sacrifice to be a rite which could be regarded as representative of all the rest, we came to the general conclusion that the basic idea of all ritual-which was to become the major theme of our enquiries-was this idea of the sacred.
However, this initial generalization was found to be wanting; we had unearthed it while studying an institution which was too special and which had not yet been stripped of its differential characteristics. We had treated it solely as a religious rite, not simply as a rite. Was our induction valid only for religious rites, on the religious quality of which it depended? Or could we extend it to all kinds of rites, whether they were religious or not? First of all we had to see if there were rites other than religious ones. This is implicitly admitted in the way people currently talk of magical rites. Magic includes, in fact, a whole group of practices which we seem to compare with those of religion. If we are to find any other rites apart from those which are nominally religious, we shall find them here.
In order to verify and broaden the conclusions of our researches, we decided to make magic the subject of a second study. If we succeed in finding ideas related to a concept of the sacred, as the basis of magic, we shall be justified in extending the conclusions which we proved to be true for sacrifice to all kinds of mystical and traditional techniques. That is because magical rites are precisely those which, at first glance, seem to be imbued with the least amount of sacred power. One can easily imagine the fascination of these studies which were to lead to a theory of ritual in general. Nevertheless our ambitions did not cease here. At the same time we were making our way towards a theory of the idea of the sacred; that was due to the fact that, while we found ideas of the same order functioning in magic, we had gained quite a different image of its meaning, its generality and also its origin.
At the same time this raised a serious difficulty and as a result we were encouraged to embark on this study. As we have said before, the idea of the sacred is a social idea, that is, it is a product of collective activities; moreover, the prohibition or prescription of certain things seemed to be in fact the result of a kind of entente. We were forced to conclude, therefore, that magical practices which derive from this idea or a similar one, are social facts in the same sense that religious rites are social facts. But this is not the normal aspect presented by magical rites. Since they are practised by individuals who are outside the social group, who act in their own interests or in the interest of other individuals, or in their name, these magical rites seem to require much more ingenuity and savoir-faire from their practitioners. In these circumstances, how is it that magic can derive from a collective idea such as our notion of the sacred, and exploit it?
Here we are faced with a dilemma: either magic is a collective idea or the notion of the sacred is an individual one. In order to solve this problem we shall have to discover whether magical rites take place in a social milieu, because if we are able to discover in magic the presence of such a milieu, we shall be able to show then that an idea of a social nature similar to that of the sacred can function in magic; it would then only be a matter of revealing that this idea did function there.
This is the third gain we promised to make from these researches. We shall pass from observing the mechanism of the rite to the study of the milieu of these rites, since it is only in the milieu, where magical rites occur, that we can find the raison d'être of those practices performed by individual magicians.
We shall not, therefore, analyse a series of magical rites, but that ensemble of magic which is the immediate milieu of magical rites. An attempt at such a description may then allow us to resolve the very controversial question of the relation between magic and religion. For the moment we are not banning any consideration of this problem, but we do not wish to dwell on it, since we are anxious to attain our ends. We wish to understand magic itself, before we explain its history. We shall leave aside for the moment-keeping it for a future study-any contribution these researches could make, in the form of new facts, to religious sociology. Moreover, we have been tempted to quit the limited sphere of our usual preoccupations and make a contribution to sociological studies in general, by showing how, as far as magic is concerned, isolated individuals can affect social phenomena.
The subject we have assigned ourselves demands different methods from those we used in our study of sacrifice. It is not possible here, or rather it would not be fruitful, to proceed by the analysis-even a very complete analysis-of a number-even a large number-of magical ceremonies. In fact, magic is not to be compared with sacrifice; it is one of those collective customs which cannot be named, described, analysed without the fear that one may lose the feeling that they have any reality, form or function of their own. Magic is an institution only in the most weak sense; it is a kind of totality of actions and beliefs, poorly defined, poorly organized even as far as those who practise it and believe in it are concerned. As a result we cannot know a priori its limits, and we are in no position to choose, with any certainty, those typical facts which could be said to represent the totality of magical facts. We must first make a kind of inventory of these facts, which will give us an opportunity of limiting-at least to some extent-the field of our researches. In other words, we ought not to try and consider independently a series of isolated rites, but consider all those things which constitute magic as a whole; we must, in sum, begin by defining and describing it. In the analysis which follows, we shall not be guided by the successive moments of a rite. The interest lies not in the plan or composition of the rite, but in the nature of magic's working methods, independently of their application: the interest lies in the beliefs involved in magic, the feelings it provokes and the agents who perform it.
