PART I

DEEP PERSPECTIVES

1

THE GLOBAL CONTEXT

THE SIGNIFICANCE OF a quest for a worldwide context for the early modern European witch trials is that it can determine what, if anything, is specifically European about those trials, and about Europe’s images of what a witch was supposed to be. It may answer the question of whether what happened in early modern Europe was something unusual, in a global setting, or simply the most dramatic regional expression of something which human beings have done in most places at most times. To embark on such a course, it is essential to establish from the beginning precisely what is being sought, and what the characteristics of the figure known in English as the witch are supposed to be. The basic usage chosen earlier, of an alleged worker of destructive magic, establishes the first and most important characteristic credited to the people who were prosecuted in the early modern European witch trials: that they represented a direct threat to their fellow humans. In very many cases it was believed that they employed non-physical, and uncanny, means to cause misfortune or injury to other humans, and very often they were accused, in addition or instead, of striking at the religious and moral underpinnings of their society. Four more distinguishing features were embodied in the figure of the witch as defined by those trials and the ideology on which they were based. The first of these four features was that such a person worked to harm neighbours or kin rather than strangers, and so was an internal threat to a community. The second was that the appearance of a witch was not an isolated and unique event. Witches were expected to work within a tradition, and to use techniques and resources handed down within that tradition, acquiring them by inheritance, initiation or the spontaneous manifestation of the particular powers to which they were connected. The third component of the European stereotype of the witch was that such a person was accorded general social hostility, of a very strong kind. The magical techniques allegedly employed by witches were never officially regarded as a legitimate means of pursuing feuds or rivalries. They were always treated with public, and usually with spontaneous, anger and horror, and often associated with a general hatred of humanity and society and with an alliance made by the witch with malignant superhuman powers loose in the cosmos: in the European case, famously, by a pact with the Christian Devil. Finally, it was generally agreed that witches could and should be resisted, most commonly by forcing or persuading them to lift their curses; or by making a direct physical attack on them to kill or wound them; or by prosecuting them at law, with a view to breaking their power by a punishment which could extend to having them legally put to death.

Few, if any, experts in the early modern European witch trials will find those five definitive components of the witch figure unacceptable; indeed, if there is anything problematic about them it is likely to be their banality. None the less, they do provide a more precise checklist of characteristics than has been employed hitherto, suitable for a comparative study covering the planet. The result of such a study is in one sense a foregone conclusion, for scholars have spoken for centuries of finding very similar figures to that of the European witch in all parts of the world, and indeed they have employed the English word ‘witch’ for those figures. Again, however, it may be suggested that more care can be taken in making the necessary comparisons, and a larger sample of material can be employed for them. Moreover, it is by no means certain that most specialists in the study of the European trials would consider such an enterprise to have any value. The story of the relationship between experts in those trials, and those in what has been called witchcraft in other parts of the world is already a long and sometimes fraught one, with a large component of estrangement. That story must be considered before this latest contribution to it can be attempted.

Historians, Anthropologists and Witchcraft: A Friendship Gone Wrong?¹

In the 1960s a global approach to the study of the witch figure was virtually the norm among British scholars, largely because most of the research published on witchcraft during the mid-twentieth century was by anthropologists working in extra-European societies, above all in sub-Saharan Africa. As British experts in European witch trials emerged at the end of the decade, they not only usually employed anthropological data to interpret European evidence, but acknowledged that their interest in the subject had been inspired partly by the reports coming from overseas.²Anthropologists reciprocated with gestures of partnership, so that their conferences and collections of essays on witchcraft routinely included papers from experts in European history.³ When Rodney Needham wrote his study of the witch as a human archetype in 1978, he used data from both African and European sources, declaring that a comparative approach was essential to the exercise.⁴ By then, however, this view was already on the wane. It had not convinced American historians, who claimed that the ‘primitive’ social groups of Africa bore little resemblance to the more complex cultures and societies of early modern Europe.⁵ Such views also affected some American anthropologists, who were already warning before the end of the 1960s that the term ‘witchcraft’ was being used as a label for phenomena that differed radically between societies.⁶ Even in Britain, at the height of collaboration between history and anthropology in the field, prominent members of both disciplines urged that such exchanges should be carried on with caution.⁷

What really doomed them was a shift within anthropology itself, as the dissolution of the European colonial empires produced a reaction against the traditional framework of the discipline, now perceived as a handmaiden to imperialism. This reaction embodied hostility both to the imposition of European terms and concepts on studies of other societies and the offering of comparisons between those societies which the imposition of the terms concerned made easier. Fashion was turning to close analyses of particular communities, as unique entities, carried on as much within their own linguistic and mental models as possible (which of course also gave added value and power to the individual scholars who claimed a privileged knowledge of those communities). This self-consciously ‘new anthropology’ was reaching British universities by the early 1970s.⁸In 1975 an American exponent of it, Hildred Geertz, published stringent criticisms of the British historian who had emerged as the most distinguished practitioner of the application of anthropological concepts to his own nation’s past, Keith Thomas. She accused him of having adopted categories constructed by the British from the eighteenth century onwards, as cultural weapons to be deployed against other peoples; and questioned in general whether cultural particulars could be formed into general concepts and compared across time periods and continents. She did not actually question the value of scholarly categories in themselves, only arguing for more care and criticism in the use of them; but Thomas made the debate an occasion to suggest that Western historians now needed to back off from comparisons with extra-European cultures and concentrate on their own societies, for which their terminology was native and so well suited.⁹

In doing so, he explicitly recognized the change in anthropology, acknowledging that its practitioners had become wary of using Western concepts to understand non-Western cultures and preferred to employ those of the people whom they were studying. He accepted that they now desired to reconstruct different cultural systems in their entirety rather than employing terms unthinkingly used by historians, such as ‘witchcraft’, ‘belief’ and ‘magic’, to make comparisons between them. In case any of his compatriots missed the point, it was being hammered home between 1973 and 1976 by an anthropologist based in Thomas’s university, Oxford, called Malcolm Crick, and with specific application to witchcraft. Crick called for the concept of the witch to be ‘dissolved into a larger framework of reference’, by relating the figures whom English-speakers called witches to others who embodied uncanny power of different kinds within a given society. He also asserted that conceptual categories varied so much between cultures that ‘witchcraft’ could not be treated as a general topic at all, and warned historians off ethnographic material, proclaiming (without actually demonstrating) that ‘English witchcraft is not like the phenomena so labelled in other cultures’.¹⁰ Historians of European witchcraft generally internalized this message, and the ever-increasing number of studies of early modern witch beliefs and trials which appeared from the late 1970s onwards limited themselves to cross-cultural studies within the European world, sometimes extended to European colonists overseas. When a very occasional scholar did try to compare European and African material, it was never somebody prominent in witchcraft studies or one who continued to publish on them.¹¹

In 1989 a review article uncompromisingly entitled ‘History without Anthropology’ concluded that anthropologists had very effectively deterred historians from taking any further interest in their work with reference to the subject of witchcraft.¹² The irony of this was that during the same period the practitioners of anthropology themselves were starting to change their minds again. In an important sense they had never abandoned the comparative approach and the Western terminology that many of them had criticized in the 1970s, because even those who described the magical practices of non-European peoples using native terms still put English expressions such as ‘witchcraft’ and ‘magic’ into their titles. For the most part they continued to put them into their introductions as well, and some made such words the framework within which the local study was introduced: they retained their value as an international semantic currency for English-speakers. By the 1990s some of the most distinguished anthropologists were starting to become more actively interested in a new collaboration between their discipline and historians of Europe. One described the fixation of her discipline on holistic fieldwork in specific small-scale societies using participant observation as an ‘academic narrowness’, which had cut it off from the history of religion.¹³ Another used both modern African and early modern European data to compare attitudes to witchcraft and leprosy as strategies of rejection, and to consider the phenomenon of witch-hunting.¹⁴ A third suggested that early modern images of witchcraft were closely related to African beliefs. In doing so she explicitly attacked the earlier assertions that the term ‘witchcraft’ lacked any validity in cross-cultural comparisons: indeed, she restated such comparisons as a duty of her discipline.¹⁵ In 1995 a British sociologist, Andrew Sanders, made a parallel challenge to those assertions, and published a worldwide survey of occurrences of the witch figure, using both historic European and modern ethnographic records.¹⁶ The most significant development in this regard was among Africanists, who called for a renewed emphasis on cross-cultural comparison in witchcraft studies. It was propelled by one of the most distressing and – to many – surprising characteristics of post-colonial states in the continent, an intensification of fear of witchcraft and attacks on suspected witches as one response to the process of modernization after independence: it will be discussed below. Anthropologists who studied this phenomenon found themselves needing to dissuade fellow Westerners from attributing the persistence of a belief in witchcraft in Africa to any inherent disposition to ‘superstition’ or ‘backwardness’ on the part of its peoples. Such a strategy called for a new emphasis on the prevalence of such beliefs across the globe, including in the relatively recent European past, and a return to a comparative method; and direct calls for that were being made by prominent Africanists by the mid-1990s.¹⁷ Typical of them was an influential study of Cameroon by Peter Geschiere, who concluded that ‘these notions, now translated throughout Africa as “witchcraft”, reflect a struggle with problems common to all human societies’. He invited anthropologists to study research into the European trials, and termed their recent neglect of this ‘even more disconcerting’ than the loss of interest by historians of Europe in African parallels. Rounding upon experts in early modern Europe who had claimed that modern African societies were totally dissimilar to those which were their own focus of study, he argued that, especially with its ruling elites of colonial European administrators and settlers, early twentieth-century Africa had been as socially and culturally complex as sixteenth-century Europe.¹⁸ By 2001 the editors of a major collection of essays on African witchcraft could introduce it by warning scholars not to restrict the study of witch beliefs to ‘any one region of the world or to any one historical period’.¹⁹ In urban centres of modern Africa, a multicultural perspective had become essential in any case: the image of witchcraft in the Soweto suburb of Johannesburg, for example, was by the 1990s a blend of ideas drawn from different native groups with some brought by Dutch and English settlers and based on the early modern European stereotype.²⁰ A rapprochement between historians and anthropologists over the issue was, however, an extremely difficult enterprise.

Despite the call made by some for a return of the comparative method, few Africanists in practice paid attention to studies of the witch figure anywhere else in the world, or in time. Those who did attempt to cite early modern European material often seemed unaware of anything published on it after the early 1970s: the burgeoning of research that had occurred since, internationally, and taking ever more sophisticated forms, had passed them by completely. As for historians of witchcraft, almost all of them had stopped reading anthropology on the assumption that they had been discouraged from doing so by its practitioners. To resume an engagement with it after more than two decades would require a large amount of additional work of unproven value, when they were already achieving apparently impressive results as a consequence of relationships with a range of other disciplines. It was quite plain by the 1990s why Africanists concerned with witchcraft might profit from a fresh engagement with European comparisons, but not even the anthropologists themselves were making a clear argument for why historians of Europe would benefit from the transaction. A concealed irony in the situation was that the newly developed cultural history of the 1980s and 1990s, which had a profound influence upon the study of European witchcraft, was itself ultimately derived partly from anthropology; but reached most historians at one or two removes from it.

