Shamans, Witches, and Healers: How Cultures Explain Misfortune

In the late 1920s, a young Oxford-trained anthropologist named Edward Evan Evans-Pritchard pitched a tent among the Azande in what was then the Anglo-Egyptian Sudan. He had come to study political organization, the kind of thing his discipline thought respectable: chiefs, courts, lineages, the machinery of authority. He stayed for the witchcraft. Across four field seasons between 1926 and 1930, he kept noticing that whenever something went wrong, a crop failed, a hut burned, a man fell ill, his Azande hosts had a ready explanation, and the explanation was witchcraft. He could have dismissed this as superstition and moved on. Instead he took it seriously as a system of thought and tried to work out its internal logic.

The book that came out of those seasons, Witchcraft, Oracles and Magic Among the Azande, published by Clarendon Press in 1937, became one of the foundational texts of modern anthropology. Its argument is deceptively simple and genuinely unsettling. What looks at first like primitive irrationality is, on close inspection, a coherent and internally consistent way of answering a question that European empirical reasoning leaves wide open: not how misfortune happens, but why it strikes this particular person at this particular moment. Every documented human society has felt the need to answer that question, and the people who answer it, the shamans, witches, sorcerers, and healers, are found everywhere the ethnographic record reaches.

Why Every Society Keeps Specialists of Misfortune

Some things in life can be managed through ordinary practical action. If you are hungry, you find food; if your roof leaks, you patch it. But a great deal of human suffering arrives without any handle to grip. The sudden death of a healthy young person, the illness no remedy touches, the run of bad luck that descends from a clear sky: these defy the practical toolkit, and every well-documented society has recognized roles for the people who deal with them.

Anthropology gives these roles four analytic names. A shaman enters altered states, typically trance, to mediate between the human world and a world of spirits, often to heal. A witch, in the technical sense that emerged from African ethnography, is someone believed to cause harm through an inborn power, frequently without conscious intent. A sorcerer harms through learned technique, spells and materials deliberately applied. A healer applies knowledge, herbal, ritual, or both, to restore health. These are not native categories from any one culture but conveniences the discipline uses to compare across cultures, and they are worth holding loosely, because real people rarely fit one box cleanly.

The Logic Hidden Inside Azande Witchcraft

Evans-Pritchard's central claim is easy to state and hard to absorb. Azande witchcraft belief, he argued, is not a primitive error waiting to be replaced by science but a different kind of explanation altogether, one that handles a question science does not even pose. Western reasoning explains the mechanism of an event, the chain of physical causes. Azande reasoning accepts that mechanism and then asks a further question that mechanism cannot answer.

To the Azande, witchcraft was a physical substance, mangu, carried in the belly of certain people and passed down through descent. A person could possess it without knowing, and its power could reach out and cause harm even while its owner slept. This is crucial to understanding the system: witchcraft was not, in the Azande view, primarily about wicked intent or dramatic curses. It was closer to a kind of inherited misfortune-generating capacity, an explanation for why bad things cluster around certain relationships, and because it was woven into the ordinary texture of life rather than reserved for spectacular catastrophes, it functioned less as superstition than as a working theory of causation that ran alongside, not instead of, everyday practical knowledge.

The Collapsed Granary, the Most Quoted Scene in the Field

The single most cited passage in the anthropology of religion appears in the second chapter of Evans-Pritchard's book, and it concerns a granary. Among the Azande, raised grain stores stood on wooden supports, and people would sit in their shade during the heat of the day. Termites ate at the supports over time, and occasionally a granary collapsed, sometimes with someone sitting underneath who was injured or killed.

Now, the Azande knew perfectly well that termites had weakened the wood. They were not ignorant of the physical cause; they would have explained the collapse in exactly those terms if you asked how it happened. But Evans-Pritchard saw that this left a real gap. The termites explain why the granary fell. They do not explain why it fell at the precise moment that this particular person was sitting beneath it, rather than an hour earlier or later when the shade was empty. Western reasoning shrugs at this gap and calls it coincidence. The Azande did not find coincidence a satisfying answer to a death, and so they filled the gap with witchcraft. The two explanations did not compete. As Evans-Pritchard put it, witchcraft belief does not replace empirical reasoning; it complements it, supplying a layer of meaning that the physics of termites leaves untouched. The supposed irrationality dissolves on contact: the Azande were answering a question we tend to suppress rather than failing to answer one we ask.

Oracles as Decision Procedures, Not Decorations

If misfortune is caused by witchcraft, an urgent practical question follows: which witch? An accusation has consequences, and the Azande did not leave the matter to guesswork. They ran a tiered system of oracles, escalating in authority, that functioned as genuine decision procedures for testing accusations and rendering judgments. The most authoritative of these was the poison oracle, benge.

The procedure was concrete. A specially prepared poison, a strychnine-bearing substance derived from a forest creeper, was administered to a fowl while a question was put to the oracle in a form that could be answered by the bird's life or death. Whether the chicken survived or died gave the verdict, and the oracle built in cross-checking: a result could be confirmed by reversing the question and dosing a second bird, so that the system tested itself for consistency. To dismiss this as mere ritual theater is to miss its social function. The oracles were not magical ornaments hung on the side of the belief system. They were the institutional machinery through which suspicion was converted into a verdict, the device that turned a diffuse sense of being wronged into an authoritative, actionable judgment, sometimes about life and death. In a society without courts of the European kind, the poison oracle did the work that a court does.

