Mike Morgan: Our eyes can play tricks, but it's not an illusion

From a lecture to the Collÿge de France by the Professor of Visual Psychophysics at City University, London

Popular books and TV programmes about the science of vision never fail to mention phenomena called "visual illusions". Why does a grey square surrounded by white look darker than an equal grey square surrounded by black? Why does a line with inward-pointing arrowheads at each end look shorter than an equal line with outward-pointing arrowheads? Why does the setting full moon look larger than the full moon at its zenith?

These images are supposed to help us question the basis for our visual perception, and perhaps they do. But what does it mean to call them "illusions?" The implication seems to be that most of the time our perception is telling us how things really are. I argue that by fostering this naive theory of vision, the pedagogy of illusions causes more harm than good.

The history of visual science is full of wonderful "illusions" that disappeared as soon as their physics was properly understood. First to go were the "optical illusions" such as the apparent bending of a stick by refraction at the boundary between air and water. Since the whole business of image formation by the lens and cornea of the eye depends upon refraction, it is a bit hard to call it an illusion.

An interesting question is whether spear-fishermen ever learn to compensate perceptually for refraction. Presumably they do, or they would catch no fish. But does this mean that they see their spear straight when it is half in the water? And if they did, would it be they who were suffering from an illusion, or ourselves?

When moving pictures were first invented they were considered a marvellous illusion. Not any more. The early films were so slow and jerky that it was easy to think of the movement they represented as illusory. Now, a movie is indistinguishable perceptually from the "real thing", so there is less temptation to question it.

From the physicists' point of view, a movie is simply a series of snapshots that contain enough information to reconstruct the trajectory of a moving object. There is no mystery in the fact that the brain is able to perform this reconstruction. The only matter requiring explanation is why we do not see the "flicker" between frames. The movie projector helps by exposing each frame twice, so that the rate of flicker is above our sensory limit.

This is no more of an illusion than the fact that we fail to see a light bulb flickering at mains frequency. An incapacity is not the same thing as an illusion. If it were, we should have to say that it is an illusion that we, unlike the falcon, are unable to see the excreta of rodents glowing in the ultraviolet.

Colour is the other great non-illusion. To the early colourists and dyers, colour mixture was a wonderful illusion. Pure red and green lights added together produce the sensation of yellow; red, green and blue added together produce white. Newton, who discovered the laws of colour mixture, was more canny. He remarked that the "rays... properly speaking are not coloured". He meant that colours are the effects of the rays upon us.

A lump of coal can be made to look white by illuminating it with a spotlight in a dark cellar. Nature performs the same trick by making the moon shine brightly at night. Again, a deeper understanding of the function of vision comes to our rescue. The function of brightness perception is not to measure the amount of light like an exposure meter. The eye does not need an exposure meter: it has no film. We see coal as black in bright sunlight because it reflects less light than surrounding objects. Brightness tells us about the reflectance properties of objects, not about the rays of light coming from the object.

So why do outgoing arrowheads make a line seem longer? I don't know, but what I do know is that we shall crack the code only when we stop calling this an illusion and enquire instead into the functions of seeing one object as larger than another.