Unsurprisingly, therefore, historians have largely ignored the opportunity for a new dialogue, and anthropologists have largely ceased to offer it. In the early 2000s the present author published two essays that drew attention to it and suggested specific advantages to experts in early modern Europe from such a comparative exercise.²¹ These have, however, been more cited than heeded. In 2004 one of the leading experts in German witch trials, Wolfgang Behringer, produced a heavyweight volume entitled Witches and Witch-Hunts: A Global History.²² It was in practice a detailed and impressive history of the European witch-hunts bracketed between two swift surveys of beliefs and prosecutions concerning witchcraft across the rest of the world. The first of these made the point that what happened in Europe was part of a global pattern, and the second of them proved that a continuation of witch-hunting was not merely a problem in contemporary Africa but in many other parts of the planet. This was a precise and fruitful application of the comparative method; but the present book seems to be the first to follow up on its achievement. The only general effect of the growing awareness of a new potential for collaboration between anthropologists and historians of witchcraft has been an apparent disappearance on both sides of assertions that such collaboration is itself inherently undesirable; which is some kind of progress. A few anthropologists have continued to make use of European material, but historians of Europe usually fail to repay the compliment.²³ A refinement of methodology is needed if any advance is indeed to be made on earlier attempts to collaborate.

Andrew Sanders was interested chiefly in the relationship between the witch figure and the pursuit of power through competitive social relationships in different parts of the world. As a sociologist, he was concerned more with the implications and consequences of a belief in witchcraft for human societies that held one than with the nature of that belief itself. Wolfgang Behringer’s aim was to show that in most parts of the world human beings have been inclined to attribute seemingly uncanny misfortune to evil magic worked by their fellows, and to illustrate the lethal consequences which such an inclination has often produced (and continues to produce). My own essays attempted to establish a coherent global model for the witch figure, with sustained cross-cultural characteristics, and proposed one based on the five characteristics delineated above as fundamental for the European concept of that figure. What will be attempted now is a more systematic application of the cross-cultural method, across the planet, checking off those characteristics one by one. It utilizes studies of beliefs concerning witches in a total of three hundred extra-European societies made between 1890 and 2013: 170 in sub-Saharan Africa; six in North Africa and the Middle East; thirty-seven in South Asia from India to China and Indonesia; thirty-nine in Australia, Polynesia and Melanesia, including New Guinea; forty-one in North America (including Greenland and the Caribbean); and seven in South America.²⁴ The predominance of Africa in the sample reflects the amount of work that has been done there by anthropologists but also the resources available to a researcher based in the United Kingdom as so many of these anthropologists were British.²⁵ There is enough data from the rest of the world, however, to provide comparison with the African material, and that exercise may now be undertaken point by point with respect to the characteristics of a European witch listed above. The societies studied are those on which anthropology published in English has chosen or been able to concentrate, being generally relatively simple and small, and consisting of tribal units. There is a dearth of information available from larger, state-based, social and political structures such as those of China and Japan, which to some extent will be made up by a sustained examination of ancient states in Europe and the Near and Middle East in the next chapter. None the less, the sample from smaller ethnic units, across the world, is large enough for a comparative exercise to promise some general insights.

Characteristic One: A Witch Causes Harm by Uncanny Means

There is little doubt that in every inhabited continent of the world, the majority of recorded human societies have believed in, and feared, an ability by some individuals to cause misfortune and injury to others by non-physical and uncanny (‘magical’) means: this has been the single most striking lesson of anthropological fieldwork and the writing of extra-European history. One prominent historian of early modern Europe, Robin Briggs, has in fact proposed that a fear of witchcraft might be inherent in humanity: ‘a psychic potential we cannot help carrying around within ourselves as part of our long-term inheritance’.²⁶ Speaking from anthropology, Peter Geschiere proposed that ‘notions, now translated throughout Africa as “witchcraft”, reflect a struggle with problems common to all human societies’.²⁷ What is valuable about these insights is that they testify to the general truth that human beings traditionally have great trouble in coping with the concept of random chance. People tend on the whole to want to assign occurrences of remarkable good or bad luck to agency, either human or superhuman. It is important to emphasize, however, that malevolent humans have been only one kind of agent to whom such causation has been attributed: the others include deities, non-human spirits that inhabit the terrestrial world, or the spirits of dead human ancestors. All of these, if offended by the actions of individual people, or if inherently hostile to the human race, could inflict death, sickness or other serious misfortunes. Wherever they appear, these alternative beliefs either limit or exclude a tendency to attribute suffering to witchcraft.

In addition, many societies have believed that certain humans have the power to blight others without intention to do so, and often without knowledge of having done so. This is achieved by unwittingly investing a form of words or a look with destructive power: in the case of malign sight, this trait has become generally known to English-speakers as ‘the evil eye’. Belief in it tends to have a dampening effect on a fear of witches wherever it is found, which is mainly in most of the Middle East and North Africa, from Morocco to Iran, with outliers in parts of Europe and India. This is because it is thought to be part of the possessing person’s organic constitution. As such, it is wholly compatible with witchcraft if the person concerned triggers it consciously and deliberately to do harm, as some are thought to do across its range. A majority of those who embody this malign power, however, are believed to do so wholly innately and involuntarily, so that they cannot in justice be held personally responsible for its effects. Protection and remedies for it mainly take the form of counter-magic, including the wearing of amulets, charms and talismans, the reciting of prayers and incantations, the making of sacrifices and pilgrimages and carrying out of exorcisms, and the avoidance or placation of the person who is locally presumed to possess it. Across the range in which it is an important component of belief, it is used to explain precisely the sort of uncanny misfortunes that are blamed elsewhere on witchcraft.²⁸

Alternative explanations for misfortune that rule out or marginalize witchcraft are found across most of the world. Before modern times, the largest witch-free area on the planet was probably Siberia, which spans a third of the northern hemisphere; a consideration of it will play a major part in Chapter Three of this book. Elsewhere in the world, societies that do not believe in witchcraft, or do not believe that it should be taken very seriously, are seldom found in compact concentrations but scattered between peoples who fear witches intensely. Although rarer than groups with a significant fear of witchcraft, they are present in most continents: the Andaman Islanders of the Indian Ocean, the Korongo of the Sudan, the Tallensi of northern Ghana, the Gurage of Ethiopia, the Mbuti of the Congo basin, the Fijians of the Pacific, the hill tribes of Uttar Pradesh, the Slave and Sekani Indians of north-west Canada, and the Ngaing, Mae Enga, Manus and Daribi of New Guinea are all examples.²⁹ The Ndembu, in Zambia, attributed misfortune to angry ancestral spirits, but the latter were seen as aroused by malevolent humans, in effect making the spirits the agents of witches. However, it was the spirits who were propitiated, by ritual, and so the witches rendered harmless and ignored.³⁰

Among peoples who do have a concept of witchcraft, the intensity with which it is feared can vary greatly, even within the same region or state. Among the ethnic groups contained within the modern state of Cameroon are the Banyang, the Bamileke and the Bakweri. The first of those believed in witches but very rarely accused anybody of being one. Those afflicted by hostile magic were believed to have brought their misfortune on themselves.³¹ The second took witchcraft seriously and made great efforts to detect its practitioners. The latter, however, were not held responsible for their actions, and were thought to lose their powers automatically on being publically exposed.³² The third people feared witchcraft intensely, hunted down its presumed operators, and believed that they remained dangerous and malevolent even when identified, so that they needed to be punished directly in proportion to the harm they were thought to have caused.³³ In neighbouring Nigeria, a clutch of tribal societies shared very similar theoretical beliefs about the existence of witches, but in practice the Ekoi dreaded them, the Ibibio and Ijo feared them moderately, and the Ibo and Yakö took little notice of them.³⁴ Likewise, a survey made in 1985 of a sample of well-studied peoples in the Melanesian archipelago found that two of them did not believe that humans used malevolent magic; five thought it a legitimate monopoly of hereditary leaders and used by them productively in order to keep order and conduct warfare; twenty-three believed that such leaders could use it but it was not respectable of them to do so; five conceived of it as a covert weapon of the oppressed, employed against unpopular leaders; eleven identified it as a means by which ordinary members of the community secretly harmed each other, but generally managed in practice to contain the tensions provoked by fears of it; and six had the same belief but were badly disrupted by the suspicions that resulted.³⁵ Among a single people, the intensity with which witchcraft was feared could vary according to the kind of settlement in which people lived. The Maya of the Yucatan Peninsula of Mexico all hated witches equally in theory during the early twentieth century, but those in villages were rarely inclined to suspect anybody of being one, while the tension was much greater in towns: in the district capital of Dzitas during the 1930s, 10 per cent of the adult population were thought to have been either perpetrators or victims of witchcraft.³⁶

The identification of a belief in witchcraft among extra-European peoples, by a European scholar, may often involve the extraction of one element from a range of native concepts of magic and of kinds of magician. The Wimbum of north-west Cameroon used three terms for occult knowledge: bfiu, the harmless employment of arcane powers for self-protection; brii, occult power malevolently used, but sometimes merely as a prank; and tfu, an inborn magical force operated under cover of darkness which could be used for both good and evil ends. Witchcraft in the European sense could embrace some forms of both brii and tfu, but the Wimbum also believed in a special strain of the latter, tfu yibi, which consisted of killing other humans magically in order to eat their flesh, and those who deployed that would correspond precisely to the early modern European witch figure.³⁷ The Nalumin of the mountains of south-eastern New Guinea distinguished biis from yakop. The former were people, mostly female, who killed others in uncanny ways, using invisible weapons while roaming in spirit-body, in order to eat the flesh of their victims in communal feasts. The latter was a technique, used mainly by women, which consisted of killing by burying personal leavings of the intended victim – food scraps, nail clippings and hair – with spells. The two methods, however, were believed to be combined at times by the same individual, and anybody thought to use either would correspond to the European figure of the witch; which is, indeed, how the anthropologist making the study interpreted them.³⁸ A final example of such equivalence is supplied by the Tlaxcala province of central Mexico, where the rural natives have feared tetlachiwike, people of both sexes who harmed by a touch or glance (the local equivalent to the evil eye or touch); tlawelpochime, people, mostly women, who sucked the blood of infants and so killed them, and caused harm to humans and their crops or livestock; tetzitazcs, men who could bring rain or hail; tetlachihuics, magicians who were believed to have powers which could be used for good or harm; and the nahuatl, a person of either sex who changed into animal shape to work harm or play tricks. The tetlachihuics were generally respected, and much employed for healing and other magical services, though sometimes murdered if they were thought to have used their abilities to kill: here as elsewhere in this book, the term ‘murder’ is used in its precise legal sense of unofficial and unsanctioned homicide. The shape-shifting nahuatl was believed to employ her or his powers of transformation to steal or rape as well as to inflict practical jokes, but did not inspire the fear and hatred that was accorded the child-murdering tlawelpochime. It was the latter to which Spanish-speaking natives gave the term bruja or brujo, meaning ‘witch’, and was regarded as inherently evil and associated with the Christian Devil.³⁹ Wim van Binsbergen, commenting on the complexities of belief in magic among Africans in 2001, could still conclude with regard to witchcraft that ‘the amazing point is not so much variation across the African continent, but convergence’.⁴⁰ Adam Ashforth, considering attitudes to destructive magic and its alleged perpetrators in the modern Soweto township near Johannesburg, decided that he had to use the terms ‘witchcraft’ and ‘witch’ because ‘there is no avoiding them’.⁴¹

Both conclusions are reproduced here, on the global scale. There are many cases of extra-European societies that have manifested, at least at the time of study, an endemic dread of witchcraft more intense than any recorded in Europe. The inhabitants of Dobu, an island group near the coast of New Guinea, had no concept of misfortune, blaming all mishaps on witches. Dobuans never went anywhere alone for fear of becoming more vulnerable to them.⁴² In the 1980s, among a small New Guinea tribe, the Gebusi, about 60 per cent of middle-aged men had killed at least one person – mostly within their own community – in revenge for presumed bewitchment.⁴³ The most notable scholar of the Tlingit of Alaska has declared that witchcraft dominated their lives, making the simplest words or action vulnerable to misinterpretation as a manifestation of it.⁴⁴ It was calculated that among the Kwahu of Ghana, 92 per cent of the population became at one point in their lives an accuser, presumed victim or suspected of witchcraft.⁴⁵ ‘Practically everyone’ in the Cochiti tribe of New Mexico was under suspicion at some time or another, and the elders had to winnow the accusations and decide which affected the good of the community and should be followed up officially.⁴⁶ In Burma in the 1970s, every village was presumed to harbour at least one woman who worked magic secretly to cause illness or death among her neighbours from personal spite.⁴⁷ Among some peoples found across Africa and Melanesia, and Australia’s Northern Territory, all deaths save those caused by murder or suicide, and most illnesses, were attributed to bewitchment.⁴⁸ Having said all this, however, most peoples who have believed in the witch figure seem to have regarded the risk factor, for most of the time, in the manner in which a modern car-driver treats the danger of a road accident.

