Strangers to Ourselves Read online

  HOW WELL DO PEOPLE KNOW WHAT PREDICTS THEIR MOOD?

  The diabetes case might seem an exception to a more general rule, namely that people's inside knowledge usually helps them understand the causes of their responses. To find out, I conducted a study with Patricia Laser and Julie Stone to see how well people understand what predicts a common response, namely their daily moods, and how this understanding compares with guesses made by complete strangers. We asked college students to keep track of their moods every day for five weeks. The students also made daily ratings of several variables that might predict their moods, such as the weather, the quality of their relationships with friends, and how long they had slept the night before. For each participant we computed the correlations between the predictor variables (e.g., length of sleep) and his or her daily mood. At the end of the five weeks people judged what they thought the relationships werefor example, how much they thought their daily moods were related to the amount of sleep they got. By comparing the actual relationships with people's estimates of the relationships, we could see how accurate people were at knowing the predictors of their moods.

  People were right about some of the predictors, such as the quality of their relationships. Most people correctly believed that this factor was correlated with their moods. Overall, however, people achieved only a modest level of accuracy. Most people believed that the amount of sleep they got was correlated with their mood the next day, for example, when in fact this wasn't true: amount of sleep was unrelated to virtually all the participants' moods.

  The next step was to see how people's level of accuracy about the predictors of their mood compared with the accuracy of complete strangers' reports. We asked a separate group of students to judge the relationship between the predictor variables and daily moods of the "typical undergraduate" at their university. These students were not told anything about the individual participants and thus knew nothing about their particular habits, idiosyncrasies, or private thoughts. All they had to go on was their theories.

  Remarkably, the guesses of these "observer" subjects were as accurate as the participants' own guesses about what predicted their mood. Like the participants themselves, the strangers guessed that relationships with others were an important predictor of mood, and they were right. Like the participants themselves, they guessed that amount of sleep was also a predictor of mood, and they were wrong. The tremendous amount of information the participants had about themselves-their idiosyncratic theories, their observations of covariation between their moods and its antecedents, and their private knowledge-did not make them any more accurate than complete strangers.

  One reason for this result might be that the participants and the strangers were using the same information, namely shared cultural theories. That is, perhaps the participants neglected to use the extra information they had, such as their private thoughts and feelings. There was evidence, however, that the participants did use private information. For example, the extremely high agreement among the strangers about the predictors of mood suggests that they were using the same, shared base of knowledge-namely, shared cultural theories. The much lower agreement among the participants about the predictors of their mood suggests that they were relying more on idiosyncratic knowledge.

  An obvious conclusion from this study (and others like it) is that when people make inferences about the causes and predictors of their responses, such as their mood, they use information that strangers do not have access to, such as their private thoughts and feelings. A less obvious conclusion is that private information both helps and hurts. It can make people more accurate than observers; I might well be right that my mood is dependent on the fortunes of a certain professional baseball team. However, I might be misled by my inside knowledge. It may seem as though my mood fluctuates more with the fortunes of my favorite team than it in fact does. A stranger who does not know that I am a baseball fan might be more accurate in using shared cultural theories about the determinants of mood. Averaging across several studies, there seems to be no net advantage to having privileged information about ourselves: the amount of accuracy obtained by people about the causes of their responses is nearly identical with the amount of accuracy obtained by strangers. 10

  If you find this argument hard to swallow-that on balance, a stranger knows as much as you do about the causes of your feelings, judgments, and actions-I confess that I do too. Think of the implications of this argument: if you are wondering about the determinants of your mood (perhaps with the aim of improving it), you might just as well ask a complete stranger as rely on the vast amount of knowledge you have about yourself and your history.

  The amount of research in this area is not huge, and, as in any area, the individual studies are open to criticisms. Perhaps Wilson, Laser, and Stone's measures of mood were inadequate, for example, or perhaps they failed to ask people about some key predictors of mood, about which they would have been more accurate than the strangers. Further, there is no way of telling how representative the kinds of responses and influences are of the ones people care about in everyday life. If a broader range of responses were studied, strangers' causal reports might not prove as accurate as people's own reports.

  It is remarkable, though, that a personal advantage over strangers' reports has been difficult to find. Further, as seen in Chapter 4, there is evidence that when it comes to judging our personalities (as opposed to judging the causes of a specific response), other people may sometimes be more accurate than we are ourselves. Although I would not yet recommend dialing a random person in the phone book to find out why you feel the way you do, we all might want to be more humble about the accuracy of our causal judgments.

  The Illusion of Authenticity

  There is a final puzzle about people's explanations of their own responses. Why don't we realize that our explanations are confabulations, no more accurate than the causal reports of strangers? One of the major points of this chapter is that people's reasons about their own responses are as much conjectures as their reasons for other people's responses. Why, then, don't they feel this way?

