Thinking about Thinking

An Interview with Daniel Kahneman
November 29, 2011

Daniel Kahneman is an extraordinarily interesting thinker. As a psychologist, he received the 2002 Nobel Prize in Economics for his work with Amos Tversky on decision-making. Here is what Steven Pinker, my previous interview subject, recently wrote about him:

Daniel Kahneman is among the most influential psychologists in history and certainly the most important psychologist alive today. He has a gift for uncovering remarkable features of the human mind, many of which have become textbook classics and part of the conventional wisdom. His work has reshaped social psychology, cognitive science, the study of reason and of happiness, and behavioral economics, a field that he and his collaborator Amos Tversky helped to launch. The appearance of Thinking, Fast and Slow is a major event.

Kahneman was kind enough to take time out of a very busy book tour to answer a few of my questions.






Much of your work focuses on the limitations of human intuition. Do you have any advice about when people should be especially hesitant to trust their intuitions?

When the stakes are high.  We have no reason to expect the quality of intuition to improve with the importance of the problem.  Perhaps the contrary: High-stake problems are likely to involve powerful emotions and strong impulses to action. If there is no time to reflect, then intuitively guided action may be better than freezing or paralysis, especially for the experienced decision maker. If there is time to reflect, slowing down is likely to be a good idea. The effort invested in “getting it right” should be commensurate with the importance of the decision.

Are there times when reasoning is suspect and we are wise to rely on our snap judgments?

As Gary Klein has emphasized (Sources of Power is one of my favorite books), true experts—those who have had sufficient practice to detect the regularities of their environment—may do better when they follow their intuition than when they engage in complex analysis. Tim Wilson and his collaborators have demonstrated that people who choose between two decorative objects do better by following their impulse than by protracted analysis of pros and cons. The critical test in that experiment is how much they will like the chosen object after living with it for a while. Affective forecasting based on current feelings appears to be more accurate than systematic analysis that eliminates those feelings.

What is the difference between the “experiencing self” and the “remembering self”?

The experiencing self lives in the moment; it is the one that answers the question “does it hurt?”  or “what were you thinking about just now?”  The remembering self is the one that answers questions about the overall evaluation of episodes or periods of one’s life, such as a stay in the hospital or the years since one left college. It involves both retrieval and temporal integration of diverse experiences. In the context of studies of subjective well-being, the happiness of the experiencing self is assessed by integrating momentary happiness over time (in the 19th century, the economist Francis Edgeworth spoke of “the area under the curve”). Experienced happiness refers to your feelings, to how happy you are as you live your life. In contrast, the satisfaction of the remembering self refers to your feelings when you think about your life.

Isn’t the remembering self just the experiencing self in one of its modes?

Of course. Thinking about your life is an experience that you have. But it is useful to distinguish these relatively rare moments from the routine emotional quality of your life. The distinction is especially important in evaluating an individual’s well-being, because the determinants of experienced happiness and life satisfaction are substantially different.

How should the split between these two points of view affect our understanding of the good life?

Some conceptions of the good life take the Aristotelian view to the extreme of denying altogether the relevance of subjective well-being. For those who do not want to go that far, the distinction between experienced happiness and life satisfaction raises serious problems. In particular, there appears to be little hope for any unitary concept of subjective well-being. I used to hold a unitary view, in which I proposed that only experienced happiness matters, and that life satisfaction is a fallible estimate of true happiness. I eventually concluded that this view is not tenable, for one simple reason: people seem to be much more concerned with the satisfaction of their goals than with the achievement of experienced happiness. A definition of subjective well-being that ignores people’s goals is not tenable. On the other hand, an exclusive focus on satisfaction is not tenable either. If two people are equally satisfied (or unsatisfied) with their lives but one of them is almost always smiling happily and the other is mostly miserable, will we ignore that in assessing their well-being?

Are there ways to get the two selves to converge? If so, would this be normative?

There is a road to convergence, but few will want to take it:  we could suggest to people that they should adopt experienced happiness as their main goal, and be satisfied with their lives to the extent that this goal is achieved. This idea implies the abandonment of other goals and values, which is surely unappealing.

The other possibility is a redefinition of momentary happiness in terms of a more global evaluation. That is the route we took in an article I published in 1997 with Peter Wakker and Rakesh Sarin. We suggested that the first step is to evaluate lives (or segments of life) as profiles of experience, and rank the lives. In the next step one would rescale momentary utility to transform the profiles so as to achieve the same ranking   So, for example, if the frequency and intensity of aesthetic experiences are important to our global evaluation of lives, then the value (or utility) of moments of aesthetic joy would be raised accordingly. The problem, of course, is that we no longer accept the subject’s own judgment of how happy they are at any moment. So this approach is difficult as well.

To what extent to do you think true self-deception (as opposed to simple bias) exists?

I don’t know how you expect to distinguish true self-deception from simple bias. Suppose you like someone very much. Then by a familiar halo effect you will also be prone to believe many good things about that person—you will be biased in their favor. Most of us like ourselves very much, and that suffices to explain self-assessments that are biased in a particular direction. You will believe these biased assessments regardless of whether they are about you or about someone else. We resist evidence that threatens our positive image of people we love. And perhaps we love ourselves more intensely than we love most (or all) others. When does this become self-deception?