would be able to draw up a limited catalogue of spirits, limited in type if not in number. This hypothetical and theoretical limitation provides us with the first hint of the collective character of our demoniacal representations. Secondly, demons may be named in the same way as gods. Since they are normally used for all kinds of purposes, the multiplicity of their services has given them a kind of individuality, and each is, individually, the subject of a tradition. Furthermore, commonly held beliefs in the magical power of a spirit being always presuppose that the spirit has given the public proof of its powers in the form of miracles or successful actions. A collective experience, or at least a collective illusion, is necessary before a demon, properly speaking, can be created. Finally, let us remember that most magical spirits are exclusively presented through ritual and tradition. Their existence is proved only after the growth of the belief which endows them with respect. Therefore, in the same way that impersonal representations of magic seem to have no reality outside collective traditional beliefs-beliefs which are held in common by a group concerned-in our view, personal representations are also collective. We even feel that the proposition will be more acceptable in this case.
The vague, multiform character of the spirit powers with which magicians have to deal is also a feature of magic as a whole. At first sight, the facts we have collected together may seem very disparate. Some tend to merge magic with technology and science, while others assimilate it to religion. In fact, it should be placed somewhere between the two, but it cannot be defined by its aims, processes or its ideas. Up to the present, our studies have shown that the subject is even more ambiguous, more indeterminate than ever. It resembles non-religious techniques in its practical aspects, in the automatic nature of so many of its
actions, in the false air of experiment inherent in some of its important notions. But it is very different from techniques when we come to consider special agencies, spirit intermediaries and cult activities. Here it has more in common with religion because of the elements it has borrowed from this sphere. There are almost no religious rites which lack their magical equivalent. Magic has even developed the idea of orthodoxy as we see in the , those magical accusations dealing with impure rites in Greco-Egyptian magic. However, apart from the antipathy which magic shows towards religion and vice versa (an antipathy, moreover, which is neither universal nor constant), its incoherence and the important role played by pure fancy make it a far cry from the image we have learnt to associate with religion. Nonetheless, the unity of the whole magical system now
stands out with greater clarity. This is the first gain to be made from our incursions into the subject and our long discussions. We have reason to believe that magic does from a real whole. Magicians share the same characteristics, and the effects of their magical performances-in spite of an infinite diversity-always betray much in common. Very different processes can be associated together as complex types and ceremonies. Quite disparate notions fuse and harmonize without the whole losing anything of its incoherent and dislocated aspects. The parts do, in fact, form a whole.
At the same time the whole adds up to much more than the number of its parts. The different elements which we have dealt with consecutively are, in fact, present simultaneously. Although our analysis has abstracted them they are very intimately and necessarily combined in the whole. We considered it sufficient to define magicians and magical representations by stating that the former are the agents of magical rites, while the latter are those representations which correspond to them-we considered them together in relation to magical rites. We are not in the least surprised that our fore-runners have preferred to
consider magic solely as a series of actions. We might also have defined magical elements in relation to the magician. Each presupposes the other. There is no such thing as an inactive, honorary magician. To qualify as a magician you must make magic; conversely, anyone who makes magic is, at least for the moment, a magician. There are part-time magicians who revert immediately to their status of layman as soon as the rite is accomplished. As for representations, they have no life outside ritual. Most of them offer little of theoretical interest to the magician and he rarely formulates them. They have solely a practical interest, and as far as magic is concerned they are expressed almost entirely through actions. The people who first reduced them to systems were philosophers, not magicians. It was esoteric philosophy which promulgated a theory of magical representation. Magic itself did not even attempt to codify its demonology. In Christian Europe, as well as in India, it was religion which classified demons. Outside ritual, demons exist only in fairy tales and church dogma. In magic, therefore, we have no pure representations and magical mythology is embryonic and thin. While in religion ritual and its like on the one hand, and myths and dogmas on the other, have real autonomy, the constituents of magic are by their very nature inseparable.
