Synthetic images

This work is licensed under the Creative Commons | © Rodney Needham. ISSN 2049-1115 (Online). DOI: http://dx.doi.org/10.14318/hau4.1.039


Synthetic images



“I have ever beleeved, and doe now know, that there are Witches.” It is little more than three hundred years since a humane and learned doctor, Sir Thomas Browne, made that decided declaration. Some years later even he acted as medical witness for the prosecution in sending two old women to their deaths for the crime of witchcraft. Since his day we in the West have not been free of the obsession, for even when the last poor creature had been burnt or put to the test, the power of the collective representation survived. Historians reverted again and again to the trials and the edicts and the fulminations; folklorists traced the continued expression of ideas about witchcraft into modern society; eccentrics claimed to be, or tried to be, witches and formed themselves into covens on desolate farms and in sedate suburbs; and the anthropologists, of course, made it into one of the stock and indispensable topics of their subject.

In these regards it can be said that we are still in the power of the idea of witchcraft, just as we resort to its dramatic power in our metaphors of moral condemnation and political castigation. If it is objected that we do not actually do anything about it, this is true in the sense that we do not arrest and torture those whom we call witches, and that we do not presume to be witches those who are the victims of witch hunts. But we do not have to do and think so in order to conform to a collective representation; for neither did most people, I can conceive, in the days when such things really were done. It was the church and the state that provided the authority and the punitive means: they were the sustainers of the institution, and in this respect were parts of it. We cannot say that the authorities responded to the common conviction that there were witches, for to the common people the fearsome notion of witchcraft was not a spontaneous apprehension but had all the autonomy, generality, and coercive force of a social fact. And if we were to say that they believed in the dread powers of witchcraft, and hence collaborated in the gruesome procedures by which the authorities tried to extirpate witches, we should be on even shakier ground. In spite of the volume of contemporary reports, we do not know what the greater number of the populace believed (on the assumption that we have a clear idea of what “believe” means), but only that the collective representation was in force.

We can well conclude, moreover, that people in the past no more positively acquiesced in the execution of witches, or played any other deliberate part in the enactment of the representation, than do we today in the operation of the institutions that govern our own lives, even when the instruments of state are called representative or the corporations are called public. If we agree that people really did think there were witches, this does not justify us in imputing any particular state of mind to them. There have been crazes and scares, but these are not generically characteristic of the institution. We ourselves know, just as objectively (let us say), that hundreds of thousands of people are killed and maimed on the turnpikes; and although we regard this as deplorable, and try to make turnpikes less dangerous or to avoid the dangers if we travel on them, the knowledge and the consequent precautions do not argue for any special kind of judgment or apprehension that is distinctively associated with turnpikes. Indeed, this is much what we find when we read ethnographic accounts of societies where witches are just as real, and also just as normal an aspect of everyday life, as are turnpikes to us. There are witches, all right, and people rely on regular precautions and techniques in order to cope with them.

I have been stressing the points in these preliminary remarks in order to throw into contrast the aspect of witchcraft that I think really important and that is to be the subject of this lecture. The more normal the idea of witchcraft is taken to be, in a society where it is in force, the more striking is that aspect. The less we assume that the institution of witchcraft involves a special inner state, the more readily we shall be able to analyze the power and persistence of the complex that defines the institution. If on the other hand, we regard the institution of witchcraft merely as a cognitive aberration, a kind of collective nightmare that can be exorcised by science, then I think we are missing some of its most interesting and instructive features. Actually, the case I intend to make first is that social anthropology, in the ways it has approached witchcraft, has tended to pass over these features and at the same time has not made a good scientific argument in any other respect.

Difficulty began with the definition of a witch, and sporadically a fair amount of literary energy went into discriminating between witch and sorcerer, the former working by some intrinsic property, the latter by recourse to material means; then in deciding whether in a certain society a mystical practitioner was the one or the other, or maybe both at the same time; and then in qualifying propositions about witchcraft according to whether a witch or a sorcerer was in question, not to mention the alternative statuses of wizard, magician, conjuror, and so on. Moreover, beyond the range of the English words there were the numerous terms in other languages that were indifferently translated as “witch” and the like, and each of these constituted a semantic problem that did not readily conduce to comparative generalizations. And against this tangled back ground anthropologists still found it possible to speak of “genuine” witches or of witches “proper,” to differ over whether witches were essentially immoral, and, expectably enough, to become diverted by one new typological consideration after another according to the ethnographic data adduced or the cast put upon them by the anthropological commentator. At the end of the day, it remains a question if there is yet a definition of a witch that is agreeable, in the rigorous acceptation seemingly required, to anthropologists as a body. If you ask whether that really matters, the answer is that it depends on the theoretical or comparative propositions that have “witch” as their subject. We shall come to some of these in a moment. For the present occasion, I shall give the word its common, or garden, meaning of someone who causes harm to others by mystical means. If you are inclined to protest that this is altogether too rough and ready, let me invite you to wait a little until I show that this vagueness of definition has no importance for the purpose I have in view.