There seems to be no functional explanation to account for the tendency of some human groups to believe in the existence of witches and some not to do so; those in both categories generally share similar societies, economies and cosmologies, and live in close proximity.⁴⁹ Likewise, there is no apparent general explanation for the varying intensity of fear of witchcraft between different peoples. In the 1960s, P.T.W. Baxter, studying those of East Africa, noted that wandering pastoralists in that region rarely accused each other of using bewitchment, even when they possessed a well-developed belief that people could do so.⁵⁰ This pattern does seem to hold true for nomads across the globe, perhaps because their mobile lifestyle and relatively small social units tend to reduce the potential for the personal conflicts that generate suspicions of witchcraft. On the other hand, not all static and deeply rooted agrarian societies have believed in witches, and not all of those that have done have feared them deeply. Moreover, even those that have taken witchcraft seriously have not done so with the same intensity at all times. Instead, witch-hunting, all over the world, has tended since records began to burgeon dramatically at particular times and die away or fall to a low level at others.

This phenomenon was confronted head-on in 2013 by the Dutch anthropologist Niek Koning, who developed a general theory of witchcraft beliefs which covered every time and place, uniting history and anthropology in a way recommended by others in his discipline since the 1990s. He suggested that small foraging bands tend to cope well with the social consequences of deceit and envy, but the adoption of agriculture much exacerbates them, leading to the development of witch-hunts. State formation, civilization and economic development abate these, and replace them with more collectivist forms of social paranoia; though demographic and economic crises can still rekindle a fear of witchcraft, as in early modern Europe.⁵¹

This broad-sweep approach is courageous and commendable, and does incorporate the truth that economic and social stress often result in intensified fear of witchcraft in societies that already possess it, as was the case at times in early modern Europe. Its determinism, however, fails to take account of too many exceptions to its rules: that small foraging bands like those of native Australia can have a pronounced belief in witchcraft; that some agrarian societies lack one; that highly developed urban civilizations such as those of ancient Rome and early modern Europe could hold big witch-hunts; and that the early modern European trials do not map simply and straightforwardly onto areas of most pronounced demographic and economic pressure, and indeed they commenced at a time of low population and relatively high incomes. All groups that do believe in witchcraft suspect certain kinds of person as being more likely to practise it than others, but the characteristics attributed to natural suspects differ greatly.

One major variable is age. In many societies, across the globe, accusations are directed mainly against the elderly, but in others they focus on the young and in many more, age is not a determining factor. It is normal for suspects to have passed puberty, because children are much more rarely involved in the social tensions between adults that generate accusations, and much less credited with power of any kind. None the less, among the Bangwa of Cameroon, children were frequently accused, and even babies could be thought culpable; and, as will be seen, there were and are other societies that associate witchcraft with the young.⁵²

Gender is another worldwide variable, witches being, at different places within each continent, viewed as essentially female, or essentially male, or of both sexes in different proportions and according to different roles. It is fairly common, also, for societies to manifest a discrepancy between the gender of their stereotypical witch and that of the people whom they actually accuse. Those making the accusations are, likewise, normally female or male or both, according to the conventions of the culture to which they belong. The same variety attaches to the social status and wealth of accusers and accused, witchcraft being viewed as a weapon employed by poor against rich, rich against poor, or between equals or competitors, or by any member of a community, according to the society concerned. There has been a common tendency across the world for suspicions to map onto economic and social tensions, so that quarrelsome or boastful individuals, or parvenus, within societies in which affability and modesty are regarded as prime virtues and economic mobility is limited, have often been considered either as obvious targets for witchcraft or as obvious practitioners of it; but several other categories of behaviour or person fall into both roles.

While being so various in such details, local concepts of the witch figure are also strongly rooted, and often seemingly impervious to the fact that neighbouring peoples could have very different ideas. There are three island groups off the north-east coast of New Guinea, close to and in regular communication with each other: Dobu, Trobriand and Fergusson. Their inhabitants are similar enough in physical, social and cultural respects to make them virtually one people. All fear witchcraft, but to Dobuans witches can be of either sex, though women are regarded as more dangerous; to Trobrianders they are mostly male; and to Fergussonians they are essentially female, and especially dangerous. An obvious question to be asked is whether the people in one of these societies find anything odd about the discrepancy between their beliefs and those of the other two. The answer seems to be completely negative, so that when Dobuans visit Trobriand, they are not afraid of the local women as witches but start to fear the men more, while the women of Fergusson frighten them even more than those at home.⁵³

Characteristic Two: A Witch is an Internal Threat to a Community

As suggested above, early modern Europeans believed that witches attacked neighbours or kin, or, exceptionally, they attacked elite figures within their own political unit such as an aristocrat or a king. Witches were therefore not imagined to be interested in harming strangers. This distinguishes witchcraft from the use of harmful magic as a weapon in conflicts between communities. Much feuding between traditional human societies, whether organized as tribes, clans or villages, has been believed by members to include a magical element, and such societies are disposed to blame misfortunes on the activities of magicians among their collective enemies. This belief is found in many parts of the world, but especially in three: the Amazon basin, Siberia, and Australia and Melanesia. It is especially prevalent in the last of these regions, although even there it is found interspersed with societies in which the threat of destructive magic is perceived to be mostly or entirely internal, as mentioned above.⁵⁴

Despite this broad dispersal of communities that expected magical danger from outside, they have been greatly outnumbered in the world by those who have feared it from within. Ralph Austen has commented that virtually all studies of rural African societies indicate that the efficacy of witchcraft is believed in them to increase in direct proportion to the intimacy between witch and victim.⁵⁵ Peter Geschiere has added that ‘in many respects, witchcraft is the dark side of kinship’, and Wim van Binsbergen that it is ‘everything which challenges the kinship order’.⁵⁶ This certainly seems true for much of Africa, although even there the degrees of kinship within which it is supposed to operate vary greatly. In polygamous societies, accusations often arose from jealousies and animosities between different wives of the same man.⁵⁷ On the other hand, such consequences were by no means certain, and there was no more inevitable and predictable a relationship between polygamy and the targets of suspicion than there was between witchcraft beliefs and any other kind of social organization. Among the Konkomba in northern Togo, who believed in witchcraft, much tension existed between co-wives, and yet accusations never arose from them.⁵⁸ The Wambugwe, living in Tanzania’s Rift Valley, thought that witches could not attack their own lineage.⁵⁹ Further north, in Kenya, the Nandi believed that witchcraft operated between in-laws, while another Tanzanian tribe, the Safwa, held that it could only be used against members of the perpetrator’s own patrilineage.⁶⁰ In Zambia the Ndembu thought that only close maternal kin were at risk, while in Sierra Leone the Kuranko viewed witchcraft as an attack on conjugal relationships, deployed only by married women against a husband or his kin.⁶¹ The Ngoni of Malawi thought that witches only attacked relatives on their mother’s side.⁶²

Nor are kin necessarily suspected of witchcraft in Africa, or indeed other parts of the world, the spectrum of favoured targets for suspicion extending through friends and neighbours to outsiders who had been allowed to settle in a community. Among the Gusii of Kenya the obvious targets were simply people who had failed to give clear evidence of their loyalty to the social group as a whole; likewise, the Nyakyusa of Tanzania suspected the generally antisocial members of their society.⁶³ One New Guinea people, the Tangu, used their equivalent word for a witch to describe all socially marginal people who had ceased to reciprocate in the social relationships of the community, whether they had begun to use witchcraft or not.⁶⁴ The Lugbara of Uganda associated witchcraft with strangers, loners, people with red or squinting eyes, the greedy and the grumpy.⁶⁵ The Quiché of Guatemala saw it in the lazy as well as the antisocial.⁶⁶ The Western Apache, on the other hand, eclectically suspected the wealthy, the elderly and strangers who had moved unexpectedly into the community.⁶⁷ Sometimes the stereotype did not actually match up to the reality, so that the Mandari of the Sudan traditionally associated witchcraft with physical filth, stealing and generally antisocial nonconformist behaviour, but admitted that most suspects were people indistinguishable from the norm.⁶⁸ The Wambugwe thought that witches had no features at all that distinguished them from anybody else, while among another Tanzanian group, the Hehe, accusations bore no relationship to sex, age or kinship.⁶⁹ The Gisu of Uganda thought that witches only attacked people of their own sex, while in Papua the Kaluli believed that they normally made victims of those unrelated to them by blood or marriage.⁷⁰ To the Mohave, whose traditional territory spanned parts of California, Nevada and Arizona, witchcraft was especially insidious because those who possessed its powers only used them to kill people whom they liked, as a compulsive and appalling consequence of genuine affection.⁷¹

In general, the comment made by Philip Mayer on Africans half a century ago holds good for human societies in general, that suspected witches and their accusers are people who ought to like one another but do not.⁷² To put it another way, as Eytan Bercovitch did after working in New Guinea, ‘The witch is everything that people truly are as communities and individuals but would rather not be.’⁷³ Suspicion of witchcraft has generally been one consequence of unmet social obligations. The circumstances under which that suspicion arises tend everywhere to be those of regular, close and informal relationships, especially those in confined and intense environments where it is difficult to express animosities in open quarrelling and fighting: which is why, for example, in southern India accusations were never made between different social castes, as they never had intimate enough relations with each other.⁷⁴ Although the consequences of allegations of witchcraft generally involved social groups, in essence they were generated by close personal relationships. In Godfrey Lienhardt’s words, ‘witchcraft is a concept in the assessment of relations between two people’.⁷⁵ A belief in it is an aspect of face-to-face human encounters.

Characteristic Three: The Witch Works within a Tradition

Around the world, it has commonly been believed that witches gain their malignant powers through training or inheritance; but there has been no general solution to the question of how this is done. Two very common responses are that the capacity to do harm is something innate in the person of the witch, or else that the witch works by the employment of magical materials. The two often overlap, in that a person who is empowered by an innate and internal force can utilize arcane forces in material objects in order to put their powers into action. Those societies that believe in witchcraft as an innate power often differ over whether it manifests because of the volition of the person concerned, or asserts control over the will and actions of that person, sometimes directly against their own inclination. It is quite common for the two kinds of witch figure, the one who operates because of innate power, and the one who needs to work by manipulation of the right tools and substances, to exist in the imagination of the same social group.

