Pixopedy
Sources · Ep. 0002 · behavior
You have never actually looked into someone's eyes. The geometry of your own retina makes it impossible - and that is the least strange thing about eye contact.
Compiled during fact-checking, before the script was written. Nothing here was reconstructed afterwards.
One line here rests on nothing but standard optics: the pupil is black because it is an aperture, light goes in and does not come back out. That is uncontroversial enough to stand without a citation, but it stands uncited. The bigger softness is in the animal-kingdom passage. Direct gaze as a status and threat signal among primates is solid, backed by Scientific Reports 2020, but the claim that butterflies grew fake eyes on their wings because a pair of eyes pointed at you means danger is still unverified. Two other passages are partial rather than settled: that wide pupils register as warmth and safety, and the hedged suggestion that cooperation, teaching and the ability to trust were the pressure that shaped our eyes. The geometry point about two eyes sitting further apart than you can focus on at conversational distance is derived, not measured. Cross-cultural gaze norms were the weakest point in the first pass and are now verified against Akechi et al. 2013 in PLOS ONE and a Leadership Quarterly review of gaze and hierarchy.
In sixteenth-century Italy, women dripped poison into their eyes. The plant was deadly nightshade. The effect was a pupil blown wide and black. And the name we still use for that plant is the compliment they paid it. Belladonna. Beautiful woman.
16th-c Italian women used Atropa belladonna drops as mydriatics; Linnaeus named the species belladonna for that cosmetic use.
They were not wrong. They had found something real about how we read each other, roughly four hundred years before anyone thought to measure it. Wide pupils register as warmth. As interest. As safety.
What was checkedWide pupils register as warmth. As interest. As safety.
Kret's work reports dilated pupils are perceived as a sign of safety and constricted pupils as threat. The three-word framing is a compression, not a direct quotation of a finding.
A comparative survey of eighty-eight primate species found that exposed white sclera was unique to us. Ours is unpigmented. Ours is more widely exposed. And the human eye is stretched horizontally further than any other primate's, which does one thing very efficiently. It advertises the direction of your attention.
Kobayashi & Kohshima studied 88 primate species; exposed white sclera unique to humans, most widely exposed. eLife 2021 adds experimental support that white sclera aids gaze-direction visibility.
A comparative survey of eighty-eight primate species found that exposed white sclera was unique to us. Ours is unpigmented. Ours is more widely exposed. And the human eye is stretched horizontally further than any other primate's, which does one thing very efficiently. It advertises the direction of your attention.
What was checkedAnd the human eye is stretched horizontally further than any other primate's.
Same study: the human eye outline is the most horizontally elongated among the species measured.
Think about what that means as a trade. Every other ape can look at the fruit, the rival, the exit, without being caught looking. We cannot. We broadcast. Our eyes leak, on purpose, because whatever pressure did this to us — cooperation, teaching, the ability to trust and be trusted — was strong enough to be worth surrendering the poker face.
What was checkedwhatever pressure did this to us, cooperation, teaching, the ability to trust and be trusted
This is the gaze-signalling / cooperative-eye hypothesis. Mearing et al. 2022 found white sclerae associated with cooperative behaviour but noted the data are consistent with all three competing hypotheses. The script deliberately hedges with 'whatever pressure' and never asserts the mechanism.
It is installed early, too. In one study, newborns between two and five days old preferred to look at faces that were looking back. Days old. Before language. Before recognition. Before anything you would call a self.
Farroni et al. tested 2-5 day old newborns; preference for faces engaging in mutual gaze.
A direct stare is one of the oldest threat signals in the animal kingdom. Predators lock on. Prey freeze. Butterflies grew fake eyes on their wings because a pair of eyes pointed at you is, across enormous stretches of biology, very bad news.
Partially closed in the follow-up sourcing pass. Scientific Reports 2020 supports direct gaze functioning as a threat and status signal in primates, and finds eye-contact tolerance predicted by social structure. The specific claim about lepidopteran eyespot mimicry remains UNVERIFIED - no source was retrieved for the butterfly element. It is well established in the literature but unchecked here.
We inherited that circuit. It runs underneath thought, through structures like the superior colliculus and the amygdala — the brain's oldest lookouts, on duty and reading eyes long before we had opinions about anyone.