You may, by the way, find it interesting to keep in mind later a point about Germanic etymology. The English word witch comes from the Old English wicce, the feminine form of wicca, which is rendered as a male magician, sorcerer, or wizard, and this is the end of the trail, since the source of these words is not understood. But in German the equivalent Hexe comes from the Middle High German hag, meaning a “fence,” “hedge,” or “enclosure”; and the Icelandic tūnriða, witch, means literally a female “fencerider.” There is thus a connection, which will acquire its point as we proceed, between witches and boundaries. It would be instructive to learn if this association is to be found in other linguistic traditions.

Another preliminary matter that ought to be mentioned is the almost universal premise subscribed to by anthropologists, that witches do not really exist: that is, no human beings have the secret power to inflict harm or do evil as witches are supposed to do. This is a curious presumption, though I do not maintain that it has led to any pragmatic difficulty in the study of witchcraft. It is odd because it is a flagrant instance of sheer prejudice of the kind that anthropologists are usually careful to avoid when dealing with other mystical institutions. They do not as a rule begin by declaring that God does not exist or that ancestral spirits are merely imaginary or that the benefits of blessing are illusory. They say that these notions are to be treated as social facts, and that there is simply no need to pronounce on their truth or falsity. But in the case of witchcraft there is no such compunction: witches just do not exist, and the question posed is hence why men universally but misguidedly think that they do. I must say this strikes me as pure dogma. I myself have no idea, empirically, whether any human beings possess a secret capacity to inflict harm by some immaterial and unseen means. This seems to me something of an open question. I have no evidence that there is such a power, whether generic or confined to a minority of individuals, and my inclination is to suspect that most probably there is none. But the one thing I am sure of, simply as a point of method, is that we ought not to base our investigation into witchcraft on an unsure (even unexamined) premise, let alone the premise that the essential attribute of witches— namely the malign power—does not exist.

The standard concession to reality made by anthropologists is that the idea of witchcraft must be related, as Philip Mayer ([1954] 1970) has put it, to something real in human experience; but the next move is none the less to fall back on another prejudice, namely, that the reality in question consists in social and psychological strains to which the postulation of witchcraft is a social response. I am not saying that correlations of the kind cannot be made (though what they are worth is a different matter), but that the presumed locus of the reality of witchcraft corresponds in the first place to the sociological predilections of anthropologists. An extreme expression of this leaning is to be seen in an assertion, by two leading authorities, that sociological analysis “must” be employed if we wish to develop explanatory formulations which can subsume the facts from more than one society.

However that may be, the notion that witchcraft accusations point to “weak spots in the social structure” has had a considerable prevalence. Related propositions are that the accusations derive from a conflict between ideal and actual social relations; that they are a means to break relationships that have become insupportable; or that they act as a safety valve by which aggressive impulses are sublimated. The ethnographic accounts presented in these terms may be very informative, and the focus on tensions and strains can certainly contribute to a true description of social life, but the approach is nevertheless subverted by a number of considerations that cannot be countered by the improvement of fieldwork.

First, it is tautological to say that witchcraft accusations point to weak spots or to difficult relationships, for it is in part the accusations themselves that characterize the spots as weak or the relationships as uneasy. Moreover, what is needed is an independent gauge of strength and smoothness by which the ethnographer could assess these qualities in the absence of their liability to accusations; but in the nature of the case the possibility of correlating witchcraft accusations and social vulnerability, as independent variables, cannot be had. And in any case a test by concomitant variation is called for: to identify “strong” points in the social structure and to check that they are free of witchcraft accusations; and to check the “weak” points and see if they are regularly the targets of accusations. These however are tasks that in the main have not been earned out in ethnographical analyses, and lacking their results we are left without satisfactory empirical proof of the alleged connection. Finally, even if the connection could be established in particular instances, the question would remain whether the weak and difficult spots were so precisely because, for some other reason perhaps, they were conventionally regarded as the loci of witchcraft. This is a question that applies similarly to the safety valve, or sublimation, hypothesis, for we cannot know to what extent the tension or the aggression is a product of the institution itself.

As matters stand, at any rate, I think it is true to say that no sociological or psychological explanation of the differential incidence of witchcraft accusations has been borne out empirically as a general proposition that is valid for witchcraft everywhere. Witches are neighbors, or else they are distant; they are relatives, or else they cannot be relatives; they are marginal, or else they are enemies within; they are lowly misfits, or else they are secure and prosperous just because of their witchcraft; they are so categorized that not everybody can be a witch, or else they are such that anyone may be a witch. Occasionally a sociological proposition is framed in even more detailed terms: for example, that witchcraft beliefs tend to be utilized in societies in which unilineal kinship principles are employed in the formation of local residential groups larger than the domestic household. The literature of anthropology is replete with propositions like these, each perhaps persuasive in its own ethnographic setting, but none survives as a key to the institution wheresoever it may be found. No wonder that a historian such as Trevor-Roper asserts that witchcraft beliefs are inseparable from the ideology of the time. But this conclusion does not allow for, even if it does not rule out, the comparativism that is proper to social anthropology and that alone may provide a general interpretation an institution that has such a global distribution. The historian’s conclusion is the ethnographer’s indispensable premise.

