Take one smart young asker of silly questions, add a smart young analyst who answers them seriously, stew over half a lifetime and what do you get? Two Nobel Prize winners and the birth of a whole new science known as behavioral economics, the study of what happens when people behave irrationally – as most of us do most of the time. For background, you might read The Undoing Project by Michael Lewis, but the short form is that people make errors, they do so systematically, and the results of those errors matter.
Consider the case of Linda, described as an intelligent, educated, liberal-minded professional woman in her thirties. From that description, nothing more, people including statisticians and economists and academics were asked to answer a silly question by choosing the more likely of two statements: first, Linda is a teacher, and second, Linda is a teacher who is an activist feminist. They overwhelmingly chose the second statement. Now clearly, if you calculate for a moment, among a thousand people answering the given description there will be some unknown number who are teachers. Of that number, some but not all will also be activist feminists. So there will be fewer activist feminists than the total number of teachers, and it must be more likely for Linda to be part of the larger group than of the smaller within it. Why, then, do people consistently choose the wrong answer?
The reason lies in the realm of human imagination. We are not computers who are fed data sets to calculate from. We evolved in a world of uncertainty, where it was important to imagine that a rustling bush might conceal a saber-tooth tiger. No hard data, just experience and imagination to supplement inadequate partial knowledge of our environment. In the case of Linda, people let their imaginations supplement the paltry description of Linda. They added detail, color, context and prejudices and stereotypes and what they had for breakfast, then answered the question with a vivid mental picture of ‘their’ Linda, who would quite possibly be an activist feminist; she might also be a teacher, but that part of the picture was much less vivid. Accordingly, the only part of the statement that mattered was the ‘activist feminist’ and the second statement seemed much more likely. Computing numerical probabilities was within their capability, especially for the statisticians, but they saw the question as belonging to personality types rather than mathematics.
You can say their answer was wrong; it is equally reasonable to say that their answer was right, but the question was wrong. Sometimes in real life it is more useful, more expedient, more creative to come up with a ‘wrong’ answer when the question asked is out of context or inappropriate or simply, well, unimaginative. Creativity springs from providing ‘wrong’ answers to old questions, seeing them in new lights, breaking out of perfectly rational, time-sanctified thought boxes. Men could have told Columbus that yes, the world is round, but its size was well-known and the westbound route to the Spice Islands was far longer and probably more dangerous than the eastbound. Had he been rational, and less convincing as a promoter, he would not have launched on his first voyage. Irrationality can be very expensive; sometimes it pays off handsomely.
Now think of ‘artificial intelligence’. Big fast computers, lots of data, sophisticated calculations of probabilities. Wonderful for the small potatoes, the known problems to be solved time after time for particular cases. But when the unknown is encountered, the probabilities are unknown, when what matters are the unrecognized unknowns, what machine can say “The question is wrong; here is my illogical, irrational, subjective, crazy suggestion of an answer”?
For easy precision a jumped-up adding machine is great. For creativity, or to stay clear of those modern-day saber-toothed tigers, I need a human mind. So take that, Hal and Watson and Siri and Alexa and the horses you rode in on.
Copyright 2017 Flight of Eagles