麻豆传媒

Daniel Kahneman, one of the world鈥檚 most influential psychologists, has made a career from challenging conventional wisdom on our choices, intuition and happiness. He talks to Liz Else about the cognitive slip-ups we make every day

You鈥檝e just written your first book for the general reader. What was your motivation?
My aim was to educate watercooler gossip about the choices and judgments of others. I am not very optimistic about people鈥檚 ability to change the way they think but I am fairly optimistic about their ability to detect the mistakes of others. If we were all more intelligent and sophisticated in thinking about each other, that would ultimately affect everyone鈥檚 behaviour. So it鈥檚 a long-term ideal.

You think there are two different systems that control the way we think and make decisions?
I talk about fast and slow systems, system one and system two. These correspond roughly to a distinction everybody is familiar with: thoughts that occur to you and thoughts that you have to generate. When you recognise an object or have a sudden feeling that you like someone, these belong to system one 鈥� fast, automatic, mostly effortless. System two is what happens when you control yourself, when you make a computation. It is slow, deliberate, effortful thinking.

Both systems have drawbacks. What are they?
System one is not quick to learn. That is, we have our habits and our ways of doing things. When you are in a situation, what associations does it trigger in your memory? We have very little control over this. Your perception of the world is not necessarily going to be accurate. Many of the so-called errors of judgement come from limitations of system one.

What about the limitations of deliberate, system-two thinking?
The main one is the limitation on its knowledge. So, for example, system one doesn鈥檛 spontaneously think statistically, but system two, unless it鈥檚 been trained, doesn鈥檛 know statistics either. So I don鈥檛 really intend to say that systems two is rational and all-knowing: it鈥檚 just our deliberate thinking with all the limitations. Its other big limitation is that it鈥檚 very slow. You couldn鈥檛 possibly live your life deliberating all the time.

Do you think it would help people to recognise these different ways of thinking?
I have been studying this for years and my intuitions are no better than they were. But I鈥檓 fairly good at recognising situations in which I, or somebody else, is likely to make a mistake 鈥� though I鈥檓 better when I think about other people than when I think about myself. My suggestion is that organisations are more likely than individuals to find this kind of thinking useful.

How can you avoid slipping up?
I鈥檒l give you an example of a mistake that people can be trained to recognise and avoid. Suppose I ask: is the average price of German cars more or less than 拢90,000? Then I ask you to estimate the average price of German cars. Your answer will be quite different from what it would have been if I had first asked: is the average price of German cars more or less than 拢12,000. It will be much higher. This is because when I give you a ridiculously high number, I make you think of expensive cars, Mercedes, Audis, and so on. If I give you a low number, you鈥檒l think of more popular cars. Your mind automatically generates a biased sample.

This phenomenon is called anchoring. It causes you to focus on a given number, and that number becomes plausible just because it鈥檚 been mentioned. A lot of what happens in negotiations are attempts of one side to anchor the thinking of the other. We published an in June this year which provides a checklist for decision-makers to find biases and errors in judgement.

What else should we watch out for?
System one produces the best coherent story possible from the evidence at hand. I describe it as a machine for jumping to conclusions. If I ask you, is X a good leader, and I say, she is intelligent and strong, you have already formed the impression she鈥檚 a good leader. I haven鈥檛 told you other things 鈥� she鈥檚 also corrupt and cruel, say. You haven鈥檛 waited for information but formed an impression on the basis of the information you had. We need that to get around in the world 鈥� we can鈥檛 live in a state of perpetual doubt 鈥� so we make up the best story possible and we live as if this story were true.

鈥淲e make up the best story possible and we live as if this story were true鈥�

We are often not aware of how little information we have, and if we鈥檙e not aware of this then we get the phenomenon of overconfidence. Confidence is not a judgement, it鈥檚 a feeling.

How does this fit in with another strand of your research, on happiness?
That evolved from previous work I鈥檇 done on the issue of rationality and whether decisions were rational. Clearly one condition for rational decision-making, when you are making a commitment that is going to affect outcomes in the future, is to know whether you will like those outcomes, whether your tastes will stay as they are now. I did some experiments with ice cream which showed that people are not good at predicting what they will like.

What happened in the ice cream experiments?
I paid students to eat a substantial helping of their favourite ice cream, while listening to a piece of music, every day for eight consecutive days. It was always the same ice cream and the same music. After the first day, they rated their experience and predicted the rating that they would give at the end of the eight days. Some became addicted to the ice cream so they started fantasising about it, and others got really tired of it. People did not know in advance whether they would become addicted or tired. That was interesting so I started asking whether people are good at remembering and evaluating the experiences they鈥檝e had in the past. They are not.

That showed up in experiments you did on colonoscopies too?
Yes, we studied the experiences of people undergoing painful colonoscopies to find out who had suffered most pain. We found something really odd: we could pretty much predict what the patients鈥� retrospective pain rating would be from looking at the average of the level of pain reported at the worst moment of the experience and at its end. We also found that the duration of the procedure had no effect on the rating of total pain.

When you ask people to think about a medical procedure, they think of two things: how bad did it get at its worst, and how did I feel when it was ending. How long it lasted plays very little role in people鈥檚 assessment of their experience, which, when you stop to think about it, is ridiculous.

You鈥檝e found that how an event ends is very important to the way we evaluate past experiences generally?
Endings of stories turn out to be enormously important in the way that people evaluate their own lives and those of others. Take the importance that we attach when we hear a story about an estranged mother and daughter and just before the mother died, they get together again. We find that touching, and think it matters a lot. But that鈥檚 very odd, isn鈥檛 it? She barely had time to experience it. Clearly we think of the lives of others, and probably our own lives, mainly as stories. We like the stories to be good stories.