There is still one other kind of explanation, however, that seems to fare better and that is not cast in terms of jural institutions and social systems. This is the view that the idea of witchcraft provides a theory of misfortune. If that termite-riddled granary (the bane of anthropological examiners) falls on you and not on someone else, just as you happened to be sitting under it, the activity of a witch provides not only an answer to the question “Why me?” but a final and complete answer. Also it enables you to do something definite, dramatic, and perhaps personally advantageous about the source of your misfortune.

No doubt the institution of witchcraft does have these occasional uses, but they do not explain why it is that this particular institution is employed in order to explain misfortune. After all, there are many ways to do that. If misfortune strikes, you can blame an inscrutable god or capricious spirits; you can concede that it is the just retribution of your own sin, or else that it is the automatic consequence of some unintended fault; you can put it down to bad luck (if your culture happens to have this Germanic concept), or more calculatingly you can ascribe it to chance (though this is an even more difficult notion). Theories of these kinds are legion, they are found also in societies that ascribe misfortune to witchcraft as well, and even people obsessed by witches do not blame every misfortune on their malign intervention. Why then should people hold to a theory that places responsibility for their misfortunes on other people? From a pragmatic point of view, it does not even seem a particularly desirable theory. If people think they are afflicted by an inscrutable god, they can at least band together in an attempt to placate their divine scourge by communal means that do not foster suspicion and set them one against another. But to blame individual human beings for riddled posts or failing crops or the attacks of unpredictable wild animals seems the most self-damaging theory of any.

All the same, the idea is remarkably prevalent in history and in world ethnography, and for all its apparent social disadvantages we have to accept that in fact this theory of blame is the way in which a great many peoples have chosen to think about the sources of their troubles. This fact in itself tells us something about the inclinations of human beings under stress, but I do not think the theory of misfortune hypothesis tells us anything interesting or revealing about the institution of witchcraft. In a sense, indeed, we might well not have expected it to do so, for this proposition too is tautologous by definition: ideas about witchcraft account for the blows of misfortune after all not for the blessings of comity.

It has taken me some little while to run through this essential introductory survey of the state of theory with regard to witchcraft, and as you see I cannot find that there is much in the way of positive results to report. This is not surprising, let alone dejecting, if only because it is a delusion to suppose that we shall do best at explaining widespread and constant social facts. We can do much better with limited and variant institutions, that is, when the weight of comparativism can least be brought to bear. If comparison is the characteristic method of social anthropology, it does not follow that we shall be very effective with it. At any rate, we have not in fact got very far in the scientific treatment of the idea of witchcraft.

It is against this background that I now want to turn to the features of witchcraft that earlier on I said struck me as interesting and as anthropologically neglected. I am not going to claim any great progress or revelation, and the matter is not one for decisive argument; but when we are so much baffled by an institution as we are in this case, an oblique attention to it, from another standpoint, may just make a difference. You will already be familiar with the features themselves, but it is perhaps when we think we are most familiar with a thing that a change of aspect can best disclose its further properties.

The aspect that I want to focus on is the image of the witch, and I shall try to make sense of its components by resort to what were introduced in the first lecture as primary factors of experience. There is, as you will see, no technique for doing this, and at points I have to rely on conjectures that are no more than plausible; but what follows is at any rate a way of thinking about witchcraft that may lead somewhere interesting.

The first reason for taking the image as the object of analysis, rather than the sociological matters that have hitherto preoccupied most anthropologists, is that amid a welter of contingent social facts (which, as we have found, have not been brought into a consistent theoretical order) this complex construction of the imagination displays a very remarkable constancy. I do not mean by this that the components of the image of the witch are always the same in number and character, from one tradition to another, but that there are characteristic features which combine polythetically (that is, by sporadic resemblances) to compose a recognizable imaginative definition of the witch.

Let us approach this representation by way of its moral component. The witch is said to do such horrible things as to eat children, practice cannibalism by secretly devouring people’s organs, commit incest, and otherwise act in a vile and malevolent manner such as only the right-minded could imagine. (Eating children, incidentally, has a special vogue: Domitian charged Apollonius with it; the Romans accused the Christians; the Christians in turn the Jews.) Such particulars are infinitely elaborated on, but these are some of the major attributes. Anthropologists have frequently remarked that such conduct is grossly abnormal, or shocking, or contrary to decent norms; but I suggest that those commentators are more exact who say that morally the witch is the very opposite of the right values of society. This is not a trite point of vocabulary, but it corresponds significantly to the estimations of members of society. Typically, moral evaluations are scalar, such that actions are judged as more or less good, more or less bad. There are extreme instances at the polar ends of the scale, but at one end these are hypothetical; even societies that recognize saints concede that these exemplars have blemishes (as the saints themselves of course have to insist) and are not absolutely good. It would be both expectable and practicable, therefore, if societies were to place witches at various points toward the opprobrious end of the moral scale; certainly they would be bad, but more or less bad, and this means that they would be allowed to possess some virtues. But in fact the witch is not given the benefit of this moral nicety: the witch is absolutely and irremediably evil, a real instance of the polar type that merits utter condemnation. The witch’s conduct is not merely contrasted with the ideal: there is nothing worse than the acts that the witch is imagined to perpetrate, so that the witch’s conduct is strictly the opposite of the ideal. This gives us our first factor in the complex, namely, the relation of conceptual opposition.

