Sunday, December 6, 2009

Patterns, Probability, and Plagiarism (Part 5 of "Cognitive Biases")

This week's installment of Cognitive Biases, the ways in which our brains distort our thinking is brought to you by the letter “C.” The series begins here.

Clustering illusion

Is the sequence below random or non-random??

OXXXOXXXOXXOOOXOOXXOO

If you thing the sequence looks non-random, you’re with the majority…but you’re wrong. The sequence has several characteristics of a random stream, an equal number of each result and an equal number of adjacent results. But people seem to expect a “random” sequence to have a greater number of alternations (O to X or vice-versa) than statistics would predict. The chance of an alternation in a sequence of independent random binary events (flips of heads or tails) is 50%, but people seem to expect an alternation rate of about 70%.

The clustering illusion is a cognitive bias that creates atendency to see patterns where actually none exist. This is why most people believe in “streaks.” When you expect greater variation in a sequence, you tend to assume that there’s a trend. But that isn’t necessarily the truth.

Conjunction fallacy

Linda is 31 years old, single, outspoken, and very bright. She majored in philosophy. As a student, she was deeply concerned with issues of discrimination and social justice, and also participated in anti-nuclear demonstrations.

Which statement is more probable?

1. Linda is a bank teller.
2. Linda is a bank teller and is active in the feminist movement.

In a 1982 study by Amos Tversky and Daniel Kahneman, 85% thought statement 2 was more probable than statement 1, but that’s wrong. The probability of two events occurring together is always less than or equal to the probability of either one occurring alone. Even if there’s a very low probability Linda is a bank teller (let’s make it 5%) and a very high probability that Linda is active in the feminist movement (95%), the chance that Linda is a bank teller AND active in the feminist movement is 5% x 95%, or 4.75%, lower than the first statement.

The conjunction fallacy happens when you assume that specific conditions are more probable than a single general one, which is a violation of basic logic. Now, one possibility is that because most people aren’t familiar with the rules of formal logic, they may assume that statement 1 (Linda is a bank teller) implies that she isn’t active in the feminist movement.

But the fallacy has been demonstrated with very educated audiences.

Another Tversky/Kahneman experiment in the early 1980s surveyed a group of foreign policy experts to determine the probability that the Soviet Union would invade Poland and the US would break off diplomatic relations in the following year. The consensus estimate was about a 4% chance. Next, another group of experts was asked the probability that the United States would break off relations with the Soviet Union the following year. They estimated only a 1% chance. This implies that the detailed, specific scenario of the first scenario all by itself made it more likely.

Cryptomnesia

Robert Louis Stevenson refers to an incident of cryptomnesia that took place during the writing of Treasure Island, and that he discovered to his embarrassment several years afterward:


I am now upon a painful chapter. No doubt the parrot once belonged to Robinson Crusoe. No doubt the skeleton is conveyed from Poe. I think little of these, they are trifles and details; and no man can hope to have a monopoly of skeletons or make a corner in talking birds. The stockade, I am told, is from Masterman Ready. It may be, I care not a jot. These useful writers had fulfilled the poet's saying: departing, they had left behind them Footprints on the sands of time, Footprints which perhaps another — and I was the other! It is my debt to Washington Irving that exercises my conscience, and justly so, for I believe plagiarism was rarely carried farther. I chanced to pick up the Tales of a Traveller some years ago with a view to an anthology of prose narrative, and the book flew up and struck me: Billy Bones, his chest, the company in the parlour, the whole inner spirit, and a good deal of the material detail of my first chapters — all were there, all were the property of Washington Irving. But I had no guess of it then as I sat writing by the fireside, in what seemed the spring-tides of a somewhat pedestrian inspiration; nor yet day by day, after lunch, as I read aloud my morning's work to the family. It seemed to me original as sin; it seemed to belong to me like my right eye.

Sometimes what seems like inspiration turns out to be memory, and you’ve committed inadvertent plagiarism, or cryptoamnesia. In a 1989 study, people generated examples (such as kinds of birds), and later were asked to create new examples and to recall which answers they had previously personally given. Between 3-9% of the time, people either listed examples previously given, or recalled as their own someone else’s thought.

Few writers would risk committing deliberate plagiarism, but the dangers of cryptoamnesia are real. It’s most likely to occur when you don’t have the ability to monitor your sources properly, when you’re away from the original source of the idea, or when the idea was originally suggested by a person of the same sex (!). It’s also likely to happen in a brainstorming session, in which you recall as yours an idea that came up immediately before your idea.

Of course, not all claims of cryptoamnesia are necessarily valid; sometimes the plagiarism was all too deliberate. But nothing else explains certain situations in which people with an awful lot to lose commit what appears to be blatant plagiarism with no upside whatsoever.

The courts have ruled that the unconsciousness of the plagiarism doesn’t excuse it; the classic (rock) case is Bright Tunes Music v. Harrisongs Music involving the similarities between “He’s So Fine” and “My Sweet Lord.”

That cost George Harrison $587,000.

Cognitive biases can be expensive.

More next week.

No comments:

Post a Comment