The Magic Instinct: Why We Tell Stories of Wizards and Witches

by

Updated 17 August 2026 · 8 min read

The Magic Instinct: Why We Tell Stories of Wizards and Witches

Every human society we know of has produced someone whose job is to deal with the invisible. The names differ and the status differs wildly, but the role keeps appearing: a person the community turns to when a child sickens for no reason, when the rain fails, when the dead will not stay put.

That much is uncontroversial. What is usually said next is where it goes wrong.

The comfortable version says these figures were healers and balance-keepers, that their magic served the community rather than themselves, and that the witch is a later corruption of something originally benign. It is a lovely story. The evidence does not support it, and the evidence is unusually good, because the earliest magic we can actually read was written down in enormous quantities by people who wanted very specific things.

The oldest magic we can read is mostly grudges

Start with the Greek world. From roughly the fifth century BCE, people wrote curses on thin sheets of lead, folded or rolled them, sometimes drove a nail through them, and deposited them in graves, wells and springs, places understood to be closer to the underworld than ordinary ground. Scholars call them katadesmoi in Greek and defixiones in Latin. Both words mean something close to “binding”.

They survive in large numbers because lead survives. And they are not spiritual. They are:

  • Lawsuits. Binding an opponent’s tongue, memory and wits before a court appearance is one of the single most common categories.
  • Business. Curses against rival tavern-keepers, shopkeepers and craftsmen.
  • Sport and theatre. Curses against charioteers, their horses, and competing performers.
  • Love, of an aggressive kind. Not “make her love me” so much as “make her unable to eat, drink or sleep until she comes to me”.

This is what magic looked like when ordinary people paid for it. Not cosmic order. A neighbour, named, and a request to ruin them.

Egypt gives the same answer from a different direction. In the Middle Kingdom, Egyptian ritualists produced execration texts, red pottery jars and bound-captive figurines inscribed with the names of foreign rulers, hostile peoples, and sometimes internal enemies. The objects were smashed and buried. The magic was the breaking. Egyptian mythology treated this as entirely normal statecraft.

None of this means magic was only malice. It plainly was not. But any account of the magic instinct that starts from “it was for the greater good” has skipped the sources.

Magic was regulated, which tells you it was taken seriously

Here is the fact that changes the shape of the whole subject. Magic was not a fringe practice tolerated at the edges of ancient societies. It was licensed, professionalised and legislated.

In Mesopotamia the āšipu was a salaried specialist with a training curriculum and a known reading list. He diagnosed supernatural causes of illness and misfortune, and he performed the counter-rituals. He often worked alongside the asû, whose methods look more like what we would call medicine. The two were colleagues, not rivals.

And the āšipu’s largest single body of work was defensive. The ritual series Maqlû, whose name means “Burning”, runs to nine tablets and was performed across a single night. Its entire purpose is to identify, neutralise and destroy a witch who is assumed to be attacking the patient. A professional class existed specifically to counter illegal magic, which means illegal magic was a recognised category with a recognised remedy.

The law backs this up. The Code of Hammurabi opens on the subject. Its second provision deals with a man accused of sorcery: he is thrown into the river. If the river takes him, his accuser inherits his house. If he survives, the accuser is put to death and the accused takes the accuser’s house instead.

That is not a society treating magic as superstition. That is a society treating a witchcraft accusation as a capital matter serious enough to need a deterrent against making one lightly.

The role recurs. The status does not.

This is the part the tidy version flattens. It is true that almost every culture produced a magic specialist. It is not true that the specialist occupied the same social position.

  • Mesopotamia’s āšipu was an establishment figure: temple-trained, employed, respected, and on the state’s side.
  • The Norse völva was something else entirely: itinerant, consulted, fed and honoured on arrival, and feared. The word itself appears to derive from vǫlr, a staff, so the title means something like “staff-bearer”. Norse mythology gives the practice, seiðr, to Freyja, and gives Odin real social trouble for taking it up, because it was understood as women’s work.
  • The Yoruba babaláwo, whose title means “father of secrets”, is an initiate of the Ifá oracle, an authority on fate, medicine and conduct. The role is honourable and has never been the accusation-word. Orunmila is the deity of the system he serves.
  • The Japanese onmyōji was a civil servant. The Bureau of Yin and Yang was part of the machinery of the state, and Abe no Seimei is its most famous graduate: a court official, not an outsider.