Magic is a living mass, formless and inorganic, and its vital parts have neither a fixed position nor a fixed function. They merge confusedly together. The very important distinction between representation and rite sometimes disappears altogether until we are left with the mere utterance of a representation which thereby becomes the rite: the venenum veneno vincitur is an incantation. The spirits which the sorcerer possesses or which possess the sorcerer may become confused with his soul or his magical powers. Spirits and sorcerers sometimes have the same name. The energy or force behind the rite-that of the spirit and the magician-is usually one and the same thing. The normal condition of magic is one involving an almost total con
fusion of powers and roles. As a result, one of its constituent features may disappear without the nature of the whole changing. There are magical rites which fail to correspond to any conscious idea. The action of spell-binding is a case in point, as well as many imprecations. Conversely, there are cases where representations absorb the ritual, as in genealogical charms, where the utterance of natures and causes constitutes the rite. In sum, the functions of magic are not specialized. Magical life is not compartmentalized like religion. It has not led to the growth of any autonomous institutions like sacrifice and priesthood. And, since magical facts cannot be divided up into categories, we have been forced to think in terms of abstract elements. Magic is everywhere in a diffuse state. In each case we are confronted with a whole, which, as we have pointed out, is more than the sum of its parts. In this way we have shown that magic as a whole has an objective reality-that it is some thing. But what kind of thing is it?
We have already gone beyond the bound of our provisional definition by establishing that the diverse elements of magic are created and qualified by the collectivity. This is our second, noteworthy advance. The magician often qualifies professionally through being a member of an association of magicians. In the final count, however, he always receives this quality from society itself. His actions are ritualistic, repeated according to the dictates of tradition. As for representations, some are borrowed from other spheres of social life: the idea of spirit beings, for example. Further research will be required, involving religion directly, if we are to find out whether this idea is the result of individual experience or not. Other representations are not derived from the observations or reflections of individuals, nor does their application allow any individual initiative, since they are remedies and formulas which are imposed by tradition and which are used quite uncritically.
While elements of a magical system are collective in nature,
can the same be said for the whole? In other words, is there some basic aspect of magic which is not the object of representations or the fruit of collective activities? Is it not, in fact, absurd or even contradictory to suppose that magic could ever be, in essence, a collective phenomenon, when, in order to compare it with religion, we have chosen, from among all its characteristics, those which set it apart from the regular life of society? We have seen that it is practised by individuals, that it is mysterious, isolated, furtive, scattered and broken up, and, finally, that it is arbitrary and voluntary in nature. Magic is as anti-social as it can be, if by 'social' we primarily imply obligation and coercion. Is it social in the sense of being, like a crime, secret, illegitimate and forbidden? This is not quite true, at least not exclusively so, since magic is not exactly the reverse side of religion, in the way that crime is the reverse side of the law. It must be social in the manner of a special function of society. But in what way should we think of it? How are we to conceive the idea of a collective phenomenon, where individuals would remain so perfectly independent of each other?
There are two types of special functions in society which we have already mentioned in relation to magic. They are science and technology on the one hand, and religion on the other. Is magic a kind of universal art or possibly a class of phenomena analogous to religion? In art or science the principles and methods of action are elaborated collectively and transmitted by tradition. It is for these reasons that science and the arts can be called collective phenomena. Moreover, both art and science satisfy common needs. But, given these facts, each individual is able to act on his own. Using his own common sense, he goes from one element to the next and thence to their application. He is free: he may even start again at the beginning, adapting or rectifying, according to his technique or skill, at any stage, all at his own risk. Nothing can take away his control. Now, if magic were of the same order as science or technology, the difficulties we
previously observed would no longer exist, since science and technology are not collective in every single essential aspect, and, while they may have social functions and society is their beneficiary and their vehicle, their sole promoters are individuals. But it is difficult to assimilate to magic the sciences or arts, since its manifestations can be described without once encountering similar creative or critical faculties among its individual practitioners.
It only remains now to compare magic with religion; and here we are faced with formidable difficulties. We still uphold, in fact, that religion in all its aspects is essentially a collective phenomenon. Everything is done by the group or under pressure from the group. Beliefs and practices, by their very nature, are obligatory. In analysing a rite which we took as a type-that is, sacrifice-we established that society was present and immanent everywhere; that society itself was the real actor in the ceremonial drama. We even went so far as to maintain that the sacred objects of sacrifice were social things, par excellence. Religious life, like sacrifice, permits no individual initiative, and invention is admitted only under the form of revelation. The individual feels constantly subordinate to forces which are outside his power-forces which incite him to action. If we are able to demonstrate that within the field of magic there are similar powers to those existing in religion, we shall have shown that magic has the same collective character as religion. All that will then remain to be done will be to show how these collective forces are produced-in face of the isolation which magicians insist on-and we shall thereby conclude that these individuals have merely appropriated to themselves the collective forces of society.
