3.10 — How Fear Became Superstition, and Superstition Became Religion

In 1948 the psychologist B. F. Skinner put hungry pigeons in boxes that delivered food at fixed intervals, regardless of what the bird did. Within a short time most of the pigeons had developed elaborate repeated behaviours — turning counter-clockwise, bobbing the head, swinging it side to side — each bird a different one, each performed intently before feeding.

Whatever the bird happened to be doing when food arrived, it did again. The food came anyway, which confirmed it.

This is not an explanation of religion, and it would be cheap to present it as one. But it is a clean demonstration of the underlying machinery: a mind built to find patterns will find them in noise, and a mind that has found one will act on it. This chapter follows that machinery from a startled ape to the temple at the centre of the first cities — as history and psychology, with respect for what people believed, and without pretending to settle whether any of it is true.

On which point, one sentence of position. This volume describes what people believed, when, why, and what it did in the world. It does not adjudicate whether God exists; that question belongs to philosophy and Volume VII takes it up seriously. Explaining how a belief arose is not the same as showing it to be false — that confusion is called the genetic fallacy, and this chapter avoids it.

The machinery in the mind

Four well-studied cognitive tendencies show up in every human population, and each contributes.

Agency detection, biased to over-fire. Rustling in the grass could be wind or a predator. The cost of assuming a predator when it is wind is a wasted moment. The cost of assuming wind when it is a predator is death. So natural selection tunes the detector to fire easily. The consequence is that we see faces in clouds, hear voices in noise, and instinctively treat unexplained events as though somebody caused them. A mind that infers agents behind rustling grass will infer agents behind thunder, disease and drought.

Cause-and-effect reasoning without a way to test it. Humans are exceptional at finding causes, which is the foundation of every technology in Chapter 3.4. But before controlled comparison existed, there was no way to distinguish a real cause from a coincidence. The ritual was performed and the rain came. The pattern is recorded and repeated, exactly as with Skinner's pigeons — and unlike the pigeons, humans then explain it, teach it, and enforce it.

Mind-reading, extended. Humans model other minds constantly — what does he want, what does she know, what will they do. This machinery does not switch off when the thing being modelled has no mind. We talk to cars that will not start and thank a river for a good catch.

Dualism, which appears to come naturally. Young children in every studied culture readily treat mind as separable from body, and this intuition is not taught. If a mind is a separate thing from a body, then it is not obvious that it ends when the body does, and belief in continuation after death follows without anyone having to argue for it.

Together these produce a mind that expects invisible agents with intentions who can be dealt with. That is the raw material. What each culture builds from it is enormously varied, and the variation is the history.

The oldest evidence

Burial with grave goods is the earliest reasonable evidence, from at least 100,000 years ago, with Neanderthal burials in the same range. Placing food, tools or ornaments with a body implies that something about the person is expected to continue, though it can also be read as grief expressed materially, and honest archaeologists say so.

Ochre on bodies from the same period. Red pigment, associated with blood and life, applied to the dead across many unconnected cultures.

Sungir, 34,000 years ago (Chapter 3.5), with tens of thousands of hand-made ivory beads buried with children. Nobody spends that labour on a hole in the ground without believing something specific about what happens next.

The caves. Deep, dark, difficult chambers with paintings, sometimes hundreds of metres from daylight (Chapter 3.5). At Chauvet, a bear skull placed on a stone block in the middle of a chamber. Not proof of religion, but not easily explained as decoration either.

Göbekli Tepe, from 9500 BCE (Chapter 3.9), where hunter-gatherers moved ten-tonne carved pillars. This is the first unmistakable evidence that shared belief could organise labour on a scale beyond the group's daily needs, and it is one of the most important facts in this entire Part.

From shaman to priest

The earliest religious specialist we can identify across many cultures is the shaman — a part-time practitioner who enters an altered state to deal with spirits on the community's behalf, treating illness, finding game, and mediating misfortune. Shamanic practice is found among foraging peoples across Siberia, the Americas, Australia, Africa and South Asia, with recurring features: induction by drumming, fasting, dancing or plant substances; a journey narrative; and a period of illness or crisis marking the practitioner's calling.

The shaman is not a separate class. They hunt and gather like everyone else. Their authority is personal and depends on results.

The priest is a different institution and it appears with surplus (Chapter 3.9). A priest holds an office rather than a gift; the office survives the person; the position is often hereditary; and the priest is supported by other people's food. That is only possible where there is a surplus to divert, which is why full-time religious specialists appear at the same moment as full-time potters, smiths and soldiers.

And once religion has an institution, it has interests — property, succession, doctrine, and a relationship with whoever holds power. Every major development in Parts 4 through 12 involving religion is a development in that institution, and it is worth keeping the distinction between a belief and an organisation clear throughout.

What the gods were about

Early gods are overwhelmingly about the things people could not control and could not survive without. Rain, sun, river, storm, fertility, disease, death, and the animals hunted or herded.

The correspondence between a society's economy and its gods is close enough to be predictive. Agricultural societies have storm and fertility gods at the centre — Indra, Zeus, Baal, Thor all wield the thunderbolt that brings rain. River civilisations deify the river: the Nile as Hapi, the Ganga as a goddess. Pastoral peoples emphasise cattle and the sky. Foraging societies more often have animal spirits and a distant, uninvolved creator.

Ancestors are a near-universal starting point, and the path from a remembered ancestor to a god is short: the recent dead are remembered as individuals, the older dead merge into a category of ancestors, and the oldest become founders and then deities. Chinese religion preserves this whole gradient visibly.