I need not say much about this, for after Hertz a great deal has been written about polarities, complementaries, and syzygies in dual symbolic classification, especially with regard to the sets of opposed values signaled by right and left. It is important, however, not merely to isolate the relation of opposition from the moral definition of the witch, but also to stress that this relation has a fundamental character that makes it apt to the purpose of relegating the witch to the point of extreme censure. In other words, people resort to opposition as the simplest and most efficacious means of classification. They do so in innumerable institutions and contexts, including moiety systems and symmetric alliance and Manichean theologies, and witchcraft is merely one example of this proclivity.

I mentioned just now that opposite values may be signaled by the opposition of right and left. This spatial expression of nonspatial values, including moral qualities, has an analogue in a second component of the image of the witch. The behavior of the witch is commonly described as inverted, or the witch is said to embody inverted values. This is not simply a metaphor (like “opposite,” for that matter), but is frequently a description of the witch’s physical posture: the witch proceeds upside down, walking on his hands as the Kaguru imagine, or presents himself backwards. This imaginary inversion makes an apt picture of the perverse nature of the witch, and in itself it is both comprehensible and telling; but once again what we encounter in this component also is simply one example of the operation of a primary factor. When the Lugbara advert to the alien nature of their neighbors, members of other tribes just over the horizon, they say that the latter normally go about upside down. (It is only when you look at them instantaneously they turn the right way up.) The Lugbara are not saying that all strange peoples are witches: they are expressing the strangeness by resorting to the image of inversion. This is in fact a very common symbolic means of marking a boundary. Similarly, many peoples imagine the world of the dead as the opposite of this, in that the spirits there speak backwards or place opposite values on things; and the Batak say also that the dead climb house-steps downwards, head first, that is, upside down. They are not saying that the spirits of the departed are witches: they are expressing the spirituality of ghosts, in opposition to terrestrial natures, by resorting to the image of inversion. The symbolic operation is the exploitation of the imaginative recourse. Inversion is an elementary mode of marking a contrast—especially at a moral or temporal boundary—and the wide extent to which it is employed, in many customs and ideologies, indicates that it is not merely a formal possibility that is available to the imagination, but that it is a positive proclivity by which men tend to be influenced in their collective representations.

Next, witches habitually go about their business at night. This looks obvious enough, for they work secretly and wish to avoid the gaze of decent people, including that of their victims. An Apache raiding party also used to travel at night, in order to gain the advantage of surprise; and burglars are supposed to operate at night in order not to be seen, either by their victims or by potential witnesses for the prosecution. But of course there is more to the witch’s association with night than a practical precaution, especially since it is doubtful that anyone does travel as a witch at night anyway. What we are dealing with is a symbolic, I dare say universal, image by which light and the absence or deprivation of light stand for opposed values and properties of innumerable kinds. The Bible is full of examples of the precellence of light and the opposite condition of darkness; mystical ideology continually resorts to the contrast; and it is a rich source of metaphor in estimating clarity of understanding and moral inclinations. As for the specific connection of witches with night, this too has analogues in other institutions. The first to come to mind is shamanism, in which it is a standard feature that the shaman holds his seance either at night or in a darkened enclosure. This is no doubt because the proceedings have to do with hidden things, with the mystical. (Recall that this word comes from the Greek mūein,“to close the eyes.”) There is no implication that a shaman is a nefarious person, acting at night in order to conceal his iniquity. Nor of course do I imply that the mystical is symbolized only by darkness, for under some aspects it is enacted in the full radiance of bright light. But the image of the witch combines both the fearsome presence of the powers of darkness and the nocturnal setting of spiritual activity.

This is sometimes symbolized by the color that is appropriate to night, namely, black. European witches used to smear themselves with soot, and witches in other cultures are on occasion distinguished by black appurtenances. In this case again, though, it is not witches alone that are symbolized by black. Our own priests wear black, so does the Mugwe of the Meru, and so also does a Gurung medium: In each instance the color, like night, stands for the mystical character of the status or the undertaking. There is no implication that all persons whose offices are symbolized by black share the horrible attributes of a witch. What particular attributes may be shared, beyond the mystical connection, is a contingent matter of local ideology. The imaginative constant is the resort to color in order to convey the significance. This general feature is nicely glossed by the example of the Kaguru witch: naturally black of skin, he is supposed to cover himself with white ashes, thus employing symbolic color in an image that is opposite to normal human appearance. So the generic factor is color, we may say, while the specific mystical significance is predominantly but not always conveyed by the color of darkness.

