Introduction: New Language, Old Obsessions

International human rights activists have recently become interested in violence perpetrated against accused witches, in Africa and elsewhere. The issue has received attention from, among others, the U.N. Special Rapporteur on Violence Against Women (2002), the U.N. Special Rapporteur on Extra-Judicial Killings (UNHCR 2009), the U.N. High Commission on Refugees (Schnoebelen Reference Schnoebelen2009; UNHCR 2004, 2009, 2011), UNICEF (Cimpric Reference Cimpric2010), and the European Parliament (Hanson & Ruggiero Reference Hanson and Ruggiero2013), as well as NGOs such as Save the Children (Molinar Reference Molinar2006), HelpAge International (2011), and Stepping Stones (Foxcraft Reference Foxcroft2009). The latter NGO was founded by Gary Foxcroft, a young Englishman, specifically to address the plight of Nigerian children accused of witchcraft; Foxcroft has also been active in founding the Witchcraft and Human Rights Information Network (www.whrin.org). International conferences have been held to address violence against witches. Christian clergy have convened committees.^{Footnote 1} Campaigns have been launched.^{Footnote 2} Money is being raised.^{Footnote 3} A novel has been published (Ormsby Reference Ormsby2013). Law review papers proliferate (e.g., Ludsin Reference Ludsin2003; Mgbako & Glenn Reference Mgbako and Glenn2011; Quarmyne Reference Quarmyne2011; Tebbe Reference Tebbe2007). And documentaries have been made to raise awareness of the issue, such as Saving Africa’s Witch Children about Foxcroft’s work in Nigeria (Gavan & van der Valk Reference Gavan and van der Valk2009) and The Witches of Gambaga about a “witch camp” refuge for accused witches in northern Ghana (Badoe, Reference Badoe2010).

Lurid tales of grotesque violence serve as a rallying cry for the well-intentioned in their efforts to curb what they call “witchcraft violence” or “witchcraft abuse” (see Foxcroft Reference Foxcroft2009), or, more recently, with a view to a catchy acronym, “witchcraft accusation and persecution” or WAP (WHIRN 2013) as in this brochure advertising a two-day course in Oxford for lawyers working on asylum cases relating to witchcraft:

A video of a group of accused witches being burned to death, probably in Kenya, has been circulating on the Internet in recent years, having been viewed more than a half million times at the time of writing (see www.liveleak.com/view?i=dae_1236854361). Refugee rights groups and asylum lawyers are also dealing with people, not only Africans, fleeing persecution as “witches” and seeking asylum in countries where witchcraft is not high on the list of public concerns and generally not well understood.^{Footnote 4}

In all these discussions, the issue of justice is presented as a matter that is so clear and unambiguous that it hardly needs to be raised: innocent people are being persecuted with “torture, starvation, abandonment, and death.” Worse still, the most affected are members of “vulnerable groups,” particularly women, children, and the elderly (see Adinkrah Reference Adinkrah2004; Federici Reference Federici2008; Mgbako & Glenn Reference Mgbako and Glenn2011; Quarmyne Reference Quarmyne2011; Secker Reference Secker2013). As one outspoken activist put it, “Witchcraft in much of the developing world—Africa, India, Southeast Asia, the Pacific, Central and South America—is at the core a human rights issue because millions of people are victimized for bogus, non-rational, non-scientific reasons” (Norman Miller, quoted in Barber Reference Barber2012). Something, surely, must be done.

Most Africans who worry about witchcraft, however—which is just about everyone at one time or another—see witchcraft as raising a very different issue of justice, one that goes beyond the simple problem of false accusation and involves harm done by witches to their victims. Perpetrators of witchcraft, from this perspective, are dangerous and powerful figures, not members of vulnerable groups, however frail-seeming they may be. Witches are said to cause illness, death, suffering, and misfortunes of all kinds. They are, in essence, perpetrators of criminal violence, albeit of a particular kind. For people who live in a world with witches, the dangers they pose are real and present. In such places the primary concern is security: how to protect against witchcraft and other forms of violence. A secondary issue is justice: what to do about perpetrators of harm when injury has been done. For people who live with the fear of witchcraft, the issue of justice in relation to witches is clear. Innocent people are being attacked, harmed, and killed. The question, then, is: how can the perpetrators best be identified, disarmed, and, if necessary, punished? Historically, this was a central problem facing political authorities. It remains so to this day.

The question I want to address in this article is: Can the rights of accused witches be protected while justice is done, and seen to be done, for their victims? That is, can there be justice for both accused witches and fearful communities in the era of human rights? Or, to put it in a slightly different form: Is it possible to protect the rights of an accused witch while still securing justice for a community that sees itself at risk of occult assault? A reading of the recent literature on the subject of “witchcraft violence” and human rights would suggest that the answer to these questions is a resounding No. From this perspective, the crimes of witches are imaginary, so the question of justice for their victims is moot.