SOURCES AND HISTORICAL BACKGROUND
For a long time, magic has been a matter for speculation. However, the studies of ancient philosophers, alchemists and theologians were purely practical in nature and belong more to the history of magic itself than to the history of those scientific studies which have been devoted to the subject. The first in this list is the work of the brothers Grimm, which inaugurated a long series of studies of which our researches are a continuation.
Today we have good monographs on most of the important magical themes. Data have been collected both from a historical as well as an analytical point of view, and we can now call on a whole range of knowledge. On the other hand a certain number of theoretical ideas have become established, including the notion of 'survival' and that of 'sympathetic magic'.
Our immediate predecessors are the scholars of the anthropological school who have already produced a sufficiently coherent theory of magic. E. B. Tylor, in his Primitive Culture, deals with the subject twice. He first associates magical demonology with primitive animism. In his second volume he mentions-and he is one of the first to do so-'sympathetic magic'; this term covers those magical rites which follow the so-called laws of sympathy. Like produces like; contact results in contagion; the image produces the object itself; a part is seen to be the same as the whole. Tylor's main aim was to show that these rites played a role in the system of survivals. In fact, Tylor offers no other explanation of magic than the one provided by his general theory of animism. George Wilken and Sidney Hartland also studied magic: the former in connexion with animism and shamanism, the latter in relation to life tokens, equating sympathetic magic with those bonds which are said to exist between a man and the object or being with which his life is bound up.
With J. G. Frazer and W. Lehmann we finally have genuine theories of magic. Frazer's ideas, as they are set out in the second edition of The Golden Bough, provide, as we believe, the clearest expression of a whole tradition to which the works of Tylor, Sir Alfred Lyall, F. B. Jevons, A. Lang and H. Oldenberg all belong. Despite divergent opinions in matters of detail, all these writers agree in calling magic a kind of pre-science; and as it is the basis of Frazer's theories we shall begin by discussing this aspect. As far as Frazer is concerned, magical actions are those which are destined to produce special effects through the application of two laws of sympathetic magic-the law of similarity and the law of contiguity. He formulates these in the following way: 'Like produces like; objects which have been in contact, but since ceased to be so, continue to act on each other at a distance after the physical contact has been severed.' One might add, as a corollary: 'The part is to the whole as the image is to the represented object.' Thus the definitions of the anthropological school tend to confuse 'magic' with 'sympathetic magic'. Frazer's ideas are dogmatic in this regard; he expresses no doubts and offers no exceptions to his rules. Sympathy is a sufficient and inevitable feature of magic; all magical rites are sympathetic and all sympathetic ritual is magical. It is true that magicians perform ritual which is akin to religious prayers and sacrifice-and not always in parody or imitation. It is also true that priests in a number of societies have a remarkable predisposition towards magical practices. But these facts, we are told, are but the encroachments of recent times and should be excluded from our general definition, which is only concerned with pure magic.
From this first proposition it is possible to deduce others. In the first place magical rites act upon their object directly without any mediation by a spiritual agent; moreover, their effectiveness is automatic. However, as far as these two properties are concerned, the first is not universal, since it is admitted that magic-in its degenerate phase, when it became contaminated by religion-has borrowed figures of gods and demons from religion. The truth of the second proposition is not affected by this, since in the cases where we have intermediaries, the magical rite acts on them in the same way as it does on external phenomena; magic forces and constrains, while religion conciliates. This last property, which seems to distinguish magic from religion in every case where there is a temptation to confuse the two, remains-according to Frazer-the most permanent general feature of magic.
This theory involves a hypothesis of much wider import. Magic, thus defined, becomes the earliest form of human thought. It must have once existed in its pure state; mankind, originally, thought only in magical terms. The predominance of magical ritual in primitive cults and folklore provide-so it is thought-strong proof in support of this argument. Moreover, it is maintained that these magical states of mind still exist among a few central Australian tribes whose totemic rites are purely magical in character. Magic is, therefore, the foundation of the whole mystical and scientific universe of primitive man. It is the first stage in the evolution of the human mind which he determined-or even conjectured. Religion grew up out of the failures and mistakes of magic. Originally man unhesitatingly expressed his ideas and their associations in concrete form. He thought he would create those things suggested to him by his mind; he imagined he was master of the external world in the same way as he was master of his own movements. But he finally realized that the world was resisting his attempts to do so. Immediately he endowed his universe with mysterious powers, of the kind he once arrogated to himself. Once upon a time man himself was god, now he peopled the world with gods. These gods were no longer bent to his will, but he attached himself to them in worship, through sacrifice and prayer. Frazer, it is true, presents these hypotheses with many careful reservations; nevertheless, he is determined to stick to them. He rounds the theory off by explaining how the human mind, following on from religion, moved off in the direction of science. Once man became capable of noting the errors of religion he returned to a straightforward application of the principle of causality. But from this time onwards it is a matter of experimental causality and not magical causality. Later we shall return in detail to different aspects of this theory.
