Eyes in the Sky: Why Belief Comes Easily and Fades Quietly

For ten weeks, a coffee room at a British university quietly became a lab. Staff paid for their drinks by dropping coins into an honesty box. Above the price list hung a small picture that changed every week. Some weeks it showed flowers. Other weeks, a pair of eyes looking straight at the viewer.

Nobody was actually watching. Still, in the weeks with eyes, people paid nearly three times as much (Bateson et al., 2006).

Now picture the great Buddhist stupas of Kathmandu, with a pair of eyes painted on each of their four sides.

I think the poster and the stupa tell the same story. One was put up by researchers for ten weeks, the other shaped by culture over centuries. Both speak to the same old part of the mind.

A brain that checks for eyes

For most of our history, humans lived in small groups where everyone knew everyone. Being caught cheating could cost you friends, food, and protection. In that world, it paid to notice when you might be seen, and to behave a little better when you were.

That sensitivity is still with us, though it has limits. Eye images do little to promote generosity (Northover et al., 2017), but across 15 experiments, they reduced antisocial behavior by about a third (Dear et al., 2019).

Two other habits help. A rustle in the dark feels like someone before it feels like the wind, because missing a real predator costs far more than a false alarm (Barrett, 2000). And we easily imagine minds without bodies, a habit that shows up early in childhood (Bering et al., 2005).

So the human mind can easily feel watched by someone who is not there.

From painted eyes to watching gods

In one study, children aged 5 to 9 played a game with rules that were hard to follow. Some were left alone, some were watched by an adult, and some were told that an invisible princess, Princess Alice, was sitting in an empty chair. Children who believed in the princess cheated about as rarely as children watched by a real adult (Piazza et al., 2011). The princess had been made up minutes earlier.

This is the heart of my argument. Supernatural watchers are a cultural invention, but one the mind is ready to receive. No one has to be taught to feel watched. Culture only has to give that feeling a face, a name, and a set of rules: an ancestor, a spirit, a god who sees what you do and even what you think.

This mattered most once societies grew large. Gossip and reputation work in a village, but poorly in a city of strangers. A god who sees everything can fill that gap.

In a study of eight communities around the world, people who saw their gods as moral, punishing, and all-knowing were fairer toward distant strangers of the same faith (Purzycki et al., 2016). Over centuries, groups with such gods may have cooperated better and spread further, taking their gods with them (Norenzayan et al., 2016).

The sensitivity to eyes comes from biology. Turning it into an all-seeing god was a cultural step. Together, they helped strangers cooperate.

When people stop believing

If our minds are this ready for watchful gods, why do so many people today not believe in them?

The popular answer is that they thought their way out. The evidence is weaker than the story suggests. An early experiment found that viewing a picture of Rodin’s The Thinker reduced belief in God (Gervais & Norenzayan, 2012), but a much larger replication of the study found no effect (Sanchez et al., 2017; see Yilmaz, 2021, for a review).

In a study of 7,771 students in 19 countries, my colleagues and I found that more reflective thinkers do believe a little less. But the link is small, and having students take a thinking test first did not lower their belief (Ghasemi et al., 2025). Reflection can question an intuition, but it rarely erases it.

Culture offers two better answers. The first is about what children see: we learn faith less from what adults say than from what they do, such as fasting or giving to the poor (Henrich, 2009). In a large American sample, growing up with few such displays predicted disbelief more strongly than thinking style did (Gervais et al., 2021).

The second is that the watchers lost part of their job. Today, courts, police, and welfare systems do much of what watchful gods once did. These secure societies are also among the least religious in the world (Norris & Inglehart, 2004). As Norenzayan (2013, p. 172) put it, “secular societies climbed the ladder of religion, and then kicked it away.”

The two answers are linked. When institutions do the watching, people lean less on religious communities. Parents may still say they believe, but they show it less, and their children see fewer displays. Belief then fades quietly over generations. Disbelief is learned much like belief (Gervais, 2024).

What genes, minds, and culture each do

Genes do not carry a belief in God or a lack thereof. They build a brain that checks for eyes, senses hidden agents, and learns by watching others.

Each mind works with what it is given. Children easily accept the watchers their culture offers. Later, they may question them.

Culture invents the watchers and decides which ones last. It turns the feeling of being watched into gods and spirits with names and stories. Those that help groups hold together survive, and people pass them on through their actions. In some places, culture has also built institutions that do the watching instead.

The next time you pass painted eyes, on a temple, a poster, or a blue evil eye charm by a front door, notice the small shift in how you feel. That reaction is very old. What you believe those eyes mean, and whether anyone is behind them, depends far more on the people who raised you and the society you live in.

References