If it is indeed the case that it is impossible to protect the rights of accused witches while serving the ends of justice for a fearful community, then the outlook for protecting accused witches as well as promoting respect for human rights in Africa is dim. For the victims of putative witches far outnumber those accused of witchcraft, and there is no sign that the plausibility of witchcraft narratives is diminishing in contemporary Africa. In this article, however, I shall argue that the question of justice need not be dismissed out of hand; that while there is a regrettable amount of suffering imposed on suspected witches, there is far more work taking place in everyday contexts as well as indigenous and formal judicial institutions that addresses the problem of witchcraft in ways that not only uphold justice and rights, but also resolve conflicts (see Kgatla et al. Reference Kgatla2003; ter Haar Reference ter Haar2007; Crampton, Reference Crampton2013).

An obsession with spectacular aspects of anti-witchcraft violence obscures attention to developments that are taking place around the continent in the management of spiritual insecurity by Africans themselves, notably in forms of enthusiastic religion and new forms of jurisprudence of the sort Jean and John Comaroff (Reference Comaroff and Comaroff2004) call “organic African jurisprudence” (see also Fisiy Reference Fisiy1998; Fisiy & Geschiere Reference Fisiy and Geschiere1990; Fisiy & Rowlands Reference Fisiy and Rowlands1990). In earlier work I wrote about the struggles of people accused of witchcraft in postapartheid Soweto, where there were no institutional forums for dealing with their problems. In one case my friend Madumo, accused of killing his mother with witchcraft, pursued a purely private resolution of his problem in the form of traditional healing to overcome what he saw as a “curse” (Ashforth Reference Ashforth2000). In another case a friend, having searched in vain for a forum wherein the conflict could be resolved, was hounded to death by an accuser (Ashforth Reference Ashforth2005). I have also written of religious enthusiasm as a form of engagement with powers deemed capable of overcoming spiritual insecurity (Ashforth Reference Ashforth2011). Here I want to explore the question of justice in contemporary Malawi, where institutions dealing with cases of witchcraft do exist and are creating new forms of local jurisprudence to handle such accusations.

The key to this emergent jurisprudence is a shared commitment on the part of judicial authorities, in both formal and informal judicial forums, to the need for demonstrable evidence to sustain accusations of witchcraft and an insistence on a distinction between being a witch and performing witchcraft. This distinction echoes one that is operative in indigenous narratives of witchcraft and finds unexpected reinforcement in a judicial interpretation of the colonial era Witchcraft Act (which is still in force). In terms of this distinction, being a witch is a matter of shame and suspicion, but not, in and of itself, a crime; causing injury and harm by occult means, however, is unambiguously criminal, though devilishly hard to prove. Insisting on proof in cases of witchcraft, as the English experience of witch trials in the seventeenth century shows (Shapiro Reference Shapiro1991), amplifies doubt in ways that can be more protective of suspected witches than an insistence on the nonexistence, and hence irrationality of believing in, witches and witchcraft. Or, to put it another way: epistemology trumps ontology when it comes to protecting African witches.

In terms of this emergent jurisprudence, the rights of a person accused of practicing witchcraft are the same as those of anyone accused of a crime: habeus corpus, fair trial, no cruel and unusual punishment, no torture, and so on. And, as the history of witch trials in England, as opposed to continental Europe, shows, without torture, confessions and thus convictions for witchcraft are hard to come by (Langbein Reference Langbein1977). Indeed, under the terms of the Universal Declaration of Human Rights and the Malawi Constitution, along with most other recent constitutions, the right of persons to identify themselves as witches, if they so desire, can be upheld in terms of the right to “freedom of belief”—even when such persons are widely despised.^{Footnote 5} Further, the widespread commitment to the necessity of evidence in support of accusations of witchcraft and the requirement of proof “beyond a reasonable doubt,” at least in the formal legal system, open avenues for resolving conflicts involving accusations of witchcraft through discourse and dialogue.

To be clear: I shall be arguing that current humanitarian efforts regarding violence against suspected witches in Africa are misplaced. I am not opposed to acts of kindness, by Africans or outsiders, in the protection of people suffering cruelty in whatever form. I also am not engaging in a critique of the universalism of human rights discourses (see Mutua Reference Mutua2002), though much of the recent writings on saving Africans accused of witchcraft, especially the children, consist of veritable caricatures of the notorious triad of victim, savage, and outsider hero. But I am suspicious of interventions such as campaigns or policy initiatives premised on a presumption of superior knowledge and wisdom in relation to matters of witchcraft. I would argue that violence against and mistreatment of suspected witches reflect a failure of institutions of security and justice. So if there is an increase in violence against suspected witches—as many human rights groups are asserting, albeit without any hard evidence—then it is unwise to advocate policies that further undermine the possibilities of collective justice. And even more interesting than these issues is the question of how Africans are dealing with matters of security and justice relating to witchcraft in local communities.