Hadjikhani et al. measure the subcortical face-processing system, naming the amygdala and superior colliculus. 'Oldest lookouts' is rhetorical framing, not a sourced claim about evolutionary age.
Researchers ran nearly five hundred people from fifty-six countries through varying durations of simulated eye contact and asked what felt right. The average preference came out at about three point three seconds. The comfortable band sat between two and five. Almost nobody wanted more than nine.
Binetti et al.: ~500 participants from 56 nations at the London Science Museum; mean preferred mutual gaze 3.3s; comfort band 2-5s; almost none preferred beyond 9s.
And notice what it is not. It did not track well with gender, or personality, or attractiveness. What it did track with was pupil response — physiological arousal, the body setting the timer without asking you.
Same study: preference was not dependent on gender, personality traits or attractiveness; pupillary change correlated with preferred duration.
It is called gaze aversion, and it is well documented in both children and adults. When a question gets harder, people break eye contact to answer it. Children begin doing it around five and are using it like adults by about eight. It is not evasion. It is bandwidth. Holding someone's face costs processing, and difficult thought needs that processing back.
What was checkedIt is called gaze aversion. When a question gets harder, people break eye contact to answer it.
Doherty-Sneddon & Phelps: gaze aversion increases with cognitive difficulty; looking away aids performance on demanding questions.
It is called gaze aversion, and it is well documented in both children and adults. When a question gets harder, people break eye contact to answer it. Children begin doing it around five and are using it like adults by about eight. It is not evasion. It is bandwidth. Holding someone's face costs processing, and difficult thought needs that processing back.
What was checkedChildren begin doing it around five and are using it like adults by about eight.
Gaze aversion is used far less by 5-year-olds, rising through early primary years and reaching adult levels by age 8.
Your pupils track the pupils of the person in front of you. Research on pupil mimicry finds this synchronization happening automatically, and finds that mirroring a partner's dilating pupils is associated with trusting them more.
Kret et al. 2015. The script says 'is associated with' rather than 'causes' - the correlational framing is deliberate.
But the same work found something colder. The effect was moderated by group membership. The trust followed the mimicry most reliably when the partner belonged to the participant's own group.
Trust followed pupil mimicry reliably only for in-group partners; the 2017 Proc B paper examines the moderation directly.
In a study of twenty healthy participants doing exactly that, people reported dissociation and hallucination. The other person's face deformed. Some saw a stranger. Some saw a monster. Some saw an animal, or a dead relative, or their own face looking back at them.
Caputo: 20 healthy participants, dyadic gaze at low illumination for 10 minutes; dissociative symptoms and strange-face apparitions including monstrous beings, unknown persons, animal faces and relatives' faces.
Sharp vision lives in the fovea, a tiny patch covering roughly one to two degrees of your visual field. At conversational distance, a person's two eyes sit further apart than that. The geometry does not allow both at once.
The central one to two degrees of the visual field (the fovea) carries the bulk of high-acuity visual information.
Sharp vision lives in the fovea, a tiny patch covering roughly one to two degrees of your visual field. At conversational distance, a person's two eyes sit further apart than that. The geometry does not allow both at once.
Assembled, not quoted. Mean interpupillary distance is ~63mm; at ~1m conversational distance that subtends ~3.6 degrees of visual angle, exceeding the 1-2 degree foveal window, so both eyes cannot be foveated simultaneously. The premise (C16) is sourced; the conclusion is inference. The script phrases it as a consequence of anatomy, never as 'a study found'.
And the thing you are aiming at is a hole.
What was checkedThe pupil is black for a plain reason. It is an aperture. Light goes in and does not come back out.
Standard optics; no source retrieved. Uncontroversial but unchecked.
Which is where the word comes from. Pupil is from the Latin pupilla, little doll, named for the miniature person you see standing in someone else's eye. Greek did the same thing with kore. Girl, doll, and pupil of the eye, all one word.
Latin pupilla = little doll, named for the small reflection of oneself seen in another's eye; Greek kore carries girl / doll / pupil in one word.
Start with the practical damage of getting this wrong. We treat eye contact as a character test. Confidence. Honesty. Respect. That reading is parochial. In many cultures, holding a superior's gaze is not respect at all. It is aggression.