The association of a witch with night is reinforced by the usual character of the witch’s animal familiar, or of the creature into which the witch can transform himself. Some of the species are the black cat in Europe, the polecat in North America, the fox in Japan, the maned wolf in Brazil, the owl in India, the hyena in East Africa, the bat in the Congo. The obvious and well-recognized attribute that they share is that they are all nocturnal. Some moreover are predators, others carrion-eaters, and yet others are menaces in the dark. These attributes are easy enough to grasp as being appropriate to the nocturnal operation and the dark character of the witch. But once more this is not a singular kind of symbolism that distinguishes only the witch. The individual species are indeed particularly apposite to the image of the witch, but the fact that the status is symbolized by animals is not. The classical demonstration of this fact in anthropology is totemism, in which the classification of men into social groups is symbolized by a parallel classification of animal species. In totemism, a clan may even be associated with a nocturnal species, but there is no usual implication that the social group is composed of witches or that its members have the character of witches. What is happening in this institution is that in discriminating social statuses the members of totemic societies are resorting, by virtue of their tradition, to an extremely common method of symbolism. Ready examples are the animal supporters, such as the lion and the unicorn, in European heraldry; the animal “vehicles,” such as the bull of Shiva, that help to identify Hindu deities; the animal-headed gods of Egypt; the birds and other species that represent nations or the states of the U.S.A.; the nicknames and emblems of military formations and sports teams; the terms of moral appraisal that ascribe the strengths and failings of animals to human beings. In all these cases, and in a great many other institutions and figures of speech, men are resorting to a mode of symbolism that has little to do with the particular social facts and a great deal to do with a natural imaginative impulsion. The animal familiar of the witch is merely one instance of the operation of this factor.

Another prominent component in the complex image that we are examining is that very often the witch is supposed to have the power of flight. There is one apparent reason for this, namely, that the action of the witch is usually considered to take place over some distance, either very rapidly or almost instantaneously, and in any case at such a range as could not be covered in the time by a normal human being. The attribute of flight is a simple way to imagine the witch at his swift work, and it is reinforced by those familiars, such as owls and bats, that as natural species are themselves actually capable of flying. That speed is a contributory feature is indicated by the example of the Kalapalo, who explain the rapidity with which a man travels (at night, incidentally) by stating that he came like a maned wolf, the creature that is associated with their witches.

We can conjecture also a negative basis to the employment of the image of flight. When we conceive the operation of some force at a distance, we have the scientific knowledge to enable us to think of a wavelike action, as in the form of a radio beam or a laser or a death ray. But in what terms could we represent such an action, on the part of a witch, if we had not this idiom of physics to rely on? The mystical task, let us consider, is to bring the malign power of one individual to bear, rapidly and over a distance, on another. Lacking the concept of wavelike force, we shall I think find ourselves conceiving that the witch comes directly into contact with his victim; and an immediate recourse is to imagine, with the inspiration of observing airborne species, that the witch flies. Sometimes this action is represented as a detached head (in certain cases with entrails hanging from it) flying through the air; examples come from as far afield as Borneo and Chile. But this image reminds us of Durer’s angels, who appear as cherubic heads equipped with wings, and these lead to a far more general significance in the imagined power of flight. It is a worldwide notion that persons of supernormal status are capable of flight: not only gods and angels and other spirits, but also men who are associated with them or who draw upon their powers; so that saints and shamans and yogis are all credited with mystical flight and sometimes (in Christian tradition, for instance) with physical levitation. Hocart has suggested furthermore that the common practice of carrying kings and other potentates so that they are never in contact with the earth is a ceremonial surrogate for the imagined gift of flight. Once more therefore we have to conclude that a prominent and universal component in the image of the witch has nothing exclusively to do with witches, but is merely one instance of a widely exploited recourse of the imagination in representing persons who possess abnormal powers.

Lastly, let me just mention one more component. I need only mention it, since it is too naturalistic a matter to be very interesting. This is the feature that a witch is supposed to emit a fiery trail or glow as he travels through the air. I am afraid all this means is that marsh gas and similar nocturnal illuminations are very commonly found in the world, and that peoples who imagine witches to fly at night are highly likely to associate their trajectories with the passage of these eerie lights. If the Trobrianders hold that the glow is emitted from the witch’s anus, this is just a dramatic touch added to the natural observation. Nevertheless, the component of the aerial light is a very common feature in the image of the witch, and it does reinforce the component of flight. It may not be symbolically so informative as other features, but it is still a contribution to the complex image of the witch. In some cases, moreover, it is a manifestation of fire, and we need no reminder of the fundamental importance everywhere of this lively symbol.

So much, and by necessity rather superficially, for certain factors that characteristically constitute the image: opposition, inversion, darkness, color, animals, flight, and nocturnal lights. I am not asserting that these features alone compose the image, or that all of them must be present, or that witches are necessarily represented by this image. The notion that I am putting forward is in fact doubly polythetic: first, in that disparate phenomena are grouped together under any one factor; second, that the disparate factors are variously and sporadically combined into the image of the witch.

In the first lecture I stressed that the primary factors are heterogeneous and that they are independent of the will. We now see these attributes manifested with special force in the synthetic image of the witch. The operative factors include the relational abstraction of opposition, the spatial metaphor of inversion, and a variety of observable phenomena of nature. The complex that they form has a global distribution and constancy that make it out of the question that each instance should be the result of individual or even traditional invention. As a collective representation, the complex is autonomous, and men have merely altered its particulars according to their circumstances—according to whether they themselves had white or black skins, or whether they had hyenas or the Japanese fox in their environment. It is as though the complex in itself, and not only the several factors out of which it is synthesized, were also primary.