Violence in a World of Witches and the New Anti-Anti-Witchcraft

The central feature of all talk of witchcraft, for those who fear the power of witches, is that it is a way for a person, or persons, to inflict harm on others. Witchcraft, in short, is a form of violence. Suppositions about matters of “witchcraft” have been found throughout all recorded human history and in virtually every place humans have been found, even if, in recent centuries, a section of the human community has prided itself, perhaps prematurely, on transcending the superstition and irrationality of talk of witchcraft.^{Footnote 6} Vast bodies of literature document varieties of witchcraft. Incalculable amounts of time have been invested in documenting putative “systems of belief” pertaining to particular human collectivities and in attempting to describe and explain their peculiarities in relation to things like “social structure.” Much ink has also been spilled speculating about the connection between gender relations and witchcraft, seeking to explain why women predominate in the ranks of putative perpetrators. In all this talk and writing about witchcraft, however, one feature always recurs: speaking of “witchcraft” on the part of people who live in a world with witches, in all the varieties of terminology that are used, is a way of construing injury as harm; that is, as a hurt deliberately inflicted on a victim, or victims, by a perpetrator or perpetrators. Witchcraft is a form of violence, albeit one perpetrated in secret with invisible—or, for want of a better word, occult—means.

For people who do not live in a world of witches, one of the most difficult issues to comprehend is the ontology of witchcraft as violence, as a variety of the multitudinous harms that humans are wont to inflict on one another. For such people witchcraft is a matter of “belief,” an issue pertaining to the domain of the cognitive.^{Footnote 7} For people living in a world with witches, however, there is no essential ontological difference between the violence perpetrated by witches and the ordinary forms of violence perpetrated by physical means. Nor are they speaking metaphorically when referring to the violence perpetrated by witches. Witchcraft is also not merely an idiom through which other forms of conflict are expressed. The violence of witchcraft, for those who live in a world with witches, is every bit as real as any other form of violence; only the means and forms differ. In Africa, most Africans would agree, witchcraft is widespread.

Nevertheless, violence against witches, of the sort animating the current surge of interest in human rights circles, is rare. Although fears and suspicions of occult assault are widespread, open accusations of witchcraft are the exception in most African communities; still more rare are actual assaults on suspected witches. A report by the Witchcraft Human Rights Information Network, an organization devoted to raising alarm about the issue, found “a minimum of 865 will have experienced violations of their human rights in 2013 due to belief in witchcraft and other malevolent beliefs” (WHIRN 2013:3). Given the scale of misery on planet earth that year, these are not huge numbers. Anti-witchcraft violence, when it occurs, is mostly perpetrated by angry mobs, usually composed of young men (see, e.g., Commission of Inquiry into Witchcraft Violence and Ritual Murders 1996). Most people deplore such attacks, seeing them as evidence of a breakdown in social institutions that should protect communities both from the harms perpetrated by witches and the violence of the attackers. Violence against suspected witches is little different from other forms of vigilante vengeance; it flourishes in the absence of institutions serving justice (Taussig-Rubbo Reference Taussig-Rubbo2011).

Many human rights activists, inaccurately, call attacks on suspected witches “witchcraft violence,” “witchcraft abuse,” or “witchcraft-based violence,” and often lump it in, even more inaccurately, with acts of violence such as the notorious muti murder phenomenon of South Africa, or attacks on albinos in Tanzania, where people are killed and body parts harvested for magical purposes (see Cohan Reference Cohan2011; Hund Reference Hund2003).^{Footnote 8} Witchcraft violence, however, is violence perpetrated by witches. Violence against putative witches should properly be termed anti-witchcraft violence. The humanitarian interest in this violence, then, should be described as “anti-anti-witchcraft” violence, although I use the term somewhat sardonically, with a nod to Clifford Geertz (Reference Geertz2000) and his stand on “anti-anti-relativism.” But I also want to raise the logical question that the term suggests: does being anti-anti-witchcraft mean being pro-witchcraft? I would suggest that it does—at least when viewed from the perspective of people living in a world with witches.

While anti-witchcraft violence is rare, the violence perpetrated by purported witches is common. In a 2012 study of “witchcraft-based violence” in Malawi conducted by the Association of Secular Humanists (ASH) and funded by the Norwegian Embassy, Charles Chilimampunga, a sociologist from Chancellor College, and George Thindwa, executive secretary of ASH, found that out of a sample of ninety-four “key informant” interviewees, ten had considered themselves victims of witchcraft attacks in the past two years. Of these, seven claimed to know the identity of the attacking witch, although only one directly confronted the assailant. Most victims and their families sought help from healers and pastors—of whom there are countless thousands in Malawi, as elsewhere in Africa, offering services of healing and protection. Of the witches who were openly accused—and the researchers found only fifteen in their sample of 638 households surveyed—eleven (73%) reported being subjected to some sort of “abuse” (Chilimampunga & Thindwa Reference Chilimampunga and Thindwa2012:61).

