Geometric

Penrose Triangle

Three beams that form an impossible closed triangle. Turn a real 3D model in the demo and find the one direction from which it becomes impossible.

This is a real, possible 3D object: nineteen cubes in three straight beams, drawn without perspective, like a technical drawing. Turn and tilt it until both sliders are in the middle and you'll be looking exactly along the line that hides the gap, at which point the open chain becomes the impossible triangle. Sculptures of the Penrose triangle work the same way.

Three good corners, one impossible object

Three straight beams, each drawn with ordinary shading and thickness, join end to end in a closed triangle. Cover two of the corners and the third looks like an unremarkable right-angled joint, the kind you'd find on a bookshelf. Each corner passes that test. But follow the whole loop and the beams can't possibly meet as drawn: at least one beam would have to be both nearer to you and further away than the beam it joins.

The demo shows how a real object can produce the same picture. It's a chain of cubes in three straight beams that don't meet. Seen from exactly the right direction, the far end of the chain hides precisely behind the start, and the open chain looks like a closed, impossible triangle. Turn it slightly and the trick is obvious.

The brain checks locally

The visual system doesn't verify the geometry of a whole scene in one go. It builds a 3D interpretation piece by piece, corner by corner, and assumes the pieces will agree, which for real objects they almost always do. The Penrose triangle is designed to exploit that: each corner gives a perfectly clear depth cue, but the three cues contradict each other once you follow them round.

The pull towards seeing a single solid object comes from Gestalt tendencies such as closure and continuity. They make you see one triangle rather than three disconnected pieces, even though disconnected pieces are the only possible reality, as the demo shows.

Invented twice

The figure has two independent origins. The Swedish artist Oscar Reutersvärd drew the first known version in 1934, built from a row of cubes. More than twenty years later, the psychiatrist Lionel Penrose and his son Roger, a mathematician who would go on to win the 2020 Nobel Prize in Physics, arrived at it on their own and published it in a 1958 paper titled "Impossible objects: A special type of visual illusion", which gave the whole category its name (Penrose & Penrose, 1958; Ernst, 1986). Their paper reached M. C. Escher, who used the idea in his lithograph Waterfall.

Impossible sculptures

Because the contradiction only exists in a single flat view, sculptors can build Penrose triangles that work from one spot. The best known is a large aluminium sculpture in East Perth, Western Australia: walk around it and it's a set of disjointed beams, but stand at the right place and they line up into the impossible triangle. The demo above is a small version of the same idea. The Penrose stairs apply the same local-versus-global trick to a staircase, and the Kanizsa triangle is a triangle your brain builds from nothing at all.

Sources and further reading

  1. Penrose, L. S., & Penrose, R. (1958). Impossible objects: A special type of visual illusion. British Journal of Psychology, 49(1), 31–33. Find this reference on Google Scholar
  2. Ernst, B. (1986). Adventures with Impossible Figures. Tarquin Publications. Find this reference on Google Scholar
Discovered / popularized by
Oscar Reutersvärd; formalized by Roger and Lionel Penrose
Year
1958
Category
Geometric & Impossible Objects

Read the science behind why this happens →