The factors constitute a steady image of the witch, but they are also to be found in the constitution of other images and institutions. For example, the relations and phenomena that we have just surveyed can be combined, with different semantic values according to the appositeness of the materials, into the image of a saint or a shaman. If mystical action at a distance is taken to be definitive, then this too is found in blessing and prayer and spiritual healing. So it is as though men have at their disposal, innately, a limited repertory of imaginative resources, and these, the primary factors, are differentially synthesized into distinct complexes representing disparate social concerns. The factors accrete to the concern, and their meaning in combination corresponds to the concern.

In the case of witchcraft, how are we to account for the synthesis, that is, for the fact that certain primary factors characteristically combine into a recognizable and comprehensible image? I am not at all sure that in principle we can hope to do so. We may in one case or another be able to see how it is that one feature or another is present or has certain value (white ashes instead of black soot, incest instead of cannibalism), but these diacritical variations are only secondary to the process of synthesis. Even where there is a naturalistic or demonstrable ground to the presence of a particular feature, still we cannot give this a causal expression, and the local explanation does not explain the synthesis. For example, if we assume that there is a patent similarity between the idea of a flying witch and the observation of a will-o’-the-wisp, the phenomenon does not explain the idea, since the idea exists where there is no marsh gas; and the adventitious occurrence of the nocturnal glow, apposite though it may be, does nothing to explain the polythetic synthesis with other factors.

Similar conclusions follow even when a causal connection can be shown to be possible between a feature of the image and some material agent. For example, it has been known for decades that European witches used an ointment containing certain chemical substances, and that they rubbed this into their bodies, apparently into the vaginal membranes and the legs, through which the substances could have entered the blood stream. These substances included hemlock, henbane, belladonna, and aconite, drugs which could conduce to hallucinations, including the sensation of flight. But this does not explain why witches are imagined to fly in societies that do not employ the drugs. Where the image is supported by drug-induced hallucinations, it does not explain the synthesis of the other factors that make up the total image. And in any case the hallucinations do not in themselves explain the factor of flight as an attribute of spirits and saints, shamans and heroes.

If we say then merely that the primary factors accrete around a social concern, what is the concern that underlies the image of the witch? It is articulated as a fear of other men, who can do evil by secret and invisible means. Certainly men have given one another ample occasion to fear other men, but this does not explain the complexity of the image of the witch or its characteristic components. Men everywhere are disposed to fear some other men, yet they adopt many ways apart from the institution of witchcraft in expressing their apprehension and in taking precautions against its object. Actually, it may not be right to say that the factors accrete around the concern, as though the concern came first and the image afterwards. All we know empirically is that the concern, in the form of a conventional unease about certain categories of persons, and the image, as a complex construction of the imagination given also in part by tradition, are found together. The true force of the metaphor of accretion is perhaps that the stated concern, namely, about malign mystical action at a distance, is the most constant definitive feature of the institution and also the theme that imparts a fit meaning to such features as trailing intestines or the invisible draining of the strength of the victims.

Moreover, to concentrate on the social concern, which people are conscious of and can express, leads to the view that witchcraft is a cognitive institution. The burden of my argument however has been that witchcraft is a complex product of the imagination and that it provides evidence of certain proclivities of the imagination. Certainly the institution has a cognitive aspect, for men reflect on their experience and they explain certain untoward events by witchcraft, but the form of the experience and the interpretation of the events are not the results of independent deliberation. If there is a “reality” to the idea of witchcraft, it is to be sought in the grounds and occasions of the image of the witch. Since the components of the image can be seen as products of factors that compose other complex images, the problematical focus is the process of synthesis that combines the components into the characteristic image. This process is not necessarily a response to experience, and we may not presume that it is otherwise determined by social facts. The synthesis may be, as I have suggested earlier, just as “primary” as are the factors that it integrates. In other words, the image of the witch is autonomous and can be conceived as an archetype of the unconscious imagination.

The concept of an archetype, in the sense of a primordial mental image, has suffered from obscurity and also from obscurantism, and it has fallen into discredit in the eyes of those for whom the work of Jung is not entirely creditable. In adverting to this concept, however, I am not trying to darken or diminish understanding but to advance analysis; and in formulating the results of comparative study by means of the word archetype I am not subscribing to Jung’s psychology as a system. In undertaking the present investigation, I was not committed in advance to a Jungian view of the unconscious. My question was, how is social anthropology possible? The outcome, so far as the institution of witchcraft is concerned, is the notion of a psychic constant in the form of an autonomous image to which the human mind is naturally predisposed. Archetype happens to be the right word to denote this complex product of the unconscious.