Four specialists, four completely different relationships to power. The instinct is universal. The institution is local, and the institution is where the interesting variation lives.

The healer and the witch are frequently the same slot

The single most useful idea in the anthropology of magic came out of the southern Sudan in the 1920s, and the article you are reading exists partly because it is so often cited and so rarely used.

E. E. Evans-Pritchard, working among the Azande, found that they drew a distinction English does not. Witchcraft, mangu, is a physical substance in the body, inherited, and it can act without its owner intending anything at all. A person can bewitch a neighbour while asleep and be genuinely horrified to learn of it. Sorcery is different: the deliberate use of medicines to cause harm, learned rather than inherited, and unambiguously a crime.

The consequence matters. Under this scheme, a witchcraft accusation is not primarily a claim about what someone did. It is a claim about what they are. And that is exactly why the accusation travels so well and lands so often on the same people: the isolated, the old, the widowed, the resented, the ones already standing slightly outside.

Which brings the two halves of this series together. The wizard and the witch are frequently not two different jobs. They are one job described by two different sets of people: the ones who need it, and the ones who fear it.

What the specialist is actually for

Strip away the local colour and three tasks recur almost everywhere.

Diagnosis. Not “what is wrong” but “why me, and why now”. Medicine can explain how an illness works. It has never been much good at answering why it arrived at this house in this month, and that is the question people actually ask. Magic answers it.

Redress. Once a cause is named, something can be done. This is the entire premise of Maqlû, and it is the reason the counter-ritual is always more elaborate than the curse.

Boundary work. Somebody has to handle the dead, the spirits, the ancestors and the outside. This is the job nobody wants and everybody needs. Hecate stands at the crossroads for a reason.

James Frazer’s The Golden Bough proposed in 1890 that magic works on two principles: that like produces like, and that things once in contact stay connected. The observation is genuinely useful and holds up across a startling range of practice. His larger scheme, in which humanity climbs from magic to religion to science, does not, and no working anthropologist defends it now. Take the mechanism, leave the ladder.

Why the stories outlast the belief

Very few readers of this article expect a curse tablet to work. The figures have not gone anywhere regardless.

Merlin survives because he is the licensed insider, the wizard as advisor to power. The āšipu, essentially, with better press. Baba Yaga survives because she is the boundary-dweller, out past the edge of the village in a house that will not stay still, and the terms she offers are always transactional. Väinämöinen survives because he does his magic by singing: knowledge as power, in the most literal available form. Romania’s Solomonari survive as the wizard-scholar who learned it from a book and can call the weather.

Each of these is one of the three jobs, wearing a costume. That is why they keep coming back. The stories are not nostalgia for a world where magic worked. They are a working record of the roles a community actually needs filling, and of what it does to the person who fills them.

The next article in this series goes back to the beginning and asks what the earliest evidence really supports, and, just as usefully, where it runs out.

Further Reading

  • E. E. Evans-Pritchard, Witchcraft, Oracles and Magic Among the Azande (1937). The source of the witchcraft/sorcery distinction, and still the clearest thing written on why accusations land where they do.
  • Tzvi Abusch, Mesopotamian Witchcraft and his edition of Maqlû. The standard work on the ritual described above.
  • Bronisław Malinowski, Magic, Science and Religion (1948). Magic as what people reach for when skill runs out and risk begins.
  • Owen Davies, Magic: A Very Short Introduction (2012). The best short survey, and current.
  • Sir James Frazer, The Golden Bough (1890). Read for the sympathetic-magic principles and the reach of the material. Its evolutionary scheme has been rejected for a century.
  • Mircea Eliade, Shamanism: Archaic Techniques of Ecstasy (1951). Foundational and contested; see the next article for why the word “shaman” does more work than it can bear.

Explore More on Gods and Monsters