Strange-face illusion · mirror gaze · dim room

The first minute still looked like me.

JM Thomas · 17 August 2026

A long gaze in low light lets the face-file slip.

Dark lime drawing of a dim room, a 25-watt lamp on the floor, a square mirror, a ten-minute timer, and a warped oval in the glass. No face.
Dim room, lamp on the floor, ten-minute timer, warped oval. No face.

I stared in the bathroom glass to become the person who already had the apartment.

Lamp on the floor, glass at arm's length

The bulb stayed out of the glass.

Caputo printed the sit in Perception in 2010 as volume 39, pages 1007 to 1008, DOI 10.1068/p6466, after he sat fifty naive people ages 21 to 29, mean 23 with a standard deviation of 2.1, in a quiet room dimmed by a 25 watt incandescent lamp placed on the floor behind the chair so the bulb stayed hidden from the eyes and from the glass.

A mirror 0.5 m by 0.5 m sat about 0.4 m in front. Face luminance sat near 0.2 cd per square meter, enough to keep the fine traits and low enough to thin the color. The illusion still showed when he raised that light from 0.2 up to 1.6.

The task was a gaze at the reflected face. The strange face usually started in less than a minute.

What they wrote after the sit

The write-up was the measure.

After ten minutes he asked each person to write what the glass had shown. The descriptions split hard across the room.

Sixty-six percent of the fifty wrote huge deformations of their own face. Eighteen percent wrote a parent with the traits changed, 8 percent still alive and 10 percent already dead.

Twenty-eight percent wrote an unknown person. Another 28 percent wrote an archetypal face, and Caputo's examples included an old woman and a child.

Eighteen percent wrote an animal face. Fantastical or monstrous beings showed up in 48 percent of the scraps.

Dark lime split. Left a bright bathroom and a locked oval. Right a dim 25-watt lamp, a ten-minute timer, and a warped oval. No face.
Left: bright glass, oval locked. Right: dim sit, oval slipped.

Blank would have been cheap

Blank would have been the cheap story.

Troxler fading, the old peripheral wash, would predict the traits go blank if you stare long enough. The write-ups did not go blank. New faces arrived with new traits.

Caputo named a break in the bind that turns separate traits into one face. Long viewing of a weak face can scramble that bind and hand back a deformed face or a scrambled one. Unknown people in the glass point at a high-level face file. Animal faces and monstrous beings, he wrote, sit outside the usual face-processing story.

A still photo in the corner of the eye can warp too. Your own face in a dim central gaze hits harder because the copy is exact in space and time, and because the thing that warps is the face you brought in.

The other looked back

All fifty felt some form of otherness.

When a new face sat in the glass, people felt watched from inside it or beyond it. Some felt a malign look and got anxious. Some felt a smile and warmed.

Deceased parents and archetypal portraits produced a silent query. Monstrous beings produced fear. Dynamic deformations, a pulse or a shrink, produced a sense that things had slipped the leash.

The glass copies your face and ties the picture to the small motions of your own head. That is a broader stack than a photo on a desk. Stay in the gaze long enough and the self-file that learned this glass in childhood can lose its lock.

I wanted the apartment in the glass

I asked the bathroom for a new me.

The aisle sells mirror work as a speech you give your reflection until the life changes. The Urbino room measured a slip. I looped the morning pep talk at the sink and never dimmed the room long enough for the file to move.

You can want a new self and still run the dim sit. Write what showed.

Write what the glass showed.

Put a lamp on the floor behind you. Sit a mirror about 0.4 m in front. Kill the overhead. Gaze at the reflected face for ten minutes. Then write what the glass showed on a dated scrap.

This page is about a dim-room gaze that lets the face-file slip. The Library is where I keep the other sits.

After the ten minutes

Does the lamp have to be 25 watts?

Caputo used a 25 watt incandescent on the floor behind the chair. He also wrote that the sit still worked when face light rose from 0.2 up to 1.6 cd per square meter. The overhead can stay off.

What do I write if the face stays mine?

Write that. The file asked for a write-up. A monster was optional. A face that holds is still a result.

Can I run this with the bathroom light on?

The measured sit was a quiet room with the lamp on the floor behind you so the bulb stayed out of the glass. A bright overhead fights the dim that lets the file slip.

What if the glass makes me anxious?

Stop the gaze. Turn a light on. The write-ups already named fear when monstrous beings showed. Anxiety that stays the night belongs with a clinician.

Is this a medical page?

A dim-room gaze in a perception note. Literature. A face that will not come back, or distress that will not leave, belongs with a clinician.

Numbers Caputo printed

Caputo G.B. "Strange-face-in-the-mirror illusion." Perception 2010;39(7):1007-1008. DOI 10.1068/p6466. PMID 20842976 · Received 15 May 2009 · In final form 11 May 2010 · University of Urbino · n=50 naive individuals, ages 21 to 29, mean 23, SD 2.1 · Quiet room, 25 W incandescent on the floor behind the observer, hidden from eyes and glass · Mirror 0.5 m by 0.5 m, about 0.4 m in front · Face luminance about 0.2 cd m-2, still present from 0.2 to 1.6 cd m-2 · Gaze at reflected face · Strange-face usually after less than one minute · 10 min session, then a written description · Huge deformations 66% · Parent with traits changed 18% (8% alive, 10% deceased) · Unknown person 28% · Archetypal face 28% · Animal face 18% · Fantastical or monstrous beings 48% · All fifty reported some form of dissociative identity / otherness · Mechanism named as a break in trait-to-face binding, Troxler fading rejected because new traits arrived · Phenomenological note, no control arm, no inferential p. Simas 2000 Perception 29:1393-1395, multiple-faces on peripheral photos. Thompson 1980 Perception 9:483-484, Thatcher illusion. doi.org/10.1068/p6466 · pubmed.ncbi.nlm.nih.gov/20842976 · journals.sagepub.com/doi/10.1068/p6466