Lehmann's work is a study in psychology, prefaced by a short history of magic. He begins by pointing out some contemporary facts. Magic, which he defines as 'the practising of superstitions'-that is, 'beliefs which are neither religious nor scientific'-exists in our society in the observable forms of spiritualism and occultism. He attempts, therefore, to analyse the principal experiences of spirits through the processes of experimental psychology and he manages to discover in it (and also, as a corollary, in magic) illusions prejudices and errors of perception caused by these anticipatory phenomena.
All these studies betray one common feature, or error. No attempt has been made to enumerate fully the different categories of magical facts and, as a result, it is doubtful whether, at this stage, it is possible to propose a scientific scheme which could embrace the whole subject. The only attempt so far made-by Frazer and Jevons-to circumscribe magic, has been spoilt by the authors' prejudices. They used so-called 'typical data', assumed the existence of a period in the past when magic existed in its purest state, and then reduced the whole to facts of sympathetic magic. However, they failed to prove the legitimacy of their selection. They ignored a considerable body of practices which are called magical by all those who perform the rites and also those who observe them; as well as incantations and rituals involving demons, properly so called. If one ignores the old definitions and sets up in their place a class of ideas and practices which are so narrowly limited that they exclude magical phenomena which only seem to be magical, we must again ask how it is possible to explain those illusions which have induced so many people to accept facts as magical which, by themselves, are not. We are still waiting in vain for an explanation of this. We may also be told that the phenomena of sympathetic magic form a natural and independent class of facts which it is important to distinguish. This may be so. In this case we should need proof that they have produced expressions, images and social attitudes which are sufficiently distinct for us to be able to accept that they clearly do form a separate class from the rest of magic. We, it should be added, believe that this is not so. In any case, it would then be necessary to make it clear that we were being given only a theory of sympathetic actions, not a theory of magic in general. In fact, nobody so far has been able to produce a clear, complete and wholly satisfactory idea of magic which we could make use of. As a result we are reduced to providing one for ourselves.
In order to succeed in this aim, we determined not to restrict our studies to one or two magical systems, but to consider the largest number possible. We do not believe that an analysis of a single system, however well chosen, would be sufficient to deduce laws, applicable to all magical phenomena, since our uncertainty about the actual boundaries of magic leads us to doubt whether we could find the totality of magical phenomena in one magical system. On the other hand, we propose studying as many heterogeneous systems as possible. In doing so we may be able to establish whether magic-no matter how it varies in relation to other categories of social phenomena from culture to culture-involves, in some degree, the same basic elements and whether it is on the whole everywhere the same. Above all we must make parallel studies of magical systems of both primitive and differentiated societies. In the former we shall find the most perfect form, the basic phenomena of magic from which others derive; in the latter, with their more complex organization and more distinct institutions, we shall find data which are more intelligible to us and which will provide insights into the functioning of the primitive systems.
We have taken care to use only the most reliable material which gives a complete coverage of magic in the society concerned. This, of course, drastically reduces the field of our observations, but it is essential to rely on facts, which as far as possible are beyond criticism. We have included the magic of certain Australian tribes; 1 a number of Melanesian societies; 2 two Iroquois nations, the Cherokee and Huron, and the Algonquin magic of the Ojibway. 3 We have also included ancient Mexican magic; 4 the contemporary system of the Straits Settlements in Malay, 5 and two of the forms magic has assumed in India-contemporary folk magic of the north-western states and the quasi-scientific form it took under the direction of certain Brahmans of the literary period known as Vedic. 6 While we are unfortunate in the quality of material in the Semitic languages, we have not neglected this subject entirely. 7 Studies of Greek and Latin magic 8 have been particularly useful in the study of magical representations and the functioning of a well-differentiated magic. We have also used well attested material taken from the history of magic in the Middle Ages 9 and from French, Germanic, Celtic and Finnish folklore.
NOTES 1 The Arunta:-B. Spencer and F. J. Gillen, The Native Tribes of Central Australia, London, 1898; Pitta-Pitta and neighbouring tribes in central Queensland-W. Roth, Ethnographical Studies among the North-Western Central Queensland Aborigines, Brisbane, 1897. G. Kurnai; Murring and neighbouring tribes of the south-east-L. Fison and A. W. Howitt, Kamilaroi and Kurnai, 1885; 'On some Australian beliefs', Journal of the Anthropological Institute, 1883, xiii, p. 185 et seq.; 'Australian medicine-men', J.A.I., xvi, p. 30 et seq. 'Notes on Australian songs and song-makers', J.A.I, xvii, p. 30 et seq. These precious documents are often incomplete, particularly as far as incantations are concerned.