AN ANALYSIS AND EXPLANATION OF MAGIC
Thus we have gradually reduced our study of magic to the pursuit of collective forces which are active in both magic and religion. Indeed, we believe that once these collective forces are found, we will have an explanation for both the whole of magic and its parts. In fact, we should never forget that magic is continuous in nature and that its elements, which are extremely interdependent, frequently seem to be little more than different reflections of the same thing. Actions and representations are inseparable to such an extent that magic could be called a practical idea. Taking into account the monotony of its actions, the limited variety in its representations, the sameness which is found throughout the history of civilization, we might also assume magic to be a practical idea of the utmost simplicity. We should, therefore, expect that the collective forces involved would be far from complex and that the methods thought up by the magician to use them would be far from complicated.
We shall try to determine these forces by first posing the
problem as to the kind of beliefs of which magic has been the object, and then analysing the idea of magical efficacity.
Magic, by definition, is believed. Since, however, we cannot separate the various aspects of magic and since they frequently merge, they cannot be the object of very clear-cut beliefs. They are all, at one and the same time, the object of the same affirmation. This includes not only the magician's power and the value of the ritual, but also the totality of, and the principles behind, magic. Just as the whole of magic is more real than its parts, so a belief in magic is generally more deeply rooted than beliefs in its separate parts. Magic, like religion, is viewed as a totality; either you believe in it all, or you do not. This can be verified in those cases where the reality of magic has been questioned. When this kind of debate first arose, at the beginning of the Middle Ages, then again in the seventeenth century and today where it is carried on obscurely, we find that discussion always turns on a single point. Agobard of Lyons, for example, was concerned only with people who brought about bad weather. Later it was the fact of impotence being caused by spells, or the aerial flights of Diana's suite. Balthasar Bekker in De Betoverde wereld (Leeuwarden, 1691) was concerned solely with the existence of demons and the devil. And in our times it is astral bodies, apparitions, the reality of the fourth dimension. But in all cases conclusions are immediately generalized, and a belief in a single case of magic implies the belief in all possible cases. Conversely one negative instance topples the whole edifice; magic itself then comes under suspicion. We have examples of obstinate credulity and deeply rooted faith crumbling before a single experience.
What is the nature of these magical beliefs? Have they anything in common with scientific ones? The latter are a posteriori beliefs, constantly submitted to the scrutiny of individuals and
dependent solely on rational evidence. Does the same hold for magic? Evidently not. We have one case, extraordinary though it may seem, of the Catholic Church upholding belief in magic as a dogma, and maintaining it with sanctions. In general, these beliefs are automatically diffused throughout society. They are separated from their origins. In this sense, magical beliefs are not so very different from scientific beliefs, since every society has its science, equally diffuse, whose principles have sometimes been transformed into religious dogmas. But while all science, even the most traditional, is always conceived as being positive and experimental, magic is a priori a belief. Magical beliefs, of course, derive from experience: nobody seeks out a magician unless he believes in him; a remedy is tried only if the person has confidence in it. Even in our own days, spirits do not admit unbelievers into their midst. Their presence is believed to render their activities null and void.
Magic has such authority that a contrary experience does not, on the whole, destroy a person's belief. In fact, it escapes all control. Even the most unfavourable facts can be turned to magic's advantage, since they can always be held to be the work of counter-magic or to result from an error in performance of the ritual. In general, they are seen to stem from the fact that the necessary conditions of the rite were not fulfilled. Cross-examination during the trial of the magician Jean Michel, who was burned alive in Bourges, 1623, showed that this poor man-a carpenter by profession-spent his whole life carrying out experiments which always failed. Once he almost achieved his aim but, overcome by fear, he ran away. Among the Cherokee, the failure of a magical rite, far from undermining the people's confidence in the sorcerer, merely endowed him with greater authority, since his offices were indispensable to counteract the terrible effect of the powers, which might return to harm the clumsy individual who had unleashed them without taking the correct precautions. This happens in all magical
experiments. Fortuitous coincidences are accepted as normal facts and all contradictory evidence is denied.
Nonetheless, there has always been a pressing urge to support magical beliefs by providing precise, dated and localized proof. In cases where a whole literature on the subject exists-in China and medieval Europe, for example-it will be found that an identical recital of facts is repeated ad infinitum from text to text. They are traditional proofs, anecdotal magical tales which are used to bolster magical beliefs, and they are pretty much the same the world over. In all this, we are not dealing with any conscious sophistry, but rather with exclusive prepossession. Traditional proof is sufficient and magical stories are believed in the same way as myths. Even in those cases where magical tales are jokes, there are very few examples of any turning out badly. Belief in magic, then, a priori is quasi-obligatory and exactly analogous to belief in religion.