Although the ASH study did not draw on a random sample of the national population, the proportion of injury to insult among people who claim to have been bewitched is instructive. Remember: for every ten persons who found themselves injured by witches, only one took direct action against the suspected witch. By this reckoning, then, the problem of violence by witches is more than ten times greater than the problem of violence against witches—unless, of course, you believe the putative victims of witchcraft are merely imagining the harms they say they have suffered.

Most communities in Malawi, as elsewhere in Africa, manage most of the time to contain suspicions of witchcraft and deflect assaults on witches. Although the dangers and damages of witchcraft are deemed material, most people deplore violence. When well-respected figures are able to speak with authority in a community, it is usually possible to prevent mob justice from flaring up against suspected witches (van Beek Reference van Beek and Haar2007). Moreover, despite seeming “vulnerable” to outsiders, witches are feared for their power to cause harm, so it is usually prudent to protect against them rather than seek to punish them. Indeed, it is not always a bad thing to be thought a member of a group with extraordinary powers—especially, perhaps, for the otherwise vulnerable.^{Footnote 9} Chilimampunga and Thindwa, for example, quote a man who felt himself to be a victim of witchcraft but was advised to do nothing against the perpetrator “because of fear of what might happen next to him” (2012:61). Most of the retaliation against witches takes place through the mediums of traditional healing, whereby healers promise to deflect the damaging forces back onto the perpetrator, and of gossip, which undermines reputations.

Many communities, however, do not have authority figures capable of resolving issues of witchcraft when accusations surface. Many are riven by conflicts between competing authorities, leaving victims of witchcraft in confusion about where to turn for help and a sense that their only hope is to take matters of justice into their own hands. Sometimes political and religious leaders have a direct interest in fomenting conflict and hostility. Witch hunts do happen. Chilimampunga and Thindwa estimate that in Malawi, a country of about 16 million, violent witch hunts took place about once a week in 2011, the year preceding their study.^{Footnote 10} Sometimes anti-witchcraft movements can sweep whole regions, causing large-scale damage and disruption, as in the Northern Transvaal region of South Africa in the 1980s when, by some estimates, several hundred suspected witches were killed by youthful vigilantes in the context of the struggle against apartheid (Ralushai Reference Ralushai1996; Ritchken Reference Ritchken1988).

There are no reliable statistics concerning the prevalence of witch killing in Africa and few explanations for the variations that occur across time and place. One of the more systematic studies, by the economist Edward Miguel (Reference Miguel2005), found that a woman over the age of fifty in the Sukumaland region of Tanzania, the area with the highest levels of witch killings in that country, had a one in five hundred chance of being killed in a given year.^{Footnote 11} Nonetheless, it is important to remember that violence against witches is the exception rather than the norm. Every community knows its witches, and they are mostly left unmolested. Knowledge of who they are circulates within communities in the form of rumor and gossip (Ashforth Reference Ashforth2005; Stewart & Strathern Reference Stewart and Strathern2004). The woman everyone knew as the witch in the neighborhood where I used to live in Soweto, for example, lived peacefully with her family and died of old age in her nineties. Being a witch, it seems, is one thing. But doing the things witches do—such as killing innocent victims or teaching the trade to children—is another.

Why Malawi?

Malawi is a good place to examine issues of justice and witchcraft. It is a peaceful and relatively stable, if very poor, democracy since the founding president, the old-school African dictator Hastings Kamuzu Banda, was ousted in 1994 in a surge of democratic sentiment catalyzed by an encyclical letter from Catholic bishops (see McCracken Reference McCracken1998). It has a court system that is reasonably functional, at least in urban areas, and unlike in South Africa, where the courts are still dominated by the descendants of white settlers, its courts are thoroughly Africanized. According to the National Victimization Survey of 2004, the justice system also enjoys a high degree of legitimacy in the public’s eyes (Kanyongolo Reference Kanyongolo2006).

After independence President Banda retained the colonial legal system founded on English Common and Statute Law. The law continues today to be shaped by English procedures, practices, and precedents, albeit in an African context where the so-called Traditional Authorities (TAs) also hold sway in local affairs and exercise considerable juridical authority (see Cammack, Kanyongolo, & O’Neil Reference Cammack, Kanyongolo and O’Neil2009; Kapindu Reference Kapindu2009). Principles of customary law occasionally intrude into the formal courtrooms, particularly at the lowest levels. Despite recent efforts at restructuring the court system to formalize the relationship between customary jurisdictions and the formal system, a great deal of confusion remains (Kaunda Reference Kaunda2011).

During the reign of President Banda (from 1964 to 1994), a system of Traditional Courts with jurisdiction over both customary and criminal cases was established in parallel with the formal courts. These courts, which were notorious for being used by the regime to punish political dissidents, were disbanded in 1994, leaving customary cases in the hands of informal chiefs’ courts. Chiefs’ courts mostly handle family disputes and land issues, along with most of the other disturbances of peaceful life in the roughly twenty-four thousand villages in the country (Schärf et al. Reference Schärf2002). Land disputes, it should be mentioned, are becoming more intense and more common in this densely populated country, often exacerbating claims of witchcraft in the process (Peters Reference Peters2002). In 2011, in an effort to better coordinate relations between the domains of customary and formal law, the government introduced legislation to establish Local Courts, once again with criminal as well as customary jurisdiction (Kaunda Reference Kaunda2011). Implementation of this legislation, however, has been slow, and the two main domains of law remain the chief’s court under the tree, and the Magistrate’s Court in the town.