  One explanation is that it is important for people to feel that they are the well-informed captains of their own ship and know why they are doing what they are. Recognizing that we are no more informed about the causes of our responses than a complete stranger is likely to make people feel less in control of their lives, a feeling that has been shown to be associated with depression.

  Another key, I suggest, is that the amount of inside information we have produces a misleading feeling of confidence, namely the sense that with so much information we must be accurate about the causes of our responses, even when we are not. Suppose you are thinking of investing in two Internet stocks, both of which you think have the same potential to increase in value. Your faith in Alpha.com is based on a visit to the company and an extended conversation with its president. Your faith in Beta.com is based on an article you read in the newspaper. Surely you will be more confident in your judgment of Alpha.com, given that your judgment is based on a great deal of firsthand knowledge. But there is no guarantee that this firsthand knowledge will lead to a more accurate judgment; in fact you may have been misled by the president's enthusiasm and exaggerated claims. Similarly, the vast amount of inside knowledge we have about ourselves increases confidence in our selfknowledge, but does not always lead to greater accuracy.

  If so, then this illusion of authenticity should be reduced by equalizing the amount of inside information people have about themselves and another person. Suppose you have an extended conversation with your best friend about her experiences on beach vacations over the years, and I now ask you why she feels the way she does about vacations at the beach. You can use a good deal of specific information about your friend other than your general theories about why people like or dislike the beach-the fact that your friend met her husband at the beach, loves saltwater taffy and wind-swept hair, and is an avid surfer. You are likely to be very confident in your re
asons for why she loves the beach. Perhaps you are not quite as confident as you are in your own reasons, given that we never have as much information about someone else as we have about ourselves. Surely, though, you will be more confident in your reasons for why your friend likes the beach than for why a stranger likes the beach.

  Although this seems obvious, keep in mind that there is no guarantee that your reasons for your friend's feelings will be any more accurate than your reasons for a stranger's feelings, because extra information does not always give people an accuracy advantage. Recall the propensity with which P. S. 's left hemisphere could explain actions by the right hemisphere, by drawing on his knowledge of shovels, chickens, and chicken coops. P. S. had lots of information from which to generate an answer, but none of it was relevant to the real reason he picked a shovel with his right hand.

  Another way to reduce the authenticity illusion is to limit the amount of inside information people have about themselves, thereby lowering their confidence in their own reasons. Of course, the mind is not like a hard disk that can be reformatted to erase all its current contents. We have private information that is more relevant to some judgments than to others, however. Suppose I asked you, "Why do you feel the way you do about the cover of this book?" and "Why do you think a stranger feels the way that he or she does about the cover of this book?" Compared to information about vacations at the beach, you probably have less personal information relevant to your judgment of something as esoteric as a particular book cover. Consequently, you are more likely to rely on general theories about why people like book covers-the same theories you will use to explain the stranger's reaction ("I don't know, I guess I like it because it's mysterious and eye-catching."). Of course, we can bring to bear personal information about virtually anything; it is possible that the book cover reminds you of your Uncle Henry or a photograph you once saw in an antique shop. Still, people use less personal information to explain some responses than they do to explain others. In these cases their reasons probably seem a little less compelling and more like the reasons they would offer for a stranger's responses.

  In short, the authenticity illusion varies with the amount of private information people use when generating reasons. But as we have seen, the accuracy of people's explanations seems not to vary much with the amount of private information they use.

  It is not welcome news, I suspect, that strangers may know as much about the true causes of our responses as we do. We turn now to a bastion of self-knowledge that is harder to assail, people's feelings and emotions. Even if we do not know the causes of our feelings, surely we know that we have them. Or do we?

  Knowing How We Feel

  We must never take a person's testimony, however sincere, that he has felt nothing, as proof positive that no feeling has been there.

  -William Jaynes, Principles of Psychology (1890)

  Ben liked to joke that he was his own invention and therefore never could be certain how he really felt about anything or anybody. I wondered whether he did not sometimes try to solve the problem, and put an end to tormenting doubts, by also inventing various experimental versions of his feelings.

  -Louis Begley, The Man Who Was Late (1992)

  As the old cliche says, "I may not know why but I know what I like." Several contemporary theories of emotion argue that whereas the mental processes that produce emotions may be unconscious, the emotions themselves are not. Affective reactions such as evaluations, moods, and emotions may be the specialty of the house of consciousness. As the quotations above indicate, however, the story is not quite so simple. Feelings are often conscious, but they can also reside elsewhere in the mental neighborhood.