One such group was the Azande of southern Sudan, who became the subject of a very famous study during the 1930s by Sir Edward Evans-Pritchard, which helped to inspire the subsequent interest in witchcraft shown by members of his discipline and created some of the methods and models for it. As one of the latter, he confined the term ‘witchcraft’ to describe the actions of people who worked harm through natural and internal abilities, and employed that of ‘sorcery’ for those who needed external means.⁷⁶ For a time his distinction was widely applied to the study of extra-European magic, and that in Africa in particular.⁷⁷ By the 1960s, however, it was coming under criticism as inapplicable to many traditional peoples, in Africa as elsewhere.⁷⁸ It has now largely been abandoned, although some anthropologists still sometimes find it relevant to the particular societies with which they are concerned.⁷⁹ What emerges from a global analysis is that traditional peoples distinguish between forms of magic in different ways, some of which map onto the division made by Evans-Pritchard and some of which do not. A classification of harmful magic into witchcraft and sorcery, according to his criteria, will therefore not be used here. None the less, it must be recognized that societies across the world have divided workers of harmful magic into categories, in which some operate more from instinct and natural power and others more from design. In Dobu, for example, women were believed to work evil in their sleep, their spirits going forth to attack those of neighbours and so do them harm, while men worked it when awake, putting curses on the belongings of victims.⁸⁰

Equally variable across the world are local answers provided to the question of whether witchcraft is voluntary or involuntary; and if it is involuntary, what implications this has for the treatment of the suspected witch. Some peoples in Africa and Melanesia have regarded it as the consequence of a literal physical malady. The Hewa of the New Guinea Highlands thought that witches had a being like a small human foetus living inside them, which craved human flesh and drove them to kill to get it.⁸¹ The Tiv of Nigeria thought that witchcraft was a substance that grew on the hearts of certain people and gave them magical powers.⁸² In southern Africa the Swazis considered it to be a virus, transmitted by mothers to children or acquired by infection later in life, which drove sufferers to join a secret witch society dedicated to murder.⁸³ In north-eastern Ghana the Mamprusi also thought it a substance in the body inherited from the mother, though virtuous people were believed to be capable of resisting and neutralizing it.⁸⁴ The Bamileke of Cameroon believed it to be an extra organ, which produced a literal blood lust, satisfied by magical attacks.⁸⁵Elsewhere in the same country, the Bangwa thought that it was generated by a substance in the gullet, with which a person was born: parents with a baby who seemed to be manifesting strange behaviour would presume it to have this affliction, and would allow it to die.⁸⁶ Inhabitants of Seram in the Molucca archipelago of eastern Indonesia were of the opinion that the power to work evil magic was generated by a hard lump in the stomach or intestines.⁸⁷

Other societies regarded involuntary witchcraft as more of a spiritual than a physical affliction, though the boundaries between the two were hazy. Among the Azande, those whom Evans-Pritchard termed witches were thought to inherit an evil spirit from a parent, fathers passing it to sons and mothers to daughters. This dwelt inside their intestines, possessed them, and needed to prey vampire-like on the life forces of non-witches. Those afflicted by one were born with it, but like some genuine hereditary illnesses it grew stronger with age.⁸⁸ The Nyakyusa thought that witchcraft was endowed by an evil entity which took the form of a python lodged in the belly of the witch, while in New Guinea the Kaluli thought that such a being lodged in the witch’s heart.⁸⁹ In parts of the Indian region of Mysore, witches were likewise believed to be women afflicted with an evil spirit that drove them to do harm.⁹⁰ Among the Gă people of southern Ghana, it was thought that the spirits that possessed witches could torment or kill their human hosts unless they placated them by murdering others; those who feared that they were in danger of becoming thus possessed would seek magical cures for the condition.⁹¹ In the Philippines, the tariff demanded by the possessing spirit was at least one murder per year, in default of which it would kill the witch.⁹² Most cultures to have credited the existence of witchcraft, however, have considered it to be as controllable, and culpable, as any other kind of human ill nature (though normally more frightening and dangerous). Even some that regarded witches as people completely possessed by evil spirits, and so no longer responsible for their actions, have often thought that to permit such a degree of possession, the individuals concerned must have been at least weak and perhaps malevolent. Nicola Tannenbaum, studying the Shan, a Buddhist tribe spanning the border between Thailand and China, noted that they treated suspected witches in much the same way as antisocial drunks: as a real danger to others, and responsible for their condition although not really responsible for particular actions.⁹³

Another variation in global perceptions of the witch figure has been between those who have regarded witches as essentially solitary or operating in partnership with the occasional friend or ally, and those who believe that witches are members of organized secret societies. A belief in such associations has been recorded across much of sub-Saharan Africa, and in the south-western United States, India, Nepal and New Guinea. The participants were generally thought to feast together, encourage and strengthen each other in their vocation, plan in concert to work evil magic, and often actually do so. The methods imagined to be employed by witches in the working of such magic, whether collectively or alone, have inevitably taken different forms in different places, but certain patterns are found commonly across the world. One is a belief that witchcraft is worked with especial ease if the witch can obtain bodily waste from the person who is the target. Among the Maori of New Zealand witches were said to kill victims by destroying their clothes, hair, nails or excrement while uttering spells.⁹⁴ The Zuñi burned all hair clippings and their neighbours the Navaho concealed all human waste, lest it be used to work magic against the former owners.⁹⁵ In Alaska the Tlingit thought that witches took food leavings or scraps of clothing from intended victims, and made them into dolls, which became the vehicles for curses.⁹⁶ Such fears, and reactions, are recorded across most of Polynesia, Melanesia, Africa, South Asia and North America. Another belief system, which is not mutually exclusive with the first, is an emphasis on the use made by witches of magical properties within objects taken from the natural world, such as special stones, plants and parts of animals. The Nyoro of Uganda thought that most bewitchment was achieved by use of vegetable matter, mixed with pieces of reptiles.⁹⁷ Another very widespread tradition, found in North America and Africa, is that witches strike by introducing magical objects into their victims’ bodies, such as stones, bones, quills or ashes, the removal of which cures the effects. Yet another belief pattern, especially common in areas of West and Central Africa and Melanesia, is that witches work through their own innate powers of evil, having no need of physical aids. A further very widespread tradition is for the witch to be assisted or empowered by a personal spirit helper, or a set of them, often in animal form; such traditions will be examined in detail in the last chapter of this book. In the Solomon Islands of Melanesia, it was thought, unusually, that the evil spirits serving living witches were the ghosts of their dead predecessors.⁹⁸ Across the Americas, Africa and Melanesia, tradition also varied with regard to the question of whether witches were expected to go about their work as their normal, physical, selves, or to travel in some spiritual form to do it while their bodies remained asleep at home. In parts of sub-Saharan Africa, South Asia, Melanesia and North America, it was believed that they could fly, which greatly enabled their ability to cover distance in pursuit of their targets, although, again, opinion varied as to whether they did so in physical or spectral bodies.

Characteristic Four: The Witch is Evil

Across the world, witches have been regarded with loathing and horror, and associated with generally antisocial attitudes and with evil forces in the supernatural world. Such a trait rules out from the category of witchcraft the sanctioned or informally approved use of magic in neighbourhood feuds. That is sometimes found: for example in the Trobriand Islands service magicians would employ their skills to harm individuals who had incurred the jealousy of chiefs or neighbours by prospering above their station in life and disrupting the usual social order. Their activities were regarded as generally justifiable.⁹⁹ Among most peoples, however, the use of magic was never regarded as a legitimate means of pursuing feuds and quarrels within communities, but as an activity distinguished by secrecy, malevolence and intrinsic wickedness. The element of secrecy was considered to deprive the intended victim of any warning of the coming attack or consciousness of what was happening, until the harm had been done. It was designed to prevent any opportunity for compromise, negotiation and reconciliation, and for defensive measures, and to shield the witch as far as possible from being called to account for the crime. Such a way of proceeding, linked to the witch figure, violates common human notions of courage, sociability and justice. In some aspects, witchcraft has been used to represent the evil inherent in the universe, manifesting through humans who are fitted by their natures to act as vessels or conduits for it. In others, it has embodied all that is selfish, vindictive and antisocial within human nature, epitomizing treachery and disharmony in societies that strive for unity and neighbourliness. Godfrey Lienhardt summed up a general rule when speaking of one African people, the Dinka: that the witch ‘embodies those appetites and passions in every man which, ungoverned, would destroy any moral law’.¹⁰⁰ Accordingly, across most of the world, it has often been believed that witch societies reverse those norms in more dramatic ways, engaging during their meetings or their acts of evil-working in such activities as incest, nudity or cannibalism. Examples of this belief are abundant. The Zuñi of the American south-west thought that witch associations were dedicated to the destruction of the human race, and entry was allowed only to somebody who had already claimed a victim by magic.¹⁰¹ Their neighbours the Hopi thought that their own local witches were the leaders of a worldwide network in which every nation was represented, the initiates of which had to keep sacrificing the lives of their relatives in order to prolong their own.¹⁰² The Yoruba of Nigeria and Gonja of Ghana held that to join the secret witch society, people were even required to kill their own children as an initiation rite.¹⁰³ In New Guinea the Abelam believed that witch power was activated in a girl if she participated in a rite whereby a group of existing local witches dug up and ate a recently dead baby.¹⁰⁴ Across Polynesia there was no apparent belief in witch societies, training in witchcraft being conceived of as an individual business passed by an experienced practitioner to a novice, but those practitioners were still believed to have a general grudge against humanity and the test of proficiency was to slay a near relative.¹⁰⁵ The Iroquois thought that the price of joining the local witch organization was to use magic to murder one’s nearest and dearest relation.¹⁰⁶ Across most of the world witches have been thought to gather at night, when normal humans are inactive, and also at their most vulnerable in sleep. The Tswana believed that they assembled in the hours of darkness to exhume corpses and use parts of them for their destructive magic.¹⁰⁷ More often, across most of Africa and Melanesia, including New Guinea, witches were expected to dig up freshly buried bodies in order to feast on them together, this being the prime motivation for murdering the people concerned. The Bemba of Zambia thought that witchcraft was the work of people who committed incest as well as murdering babies.¹⁰⁸ Nudity was a common attribution of witches, not merely because it transgressed social norms but because it stripped away their everyday identities. In the Solomon Islands, witchcraft was attributed to women who met at night to take off their clothes and dance.¹⁰⁹ The Agariyars of Bengal thought that women became witches by going to the cremation ground at midnight, removing their clothing, sitting on the ground and speaking incantations over cremated ashes.¹¹⁰ Children among the Lala of Zambia were told not to go out naked lest they be mistaken for witches, while in the Lowveld region of the Transvaal the same fate met, and still in places meets, any woman seen without clothing out of doors, even in her own yard.¹¹¹ On Flores in the southern Indonesian island chain, it was said that a person could attract a possessing spirit, which conferred the power of witchcraft, simply by running nude in the open air.¹¹² Witches among the Kaguru of Tanzania were not merely thought to operate naked, but to walk upon their hands and to smear their normally black bodies with ashes to turn them white, in further rites of reversal.¹¹³ The Amba of western Uganda thought that they rested by hanging upside down in trees and ate salt when thirsty (as well as embracing the usual nudity and cannibalism).¹¹⁴ In the Philippines, they were also supposed to hang upside down like bats, as well as having no sense of physical modesty.¹¹⁵ Zulu witches were said to ride naked on baboons at night, facing backwards.¹¹⁶ Those imagined by the Western Apache removed their clothing for all-night dances around bonfires, holding parts of exhumed corpses, as part of which rites men deliberately copulated with menstruating women.¹¹⁷ In these senses, the witch figure has represented an attempt to imagine how human beings can continue to live within communities while secretly rejecting and attacking all of their moral constraints, striking at all the imperatives that bind their societies together and make them functional. In societies where the expression of aggression and resentment is customarily repressed in the name of communal solidarity and harmony – and these are very common among traditional peoples – the witch figure provided a kind of human being whom it was not only proper but necessary to hate actively and openly.