Malawi has been the subject of some classic writings on the issues of law and custom in Africa (e.g., Chanock Reference Chanock1985), which provide the historical perspective to the cases discussed here. It is also the scene of an ongoing project being implemented by the ASH and generously funded by the Norwegian government to protect the rights of witches by “sensitizing” various authorities, such as magistrates, police, and traditional leaders, to the prohibitions against anti-witchcraft activities contained in the colonial-era Witchcraft Act. In other words, the ASH has given its support to an approach that is diametrically opposed to the one that I advocate here. Convinced that witchcraft does not exist, and intent on protecting the rights of the accused, its project seeks to make it impossible for any institution in the country to respond to the desire for justice among people who feel themselves victims of witchcraft. I would advocate an approach that respects the desires of communities for justice and security in relation to witchcraft while strengthening the commitment to evidence among those called on to adjudicate cases.

Before we address the implications of these new forms of anti-anti-witchcraft activism, let us examine some contemporary witchcraft trials from Malawi’s courts. My cases are drawn from the records of proceedings in the Balaka Magistrate’s Court.^{Footnote 12}

Two Recent Witchcraft Cases

Truth and Justice in a World of Witches: The Problem of Proof

At the heart of Magistrate Banda’s judgments is a distinction between being and acting, between the witch as a particular kind of being and “practicing witchcraft” as a distinct form of action. Both Mrs. K. and the elder Mr. Kasito were found by the magistrate to probably be witches. In neither case, however, was there sufficient proof that they were guilty of practicing witchcraft in the specific form of teaching children. The distinction Magistrate Banda deployed in his judgments resonates with common usage in discourse pertaining to witchcraft in Malawi as in most other places where people live in a world with witches. In the Malawi vernacular of Chichewa, for example, mfiti refers to a particular kind of person—human, superhuman, and subhuman at the same time—inhabiting a particular mode of being. This is the witch, or sometimes “wizard” (when the speaker takes pains to establish gender), distinguished by virtue of belonging to a different category of being from otherwise ordinary humans. The key to this mode of being is the ability to operate within a space ordinary persons experience as the domain of dreams. Matsenga, on the other hand, refers to witchcraft as a mode of action perpetrated by an ordinary kind of person who does not necessarily possess the forms of personhood indicative of the mfiti. Anthropology knows this as the distinction between the witch and the sorcerer (see Evans-Pritchard Reference Evans-Pritchard1937), though this is misleading as it emphasizes a distinction between two different modes of action rather than between acting and being.

As a form of action, matsenga witchcraft is typically said to derive from specialized knowledge of the powers inherent in material substances. In principle, anyone—at least, any adult—can practice this form of witchcraft; children are generally not said to be involved. All one needs is the knowledge and the material, variously known as mankhwala, charms, herbs, medicines, or muti, which are presumed to be purchasable from specialists. Herbalists ply their wares openly in markets and on the side of the road outside hospitals. For the most part they are presumed to be in the business of healing, though there is always suspicion about whether the knowledge involved in healing may also be deployed for more nefarious purposes.

Nobody publicly proclaims any involvement in the trade of witchcraft substances, although no one doubts that the trade is widespread. The putative knowledge underlying matsenga (as well as cognate forms of sorcery across the continent) is often referred to as “African science,” which is presumed to be the inverse of the healing “science” practiced by traditional healers. Without getting too involved in an exposition of systems of belief, however, we should note that it is equally plausible to suppose that Satan imparts witchcraft knowledge to his followers, or that “traditional” practices are really demonic manifestations, as many African Christians readily assert.

Stories about matsenga typically revolve around a perpetrator motivated by jealousy and hatred to harm a victim known to him or her personally. They are stories of murder and attempted murder, of violence perpetrated on victims by their enemies. Perpetrators are typically in relatively close proximity to, if not intimacy with, the victims—typically relatives, neighbors, coworkers, and schoolmates. Victims of this kind of mystical violence get sick, die, suffer accidents, lose jobs, or experience one of the other misfortunes that we humans wish upon each other in the daily course of living together. A perpetrator of this kind of witchcraft has not necessarily gone completely over to the dark side, so to speak. He or she might use charms to kill a neighbor who has provoked feelings of jealousy while remaining a valued coworker, a loving father, and an otherwise upstanding citizen.