These beliefs hold for the sorcerer as well as for society. But how is it possible for a sorcerer to believe in magic, when he must constantly come face to face with the true nature of his methods and their results? Here we must confront the serious problem of fraud and simulation in magic.
In order to deal with this question, let us take the case of the Australian sorcerers. Of all magical practitioners, there are few who seem so firmly convinced of the efficacy of their ritual. Yet keen observers have attested that the sorcerer has never-nor believed he has ever-seen any automatic effects of his actions in rites practised under normal conditions. Let us look at the methods of black magic. In Australia they may be reduced to three main types and are practised either concurrently or separately in the various tribes. The first type-and the most widespread-is sympathetic magic proper, whereby an object which is believed to be part of a person, or to represent him, is destroyed. These objects may include left-over food, organic remains, footprints, images. It is impossible to imagine that the
magician ever believed, by virtue of any experiment, that he was really killing somebody by burning bits of food mixed with wax or fat, or by piercing an image. Our suspicion that the illusion is only a partial one is confirmed by the rite mentioned by B. Spencer and F. J. Gillen, involving first the piercing of an object representing the soul of the victim, and secondly the throwing of the object in the direction of his dwelling. The second type of ritual, practised primarily by the southern, central and western tribes, involves the removal of the fatty parts of a person's liver. It is believed that the sorcerer approaches his sleeping victim, cuts open his side with a stone knife, removes the fat and closes the wound before leaving the spot. The victim dies slowly, unaware of anything untoward having happened. Quite clearly this rite could never have actually been carried out. A third type, practised in the north and central regions is known as 'pointing the bone'. The sorcerer is believed to hit his victim with some fatal substance. In fact, however, the weapon is not even thrown in some of the instances cited by W. E. Roth. In others, it was thrown but from such a distance that it would be impossible to imagine that it ever arrived or transmitted, through contact, the fatal wound. Often it is not even seen to leave the magician's hand and certainly never seen to arrive immediately after having been thrown. Although many of these rites would never have been completely realized and although the effectiveness of many others can never have been proved, they are nevertheless in current usage, as has been shown by the best witnesses and by the existence of numerous objects signifying the tools of their magic. It must be accepted that the sorcerer sincerely, though willingly, believes his gestures to be a reality and the beginnings of an action to be complete surgical operations. The ritual preliminaries, the gravity of each move, the intensity of the dangers undergone (the rite involves approaching an enemy's camp where, if found, he would be killed on sight) and the seriousness of the whole performance reveals a genuine will to believe in it.
However, it would be very hard indeed to imagine an Australian magician opening up the liver of his victim without causing instant death.
However, along with this 'will to believe', there is plenty of proof of actual belief. The best ethnographers confirm that the magician deeply believes in the success of his sympathetic magic. In assuming cataleptic and nervous states, he may truly fall prey to all kinds of illusions. At all events, while the sorcerer may have only a mitigated confidence in his own rites and is doubtless aware that the so-called magical poisoned arrows, which he removes from the bodies of people suffering from rheumatism, are only pebbles taken from his mouth, the same sorcerer still has recourse to another medicine man when he himself falls ill. And he will either be cured or allow himself to die, according to whether his doctor condemns him to death or pretends to save him. Thus, while there are some people who do not even see the poisoned arrows depart, there are others who see them arriving at their destination. They arrive as whirlwinds, flames cleaving their way through the air, or as small pebbles, which the medicine man extracts from his body, yet the patient knows full well that they have not been removed from his body. The minimal sincerity which the magician can be accredited with is, at any rate, that he does believe in the magic of others.
This holds true for systems of magic outside Australia. In Catholic Europe we have at least one case where the confessions of a witch were not forced out as a result of the judge's inquisition. At the beginning of the Middle Ages the canonical judge and the theologian refused to accept the existence of the flights of witches in Diana's suite. But the witches, victims of their delusions, continued to boast about them, to their own detriment, finally imposing their fantasies on the Church. These untutored, yet intelligent people, like witches everywhere, easily misled and prey to nervousness, held their beliefs with a sincerity and tenacity which was incredibly strong.
Nonetheless, we are forced to conclude that there has always been a certain degree of simulation among these people. We are in no doubt that magical facts need constant encouragement and that even the sincerest delusions of the magician have always been self-imposed to some degree. A. W. Howitt relates, with reference to the pieces of quartz which the Murring sorcerers draw from their mouths-the initiating spirit packs them into their bodies-that one of the sorcerers told him: 'I know all about it. I know where they come from'. We have other confessions, no less cynical.