The coffer illusion: What you see depends on your experiences
Which do you see first? Rectangles or circles?

[Source: The Illusions Index]
Psychological scientist Anthony Norcia of The Smith-Kettlewell Eye Research Institute created the coffer illusion for the 2006 Illusion of the Year contest sponsored by the Neural Correlate Society. (Here is Norcia’s submission, which made their top ten list for that year.)
If you see the rectangles, you understand why it’s called the coffer illusion. Or at least you understand the coffer part if not yet the illusion part. A coffer is a chest. A coffered ceiling is made up of recessed rectangular panels such that it looks like a series of chests.
However, if you see the circles, all this business about rectangles and chests will mean nothing to you.
Before we go any further, let’s make sure everyone can see both the rectangles and the circles.

Now scroll back up to the original coffer illusion image. Can you now flip back and forth between rectangles and circles?
In preprint version 3, Ivan Kroupin and colleagues (2025) wondered if living in a world dominated by rectangles, e.g., rectangular buildings, would predispose viewers of the coffer illusion to see rectangles first. In contrast, they wondered if living in a world dominated by circles, e.g., round huts, would predispose viewers to see circles first.
The researchers asked online participants residing in the United States and the United Kingdom to report what shapes they saw. Researchers visited two locations in northern Namibia to ask participants in person. One location, Opuwo, is semi-urban while the other, Himba, was rural.
Once participants identified a shape (rectangles or circles), they were asked if they saw any other shapes. Participants could take as long as they wished to respond.
US & UK | Semi-urban Opuwo | Rural Himba | |
Only rectangles | 81% | <1% | <2% |
Rectangles and then circles | 17% | 13% | <3% |
Circles and then rectangles | 3% | 67% | 48% |
Only circles | 0% | 19% | 48% |
(Numbers may not add up to 100% due to rounding.)
The researchers provide a wonderful example of the different lenses the participants and the researchers brought to the coffer illusion.
When the researchers asked one participant
“’What do you see here?’ our participant responded “Houses.’ This was a surprise to both the experimenter and translator since we assumed that she meant rectangles, and seeing rectangles first is very rare in the rural group (~4% in the final sample). However, when we asked her to point to the things she saw, she picked out all of the circles in the image. We asked her to draw what she saw and she drew a neat circle in the sand (n.b. this request to draw the image was not part of the final protocol). Clearly, the geometry we (the experimenter and translator) assumed to correspond to ‘houses’ is very different from the one assumed by our participant. Traditional Himba houses are circular” (p. 5).
The researchers report similarly stark differences in the perception of three other illusions: curvature blindness, café wall, and Gestalt shapes. All the results were fascinating, but the results that really struck me were from the Gestalt shapes illusion.

Participants were asked “What do you see here?” If they didn’t name a shape, the researchers prompted with “Do you see a [name of shape] here?”
US & UK | Semi-urban Opuwo | Rural Himba | |
Reported shape without prompting | 93% | 38% | 10% |
Identified the shape after it was named | 6% | 23% | 12% |
Do not see the shape at all | 1% | 38% | 78% |
The researchers note that “Urban environments are filled not only with angles but with standardized geometric shapes: perfect circles, triangles, and squares in everything from logos to packaging. Indeed, we explicitly teach our children to identify a range of these shapes in the visual environment” (p. 8).
The researchers note that while those living in the Himba village are more likely to see circular housing, that did not mean they were more likely to see the Gestalt circle. They wonder if seeing Gestalt shapes has less to do with squares, circles, and triangles, and more to do with a lack of exposure with inferring shapes from partial images.
The examples of cross-cultural differences in perception provided in this paper are just another set in a long line of examples (Phillips, 2019). It is, however, an excellent reminder that our environments shape what we perceive, and that you and I may not see things the same way.
References
Kroupin, I., Davis, H. E., Paredes Lopes, A. J., Konkle, T., & Muthukrishna, M. (2025). Visual illusions reveal wide range of cross-cultural differences in visual perception. PsyArXiv. https://doi.org/10.31234/osf.io/gxzcp_v3
Phillips, W. L. (2019). Cross‐cultural differences in visual perception of color, illusions, depth, and pictures. In K. D. Keith (Ed.), Cross-Cultural Psychology: Contemporary Themes and Perspectives (2nd ed., pp. 375–397). Wiley Blackwell.