The mfiti witch, on the other hand, is addicted to evil and abominable practices. To be an mfiti is to be a particular kind of man or woman. The mfiti kills for the love of killing, or to sate a lust for feasting on human flesh. He or she kills to repay commensal gifts of the flesh of other witches’ kin. These witches want to bring misery and death to others simply because they can. The mfiti has extraordinary powers, capable of amazing feats like flying through the night on a witchcraft airplane while leaving a physical body asleep in bed at home. These witches enjoy the company of fellow witches and seek to convert others to their vile practices, including their abominable sexual predilections, picking on children since they are least able to resist. The mfiti, in this way, occupies a distinct status and represents a particular mode of being: at once human and not human, both superhuman and subhuman. Yet—and this point is crucial for the attribution of criminality—a person might be a witch and yet not be practicing witchcraft, or even know she is such a creature. The mfiti, then, is not the sort of being to which the notion of “human rights” applies without a considerable degree of ambiguity.

If establishing that a person is—or at least is generally considered to be—a witch is relatively straightforward, proving that a particular misfortune is the result of such nefarious actions is another matter. So, how might the truth of accusations regarding the practice of witchcraft be proved? From observations of Magistrate Banda’s court, as well as others, proof of the practice of witchcraft would seem to have some of the following elements.

First, since witchcraft is not a victimless crime, proof of witchcraft requires testimony or a complaint by, or on behalf of, a victim or victims. In the first case above, for example, Masautso Kasito, who lodged the complaint with the police about witchcraft in his family, was suffering pains the herbalist had told him were caused by witchcraft. His brother, the same healer told him, had been killed by witchcraft. Young children in his family were telling of being taught witchcraft. There were plenty of victims, past, present, and future. Witchcraft is generally detected retrospectively. Once misfortune strikes, the questions, classically framed by Evans-Pritchard, arise: Why me? Why now? When these are posed in a context in which witchcraft is a viable hypothesis, the possibility of occult assault must be considered.

The existence of a victim who has suffered harm, while a necessary condition for a witchcraft narrative to be framed, does not, in and of itself, constitute proof. When the witchcraft hypothesis is presented, the next question is: who is the likely perpetrator? Since witchcraft is rarely the work of strangers, the usual suspects readily emerge: enter the elderly Mr. Kasito. But even the suffering of the victim and the existence of a likely suspect are not sufficient for the matter to rise above the level of mere quarrel to enter the domain of crime. Whether or not the suspicion of witchcraft is confirmed by the actions of a witch will depend on the evidence.

Although witchcraft is practiced in secret, there can, in theory, be direct evidence of its use. Suspicious substances may be found in places where they don’t belong. Unusual behavior, such as tiredness in the morning (which could be interpreted as resulting from flying at night), or loss of appetite (from eating human flesh), or being caught naked after dark in a neighbor’s yard (which could be taken as resulting from the crash of a witchcraft airplane) are all, in principle, open to verification, though the interpretations might be contestable. From time to time people engaged in a quarrel have been known to utter threats of occult violence, the standard wording of which is “you shall see.” Such a threat, publically witnessed, followed by untoward events among those subject to it, would be an obvious clue. Since jealousy, anger, and hatred are considered the primary motives for perpetrating harm by means of witchcraft, a history of enmity and quarrel between individuals or families could be quarried for traces of malicious motives. In close-knit communities where people live with witches, moreover, everybody comes to know who they are. Over time, gossip, rumors, suspicions, and accusations accumulate so that when misfortune strikes, there is rarely any difficulty in identifying the suspects (Ashforth Reference Ashforth2005; Elias Reference Elias1956; Stewart & Strathern Reference Stewart and Strathern2004). Proving that the person suspected of witchcraft actually performed the deed, however, is tricky. As Magistrate Banda insists time and again in his court, mere testimony, or “belief,” is not sufficient proof in court. And the testimony of children, though necessary in cases such as those discussed above, is inherently problematic and never sufficient.

Although the means by which witches act are invisible to ordinary mortals, healers are generally considered to have special abilities to perceive the actions of witchcraft. Healers’ claims, however, are eminently contestable. While almost everyone accepts the general possibility of healers being able to see the unseen, every particular claim is open to question. Testimony by one healer, moreover, can be undermined by that of another. Since suspicions and accusations of witchcraft typically arise within intimate social networks (Geschiere Reference Geschiere2013), most people agree that “objectivity” in the evidence of ritual specialists is secured by recruiting a healer from far outside the circle of familiarity with any particular case. Greater confidence in the testimony of healers also arises from concordance among the readings of several unrelated healers. In the case of the Kasito family, while no healers testified in court, several witnesses gave reports of the results of divination. These reports were given considerable weight by Magistrate Banda in exonerating the second defendant, Annie. In his jurisprudence, however, the testimony of a healer alone would be insufficient to convict a person accused of witchcraft unless the evidence was independently corroborated.