Sourced in a follow-up pass at the user's request; the script wording was not changed. Akechi et al. 2013 (PLOS ONE, 20 Finnish and 20 Japanese adults) found that physiological responses to eye contact were comparable across cultures, but evaluative judgements diverged: Japanese participants rated direct-gaze faces as more angry, less approachable and more unpleasant than Finnish participants did. The authors attribute this to differing cultural display rules rather than a physiological difference - which is precisely the script's point that the Western reading is parochial rather than universal. The hierarchy element ('a superior's gaze') is covered by S20, a Leadership Quarterly review reporting that people readily avert gaze from dominant others and that subordinates avert their eyes in deference to a superior, with the effect strongest in hierarchy-oriented cultures. Note the script says 'aggression' where Akechi measures perceived anger and unapproachability; that is a close but not identical construct.
And for autistic people, the cost is measurable. Brain imaging found that when participants were free to look wherever they wanted, activation looked broadly similar across groups. But when gaze was constrained to the eye region, participants on the autism spectrum showed abnormally high activation in those same subcortical structures — the amygdala, the superior colliculus. The alarm, running hot.
Hadjikhani et al. 2017: comparable activation on free viewing; overactivation of amygdala and superior colliculus in the ASD group when gaze was constrained to the eye region.
In sixteenth-century Italy, women dripped poison into their eyes.
The plant was deadly nightshade. The effect was a pupil blown wide and black. And the name we still use for that plant is the compliment they paid it. Belladonna. Beautiful woman.
They were not wrong. They had found something real about how we read each other, roughly four hundred years before anyone thought to measure it. Wide pupils register as warmth. As interest. As safety.
What they could not have told you is why. Or why the same gaze that feels like being held by one person feels like being interrogated by another. Same eyes. Same act. Opposite sensation.
Your face is running two contradictory programs, and it never told you.
Start with something strange about your own eye.
Look at a chimpanzee's, or a gorilla's, or almost any primate's. The sclera, the part you would call the white, is not white. It is pigmented, dark, and it blends into the iris. You can see the animal is looking. You cannot easily see where.
A comparative survey of eighty-eight primate species found that exposed white sclera was unique to us. Ours is unpigmented. Ours is more widely exposed. And the human eye is stretched horizontally further than any other primate's, which does one thing very efficiently. It advertises the direction of your attention.
Every other primate on that list kept its camouflage. We gave ours away.
Think about what that means as a trade. Every other ape can look at the fruit, the rival, the exit, without being caught looking. We cannot. We broadcast. Our eyes leak, on purpose, because whatever pressure did this to us — cooperation, teaching, the ability to trust and be trusted — was strong enough to be worth surrendering the poker face.
It is installed early, too. In one study, newborns between two and five days old preferred to look at faces that were looking back. Days old. Before language. Before recognition. Before anything you would call a self.
So far, this is a lovely story about connection.
Here is the problem.
A direct stare is one of the oldest threat signals in the animal kingdom. Predators lock on. Prey freeze. Butterflies grew fake eyes on their wings because a pair of eyes pointed at you is, across enormous stretches of biology, very bad news.
We inherited that circuit. It runs underneath thought, through structures like the superior colliculus and the amygdala — the brain's oldest lookouts, on duty and reading eyes long before we had opinions about anyone.
So here is the situation you are actually in. You are carrying a broadcasting organ, evolved to be read, mounted directly above an alarm system that treats being read as danger.
Two instructions. Written at different times. Never reconciled.
And that is the answer to the question we started with. The reason a lover's gaze feels like warmth and a stranger's feels like exposure has almost nothing to do with the eyes. The eyes are identical. What differs is which of those two systems gets there first — and that gets decided by safety, before you have looked at anything at all.
You do not decide it. You arrive already having decided.
Once you know that, the strange rules start to make sense.
Researchers ran nearly five hundred people from fifty-six countries through varying durations of simulated eye contact and asked what felt right. The average preference came out at about three point three seconds. The comfortable band sat between two and five. Almost nobody wanted more than nine.
Three seconds. That is the whole budget.
And notice what it is not. It did not track well with gender, or personality, or attractiveness. What it did track with was pupil response — physiological arousal, the body setting the timer without asking you.
Then there is this.
You look away in order to think.