There is in particular a distinction of method to be made. Typically, I take it, Jung’s ideas on archetypes were derived from correspondences between images in the dreams or fantasies of patients and in alchemical manuscripts and hermeneutic sources. For the social anthropologist, on the other hand, the major sources are the reports of world ethnography, an incomparably vaster field of evidence and one that the comparativist nevertheless manages to comprehend as the widest testimony to the collective forms of human experience. As for the present investigation, moreover, there is a marked contrast between Jung’s studies and the approach that I have outlined. I think I am right in saying that Jung’s analyses of what he regarded as archetypes were not only culturally limited but were also highly particularistic, and even idiosyncratic, in that each archetype was interpreted in terms of its own proper significance. My own method, though, has been to investigate the archetype of the witch by analysis into primary factors that are not exelusively associated with witchcraft. If I have on occasion referred to the products of these factors (particular colors, sides, sounds, textures) as semantic units, this was with the explicit gloss that the units could carry variable and even opposite meanings from one cultural context to another. In postulating a repertory of primary factors, I have presented a possibility that so far as I know Jung was not concerned with; and by concentrating on the principles of synthesis that constitute primary factors into archetypes and that discriminate one archetype from another, I have taken what appears to be a quite different direction of research.

Nevertheless, both of these approaches to the topic of archetypes are exercises in depth psychology. If the evidence of the social anthropologist consists in the first place of social facts and collective representations, this affords him only an advantage; for as social facts these representations have a relative autonomy, a generality, and a coercive force that objectify them for the purposes of analysis and comparison. Durkheim was quite right, I think, in propounding the notion of an entirely formal psychology that would be a sort of common ground for individual psychology and for sociology, and in suggesting that the comparative study of collective representations might seek the “laws of collective ideation” or of “social mentality” in general. This enterprise would lead, hypothetically, to the establishment of propositions that were valid equally for the individual and for the collective, with an exclusive substantive locus in neither.

The essential, if we are to say anything fundamental about human nature, is to find the common term of individual and collective. I have spoken here as though “the imagination” were in effect such a common term; and in analyzing the image of the witch I have alluded to properties that can be common both to individual consciousnesses and to collective representations. But these properties are such as, empirically, must inhere in some locus or entity or system of phenomena that is common to both types of manifestation of the imagination. According to received ideas, which I have no reason to question in this instance, this can reside only in the human organism; and the plainest and most economical inference is that this means the brain. The present state of knowledge concerning this organ, the most complicated natural system known, is, I gather, still relatively superficial and partial; and I am not relying on neurophysiology when I say that I interpret the primary factors and the modes of synthesis as spontaneous manifestations of properties of the brain. The strength of this accommodating hypothesis is that it accounts for the global distribution of the characteristic image of the witch, and at the same time for the constancy of the factors as apparently innate predispositions. It is on these premises that I find it convenient to regard the archetypes as vectors of consciousness. There is nothing in principle that is particularly difficult to accept in this metaphor. All it necessarily implies, in the first place, is that the brain responds to percepts differentially; for instance, that it responds especially to red rather than to other hues, or to percussion as distinct from more mellifluous sounds. There is some experimental support for this assumption, in addition to the weight of ethnographic indications. In some cases, in connection with certain primary factors or archetypes (such as percussion, the half man), it is moreover possible to suggest empirical grounds for their presence or their prominence.

Among the heterogeneous primary factors there falls a line that divides them into two types: those that are abstract (such as the relation of binary opposition) and those that are perceptual (such as color, texture). There is an inviting correspondence here with the contrasted functions of the cerebral hemispheres. It is tempting to link the respective factors causally to these functions: the abstract factors to the left hemisphere, the perceptual ones to the right. A plausible hypothesis is that the comparatively restricted number of abstract factors has to do with the analytical function of the left hemisphere, and the more extensive range of the perceptual factors with the function of the right. A conceivable process is that these contrasted cerebral functions combine in an imaginative tropism, a synthetic response to natural foci of attraction among phenomena, whether social or physical, and that the product is the archetype.

With these considerations we are far beyond the present limits of proof, but I am not suggesting that it is essential to fall in with speculations of this kind in order to assent to to the method that I have proposed. I suggest only that the dichotomy among the primary factors, and their synthesis into the archetype, are consistent with current opinion about the lateral functions of the brain; and that this ultimate locus is consistent with the unconscious generation of the archetypes and with the likelihood that these complex images are the products of genetically inherited predispositions.

What may lie at a deeper level of probing into the imaginative operations of the brain, in the form of neuroelectrical events, is a tantalizing sequel among further questions that propose themselves, if hardly one that a comparativist can have an opinion about. I mention it, however, in order to put us on our guard against inappropriate preconceptions when, as is almost inevitable, we do speculate on the cerebral grounds to collective representations. I am alluding to Wittgenstein’s (1967: §608–09) salutary comments:

No supposition seems to me more natural than that there is no process in the brain correlated with associating or with thinking; so that it would be impossible to read off thought processes from brain-processes. I mean this: if I talk or write there is, I assume, a system of impulses going out from my brain and correlated with my spoken or written thoughts. But why should the system continue further in the direction of the centre? … It is … perfectly possible that certain psychological phenomena cannot be investigated physiologically, because physiologically nothing corresponds to them.

