Imagine a gorilla asleep in a nest high in a tree, suddenly standing up and taking a few steps. 

In a cartoon, the scene might be funny. In nature, it would probably be fatal. Humans, by contrast, sleepwalk often enough for the behaviour to feel familiar, if still disturbing.

We know that dogs may paddle their paws during dreams and cats may twitch their whiskers, but true sleepwalking—standing, walking and navigating obstacles—has not been clearly recorded in other animals.

David R. Samson, an evolutionary anthropologist at the University of Toronto, argues that this strange human habit is closely tied to where our ancestors slept.

Sleepwalking usually occurs during deep non-REM sleep, when the brain is not fully awake.

Samson describes it as a kind of "state dissociation": the systems that control movement and arousal become active, while those responsible for judgment, memory and reflective awareness remain in a sleep-like state.

As a result, the body can open doors, climb stairs or walk through a dark hallway before the waking mind has fully arrived.

For most primates, such a glitch would have been disastrous.

They sleep off the ground, on branches, in tree holes or on raised platforms, where even a few unconscious steps could mean a deadly fall.

Natural selection would therefore have punished sleepwalking severely.

Human beings, however, changed the rules when they began sleeping on safer ground, protected by shelters, fire, bedding and other people keeping watch.

In this "sleep shell," as Samson calls it, the immediate cost of getting up while asleep was reduced.

A flaw that would have been removed in tree-sleeping primates could survive in human camps.

This is an example of relaxed selection, in which a harmful trait can persist because the environment no longer weeds it out.

Sleepwalking did not evolve because it was useful. It remained possible because our social and physical surroundings made it less dangerous.

The experience is unsettling because it creates a gap between outside behaviour and inner awareness.

A sleepwalker may look purposeful, even urgent, while having no memory of the event later.

Samson recalls a friend who once charged at him in a dark hallway, struggled briefly, then returned to bed.

The next morning, the friend remembered only a dream in which he was trying to save Samson from falling timber.

From the outside, the behaviour seemed alarming; from the inside, it followed the emotional logic of a dream.

Genetics also matters. Sleepwalking tends to run in families, although there is no single "sleepwalking gene."

It appears more often in children, whose deep sleep is stronger and whose sleep-wake systems are still developing.

Sleep deprivation, fever, stress, alcohol, certain medicines and disorders such as obstructive sleep apnea can all raise the risk.

Sleepwalking, then, is not an evolutionary superpower but a revealing vulnerability. It shows that the sleeping brain is not a single switch that turns off the body. Instead, different systems can wake at different speeds.

Humans may be the species that walks in its sleep not because we are uniquely strange, but because, for much of our history, we slept together safely enough for this strange error to survive.