by Gary Marcus

September 2, 2009

This is a cute and clever little book (178 pages in a quarto volume) presenting symptoms of evolutionary psychology. The brain is not a perfect all-purpose computing machine; it is instead an organ that developed piece by piece over the course of millions of years, adapting in a completely ad hoc fashion to new environmental requirements. It is without doubt a gigantic kluge, a patchwork of alterations, fixes, corrections, and adjustments piled one on another. While the harsh logic of natural and sexual selection has enforced its overall effectiveness, there remain many tiny symptoms of the haphazard process by which it evolved. This book presents a compendium of those symptoms.

The most obvious of these are optical illusions, of which there are many. Psychologists have devised hundreds if not thousands of these as a means to test the performance of the visual system, establishing exactly how the visual system can confuse images. The most common of these are the deceptive figure-ground relationships, in which two images identical in color, shade, or shape are placed in front of two very different backgrounds. The result is that the eye perceives the foreground images to be different, even though they are identical. This happens because the visual system takes into consideration the entire visual field to interpret its components. That's good design: the information in the background of the image usually provides cues that help interpret the foreground. The system works very well with almost all natural images. But it is possible to devise unnatural images that confuse the visual system, such as optical illusions.

But there are many other kluges. Attention, for example, can blind us just as readily as it can help us see: Watch this to see. Human memory is completely unreliable for many purposes, because our memories are not stored like data on a hard disk, but instead they are compressed in terms of interpretations based on our current knowledge. If I were to show you, for example, a photograph of some weird shapeless sea creature, and then ask you to describe it a week later, you would probably describe something rather different from the original image, reflecting your memory as interpreted by visual imagery you are familiar with: "It was like a duck-billed dragon." Worse, we bring considerable prejudice to our memories. We note closely those things that we expect, and reject those things we do not expect (this is called "confirmation bias").