2 The Banks Islands, Solomon Islands and the New Hebrides-R. H. Codrington, The Melanesians, their Anthropology and Folklore, Oxford, 1890; as well as this capital study we have used a certain number of ethnographical works, including those of M. Gray on the Tanna (Proceedings of the Australian Association for the Advancement of Science, January 1892); cf. Sidney H. Ray, 'Some notes on the Tannese', Internationales Archiv für Ethnographie, 1894, vii, p. 227 et seq. These writings are of interest since they provide information on the subject of mana, but they are incomplete so far as details on ritual, incantations and the general system of magic and the magician are concerned.
3 Among the Cherokee we have proper texts, ritual manuscripts written by magicians, in Sequoyah characters; J. Mooney has collected almost 550 formulas and rituals and has often succeeded in obtaining some of the best commentaries: The Sacred Formulas of the Cherokees, 7th Annual Report of the Bureau of American Ethnology, 1891; Myths of the Cherokee, 19th Ann. Rep. Bur. Amer. Ethn., 1900. For the Huron we have used only the excellent material of J. N. B. Hewitt on the orenda and we give an account of them later on. Ojibway pictograms (Algonquin), depicting initiation in diverse magical societies have also been of great value. Both written texts and figures are included in the work of W. J. Hoffman, The Mide'wiwin of the Ojibwa, 7th Ann. Rep. Bur. Amer. Ethn., 1887.
4 For Mexican magic see the illustrated manuscripts in Spanish and Nahuatl made for Sahagun, published, translated and commentated by E. Seler, 'Zauberei und Zauberer im Alten Mexico', in Veröff a.d. Kgl. Müs. f. Völkerk, vii, 2, pp. 2-4, in which the material is excellent if brief.
5 The book by W. W. Skeat, Malay Magic, London, 1889, contains excellent factual reports, well analysed and complete, and observed by the author himself or collected in a notable series of magical manuscripts and treatises.
6 The Hindus have left us an incomparable body of magical texts: hymns and formulas in the Atharva Veda, ed. R. Roth and W. D. Whitney, 1856; edited with commentaries by Sâyana, Bombay, 1895-1900, 4 vols; translated by A. Weber, Books I-VI in Indische Studien, vols 11-18; translation by V. Henry, Books VII-XIV, Paris, Maisonneuve, 1887-96; translation with commentary and a choice of sacred songs, M. Bloomfield, 'Hymns of the Atharva-Veda', in Sacred Books of the East, vol.42, Oxford, 1897; ritual texts of the Kauçika-Sutra (ed. Bloomfield, J. Amer. Oriental Soc., 1890, xiv: partial translation with notes by W. Caland, Altindisches Zauberritual, Amsterdam, 1900; A. Weber, 'Omina und Portenta', in Abhdl d. Kgl. Ak. d. Wiss., Berlin, 1858, pp. 344-413). However, it should be pointed out that we are aware that these inaccurately dated texts present only a single Hindu tradition, a literary tradition of a single Brahmanical school, belonging to the Atharva-Veda. It therefore does not cover all Brahman magic, any more, of course, than it represents all the magic of ancient India. For modern India we have mainly relied on the collection by W. Crooke, The Popular Religion and Folklore of Northern India, 2 vols, London, 1896. It has a certain number of gaps, above all as far as details of ritual and textual formula are concerned.
7 For Assyrian magic we possess some exorcism rites only: C. Fossey, La Magie assyrienne, Paris 1902. We have only fragmentary material on Hebrew magic: T. W. Davies, Magic, Divination and Demonology among the Hebrews, Leipzig, 1898; L. Blau, Das altjüdische Zauber-wesen, Strassburg, 1898. We have not included any discussion on Arab magic.
8 One of the authors has already provided an account of the value of Greek and Latin sources-H. Hubert, 'Magica', in Dictionnaire des antiquités grecques et romaines, C. V. Daremberg and E. Saglio, vi, Part 31, p. 9 et seq. We have preferred to rely on magical papyrus materials, which provide, if not details of whole rituals, at least exact indications of a certain number of rites. We have used the texts of alchemists (P. E. M. Berthelot, Collection des anciens alchimistes grecs Paris, 1887). We have used material in magical tales and stories only with great caution.
9 Our study of magic in the Middle Ages has been greatly facilitated by J. Hanson, Zauberwahn, Inquisition und Hexenprozess im Mittelalter, Munich, 1900, and Quellen und Untersuchungen zur Geschichte des Hexenwahns und der Hexenverfolgung im Mittelalter, Bonn, 1901.