Ensuring the truth of testimony is a challenge in all jurisdictions, and courts frequently resort to supernatural sanctions for support. In many parts of Africa “ritual” oaths, in which individuals attest to their honesty at the risk of punishment from supernatural beings, either upon their own persons or those of their families and clans, have often been used as a proof of honesty and have been accorded centrality in judicial practice (Kirk-Greene Reference Kirk-Greene1955). I have not encountered such oaths in accounts of hearings in Malawian Chiefs’ Courts, however, and Gluckman (Reference Gluckman1955), in his classic study of Barotse law in the 1940s in neighboring Zambia, remarked on the absence of oaths or other supernatural sanctions underpinning claims to the veracity of testimony in courts. Nor is oath-taking a major feature of truth-speaking in the South African context, although early Zulu ethnographies do point to forms of invoking the ancestors as a way of authorizing the veracity of speech (Bryant Reference Bryant1949). In the Magistrate’s Court, however, oaths are administered on the Bible, and in Africa as elsewhere such a pledge engages a person in a potentially dangerous relationship with a supernatural being who will dispense punishment if the truth is not told. In the case of the Kasito family, two of the children were ruled incapable of taking oaths because they were too young to “know God,” although even if they had been judged capable of testifying under oath, their testimony would have required corroboration by others before the magistrate would have considered it proof.

Perhaps the most effective proof of innocence or guilt in matters of witchcraft is the suspected witch’s response to a trial by ordeal of the sort that Bartret Kasito claimed he had undergone. Ordeals were widespread in this part of the world before being banned by colonial law (Chanock Reference Chanock1985), though evidently the ban, still in effect, is less than comprehensive, if the elder Kasito is to be believed. In medieval Europe ordeals involved the grasping of a red-hot iron or the dunking of the suspect in a pool of water; burning or drowning was taken as a sign from God of the suspect’s guilt (Pollock & Maitland Reference Pollock and Maitland1911). In African ordeals, similarly, ritual experts administer a special substance to the accused, whose consequent death or survival signifies the unquestionable judgment of relevant spiritual beings (Douglas, Reference Douglas, Middleton and Winter1963; Marwick & Middleton Reference Marwick and Middleton1967).

The beauty of the ordeal in relation to witchcraft is that it both decides the truth and executes punishment in one move. Suspects who survive, moreover, are restored from the social death that often befalls suspected witches and they can be reintegrated into the community. Colonial authorities worked hard to outlaw such ordeals, succeeding in reducing their prevalence and public scale by subjecting those responsible for them to charges of manslaughter (Chanock Reference Chanock1985). While the experience of Bartret Kasito suggests that ordeals are not quite things of the past, the fact that they are illegal under the terms of the Witchcraft Act, along with the fear of resulting murder charges in the event of the suspect’s death, must surely restrain healers who would administer them. Nonetheless, the ordeal, along with confession, remains a touchstone of objective proof of witchcraft, from the point of view of those who live in the world with witches. Yet even ordeals are open to doubt. As the experience of Bartret Kasito shows, what seemed to him to be a clear-cut ordeal was portrayed in the testimony of one of his sons as a form of cleansing ritual, a way of removing and neutralizing witchcraft, and Magistrate Banda made it clear that the evidence of ordeals had no place in his court. The fact that the old man survived an ordeal was noted, but it was not considered relevant.

Finally, and most important, proof of witchcraft, as we have seen, is obtained legally only with the confession of the accused witch to having practiced witchcraft. Because of the secrecy enshrouding witchcraft, in the absence of confession, doubt as to a suspect’s guilt can always be raised. Even the great fifteenth-century manual of the Inquisition, Malleus Maleficarum, noted that a witch should not be put to death before she confesses (Institoris, Sprenger, & Summers, Reference Institoris and Sprenger1928 [1489]), even if that confession had to be obtained through torture. Confession thus remains the ultimate proof of witchcraft, although few witches willingly confess unless they can be confident of forgiveness. For this reason, most witchcraft confessions in contemporary Africa take place in the relative safety of Christian church congregations, where they thrive because renunciation of a former evil life can add credence to claims of deliverance and salvation. In the cases described above, the imperative of confession for the proof of the witch’s identity is underlined by the fact that Section 6 of the Witchcraft Act outlaws “pretending” to the powers of witchcraft, thus requiring confession in order to secure conviction. To repeat Magistrate Banda’s argument: to say “I am a witch” is a crime, not because being a witch is a crime, but because saying so is.

Yet even though confession serves as the ultimate touchstone for truth in relation to witchcraft, confessions in witchcraft cases—as in others, for that matter—are notoriously unreliable (Kassin Reference Kassin2008; Kassin & Wrightsman Reference Kassin and Wrightsman1985). Even when not extracted under torture, seemingly voluntary confessions can be false; in initiatives undertaken by the U.S. Innocence Project, for example, false confessions were shown to play a role in some 25 percent of cases in which individuals eventually were exonerated with the use of DNA evidence (Innocence Project 2013). In addition to the reasons that suspects in ordinary criminal cases might confess to crimes they did not commit, witchcraft confessions have two other dimensions of complexity. First, given that witchcraft activities take place in a domain not accessible to ordinary consciousness, a person might, in principle, possess occult powers unknowingly and commit acts of witchcraft without being aware of such activities. Second, as the sad case of Mrs. K. demonstrates, it is impossible to prove that one is not a witch. A timely confession, in the hopes of leniency, is thus the optimal strategy for an accused witch, regardless of the truth of the accusation.