Why religions grew as societies grew

Here is the pattern that most needs explaining, and it is well documented.

Small-scale societies typically have spirits that are local, morally uninterested and concerned with ritual correctness rather than ethics. A spirit is offended by a broken taboo, not by dishonesty.

Large-scale societies typically have gods who are powerful, watchful, and care whether you cheat your neighbour.

The proposed explanation is about cooperation between strangers. In a group of fifty, reputation does the work — everyone knows who cheats. In a city of fifty thousand, most of your dealings are with people you will never meet again, and reputation cannot police that. A widely shared belief that a god sees everything, including what nobody else saw, and cares about fairness, supplies a monitor where social monitoring has failed.

The evidence. Cross-cultural surveys find moralising high gods far more common in large, complex societies. Experimental work finds that people primed with religious concepts behave more generously and cheat less in anonymous games — though this literature has had replication problems and the effect sizes are argued. And the historical sequence in several regions shows moralising gods appearing after, not before, large states.

Two honest cautions. The causal direction is disputed — one large analysis of the historical database argued that big gods follow the growth of complex societies rather than enabling it. And this is a functional account of why such beliefs spread, which says nothing about whether they are true.

What is not disputed is the effect. Shared belief allowed cooperation, trust and coordinated effort among strangers on a scale nothing else achieved. Trade over long distances requires trusting a stranger; oaths sworn before gods were the enforcement mechanism for thousands of years. Courts, contracts and credit all began as religious instruments.

The flood stories

Flood narratives appear independently in Mesopotamian, Hebrew, Greek, Indian, Chinese and Mesoamerican traditions, among others. The Mesopotamian version is the oldest written — in the Epic of Gilgamesh and earlier in the Atrahasis story — and its resemblance to the biblical account is close enough that most scholars regard the Hebrew narrative as drawing on the older Mesopotamian tradition, given the region's shared cultural sphere.

India's version is the Matsya story: Manu is warned by a fish, later identified with Vishnu, builds a boat, and survives a deluge that destroys everything, preserving the seeds of life. It appears in the Shatapatha Brahmana and later in the Puranas.

Three sober explanations, which do not compete.

Real local floods. Mesopotamia floods. Archaeologists have found substantial flood deposits at several Sumerian cities. To someone whose world is the floodplain, a flood that covers it is a flood that covers the world.

Post-glacial sea level rise. Chapter 3.6 gave the numbers: 120 metres over ten thousand years, at times four metres a century, drowning the coastal plains where most people lived. That is slow by the standards of a myth and fast by the standards of a lifetime, and it happened everywhere at once, which fits the universality of the stories better than any single event.

And the narrative logic. A story about surviving an unsurvivable catastrophe is a good story with a clear moral, and such stories are remembered and retold whether or not they began in one event.

What the recurrence does not support is a single global flood, for which there is no geological evidence and against which there is a great deal — the sedimentary record, the survival of unbroken tree-ring and ice-core sequences, and the continuity of civilisations in China, Egypt and elsewhere through the periods proposed.

Superstition, then and now

Superstition and religion are not cleanly separable, and drawing the line is usually a way of dismissing somebody else's practice while protecting your own. What can be described precisely is the mechanism, which is the same in both and which has not gone anywhere.

Reinforcement without testing. The cricketer who wears the same socks after a century. The exam sat with a particular pen. The lucky number.

Illusion of control. People bet more, and throw dice harder, when they are allowed to throw them themselves. The stronger the uncertainty and the higher the stakes, the stronger the ritual — which is why the anthropologist Bronisław Malinowski found that Trobriand Islanders performed elaborate magic for dangerous deep-sea fishing and almost none for safe lagoon fishing. That finding has held up across many later studies.

Confirmation bias. Hits are remembered and misses forgotten, which is what sustains astrology, most predictions, and every claim of prophecy.

Costly signalling. Practices that are expensive, painful or restrictive are effective badges of group commitment precisely because they are hard to fake — which is one reason demanding religious groups tend to be more cohesive and longer-lived than undemanding ones.

Where this shows up in your life

In how you evaluate a claim. The four tendencies at the top of this page are running in your head right now. They are why a run of good luck feels like a pattern, why a coincidence feels meaningful, and why the story that a person caused an event is always more satisfying than the story that no one did. Knowing the machinery does not switch it off, but it does let you ask the second question.

In respecting what religion did. It is easy, from inside a society with courts, insurance, hospitals and police, to treat pre-modern belief as ignorance. It was the operating system. It supplied the moral code, the calendar, the law, the medicine, the history, the art, the charity, the marriage system, and the only available explanation for suffering. Societies that had it functioned; the alternatives were not obviously available. Chapter 8.8 is about how one particular alternative — a method for checking claims — was constructed, and it took a very long time.

And in what has replaced it, which is worth naming. The cognitive machinery does not go away when the theology does. It attaches to conspiracy theories, to political movements, to wellness products, to markets and to national myths. Chapter 16.4 is the practical chapter on telling a checked claim from an unchecked one, and it is the most useful thing in this volume.

What the next page covers

Everything so far has been transmitted by memory and speech, which means it could be lost in a generation and could not be verified. Chapter 3.11 covers the invention that fixed that — how writing was invented, more than once, and why it began as accounting rather than literature; how numbers and counting developed separately; why the alphabet was a genuine democratisation; and what happens to a society's memory, law and science once claims can be written down, stored and checked against.