Shamans, Trance, and the Limits of a Grand Theory

If the witchcraft side of this story belongs to Evans-Pritchard, the shamanism side belongs, in the first instance, to the historian of religion Mircea Eliade. His Shamanism: Archaic Techniques of Ecstasy, published in French in 1951 and in English by Princeton in 1964, gave the discipline a vocabulary for a phenomenon reported across an enormous range, from Siberia, where the word shaman originates, to the Americas and beyond. Eliade defined shamanism around ecstatic trance, the practitioner's spirit journeying to other worlds to retrieve a lost soul, negotiate with spirits, or escort the dead.

Eliade's synthesis was influential, and it remains a useful entry point, but contemporary anthropology has had to walk a good deal of it back. The problem is essentialism: Eliade treated shamanism as a single, unified, archaic essence underlying a vast diversity of local practices, smoothing over the real differences between, say, a Siberian healer and an Amazonian one as if they were variants of one timeless type. Modern ethnography is more cautious. It treats shaman less as a thing in the world than as a comparative label that helps us notice family resemblances, while insisting that the actual practices are historically specific, locally shaped, and not reducible to one ancient template.

Still, one recurrent pattern is robust enough to survive the skepticism. Across many societies, the path to a shamanic vocation tends to run through three stages. It begins with a call, often experienced not as ambition but as affliction, an unsought illness or crisis that resists ordinary cure and is read as a summons. It proceeds through an initiation under an established practitioner who teaches the techniques of trance and spirit-mediation. And it ends with the assumption of a mediating role between worlds. This is not Eliade's grand essence so much as a sturdy ethnographic finding about how such specialists tend to be made: through suffering, apprenticeship, and a transformation that the community recognizes.

From Spirit Worlds to Clinics: The Modern Frontier

The four categories can be arranged on two axes, which helps clarify how they relate. One axis runs from helpful to harmful in intent; the other distinguishes the mechanism, whether trance, inherited substance, learned technique, or applied knowledge. A shaman is helpful and works through trance, a witch is harmful and works through inherited substance, a sorcerer is harmful and works through learned technique, and a healer is helpful and works through applied knowledge. But the grid is a starting point, not a verdict, because most real figures occupy more than one cell. A shaman may also heal with herbs, a healer may be suspected of sorcery, and the same person can be a community's protector in one season and its scapegoat in the next.

The liveliest contemporary reworking of all this happens in medical anthropology, where the old questions return in clinical dress. Arthur Kleinman's Patients and Healers in the Context of Culture, published in 1980, introduced the concept of the explanatory model: the framework a sick person and a healer each bring to an illness, the story each tells about what it is, what caused it, and what should be done. The insight reorganized how the healing encounter is read everywhere, including in Western hospitals, because it showed that doctor and patient often work from different explanatory models and talk past each other without realizing it. This is also where the shaman's apparent effectiveness becomes legible to science, through the placebo response, the well-documented finding that the meaning and ritual surrounding a treatment can produce real physiological change. A healing rite that mobilizes a patient's belief and the community's support is not doing nothing.

Accusation as a Map of Where a Society Hurts

The misconception this whole tradition exists to retire is the idea that witchcraft accusations are random superstition, lightning that strikes anyone. They are not random. Consider the largest sustained persecution the documented record contains, the European witch craze of roughly 1450 to 1750, in which scholarly estimates of those executed range from about forty thousand to about sixty thousand. The accused were overwhelmingly women, and historians read the craze as a deeply gendered event bound up with the anxieties of the transition to early modern society: conflicts over property, healing knowledge, and the place of women who lived outside male authority.

The same patterning appears across otherwise unrelated cases. In the Azande material, in colonial Salem, in modern central African communities, and in episodes in Papua New Guinea, accusations track recognizable social tensions: gender, generational power, property and inheritance, marginality, and the strains of rapid economic change. The witch is frequently the widow, the old woman, the in-law, the neighbor who has done unexpectedly well or unexpectedly badly, the person standing at a point of friction in the social fabric. This is why an anthropologist learns to read a pattern of accusations as a diagnostic instrument. Where the accusations fall tells you where the society is under strain, which relationships are dangerous, which transitions are unmanaged. The accusations are a map of where the community hurts.

Key Takeaways

Every well-documented society maintains specialists for managing the misfortunes that ordinary practical action cannot fix, and anthropology names them with four comparative conveniences, shaman, witch, sorcerer, and healer, distinguished by whether they help or harm and by their mechanism, while recognizing that most real figures blur those lines. Evans-Pritchard's 1937 study of the Azande remains the canonical demonstration that witchcraft belief is not primitive error but a structured explanation for the question physics cannot touch, why this person and why now, captured in the collapsed granary where termites explain the fall but witchcraft explains the timing, and adjudicated through the poison oracle benge, which served as a genuine decision procedure. Eliade gave shamanism its cross-cultural vocabulary around ecstatic trance and the recurrent vocational path of call, initiation, and mediation, though his essentializing has since been tempered, while Kleinman's explanatory models and the placebo response carry these themes into medical anthropology. Across the European craze of 1450 to 1750, where forty to sixty thousand mostly female victims were executed, and cases from Salem to central Africa to Papua New Guinea, accusations prove not random but patterned, tracking gender, power, property, and economic upheaval so reliably that the anthropologist reads them as a map of where a society is in pain.