Characteristic Five: The Witch Can be Resisted

The belief that witches can be resisted by their fellow humans is also found worldwide, in the three main forms which it took in Europe. One of these was to protect oneself or one’s dependants and property by using benevolent magic, which could turn away spells and curses; if the latter seemed to take effect, then stronger magic could be employed to break and remove the effects of bewitchment; and perhaps to make the witch suffer in turn. The Dowayo of Cameroon put sharp thistles or porcupine quills on the roofs of their homes, and spines and spikes around their fields and threshing floors, to ward off evil spells.¹¹⁸ The Navaho had a wide range of objects and techniques which were said to proof the owner against witchcraft, including songs, prayers, stories, consecrated artefacts, paintings and plants.¹¹⁹ In northern India, the performance of blood sacrifices or the deployment of tamarind or castor oil plants was thought effective.¹²⁰ Across Polynesia, protective rituals were enacted to safeguard people against witchcraft, and if these apparently failed, others were used to counter-curse the witch.¹²¹ The Vugusu and Logoli of western Kenya usually responded to the threat of witchcraft by avoiding presumed witches and carrying out counter-magic against them.¹²² On the Melanesian island of Gawa, suspected witches were never publicly accused and no mechanism existed for trying them, the population depending instead on defensive magic.¹²³ The Gaya of northern Sumatra treated bewitchment with exorcism, designed to send back the evil spirit that caused the complaint to the witch who had originally sent it out.¹²⁴ Most societies that believed in witches have contained service magicians who were regarded as expert in such remedies and could provide them to others as a duty or for payment. Indeed, this activity is embodied in the common English term for such a magician (usually in a non-European and tribal context) of ‘witch-doctor’, which was first popularized by a best-selling book by the famous Victorian British explorer Mary Kingsley. It has sometimes been mistaken as meaning a witch who is a doctor, but it signified instead a doctor who specialized, at least some of the time, in curing the damage done by witches: Kingsley’s own definition was a ‘combatant of the evils worked by witches and devils on human souls and human property’.¹²⁵ Under whatever name, the breaking of bewitchment has been, worldwide, one of the most commonly found and important functions attributed to service magicians.

The second widespread remedy for bewitchment was to adjust the social relations that had created the suspicion of it. This could take the form of persuading or forcing the witch into removing the spell that she or he had placed, and so its destructive effects. Among the Azande, when a service magician or chief had decided that a malady was the result of bewitchment, then the next step would be to ask the alleged culprit to lift the spell. The same pattern was found in Botswana, with the Tswana.¹²⁶ Among the Gusii, the first reaction to a suspicion was to employ private magic to break the hostile spell, and the second to sever all relations with the suspected witch, to deprive the latter of those contacts with their victim(s) that had made bewitchment possible.¹²⁷ In the Tonga Islands of Polynesia, it was believed that the only way to cure bewitchment was to persuade or force the witch to remove it.¹²⁸ The Yakö of eastern Nigeria thought that suspicions were best dealt with in private, by asking the suspect to desist from bewitchment.¹²⁹ In Ghana the Ashanti blamed the act of witchcraft rather than the person perpetrating it, so the presumed witch was forgiven after making a public confession (which was presumed to break the bewitchment) and paying a fine or enacting a penance.¹³⁰ The Tangu of New Guinea expected an unmasked witch to pay compensation to the victim, after which the matter was closed.¹³¹ On Dobu, a service magician was hired to identify the source of bewitchment, usually by gazing into water or a crystal. A suspect would be accused as a result, and required to recall the curse placed upon the victim; and when this was apparently done, both diviner and alleged witch would be paid by the victim. Such faith was placed in this process that, if the victim still failed to recover, a new curse and witch were presumed to be the cause.¹³² In Cameroon the Bamileke, who thought that witchcraft was the involuntary consequence of an extra organ in the body, likewise believed that public exposure as a witch automatically destroyed the power of the growth and so the accused was both disarmed and reintegrated into society.¹³³ The Lisu of the northern Thai highlands feared witchcraft acutely but relied on service magicians or private counter-magic to keep them at bay. If this failed, then a suspected witch would be accused and made to pay compensation and retract the spell; people very rarely killed those whom they blamed for bewitchment, for the good reason that people who murdered witches were thought to become witches themselves, by contagion.¹³⁴

The third remedy was to break the power of the witch with a physical counter-attack, which could take the form of direct action, such as a severe beating or murder, or intimidation that ran the person concerned out of the neighbourhood. In most societies, however, a formal and legal remedy was preferred to this sort of private action, by which the suspect was prosecuted before or by the whole community, and if found guilty was subjected to such punishment as it appointed. In many cases the identification of the culprit was assisted or carried out by the same kind of magician as that which provided counter-magic against witchcraft. In Central and Southern Africa, the ability to detect witches was also believed in several places to be inherent in chiefs, as one of that concentration of semi-mystical qualities that gave them the right to lead. In central India the same power was attributed to holy men. Across much of the world, oracles and special rites were employed to find the guilty party when witchcraft was suspected. The Dangs of western India would drop lentils named for each adult male villager into a vessel of water: the one that floated would be that of the husband of the witch.¹³⁵ Service magicians of the Lala of Zambia found witches by gazing into a bowl of consecrated water, throwing an axe handle into ashes or watching horns stuck into the ground.¹³⁶ The Nyoro of Uganda tossed cowrie shells into a mat and interpreted the pattern they made, while in the same country the Gisu asked questions of pebble patterns in a swung dish.¹³⁷

Once under suspicion, people were commonly forced to undergo an ordeal to demonstrate innocence or guilt. The traditional witch-finding society among the Nupe of northern Nigeria forced suspects to dig the ground with bare hands: if they bled, they were deemed to be guilty.¹³⁸ The Dowayo would make them drink beer in which a poisonous sap had been mixed. A person who died or produced red vomit as a result was deemed guilty, while those who produced white vomit, and lived, were exonerated.¹³⁹ Different forms of this poison ordeal were found across Central Africa, from Nigeria to Zambia and Madagascar, and its consequences depended on how toxic a potion was made. The Lele herded suspects into pens for testing, and the drink administered killed many of them.¹⁴⁰ The same test was used in north-western New Guinea, where those who vomited the poison were declared guilty and put to death: as it was quite difficult to survive the poison without bringing it up, this was an ordeal heavily weighted against the person submitted to it.¹⁴¹ In Africa from Ghana to the islands off the Tanzanian coast, a chicken had its throat cut or was given poison in front of a suspect, whose guilt or innocence was determined by the final posture of the dying bird. The danger in which the accused was placed could be manipulated by deciding how many such postures counted as proof of innocence: in much of Nigeria during the 1940s and 1950s, the odds were heavily weighted against acquittal by the ruling that only one position did.¹⁴² A standard test for witchcraft on Flores, in the southern Indonesian island chain, was to have to pick a stone out of boiling water: the guilty would blister.¹⁴³

Once a person was identified as a probable witch, torture was sometimes used to extract a confession: in India the Dangs commonly swung an accused person upside down over a fire.¹⁴⁴ Across much of the rest of India and in Burma suspects were flogged with wood from a sacred tree.¹⁴⁵The Navaho of the south-western United States preferred to tie them up and starve them of food and shelter.¹⁴⁶ How severe a penalty was imposed on those convicted of witchcraft depended both on local attitudes to it and the perceived extent of the damage done by the presumed witch. To societies that prescribed the death penalty for murder or other serious crimes against the person, it was logical to apply it to people convicted of inflicting death or ruinous damage by means of magic. Most peoples who have traditionally believed in witchcraft have killed at least some of those formally convicted of it. In communities that greatly feared witchcraft, the body counts achieved could be considerable. It was said that in pre-colonial days every village of the Bakweri of Cameroon had its witch-hanging tree.¹⁴⁷ Among the Pondo of South Africa, the rate of execution ran at one per day on the eve of the British conquest and this number did not include those who fled when accused, or were fined.¹⁴⁸ A British official serving in India during the early nineteenth century estimated that about a thousand women had been put to death for alleged witchcraft on the northern plains during the previous thirty years: a rate of mortality far more serious than that caused by the more notorious local practice of sati, or widow-burning.¹⁴⁹ The rupturing of British rule over India in the rebellion of 1857 permitted a great witch-hunt, with lethal effects, to occur among the tribes of northern India.¹⁵⁰ Before British colonialism arrived, the Nyoro allegedly burned many of their people alive as witches, while before the Germans conquered them, the Kaguru clubbed to death those convicted of witchcraft and left them to rot in the bush, and the Pogoro burned them alive.¹⁵¹The Greenland Inuit cut the bodies of those executed into small pieces to prevent their spirits from haunting the living.¹⁵² Likewise, the Northern Paiute of what became Nevada and Oregon stoned convicted suspects to death and then burned the corpses.¹⁵³ A Jesuit missionary working among the Huron of Canada in 1635 noted that they often murdered each other or burned each other alive on the testimony of dying men who accused the victims of having caused their fatal illness by magic.¹⁵⁴ On Flores, the penalty for witchcraft before the Dutch conquest was to be buried alive, and this apparently occurred regularly. On another Indonesian island, Sulawesi, the Toraja people submitted accused witches to ordeals that allowed virtually no proof of innocence, and then beat them to death. Young boys were encouraged to participate in this to prove their courage.¹⁵⁵ Before being ruled by the British, the tribes of what is now Botswana avenged deaths by presumed witchcraft either by allowing the bereaved relatives to kill the family of the suspected witch or by having the local chief try the suspects and execute those convicted: there were twenty-six such trials among the BaNgwatetse alone between 1910 and 1916. The former execution places of witches were still pointed out to British visitors to the region in the 1940s.¹⁵⁶ The Kaska, who lived on the border between Canada and Alaska, had no concept of magical cures that could be used against witchcraft, and so the only known remedy was to deal with the witch, who in that society was usually thought of as a child. This belief led to persistent killings in the first two decades of the twentieth century, often by the families of the youngsters accused.¹⁵⁷

Across the world, traditional peoples have often manifested the pattern of sudden upsurges in witch-hunting among populations hitherto or for a long time characterized by little of it. In general, people who have traditionally feared witchcraft tend to accuse neighbours of it much more frequently in times of economic pressure and/or of destabilizing economic, political and cultural change; but it is also true that such times do not automatically and necessarily produce an increase in accusations. When such an upsurge has occurred, it has tended to rebound on the social order in three different ways: to confirm the authority of the traditional leaders and society; to enhance the power of an individual member of the traditional elite; or to enable a new social group to seize authority. In Africa, Lobengula, king of the Matabele, Ranavalona, queen of the Malagasy, and Shaka, king of the Zulus are examples of nineteenth-century leaders who reinforced their hereditary authority by waging war on alleged witches. Shaka once summoned almost four hundred suspects to his court at once, and killed them all, while under Ranavalona about a tenth of her subjects were forced to endure the poison ordeal to test for witches, and a fifth of those died. Lobengula presided over an average of nine to ten executions per month, mainly of relatively powerful men. In nineteenth-century North America the Navaho chief Manuelito executed more than forty of his political opponents on charges of witchcraft, and a generation earlier the Seneca chief Handsome Lake established himself as a religious leader by directing a persecution of it.¹⁵⁸ Such figures sometimes used witch-hunting to defend traditional ways against innovation: in the eighteenth-century Ohio Valley the Shawnee prophet Tenskwatawa instigated it against Christian converts in his tribal confederacy.¹⁵⁹ Political use of the mechanism could be deployed collectively as well as by particular rulers and prophets: thus, in the seventeenth century the north-eastern Algonquian tribes of North America made witchcraft accusations their main means to establish new territorial boundaries to service the fur trade developing with European settlers.¹⁶⁰ On the other hand, some strongly based and long-established regimes chose to discourage witch-hunts as part of the demonstration of their authority. When a panic swept twelve provinces of China in 1768, that itinerant magicians were cursing people (especially male children) to death in order to enslave their souls, the imperial judges quashed the convictions imposed by local courts, although mobs had murdered some suspects before they could be arrested.¹⁶¹

