What do optical illusions reveal about how we see?

Optical illusions reveal that seeing is not a photograph but an interpretation. Your visual system makes assumptions about the world, and an illusion is a picture built to make one of those assumptions fail. Scientists call the idea "unconscious inference", first suggested by Hermann Helmholtz in the 19th century.

Take the checker shadow illusion, published by the MIT vision scientist Edward Adelson in 1995. A checkerboard is partly in the shadow of an object, and square A looks darker than square B. But within the image they are of identical brightness, as if printed with the same ink. Your brain compensates for the shadow, so what you perceive is the surface colour, not the raw light.

Or take the Müller-Lyer illusion, devised in 1889 by the German sociologist Franz Carl Müller-Lyer. Two lines of equal length look unequal depending on whether their fins point in or out: the shaft with two tails looks longer. A well-known explanation, by Richard Gregory, is that the visual system has learned that fins pointing in look like the nearer corner of a room, and fins pointing out like a farther one. Another, simpler, example: mountains seem much nearer in clear weather, because haze is the depth cue your brain expects to see.

The same machinery that fools you makes your vision work. It is why a person walking away from you does not look like they are getting shorter. Illusions are the cases where a good rule meets a trick.

And not every answer is settled. Even the long-standard explanation for the Hermann grid, the grey ghosts at the crossings, has been disproved.