Conclusion: The Witch Trial as Restorative Justice

The human rights approach to witchcraft accusations denies their validity and forecloses the possibility of a trial, fair or otherwise. While there is much to be said for a bracing rationalism in all aspects of life, evidence from Africa over the past couple of centuries shows no sign that witchcraft narratives lose their plausibility as a result of people being told that witches do not exist. Rather than insisting that such accusations are, by definition, baseless, a better approach is the one exemplified by Magistrate Banda: hearing the evidence and deciding the case on its merits. By making the distinction between the person of the witch and the action of witchcraft, moreover, the magistrate was able to identify possible sources of perceived danger for the community while insisting on a high standard of proof in relation to judgments of crime.^{Footnote 20}

In this, the magistrate is not alone, for there is a widespread commitment in Africa to the need for evidence in relation to witchcraft accusations, both among authorities and the public more generally. And while it is relatively easy to identify a perceived witch simply by adjudicating the plausibility of narratives relating to a person’s identity, it is much more difficult to use evidence to prove the actuality and efficacy of purported occult actions. Indeed, the growing literature in forensic psychology relating to the testimony of children and the phenomenon of false confession increases our understanding of the difficulty of arriving at definitive proof in such cases.

In practice, when a magistrate hears a witchcraft case with a view to judging the evidence, he is allowing all concerned to tell their stories in ways that facilitate the airing of grievances and dissipation of anger without causing serious harm to the accused, bringing about, in the process, a version of the restorative justice for which African jurisprudence is celebrated. Contrary to the blanket application of the universal principle of “human rights”—under which any accusation of witchcraft must always be presumed false—Magistrate Banda applied a form of casuistry in which each case is considered on its merits, evidence is weighed, and precedents are considered (Jonsen & Toulmin Reference Jonsen and Toulmin1988). If witches were accorded the right to a fair trial of the sort Magistrate Banda conducts—in which the evidence is presented in open court and the verdict is based on reasoning from law and precedent—the accused would have a measure of protection from violent mob justice, as would the community from the fear of witchcraft. Like chiefs adjudicating witchcraft cases in customary justice forums, Magistrate Banda was less concerned with imposing punishment than with achieving a sense of security and justice, if not harmony, in the community. As he told me in our 2013 interview, he has a duty as a magistrate to “advise” or to “give advice” from his position on the bench, not merely to impose the “technicalities” of the law.

The concept of advice generally has a much greater weight in African contexts than it has in contemporary Western usage, more akin to the obligatory “advise and consent” clauses of various national constitutions than the optional take-it-or-leave-it character of the “advice columnist.” Parents, relatives, siblings, and friends, among others, all have obligations to give advice in particular contexts, often with quite formal rules establishing who can speak on what, how, and when. For example, in the Malawian context, marriage counselors known as Ankhoswe—in this region, usually maternal uncles of the husband and wife—after advising on the suitability of the union prior to marriage, are charged with giving advice in the event of marital discord. Failure to heed advice, similarly, is classed as a much more weighty infraction of societal norms than the mere choosing of a different option. It is usually construed not only as the deliberate choosing of the wrong path, the road to immorality, but also a souring, if not rejection, of the proper relationship that ought to subsist between the advisor and advisee, and thus their larger community. By patiently listening to the details of complex conflicts and then weighing the evidence of culpability and blame, Magistrate Banda gave advice that brought about justice.

Magistrate Banda’s court provides a good example of a situation that I suspect is quite common across Africa when courts are confronted with witchcraft cases.^{Footnote 21} African courts since independence have been dealing with issues of concern to local people that are inadequately dealt with in legislation, notably in relation to marriage and inheritance but also in cases pertaining to witchcraft (Van Gijsegem Reference Van Gijsegem2008; Veitch Reference Veitch1971). Since refusing to hear such cases would be impolitic, and since principles of legal evidence make conviction difficult, Magistrate Banda in effect conducts two trials simultaneously. The formal trial adheres to the letter of the law; the verdict is recorded accordingly with the possibility of appeal explained to all parties. The informal trial does not contravene the law, but enshrines somewhat different principles of truth and justice, such as those outlined above.

Drawing on his legal status as magistrate, Damson Banda uses his authority to resolve witchcraft cases by carefully considering the evidence and assigning guilt, albeit not in a formal appealable verdict but rather in a personal opinion delivered from the bench with all the weight of the magistrate’s voice. In this way he uses his authority to resolve the real case, and he uses the case to bolster his authority, his ability to speak the truth from the bench. From the perspective of the new anti-anti-witchcraft activists, these informal witch trials are perpetrating an injustice on people such as Mrs. K. or Masautso Kasito. From the perspective of the community, however, justice is served. Just as important, perhaps, the need for violence is averted.