This famous illusion plays with our expectations about light and shadow. It demonstrates an important feature of vision called lightness constancy: our visual system tries to perceive the lightness of a surface as relatively stable even when the amount of light reaching it changes.
Look at squares A and B. They appear to be different shades of grey — but they have the same luminance. Luminance refers to how much light an object reflects back to you objectively (how many photons are reflected back into your eye), whereas lightness is our perception or subjective interpretation of how light or dark an object is; brightness, similarly, is the subjective interpretation of the intensity of the object. In this illusion, tiles A and B look different in lightness, but their luminance is actually the same! So why do they look different?
Part of the answer is Simultaneous Lightness Contrast (SLC): the appearance of a surface is influenced by the luminance of the surfaces surrounding it. A grey patch surrounded by lighter areas can appear darker, while the same grey patch surrounded by darker areas can appear lighter.
But Adelson’s illusion goes beyond simple contrast. The visual system doesn’t simply measure how much light reaches each point in the image. It also tries to work out what is causing the pattern of light and dark. In this image, the cast shadow provides a powerful contextual cue. Your visual system interprets the shadowed region as receiving less illumination and therefore adjusts its estimate of the surfaces within it.
This is an example of lightness constancy. In the real world, this ability is extremely useful. A white wall in shadow can send less light to your eyes than a black object in bright sunlight, but you still need to perceive the wall as white and the object as black. The illusion exploits that normally useful process. The image provides cues suggesting that the two squares are under different lighting conditions, even though the target squares have identical luminance. Your visual system therefore interprets them as having different surface lightness.
There are several proposed explanations for how this happens. Low-level mechanisms, including centre-surround responses and lateral inhibition in early visual processing, contribute to how local contrast is encoded. But these mechanisms alone cannot fully explain the checker-shadow illusion. Higher-level and mid-level processes involving scene structure, edges, shadows and assumptions about illumination also appear to play a role.
This connects to a theory proposed by Hermann von Helmholtz: perception can involve unconscious inference. Rather than simply recording the light entering your eyes, your visual system uses sensory information and prior experience to infer what is most likely happening in the world. In this case, your brain makes a perfectly reasonable assumption: “That square is in shadow, so it must actually be lighter than it looks.” And that’s exactly what makes the illusion so convincing.