In Africa witch-finding movements were common in the colonial period, affecting much of the western and central parts of the continent, and functioning partly as a response to the prohibition or extreme modification of traditional trials for witchcraft by the European administrations. It is also possible that colonial rule, by shattering tribal institutions and moral codes, increased the instability in which a fear of bewitchment often flourishes.¹⁶² The Lele were caught up in no less than five witch-hunts between 1910 and 1952.¹⁶³ Typically, they were conducted by young men who toured regions, crossing tribal boundaries and claiming the power both to detect witches and to render them permanently harmless. The latter process usually took the form of forcing suspects produced by communities to deliver up the materials with which they were supposed to work their magic, for destruction, and administering a drink or ointment to them, or a particular rite, which was supposed to remove their ability to bewitch. Likewise, in western India, the ‘Devi’ religious revival of the 1920s included the detection and banishment of witches from villages as part of its remit.¹⁶⁴ Such movements originated from outside traditional structures of authority and custom but generally worked within them. Even under colonial rule, however, witch-hunters sometimes emerged who provoked a rejection and punishment of the familiar native elites or religions. The Atinga witch-finding cult in West Africa was conveyed by devotees of a single shrine in northern Ghana, who destroyed other traditional cult centres as they travelled.¹⁶⁵ The Nyambua, their equivalent in Nigeria, denounced established chiefdoms as well as witches.¹⁶⁶ Sometimes, also, such movements blended with anti-colonial feeling, or even with outright rebellion: the Maji Maji uprising against German rule in Tanganyika in 1905–6 was led by a prophet who termed himself a ‘killer and hater’ of witches, and indeed ordered the death of anybody who refused the ‘medicine water’ that he administered to destroy evil magic.¹⁶⁷

This pattern has become much more common since the removal of European rule, as Africa has undergone programmes of self-conscious modernization that have produced major social change.¹⁶⁸ Witch-hunting has often been prominent both in revolutionary movements which directly opposed and helped to end colonialism or white supremacy, and in the successor states, under native regimes, which emerged out of the former colonies. The groups of young men who attacked suspected witches in parts of the Transvaal during the 1980s were also those who led resistance to the system of apartheid, portraying the white government which both upheld apartheid and forbade witch-hunting, as the protector of witchcraft. After the establishment of black majority rule, they still found themselves marginalized by the new regime, and so continued their role as local defenders of their people, in the face of a largely alien central government, with persecution of witches still part of that role.¹⁶⁹ Closer to the main centres of population in the new South Africa, in the Soweto township, the daily fear of witchcraft was reported as ‘tremendous’ by the early 1990s, and it was said that ‘every older woman, especially if eccentric and unpopular, lives with the risk of being accused of witchcraft’.¹⁷⁰ Among the Mijikenda of the Kenya coast, independence was followed by an upsurge of accusations and of violence against suspects, with tribal and national administrative leaders uniting to promote a particular healer as a witch-finder.¹⁷¹ From the 1970s direct and public accusations of witchcraft increased in Zambia, and with them the use of expert witch-finders, who were ubiquitous in rural areas by the 1980s.¹⁷²

In the war of independence, which established native rule in Zimbabwe, the guerrillas assumed the traditional role of chiefs as witch-detectors, usually with the full support of local communities, and put those detected to death if those communities desired it. Unsurprisingly, the victims were often allies of the white government.¹⁷³ After independence had been achieved in the country, during the early 1990s, a local hunt was conducted by a spirit medium obtained from a government-sanctioned National Traditional Healers’ Association, who detected witches by making suspects step over his walking stick.¹⁷⁴ Both sides in the Angolan civil war of the early 1990s, which followed the collapse of Portuguese rule, put alleged witches to death as an aspect of their attempts to enhance their popularity and claims to legitimacy; one tended to burn them alive and the other to kill them after making them dig their own graves. Refugees expressed outrage at the abuse of the activity, by targeting political opponents (and their children) as witches, but not at their execution.¹⁷⁵ In those parts of the world in which native people were ruled for a time by European powers, a feature of the persecution of alleged witches was the manner in which selected features of Christianity were borrowed from the colonial rulers and integrated with traditional concepts of the witch. This was a natural enough process in Latin America, where for more than two centuries the ruling Europeans themselves feared witchcraft and outlawed all kinds of magic. Two parallel systems of witch-hunting thereby met and blended, with the early modern European stereotype of witchcraft as a form of Devil-worship infiltrating indigenous ideas and taking up permanent residence among them.¹⁷⁶

The process continued in Africa in the twentieth century under a very different colonial system, in which the official attitude to witchcraft was one of disbelief. Here the Bible, in early modern translations which affirmed a disapproval of witchcraft and ordered its suppression, often acted in its own right to confirm native beliefs: ironically, Christianity therefore had the effect of reducing the credibility of ancestor spirits and land spirits, against which the missionaries preached, and so of producing a tendency to blame witches alone for uncanny misfortunes. The easy relationship that could be made among traditional peoples between Christianity and witch-hunting is replete with examples. When the Malagasy queen Ranavalona created an intolerant religion to bond together her nation in the 1830s, which persecuted alleged witches and Christians alike, she did so using early modern European Christian models.¹⁷⁷ A generation earlier, Handsome Lake’s hybrid religion, which he introduced to his branch of the Seneca of upstate New York, added Christian angels and devils to native spirituality, reinforcing an existing fear of witchcraft.¹⁷⁸ In the 1920s native members of the Jehovah’s Witness movement in Central Africa got the idea that baptism by total immersion in water could detect witches. One of the proponents of it, who came to call himself the ‘Son of God’, was executed by the British after he had been found responsible for the killings of over a score of people in their territory in what was then Northern Rhodesia, and for the indictment of almost two hundred more in the Belgian Congo.¹⁷⁹ He was followed in the next decade by a man who had been schooled by Seventh-Day Adventists and decided to found his own church in Northern Rhodesia, which included the exposure of witches in its remit.¹⁸⁰ The Bamucapi witch-hunters, who spread across Central Africa from Lake Nyasa to the Congo basin in the 1930s, wore European clothes and preached ‘the word of God’ like white missionaries.¹⁸¹Also between the world wars, a woman founded the Déima movement on the Ivory Coast after contact with Protestant Christianity had convinced her that she was an expression of the will and word of the Christian deity: she claimed to detect witches on sight. In the 1950s it was the missionary activities of the Salvation Army that triggered the Munkukusa or Mukunguna movement in the Congo basin, in which the Bible and cross were prominent symbols. Later in the decade a Protestant United mission in Northern Rhodesia baptized and instructed a woman who claimed a divine commission to preach against witchcraft. She set up her own church organization, which came to include 85 per cent of the population in her home district.¹⁸² The establishment of Zionist churches in the Northern Province of South Africa enhanced fear of witches in that area, while among the Zulu some leaders of the same denomination became notable witch-hunters. Such churches also produced a hunt in Zambia during 1988–9, led by a prophet called Moses.¹⁸³ When some of the Tangu of New Guinea were converted to Christianity, they immediately identified witches with the Devil, and exactly the same thing happened among the Ewe of Ghana.¹⁸⁴ A notable witch-hunter in Malawi in the years around 1960 had learned his ideas in a Presbyterian church, while in Zambia by the 1960s prophets from Pentecostal churches were very prominent among the magicians who detected the sources of evil magic.¹⁸⁵ The leader of the Catholic Action movement in the Zambian capital of Lusaka in the 1970s was a woman who claimed to possess servitor spirits and to have the power to detect witches by reacting physically to their presence.¹⁸⁶ When many of the Lele converted to Roman Catholicism in the late twentieth century, they promptly declared the native religion to be that of Satan and its priests witches. The young in particular proved amenable to conversion, as an opportunity to turn upon their elders, and some of the new Catholic priests among them became avid witch-hunters, employing torture to gain confessions.¹⁸⁷ By the opening of the twentieth century hundreds of community churches in the Congolese capital of Kinshasa were committed to a struggle against witchcraft, as a satanic force.¹⁸⁸ In 2005 it was estimated that Africa now had hundreds of thousands of ‘prophets’ attached to native denominations of Christianity who claimed the inspiration of the Holy Ghost and other spirits to detect the hidden causes of misfortune, especially witchcraft.¹⁸⁹ The movements to eradicate witchcraft under colonial rule were generally bloodless, because the use of serious violence would have encouraged a hostile response from the European administrators who officially disbelieved in the threat from witches. This is what happened to the Atinga cult in Ghana and Nigeria in the 1940s and 1950s, which tortured and sometimes killed suspects who refused to confess.¹⁹⁰ The ending of foreign rule, however, opened the way for the return of widespread physical attacks on suspects, often leading to death. When Belgian rule collapsed in the Congo during the 1960s, the Lele immediately reintroduced their traditional poison ordeal, and hundreds died.¹⁹¹ In northern Uganda, the end of British rule was followed by a resumption of witch-hunting by chiefs, with considerable popular support. Suspects were tortured by being made to sit or walk naked on barbed wire, exposed to termite bites, beaten, made to drink their own urine, or having pepper put into their eyes.¹⁹²

In the Northern Province (now the Limpopo Province) of South Africa, witchcraft seems to have been relatively little feared in the first two-thirds of the twentieth century, and accusations ran at a proportionately low rate: the highest level was recorded among the Lobedu, of fifty in the course of the 1930s, punished by exile. The social, political and economic instability that accompanied the breakdown of the system of apartheid, however, led to an escalation of tensions among neighbours, which resulted in 389 known witch-related killings in the province between 1985 and 1989 alone.¹⁹³ In the 1990s recorded cases of such murders there totalled 587, but this was judged to be a serious underestimate due to a fear of reporting such incidents to the authorities: it was known that forty-three people had been burned alive in one action in just the Lebowa district.¹⁹⁴At Soweto, witchcraft-related murders were rarer, but still took place at times in the 1990s, the victims being burned to death by a mob of young people who termed the process ‘democratic’.¹⁹⁵

