Do Optical Illusions Work the Same for Everyone?

It's natural to assume that an illusion works on everyone in the same way, like a reflex. That's not quite true. The basic machinery behind illusions, such as grouping, contrast processing and depth inference, is shared by nearly all sighted adults. But how strongly a particular illusion catches a particular person varies with culture, experience and individual differences in the brain. That makes illusions a useful way of asking how much of perception is built in and how much is shaped by experience.

Culture and the "carpentered world"

Some of the earliest cross-cultural research on perception used the Müller-Lyer illusion, in which arrow-like fins make identical lines look different lengths. In the 1960s, Marshall Segall, Donald Campbell and Melville Herskovits compared people living in environments with very different amounts of straight-edged, rectangular architecture (Segall, Campbell & Herskovits, 1963). Their conclusion, which became known as the "carpentered world" hypothesis, was that people raised among right angles and rectangular buildings show stronger Müller-Lyer effects, because they've learned to read the fins as corners.

The interpretation has been debated ever since, but the broader lesson has held up, and it's become a standard example of a wider problem. Much of psychology is based on people from Western, educated, industrialised, rich and democratic ("WEIRD") societies, who aren't always typical of humanity as a whole (Henrich, Heine & Norenzayan, 2010). A more recent example comes from the Ebbinghaus illusion: the Himba of northern Namibia are much less fooled by it than English participants, perhaps because they pay more attention to local detail than to the scene as a whole (de Fockert et al., 2007).

Knowing doesn't switch it off

Training does change some kinds of perception. Radiologists, for example, become very fast at spotting what they're looking for in scans. But training rarely removes an illusion. Vision scientists who have studied the checker shadow for years still see the two squares as different shades, because the effect happens at a level of processing that knowledge can't reach. Expertise improves speed and accuracy within a trained skill; it doesn't change the low-level machinery. Even experts can be caught out by what they aren't looking for, as the radiologists who missed a gorilla hidden in a lung scan showed (see the invisible gorilla).

Differences between individual brains

How strongly an illusion works also varies between people for reasons in the brain itself. People whose primary visual cortex has a smaller surface area tend to experience stronger Ebbinghaus and Ponzo illusions (Schwarzkopf, Song & Rees, 2011), and how strongly people experience the McGurk effect varies widely and stays stable over time.

Autism is a well-known example where the story turned out to be more complicated than it first seemed. Francesca Happé reported in 1996 that autistic children were less susceptible to several classic illusions, which fitted the idea that autistic perception focuses more on detail and less on the whole (Happé, 1996). But when Danielle Ropar and Peter Mitchell tested people with autism and Asperger's syndrome on a computer task, they found them about as susceptible as everyone else (Ropar & Mitchell, 1999), and most later studies have agreed, with a few exceptions depending on the illusion. It's a good reminder that a striking early result needs to be replicated before it becomes accepted as fact.

Age, tiredness and attention can shift how strongly an illusion works, too, generally by changing how reliable the incoming signals are or how heavily the brain leans on its expectations. I've written more about that in predictive processing and the brain's shortcuts.

Measure your own

I've built tests on this site so you can see where you fall yourself. The Müller-Lyer, Ebbinghaus and Ponzo pages measure how far your size judgements get pulled, and compare you with other visitors once enough people have taken part. The motion aftereffect page times your waterfall illusion, and the polls on the dress, the spinning dancer and the Necker cube show how visitors split.

A shared toolkit, tuned differently

None of this means illusions are arbitrary. The mechanisms behind them are close to universal, which is why illusions like the café wall fool almost everyone. What varies is how strongly each mechanism is applied: how heavily a given expectation is weighted, how automatically a grouping cue gets used, and how much experience has tuned a particular inference. We all have the same toolkit, but not everyone's is set up in quite the same way, and that variation is itself evidence of just how flexible human perception is.

Sources and further reading

1. Segall, M. H., Campbell, D. T., & Herskovits, M. J. (1963). Cultural differences in the perception of geometric illusions. Science, 139(3556), 769–771. Find this reference on Google Scholar
2. Henrich, J., Heine, S. J., & Norenzayan, A. (2010). The weirdest people in the world?. Behavioral and Brain Sciences, 33(2–3), 61–83. Find this reference on Google Scholar
3. de Fockert, J., Davidoff, J., Fagot, J., Parron, C., & Goldstein, J. (2007). More accurate size contrast judgments in the Ebbinghaus illusion by a remote culture. Journal of Experimental Psychology: Human Perception and Performance, 33(3), 738–742. Find this reference on Google Scholar
4. Happé, F. G. E. (1996). Studying weak central coherence at low levels: Children with autism do not succumb to visual illusions. A research note. Journal of Child Psychology and Psychiatry, 37(7), 873–877. Find this reference on Google Scholar
5. Schwarzkopf, D. S., Song, C., & Rees, G. (2011). The surface area of human V1 predicts the subjective experience of object size. Nature Neuroscience, 14(1), 28–30. Find this reference on Google Scholar
6. Ropar, D., & Mitchell, P. (1999). Are individuals with autism and Asperger's syndrome susceptible to visual illusions?. Journal of Child Psychology and Psychiatry, 40(8), 1283–1293. Find this reference on Google Scholar