It is called gaze aversion, and it is well documented in both children and adults. When a question gets harder, people break eye contact to answer it. Children begin doing it around five and are using it like adults by about eight. It is not evasion. It is bandwidth. Holding someone's face costs processing, and difficult thought needs that processing back.
Which means the sentence "look at me when I'm talking to you" is, mechanically, a demand to stop reasoning.
We say it to children. We say it in courtrooms. We say it in interviews, and then we grade people on how well they comply.
The traffic goes both ways, too, and none of it is under your control.
Your pupils track the pupils of the person in front of you. Research on pupil mimicry finds this synchronization happening automatically, and finds that mirroring a partner's dilating pupils is associated with trusting them more.
But the same work found something colder. The effect was moderated by group membership. The trust followed the mimicry most reliably when the partner belonged to the participant's own group.
So the machinery that lets you feel close to someone by matching them, involuntarily, below the level of awareness, checks who they are first.
And if you push the whole system too far, it comes apart.
Take two people. Put them in low light. Have them hold each other's gaze for ten minutes.
In a study of twenty healthy participants doing exactly that, people reported dissociation and hallucination. The other person's face deformed. Some saw a stranger. Some saw a monster. Some saw an animal, or a dead relative, or their own face looking back at them.
Ten minutes. No drug. Just sustained looking.
Whatever eye contact is, it does not get clearer the longer you hold it open. Push it past a few seconds and perception starts inventing.
Now the part I have been building toward.
You have never looked into someone's eyes. Not once. It is not physically possible.
Sharp vision lives in the fovea, a tiny patch covering roughly one to two degrees of your visual field. At conversational distance, a person's two eyes sit further apart than that. The geometry does not allow both at once.
So you flick. One eye, then the other, or some compromise point in between. What you experience as steady, continuous mutual gaze is assembled out of alternating glances your brain smooths into a single act, the way a flipbook smooths still drawings into motion.
And the thing you are aiming at is a hole.
The pupil is black for a plain reason. It is an aperture. Light goes in and does not come back out. There is nothing behind the darkness to find. The window to the soul is an opening in a wall.
But get close enough, and there is one thing visible in it.
A small, bright, upside-down figure. You.
Which is where the word comes from. Pupil is from the Latin pupilla, little doll, named for the miniature person you see standing in someone else's eye. Greek did the same thing with kore. Girl, doll, and pupil of the eye, all one word.
Two civilizations looked into the deepest part of another human being, found themselves, and named it after the doll.
I owe you the strongest argument against where this is heading.
It would be easy to land somewhere cheap. Connection is an illusion, you are only seeing your own reflection, none of it is real. That reading is false, and it is worth showing why.
Start with the practical damage of getting this wrong. We treat eye contact as a character test. Confidence. Honesty. Respect. That reading is parochial. In many cultures, holding a superior's gaze is not respect at all. It is aggression.
And for autistic people, the cost is measurable. Brain imaging found that when participants were free to look wherever they wanted, activation looked broadly similar across groups. But when gaze was constrained to the eye region, participants on the autism spectrum showed abnormally high activation in those same subcortical structures — the amygdala, the superior colliculus. The alarm, running hot.
So when someone will not meet your eyes, one live possibility is that they are managing a real physiological load. Not hiding anything. Not lying. Turning down a volume you cannot hear.
And that is where the cynical reading collapses. If eye contact were genuinely a window, if you could see in, then avoiding it would be evidence of something to hide. But there is nothing in there to see. Which means the entire apparatus of judgment we built on top of it was never measuring honesty at all.
It was measuring compliance.
So what is actually happening, when someone looks at you and you feel known?
Not transmission. There is no channel. What is happening is that two nervous systems, each unable to see inside the other, are running an enormous amount of inference on a very small amount of light. Pupil width. Timing. How long before the look breaks. Whether it comes back.
And the fact that it is inference does not make it less. It may be the only kind of intimacy there has ever been. You are not receiving someone. You are constructing them, continuously, and letting them construct you.
The Romans named the deepest part of the eye after the little reflection of whoever was standing in front of it.
They may have been more accurate than they knew.
So here is what I will leave you with. Tonight, with someone you love, hold it past the three seconds. Past the point where it starts to get uncomfortable.
Then ask yourself, honestly, which of the two you are afraid of.
That they will finally see something in you.
Or that you will notice there was never any way for them to.