Both Malawi and Cameroon have reintroduced laws that allow the trial and conviction of people for alleged witchcraft. In Cameroon service magicians are accepted as expert witnesses by judges, and their testimony valued above the protestations of innocence by the accused. The latter are commonly treated as having no human rights and are sometimes beaten to death by police attempting to extract confessions. Concrete proof or a confession is not required for conviction and the prison sentences imposed are heavy – up to ten years – but at least those found guilty are not put to death. Tanzania has refused to permit a revival of the legal prosecution of witches, and the result has been an epidemic of lethal vigilantism. At least 3,333 murders of suspected witches were recorded on the mainland of the country between 1970 and 1984, two-thirds among one people, the Sukuma.¹⁹⁶ By 1991 a ghetto had been created in the old capital of the Mamprusi of Ghana, in which 140 women had been permanently confined on suspicion of witchcraft, to live in poverty: the space operated both as a prison and as a sanctuary in which they were safe from their accusers.¹⁹⁷In 2007 the president of Gambia, Yahya Jammeh, sent a division of his personal bodyguard to join local police in rounding up over 1,300 suspected witches from one district of his country. They were taken to detention centres and dosed with a potion expected to remove their powers, which made many ill. Three years later a major hunt swept southern Nigeria, directed at children and driven by ministers from native Christian churches who offered to exorcize the accused and render them harmless. The young victims were often detained and tortured to induce them to confess, and then abandoned by their families after exorcism; and all this occurred despite the existence of a new national law forbidding accusations against children. By 2012 the panic regarding child witches had spread to Congo, and twenty thousand children were said to be living on the streets of the capital, Kinshasa, because they had been expelled from their homes.¹⁹⁸ By 2005 at least half a million people had emerged as self-proclaimed experts in dealing with the problems of bewitchment in South Africa alone. If Christianity had easily been assimilated into traditional beliefs regarding witches, and served to reinforce them, then so has modern technology. Indeed, as the anthropologist Adam Ashforth has emphasized, science has become the ‘primary frame of reference’ for interpreting witchcraft in some South African townships, as quantum physics, cell phones, digital imaging, cloning and artificial life are all more compatible with a magical view of the universe than that of the preceding machine age.¹⁹⁹

In other areas of the world, informal and illegal violence against presumed witches has also reached, or been maintained at, serious levels in recent times. During the 1960s one small Mexican town inhabited by Maya had a homicide rate fifty times that of the USA and eight times that of the Mexican average, and witchcraft was the motive in about half of all cases.²⁰⁰ In north-eastern India, there were twelve witchcraft-related murders in the Maldo district in 1982 alone, and over sixty in the Singhbhum district during four years of the 1990s.²⁰¹ A Bolivian villager was tortured and exiled in 1978 by an informal communal tribunal for allegedly using magic to suck life out of neighbours while they slept; five years later, another such group burned a man to death for the same offence.²⁰² Among the Ambrym Islanders of Central Melanesia, fear of witchcraft, and the homicides that it generated, had reached what was described as ‘critical levels’ in the late 1990s.²⁰³ By the 2010s other parts of Melanesia had become as severely affected, as a result of collapsing traditional social and cultural systems, declining health services, worsening poverty, and increasing lifestyle diseases and premature deaths. Violence against suspects was (as recently in southern Africa) mainly conducted by impoverished young men seeking to achieve value in the eyes of their communities, and was becoming more public as well as more extreme. In New Guinea a young woman was burned alive in 2013 in front of hundreds of onlookers, including police, and two other women publicly tortured and beheaded on Bougainville Island in the Northern Solomon archipelago. In 2014 two men were publicly hanged in a community hall in Vanuatu.²⁰⁴ Nor is legal action against witchcraft missing from the world outside Africa, above all in Islamic states. During the period between 2008 and 2012, laws against magical practices of all kinds were more strictly enforced in Afghanistan, the Gaza Strip, Bahrain and Saudi Arabia. In that period Saudi Arabia executed several people for such offences, mostly foreigners and mostly by beheading. A woman was murdered as a suspected witch in Gaza in 2010. The Saudi government trains employees not only as witch-hunters but in rituals to destroy the effects of witchcraft, while one recent president of Pakistan, Asif Ali Zardari, sacrificed a black goat nearly every day to ward off its effects.²⁰⁵ In Indonesia courts have become increasingly willing to try acts of magic as crimes, or at least as antisocial behaviour, judges often seeming to believe in it and their actions in doing so being popular.²⁰⁶

It is possible to make a theoretical case that witch-hunting may, at least at times, serve a positive social function. In some contexts it may reinforce cultural norms, and so communal solidarity, by discouraging aberrant or antisocial behaviour. The identification of witchcraft with jealousy, greed and malice can serve to strengthen attachment to the countervailing virtues, and discourage the expression of animosity. It can be used to enforce economic obligations and reduce competition in favour of co-operation. In other contexts, it can be a midwife to change, in that anti-witchcraft movements have often legitimized or reinforced the power of new groups. Accusations have sometimes provided a means by which disempowered individuals, such as children or women, can attract attention and respect, and intimidate people normally in superior positions to them. They can articulate otherwise unspeakable fantasies, reveal and represent destructive impulses, and identify and express tensions within families and wider social groups, blasting away unsustainable relationships. Measures against presumed witchcraft have enabled humans to act purposefully in the face of adversity. It was for these reasons that an influential school of thought among anthropologists has held that witchcraft accusations functioned as instruments of social health rather than as symptoms of malfunction.²⁰⁷

Others, however, have held a different opinion,²⁰⁸ and that is the one favoured here. It emphasizes that all these positive functions of belief in witchcraft have only acted to strengthen societies, or to enable them to adjust more effectively to changing circumstances, when the rate of accusation has been low and sporadic, and subjected to firm controls. In many cases this situation has not obtained, and suspicions and accusations have not resolved fears and hostilities, but aggravated them and represented obstacles to peaceful co-operation. At worst, they have torn communities apart and left lasting traumas and resentments, or greatly compounded the suffering consequent on adjustment to new economic and social developments. Most societies that have believed firmly in witchcraft have regarded it as a scourge and a curse, of which they have longed to be rid; but the only way in which they have been able to conceive of bringing about this happy result has been to destroy the witches. Such attempts have tended to reinforce vividly a consciousness of the threat from witchcraft, and so perpetuate fear of it, and make future witch-hunts likely, even if they have managed – often at a grim human cost – to reduce that which had existed at the moment.

Further Reflections

Anthropological research permits some other insights into the manner in which the stereotype of the witch can be constructed and maintained, which are not so readily available to a historian. One such insight – accessible to scholars working in relatively small, self-contained societies where they can themselves question inhabitants in detail – is that cosmologies do not have to be coherent mental constructions. Very frequently, traditional peoples have been shown to have believed in different kinds of supernatural entity, including deities, terrestrial spirits, animal spirits and ancestral spirits, without any clear idea of how they interrelated or could be distinguished, or of the precise relationship between each of them and witches. What mattered to the humans concerned was the presumed effect that these kinds of entity had on the human world, and what could be done to encourage, deter or counteract that according to its degree of utility and benevolence. In subsistence societies it was the practical consequences of dealings with spirit worlds that were the real issue, often being literally a matter of life and death. The fact that the theoretical origin and operation of witchcraft, and the stereotypical nature of the witch, seemed in some places to be – in the perception of the European scholar – at odds with general preassumptions about the workings of deities and spirits, did not seem to trouble the individuals whose beliefs were being recorded.²⁰⁹

Another luxury permitted to anthropological research is to observe at first hand the manner in which beliefs in a given society can mutate in changing social and mental environments. The summary of extra-European beliefs concerning witchcraft made above may have given an impression of their being more or less static, the stereotypes of the witch held by particular peoples remaining broadly unchanged over time and little affected by contact with other cultures. In essence, that impression does seem to reflect reality, but there are a few qualifications to be made to it. On the whole, the image of the witch held by a particular human society alters in detail, with changing circumstances, while remaining the same in basics. It has already been noted that extra-European peoples in various different parts of the world have assimilated forms of Christian theology into their traditional beliefs about witches, and other and more specifically local additions of the same sort have been recorded by anthropologists. It is not uncommon for traditional peoples to accord new powers and modes of operation to witches, because of changing circumstances or contact with the ideas of other cultures. In the late twentieth century, the idea spread through parts of West and South Africa that witches were turning victims into zombies, to labour for them and increase their wealth; in Ghana this took the different form that they were changing humans into animals or plants and then selling them as such.²¹⁰ Much earlier, in some areas of East Africa, particular tribes acquired from Arab traders the idea that witches controlled evil spirits (sometimes by buying them from the Arabs themselves), while the mixed-race inhabitants of the Lebowa district of Transvaal adopted the idea that witchcraft employed spirits in animal form from the neighbouring Zulus.²¹¹

It was rarer for the stereotype of what a witch should be to change, but that sometimes occurred. As the Giriama of Kenya moved from fortified communities into dispersed homesteads in the late nineteenth century, greater discrepancies of wealth appeared among them, and it was the newly enriched who became particular targets for suspicion.²¹² In the Gwembe Valley of southern Zambia, witches were traditionally male relatives of their victims but not the latter’s parents. From the 1980s, however, a declining economy produced a younger generation less wealthy than the older, who turned on their fathers with accusations of witchcraft as one product of the ensuing tensions.²¹³ In the lower Congo Valley, witches had stereotypically been elderly, but in the capital city of Kinshasa during the 2000s, as said, children and teenagers came to be blamed for all misfortunes instead.²¹⁴ Most rarely of all, it seems that some peoples who had no traditional fear of witchcraft could acquire one, or those who had feared it little could become severely afraid. Among the Kerebe of the Lake Victoria border of Tanzania, it seems that before the early nineteenth century uncanny misfortune was attributed to chiefs, who were credited with wielding a legitimate magical power over people to discipline them. Then a new trading economy disrupted the power of both chiefs and communities, producing a new competitive and individualist society in which fear of witchcraft became rife.²¹⁵ The Kaska, on what became the border between Alaska and Canada, seem to have acquired a belief in witches in the late nineteenth century, when their society underwent drastic change, endangering its very survival, because of European conquest. They therefore took on the belief from their western neighbours the Tlingits, among whom witch-hunting was traditional.²¹⁶

Ethnographic fieldwork also allows some answers to the question of how far, and in what sense, witchcraft has ever been a ‘real’ phenomenon. Anthropologists all over the world have reported similar experiences, in finding that peoples who believed in witchcraft would, when their trust and confidence had been gained, talk avidly about who witches were, and what they were supposed to do. It was virtually impossible, on the other hand, to interview somebody who actually claimed to be a witch and to act out the role expected of one. It is equally true, however, that witch accusations among traditional peoples have regularly produced confessions, especially after the accused had been found guilty by their community. Fear and despair, and a hope to win mercy and forgiveness by a show of penitence, may well have produced many such responses. On the other hand, it is also credible, and perhaps logical, that some of the accused actually had tried to curse neighbours or relatives when moved by anger, jealousy or malice, using formulae and materials associated with witchcraft. This is, however, remarkably hard to prove.²¹⁷ Anthropologists have noted from first-hand observation that when a witch-finding movement passed through a district, the people whom it convicted and forced to surrender their materials of witchcraft certainly produced objects in response. These were, however, of a kind also associated with positive magic, such as that intended for protection and healing.²¹⁸ A scholar working in New Guinea commented on how destructive magic was worked there by wrapping up physical waste products of the intended victim with bark, leaves and stones over which a spell was recited. She added that these bundles were sometimes genuinely made, but did not enlarge on the circumstances.²¹⁹ The Gusii of the south-west Kenya highlands believed that witches were usually women, and ran naked at night carrying a pot of burning vegetable matter. One man told the anthropologist Robert Levine that as a child he had seen a female neighbour hurrying home nude at dawn with a firepot, while Levine was also told that women had confessed to witchcraft and brought human remains out from their homes during a witch-finding movement. This testimony, however, remained unproven, as did that made to an anthropologist by informants among the Barotse in what is now Congo: that human bones were often found in the homes of suspects of witchcraft, which must have come from graves.²²⁰ Some of the most convincing and disturbing evidence for the actual practice of magic with the intention of harming others comes from the recent escalation of fear of witchcraft in Africa. In Soweto in the 1990s, magical healers admitted that clients regularly asked them for spells with which to kill, and there seemed to be a black market in witchcraft equivalent to that in drugs in other parts of the world.²²¹ Proven cases have occurred elsewhere in South Africa of people being killed so that their body parts could be used in evil magic.²²² Among the Kamba of Kenya, magicians who normally market their powers for benevolent purposes are known often to sell the materials for curses, especially for the pursuit of neighbourhood feuds; though the main product concerned overlaps with literal poison, being a potion slipped into food.²²³