  The Incorrigibility of Feelings

  Of all the issues I discuss in this book, the idea of unconscious feelings is perhaps the most controversial. In fact some philosophers and psychologists reject this idea out of hand, arguing that an "unconscious feeling" is an oxymoron. Suppose I were to tell you honestly that I feel a sharp pain in my left knee right now. Do you believe me? "What a strange question to ask," you might think. "As long as he is not joking or lying, then of course he must be experiencing the pain he describes." If this is what you thought, you are in good company. Quite a few philosophers, including Descartes and Wittgenstein, have argued that reports about sensations and feelings are incorrigible. Simply put, people's beliefs about their feelings cannot be doubted. If I say my knee hurts, then it does and that is all there is to it. I am the final authority on my sensations and feelings, and you have no grounds on which to doubt me.

  Or do you? Consider this example, from a short story by Mary Kierstead. Two cousins are reminiscing about their childhood summers on the family farm, when their thoughts turn to Topper, the resident pony:

  "You know, it wasn't until I was about thirty that I realized that I'd always hated that goddamn pony. He had a mean disposition, and he was fat and spoiled. He would roll on me, and then step on my foot before I could get up."

  "And he bit you when you tried to give him lumps of sugar," Kate added. It wasn't until Blake said it that Kate realized that she, too, had always hated Topper. For years they had been conned into loving him, because children love their pony, and their dog, and their parents, and picnics, and the ocean, and the lovely chocolate cake.'

  Suppose we had asked Blake and Kate, when they were twelve, if they loved Topper. "Of course we do," they would have answered honestly. But Blake and Kate are now convinced that they never loved the beast. They believed that they did, but in truth they hated him. If so, then Cartesian and Wittgensteinian incorrigibility is false. People can be wrong when they honestly report a feeling ("I love Topper").

  A longstanding philosophical debate over the incorrigibility issue often focuses on interesting conundrums. For example, suppose my knee started hurting at 2:00, when I banged it on the corner of my desk. At 2:05 I get a phone call and during the conversation, I do not notice that my knee hurts. When I hang up the phone at 2:10, my knee hurts again. What happened to the pain during the phone call? Is it possible to be in pain but not to know we are in pain? Or did the pain stop while I was on the phone and restart afterward?2

  Although I think the incorrigibility argument is wrong, there are two good reasons why it has persisted: the measurement problem and the theory problem. The measurement problem is that even if people can be wrong about their feelings in principle, we have no way of knowing if and when this is the case, because we do not have a pipeline to people's feelings that is independent of their self-reports. The theory problem is the question of how and why the mind would be organized in such a way that people can be wrong about their feelings. Why on earth would humans be built this way? Although both the measurement and theory problem are formidable, I believe they can be overcome.

  THE MEASUREMENT PROBLEM

  What kind of independent criterion could we call upon to doubt people's statements that they love someone or that their knee hurts? Short of the fanciful Inner Self Detector, there is no perfect, independent measure of such internal states such as how much pain I am feeling in my knee or how much Blake and Kate love Topper. There is no physiological "pain detector," for example, on which a dial points to the precise amount of pain someone feels in his or her knee.

  But the fact that it is difficult to prove that a self-reported feeling is inaccurate is no reason to accept the incorrigibility argument. This would be like saying that there are no other planets outside our solar system, because we did not (until recently) have powerful enough telescopes to observe them, or that when I take off my glasses, nothing exists beyond the few feet I can see. We should not let measurement issues drive theoretical ones.

  Furthermore, although there are no error-free, independent measures of people's internal states such as the Inner Self Detector, there are grounds on which we might at least be highly suspicious of the accuracy of people's reports about their feelings. It might be clear from a person's behavior, and how that behavior is interpreted by ot
hers, that he or she possesses a nonconscious feeling. A number of writers on this topic have used jealousy as an example. Sam observes his wife chatting with an attractive man at a party. The man asks his wife for a dance and she accepts. On the way home, Sam is curt and remote toward his wife. When she asks if anything is wrong, he sincerely replies, "No, I'm just tired." Sam truly believes he is not jealous, even though anyone who observed his behavior would say otherwise. The next day Sam recognizes that he did feel threatened by his wife's attention to the other man.'

  This example highlights another way in which we might doubt people's reports about their feelings: when people themselves later acknowledge that they were wrong about what they were feeling. The fact that Sam later agrees that he was jealous, and that Blake and Kate acknowledged, years later, that they had always hated Topper, is not definitive proof that they had been wrong about their feelings. After all, their reconstructions of their feelings could be in error. Examples such as these, however, meet what we might call the "strong suspicion" criterion. The fact that observers disagreed with the person about his or her feelings (e.g., all the partygoers but Sam believed that he was jealous), and the fact that Sam later believed he had been jealous are strong grounds on which to doubt the veracity of his original denial of jealousy.