For the more ascertainable present, let me conclude this lecture by resuming some of the main points. We have moved from the sociology of mystical crime, by way of induction from the global comparison of social facts, to certain apparent properties of the human brain. These properties are seen as determinants in the constitution of collective representations, not only of witchcraft but in effect universally and with regard to numerous other institutions. As such, they act as initial limits to the imagination, rather as do certain logical constraints on the forms of discursive reasoning. This conclusion runs contrary to the received idea of romantic individualism that, however constrained we may be by the dictates of society and the puny capacities of the human organism, we have an unfettered liberty in the exercise of the imagination. At the level of collective representations, our investigation of the image of the witch tends to reduce the scope and assurance of this comfortable assumption.

Select bibliography

Baal, J. van. 1977. “Review of Le symbolisme en général, by Dan Sperber.” Bijdragen tot de Taal-, Land- en Volkenkunde 133: 163–65.

Beidelman, T. O. 1963. “Witchcraft in Ukaguru.” In Witchcraft and sorcery in East Africa, edited by John Middleton and E. H. Winter, 57–98. London: Routledge and Kegan Paul.

Benveniste, Emile. 1966. Problèmes de linguistique générale. Paris: Presses Universitaires de France.

Clark, Kenneth. 1960. The nude: A study of ideal art. Harmondsworth, Middlesex: Penguin Books.

Durkheim, Emile. 1901. Les règles de la méthode sociologique, 2nd edition. Paris.

Durkheim, Emile and Marcel Mauss. (1903) 1963. Primitive classification. Translated from the French and edited with an introduction by Rodney Needham. Chicago: University of Chicago Press.

Eliade, Mircea. 1964. Shamanism: Archaic techniques of ecstasy. Bollingen Series, 76. New York: Pantheon Books.

Freud, Sigmund. 1973. New introductory lectures on psychoanalysis. The Pelican Freud Library, vol. 2. Harmondsworth, Middlesex: Penguin Books.

Hampshire, Stuart. 1959. Thought and action. London: Chatto and Windus.

Hocart, A. M. 1970. The life-giving myth. Edited by Lord Raglan. Second impression edited with a foreword by Rodney Needham. London: Methuen.

Jacobi, Jolande. 1959. Complex/archetype/symbol in the psychology of C. G. ]ung. Translated by Ralph Manheim. London: Routledge and Kegan Paul.

Lafitau, Joseph François. 1724. Moeurs des sauvages amériquains comparées aux moeurs des premiers temps. 2 vols. Paris.

Lloyd, Geoffrey. 1966. Polarity and analogy: Two types of argumentation in early Greek thought. Cambridge: Cambridge University Press.

Mayer, Philip. (1954) 1970. “Witches. Inaugural lecture, Rhodes University, Grahamstown, South Africa.” Reprinted in Witchcraft and sorcery, edited by Max Marwick, 45–64. Harmondsworth, Middlesex: Penguin Books.

Needham, Rodney. 1972. Belief, language, and experience. Oxford: Basil Blackwell; Chicago: University of Chicago Press.

———. 1973, ed. Right & left: Essays on dual symbolic classification. Chicago: University of Chicago Press.

———. 1974. Remarks and inventions: Skeptical essays about kinship. London: Tavistock; New York: Barnes & Noble.

———. 1975. “Polythetic classification.” Man 10: 349–69.

———. 1976. “Skulls and causality.” Man 11: 71–88.

———. 1978. Essential perplexities: Inaugural lecture, University of Oxford. Oxford: Clarendon Press.

———. 1978. Symbolic classification. Santa Monica, CA: Goodyear Publishing Co.

Rycroft, Charles. 1975. “Freud and the imagination.” New York Review April 3, 26–30.

Waismann, Friedrich. 1968. How I see philosophy. Edited by R. Harre. London: Macmillan.

Whitehead, Alfred North. 1927. Symbolism: Its meaning and effect. Cambridge: Cambridge University Press.

Wittgenstein, Ludwig. 1967. Zettel. Oxford: Basil Blackwell.

———. 1969. The blue and brown books. Oxford: Basil Blackwell.


Rodney NEEDHAM (1923–2006) was a British social anthropologist and Professor of Social Anthropology at Oxford Unviersity. An ethnographer of the Penan of Borneo and the Siwang of Malaysia, Needham was also a major figure responsible for the popularization of French structuralism in British social anthropology, having translated many of the theoretical movement’s foundational texts into English. A prolific writer, Needham’s theoretical works on symbolism, causality, belief, kinship, ritual, classification, and others are now classic and remain influential among new generations of anthropologists. Among his many publications are Belief, language and experience (1972, University of Chicago Press), Primordial characters (1978, University of Virginia Press), Reconnaissances (1980, University of Toronto Press), and Exemplars (1985, University of California Press).


Publisher’s note: This lecture is a reprint of Needham, Rodney. 1978. “Synthetic images.” In Primordial characters, 23–50. Charlottesville, VA: University of Virginia Press. We are grateful to Tristan Needham for extending permission to publish this work. We remind the reader that we retain the style of the original publication.