BOOK REVIEW : Contrasting Logic and Natural Intelligence

The Improbable Machine: What the Upheavals in Artificial Intelligence Research Reveal About How the Mind Really Works by Jeremy Campbell (Simon and Schuster: $19.95; 334 pages)

The subtitle of this wonderful book says it all: “What the upheavals (read: failures) in artificial intelligence research reveal about how the mind really works.” Computers are logic machines. But people don’t think that way. That’s why all efforts to get computers to think have failed.

Jeremy Campbell, a science writer of imagination and skill, here puts to rout the notion that reason--logic--is how we decide things. It is not. There is substantial experimental evidence in cognitive science that shows that the way we think is very dependent on what we’re thinking about.

“Humans are not general-purpose logicians,” Campbell writes. “What they know has a powerful effect on how they think. Something as seemingly serene and detached as reason turns out to be intimately entangled with the business of being and acting in the world . . . What gives logic its tremendous power--its indifference to content--is not what gives natural intelligence its power.”

Therein lies the difference between brains and machines. Logic--the step-by-step method of Euclid--works well only in highly formalized situations where the facts are complete and clean. But that is not the real-world situation at all.

In the real world, we have mind sets, belief systems, biases which enable us to fill in the gaps and make sense of the information that constantly flows into our senses by the bushel. In most cases, we don’t reason things out step-by-step. Our perceptions and our conclusions occur simultaneously. We recognize our aunt in a flash, even if she’s changed her hairdo or is a wearing a new hat (something computers can’t do).

“The serial computer is most at home in a world very unlike our own, a world of well-defined puzzles having clear solutions, exhaustive descriptions, lists, explicit rules, literal meanings, perfect information,” Campbell writes in his introduction to “The Improbable Machine.”

“The world in which the brain evolved is a great deal more untidy and less circumscribed, full of ambiguity, deceit, problems that sprawl in ungainly fashion and have no firm boundaries, words with multiple meanings, faces that send enigmatic messages, information that is incomplete or contradictory, answers that breed endless families of further questions.”

This book is more about the brain and how it thinks than about computers and why they don’t, though Campbell makes clear that many computer scientists have turned to “connectionism”--neural networks--in an effort to simulate how the brain actually works.

But his focus is on the brain and the far-reaching implications of the realization that our thought processes are anything but logical. The need to get along in the world has made brains very pragmatic. It forced our ancestors and forces us to invent structures--”schema,” Campbell calls them--for interpreting sense data.

Once those structures are in place, it is very hard to dislodge them by argument, reason, facts, logic or any “rational” device. We overvalue evidence that supports our world view and explain away or dismiss evidence that contradicts it. Liberals never persuade conservatives, and conservatives never persuade liberals.

Our brains have a need to explain things, and they resist the notion that some things cannot be explained. Campbell knows, for example, that the newspaper articles that appear every day purporting to explain the stock market’s activity the day before are bunk.

“Explanations are a means of bringing an uncertain world under control,” he writes. “They make random events seem to reveal a hidden, rational order, and tame the wild beasts of mere chance. Every day, experts on the stock market come up with at least one plausible explanation of why the index went up or down the previous day, with definite causes neatly producing logical effects. Yet if there is an underlying order to the gyrations of the stock market, economists have been unable to discover it.”

And that’s just the start of it. We make up “good stories” about the world, Campbell says, and they “ride roughshod over the principles of logic and the laws of probability, immunizing the mind against surprise, and consequently blocking the path to wisdom.”

I cannot praise this book enough. It draws on a wide and fascinating array of sources to show how illogical we really are, and it cuts through the gobbledygook that usually passes for discourse.

It is ostensibly about the mind, which is embodied in the brain, but it is really about how we know. As a result, it has implications in every area of knowledge and thought, which includes quite a lot.