If it is certain, therefore, that some people do try to work destructive magic within their own communities at the present day, and that some are likely to have done so among tribal societies in the pre-colonial past, it is harder to find evidence that some of these actively attempted to live up to the broader image of what a witch was held to be. Certainly, the existence of any of the horrific cannibalistic and amoral witch societies in which many traditional peoples have believed remains entirely unproven, as do the serial murders with which their members were credited. Margaret Field, working among the Gă on the coast of Ghana, interviewed over four hundred women who had been accused of witchcraft, one of whom claimed to have killed fifty people, including her own brother and seven of her children, while another confessed to having caused the deaths of four of her children and one of her grandchildren. Field was left unable to decide whether they had committed any of these crimes in reality, or whether they had only dreamed of doing so, and of gathering with fellow witches.²²⁴ One apparently unequivocal and credible first-hand testimony of active witchcraft in a traditional society was provided to an American visitor by an old Tlingit woman, who described how she heard a Christian missionary preach, and decided that his Devil was stronger than his God. She accordingly became a Satanist, as part of which she stole the hair and pieces of clothing of certain people, including children, and put them to rot in the tomb of a shaman according to one reputed method of destructive magic. The people concerned died, and she felt herself responsible and subsequently confessed herself to be a witch.²²⁵ This all sounds very real, and perhaps was, but it is hard to prove with absolute certainty that this was not also the result of dreams or fantasies. Between 1958 and 1962 a number of Shona women, who had confessed volubly to witchcraft appeared before magistrates in what was then the British colony of Rhodesia. In particular, they claimed to have met naked in the bush at night, called up evil spirits, and travelled through the air or ridden hyenas to the homes of neighbours to bewitch them to death and then eat their flesh. Under questioning, it was discovered that they had dreamed of doing this, and then compared their experiences with others, so that their individual stories were polished into a common and mutually corroborative form. As their cultural tradition was that it was the spirits of witches that left their bodies at night to work evil, there was no obvious discrepancy with the actual sensations of sleep and dream, and the confessions could be made with complete personal belief. In the short term, a reputation for being a witch could enhance the status of a woman in Shona society, in which females were usually repressed.²²⁶

Even witchcraft beliefs that rested on dream or fantasy, however, could still be lethal. In 1942 an American medical doctor called Walter Cannon took an interest in reports, drawn from South America, Africa, Australia, New Zealand, Polynesia and the Caribbean, of tribal people falling sick, and often dying, simply because they thought themselves bewitched. He suggested that the individuals concerned responded to the belief with a sustained terror that made eating and sleeping difficult, weakening the body even while it was flooded constantly with adrenalin. This forced down blood pressure and put stress on all organs, damaging the heart in particular and making any normally sustainable weakness dangerous.²²⁷ Subsequent medical studies served to confirm the reality of the phenomenon of ‘death from suggestion’, broadening it out to a realization that it can result from excessive stimulation of any system of the human body, and that a loss of hope can seriously reduce the capacity of that body to deal with any potentially pathogenic processes.²²⁸ Claude Lévi-Strauss built on the earlier of these studies to construct a classic essay emphasizing the critical role played by absolute belief in the efficacy of magic.²²⁹

This perception can pose one part of a double dilemma for a modern Western liberal rationalist. If a belief in witchcraft means that witchcraft can, in effect, kill, then are not societies with that belief justified in having criminal penalties for it? This challenge is compounded by the other aspect of the dilemma, the question of whether Western societies, in a multi-cultural, multi-ethnic world, should show respect for the differing traditions of others, and accept that witch-hunting is intrinsic to their identity and world view, and so none of the business of outsiders. Indeed, it could be argued that they should recognize that it may in fact be appropriate to their needs. This dilemma was thrown sharply into relief in the late 1990s by a debate in South Africa kindled by the report of the Ralushai Commission, a panel appointed by the government of the Northern Province to consider solutions to the burgeoning amount of witchcraft-related violence in the province following the end of white rule.²³⁰ The members, academics and magistrates, were almost wholly drawn from native peoples. Their report, issued in 1996, argued for a new approach that judged Africans according to African understandings of reality, embodying the ideas that witchcraft was objectively real and that a belief in it was a hallmark of traditional African identity. One member of the commission, Professor Gordon Chavunduka, spoke for the literal reality of most of the characteristics of witches as portrayed in native tradition, including membership of initiatory societies dedicated to evil; he only expressed a lack of certainty that witches rode hyenas.

The report called for cases to be tried henceforth in customary courts, by chiefs acting with the advice of service magicians, who could impose prison sentences or fines on those convicted, and lesser penalties for those who brought false or unreasonable accusations. Such suggestions, and those of the report itself, foundered on concerns about whether the courts would be regarded as bewitched if they failed to convict, how guilt could be empirically established, and whether such legal action would tone down or worsen fear of witchcraft. The greatest stumbling block was the problem of agreeing on a set of professional criteria that would regulate and evaluate traditional benevolent magic, and so make its practitioners appear competent as expert witnesses. Moreover, there seemed to be no easy way of extending such a law to cover white South Africans, who do not believe in witchcraft, or of exempting them from it without establishing a new kind of apartheid. In the end, the central government of South Africa decided to ignore the report, encouraging instead local reconciliation procedures to deal with suspicions of witchcraft, procedures reputed to have reduced violence since the 1990s.²³¹ The present book is openly and wholeheartedly in favour of this policy, and of the concomitant and longer-term one, however expensive and onerous it may be, of worldwide state-sponsored educational programmes to persuade people out of a belief that destructive magic is effective irrespective of the credulity of the victim.²³² If it is true, as it seems to be, that people who are utterly convinced that they have been genuinely cursed or bewitched can suffer physically, and even die, as a result, then the only really sure way of rendering them safe is to remove that conviction and halt the effect. This same process of re-education would also provide a long-term absolute remedy to the desire itself to curse, and with it to the murders designed to acquire human body parts for use in destructive magic. It would not, on the other hand, necessarily call into question the use of magical operations intended for benevolent purposes, as the same effect, of belief in magic often rendering it potent, could still apply among those involved in the processes; but for good, and with the understanding that the voluntary complicity of the human subject of the operation would be needed for its success. By such a process, however difficult, laborious and protracted, the world may eventually be delivered from an ancient horror, which has caused much division and misery through the millennia for peoples who have conceived and nurtured it. That should be an ambition as important and meritorious as the eradication of smallpox and polio.

Conclusions

It should be plain enough by now that the five basic characteristics of the early modern European stereotype of a witch can all be found around the globe, although not among all of its inhabitants. It may therefore be worthwhile to emphasize the two respects in which Europe stands out as anomalous. The first is that it was the only continent in which natives developed the common equation between witchcraft and essential evil into the idea that it represented an organized heretical anti-religion, dedicated to the worship of an embodied principle of evil in the cosmos. This was because the dominant religion of medieval and early modern Europe was Christianity, which during this period placed an unusually heavy emphasis on a polarized opposition between utterly good and utterly bad powers in the universe, of which its own god represented the former, and ultimately the more potent. The European development of witchcraft beliefs thus represented a natural concomitant of this unusual theology, though not a necessary one. It was to have knock-on effects upon the rest of the globe, as the European Christian idea of the satanic witch was communicated to peoples who were conquered by Europeans or received them as missionaries, as described.

The other extraordinary feature of Europe was that it became the only area in the world to contain societies which had traditionally believed firmly in the reality of witchcraft and yet which came spontaneously to reject that belief, at least in official ideology. This again had profound effects on the remainder of the globe, as European colonial administrators enforced that formal disbelief on traditional peoples to whom it came as a shocking, unwelcome and alien concept, with consequences that have been discussed. There are, it is true, qualifications and limitations inherent in this modern European scepticism. One of these is that a very long and arduous campaign of education and enforcement was needed, all over the Continent, to persuade most ordinary people of the truth and utility of the official change of attitude. It extended in most countries from the eighteenth to the twentieth centuries, and is even now not wholly complete.²³³ Another qualification is that an active fear of witchcraft has recently been reintroduced to the West among communities of immigrants from ethnic groups, especially African, who have traditionally harboured, and retained, it. In the first twelve years of the twenty-first century, this became a concern of the British police, who investigated eighty-three child abuse cases provoked by suspicion of witchcraft, including four murders, and still thought the problem significantly under-reported. The Metropolitan division set up a special task force, Project Violet, to tackle it.²³⁴

The early modern European image of the satanic witch lives on in its homeland, moreover, in a secularized form. The panic over satanic ritual abuse of children that erupted in North America during the 1980s and crossed the Atlantic to Britain at the end of the decade, was, as Jean La Fontaine has demonstrated in detail, based firmly on the early modern construct of an international, devil-worshipping sect concealed within Western societies. It was, however, repackaged in a form suitable for rationalists, such as many of the social workers (and in America, teachers and police as well) who became persuaded of its truth. This required no literal belief in the existence of Satan or of magic, merely a continued credence in that of well-organized groups of practising Satanists who were dedicated to the committal of antisocial and criminal acts, and so deserved to be exposed, supressed and punished. This credence was enough to produce some dreadful miscarriages of justice, on both sides of the Atlantic, before careful investigation revealed a complete lack of evidence for such a Satanist conspiracy.²³⁵ However, some of those who propagated the panic over alleged satanic ritual abuse, and most of those who did so at its formative stage, were fervent evangelical Christians of a traditional kind, with a very literal belief in a Devil. The same belief is a hallmark of another relatively recent development, the Christian ‘deliverance ministry’ in Canada and the United States, which has depended on a straightforward credulity in demonic possession, sometimes accompanied by one in satanic witches.²³⁶ The fact that members of this have not so far extended their activities into calls for renewed witch-hunts may be attributed to their ability to draw a line between private conviction and public policy; but it may also depend to some extent on the lack of any willingness on the part of governments to pay them heed. Furthermore, these alterations in Western culture have, in turn, effects on other parts of the world. For most of the nineteenth and twentieth centuries, European missionaries to extra-European peoples tended to discourage traditional beliefs in witchcraft, as an aspect of backwardness and barbarism; although, as said, this could be undermined by the fact that traditional translations of the book which they distributed as the word of their deity encouraged such beliefs. In recent years, some American missionaries visiting African peoples have, however, begun to encourage a literal acceptance of the existence of demons and of witches, and to reinforce the resurgence of witch-hunting.²³⁷

For the purposes of the present book, the most significant outcome of a survey of extra-European beliefs with regard to witchcraft is the value that it may have for an understanding of early modern mindsets and the witch trials that these generated. From the worldwide patterns that it has revealed, one would expect early modern Europe to manifest distinct fluctuations over time in the intensity of witch-hunting, linked to economic, social and political change. It would also be reasonable to expect distinctive regional variations in the nature of trials, both in their number and intensity and in the nature of the persons accused, with regard to status, age and gender. Another natural expectation would be for Europeans to distinguish between different kinds of magical practitioner, not merely with regard to the benevolence or malevolence of their operations but also the nature of those. It is hardly going to be news to anybody at all aware of the results of research into the subject to confirm immediately that all of those expectations are in fact correct. What may be novel, and is another outcome of a global perspective on the subject, is to enquire whether such differences in European belief and practice may be rooted in ancient ethnic and cultural differences, corresponding to those between particular tribes, polities and language-groups; and also what difference historical change made to those ancient traditions. That enterprise will be the purpose of the rest of this book.