Online Book Reader

Home Category

Being Wrong - Kathryn Schulz [57]

By Root 1057 0
the goal of most models of optimal human cognition proposed by most thinkers throughout most of history, and it is the goal behind our own broadly shared image of the ideal thinker. Some of these models, like Descartes’, sought to curtail error through radical doubt. Others sought to curtail it through formal logic—using valid premises to derive necessarily valid conclusions. Others, including our shared one of the ideal thinker, seek to curtail it through a kind of general due diligence: careful attention to evidence and counterevidence, coupled with a prudent avoidance of preconceived notions.

By these standards, the cognitive operating system we actually come with is suboptimal. It has no use for radical doubt. It does not rely on formal logic. It isn’t diligent about amassing evidence, still less so about counterevidence, and it couldn’t function without preconceived notions. It is eminently capable of getting things wrong. In short, our mind, in its default mode, doesn’t work anything like any of these models. And yet—not despite but because of its aptitude for error—it works better than them all.

Here, I’m going to prove it to you. Actually, I’m going to get you to prove it to me. Below you will find a very brief multiple-choice test. Please take it. If you still have that four-year-old handy from the last chapter, have her take it, too. You can reassure her that these aren’t meant to be trick questions, so neither of you should overthink anything. Here goes:

What is behind the shaded rectangle?

A.

B. I have no idea.

2. You are traveling in the country of Quine. A speaker of Quinean shows you this picture and says, “This is a Gavagai.” What is a Gavagai?

A. A rabbit.

B. How should I know?

3. Complete the following sentence: “The giraffe had a very long ______.”

A. Neck

B. I’m stumped.

Congratulations; you got all three answers right. You also found this quiz so easy that you don’t think congratulations are called for. But here’s the thing: this quiz is not easy. At least, it is not intrinsically easy. True, you could do it, and so can I, and so can any reasonably attentive four-year-old. Yet computers—which can calculate pi out to a thousand digits while you sneeze—are completely stymied by problems like these. So here is a non-easy question for you: Why is something that is so effortless for a person all but impossible for a machine?

To get a sense of the answer, consider just a tiny fraction of the possibilities a computer has to contemplate when taking this quiz*:

What is behind the shaded rectangle?

A.

B.

C. nude pictures of Lindsay Lohan

2. You are traveling in the country of Quine. A speaker of Quinean shows you this picture and says, “This is a gavagai.” What is a Gavagai?

A. grass

B. rabbit plus grass

C. dinner

3. Complete the following sentence: “The giraffe had a very long ______.”

A. tongue

B. flight from Kenya

C. history of drug abuse

These answers strike us as patently absurd. But computers don’t have a patent-absurdity monitoring function; they can’t rule out these answers (or millions of others like them) because nothing in logic prevents such answers from being right. For that matter, nothing in life prevents such answers from being right, either. It is perfectly possible that someone, somewhere—a zookeeper, a veterinarian, a children’s book author, someone enjoying a long history of drug abuse—has uttered a sentence about a giraffe’s long flight from Kenya. Likewise, it is perfectly possible (although perhaps less probable than some readers would like) that there are naked pictures of Lindsay Lohan behind that shaded rectangle. For that matter, it’s perfectly possible that there is a naked picture of you behind that rectangle—or of one of the other 7 billion people on the planet, or two pictures of three people, or ten pictures of seven people, or…you get the point: there are an infinite number of logically valid, theoretically possible answers to all these questions. Computers recognize this, and therefore can’t answer

Return Main Page Previous Page Next Page

®Online Book Reader