Strangers to Ourselves

by Timothy D Wilson

  Bern's theory caused quite a furor, partly because it seems, on the basis of simple introspection, so absurd. When I stub my toe I know right away that I feel pain; I do not have to observe myself hopping around the room howling to figure it out. When I haven't eaten in a while, I do not have to watch myself go to the refrigerator and make a sandwich to realize that I am hungry. It's like the old joke about two behaviorists who have just made love. One says to the other, "I know it was good for you, but was it good for me?" This is funny precisely because it is so silly to assume that people do not directly experience feelings such as their own sexual pleasure.

  Bern acknowledged, however, that there are times when we know directly that we feel pain or love or sexual pleasure and do not need to observe our behavior to figure this out. The point is that there are many other times when it is not so clear how we feel, and it is then that we are forced to be outside observers of our behavior, so that we can decipher our feelings, attitudes, and traits.'

  SELF-REVELATION OR SELF-FABRICATION?

  Despite years of research on self-perception theory, there is an enduring question: Is the self-perception process one of self-revelation, whereby people come to know better their true feelings by observing their behavior, or one of self-fabrication, whereby people infer internal states that did not exist before?

  When Sarah met Peter at a party, for example, she did not think she liked him very much; in many ways he was not her type. She found herself thinking about him a lot, however, and when Peter telephoned and asked her out on a date, she said yes. Now that she has agreed to the date, she discovers that she likes him more than she knew. This is an example of self-perception as self-revelation, because Sarah uses her behavior to bring to light a prior feeling of which she was unaware, until she agreed to go out with Peter.

  Another possibility is that Sarah really did not like Peter all that much when she first met him. She felt obligated to go out with him because he is the son of her mother's best friend, and her mother thought they would be a good match. Sarah does not fully realize that this is the reason she said yes, however, and mistakenly thinks, "Hm, I guess I like Peter more than I thought I did if I agreed to go out with him." This would be an example of self-fabrication: Sarah misses the real reason for her behavior (the desire to please her mother), and infers that she feels more positively toward Peter than she did before.

  The difference between self-revelation and self-fabrication is crucial from the point of view of gaining self-knowledge. Inferring our internal states from our behavior is a good strategy if it reveals feelings of which we were previously unaware. It is not such a good strategy if it results in the fabrication of new feelings.

  Self-fabrication would not be an issue if people were adept at knowing exactly why they behaved the way they did. If Sarah recognized that the reason she agreed to go out with Peter was a sense of family obligation, she would not make the mistake of thinking she liked Peter more than she did. As we have seen, however, people are not always skilled at knowing exactly why they respond the way they do and often make precisely this kind of mistake.

  In fact almost all the experiments on self-perception theory are examples of self-fabrication, whereby people misunderstand the real reason for their behavior and make mistaken inferences about their internal states. Many of these studies are ones in which people, like Sarah, underestimate the power of the situation over their behavior and mistakenly infer that they did what they did because of their inner feelings or attitudes. In a study conducted at Yale University, for example, students agreed to go to a street corner and gather signatures on a petition to reduce air pollution in New Haven. In one condition participants heard an experimental accomplice also agree to the request, and remark that he did so because he "wouldn't mind convincing people about something I really believe in." What, if anything, should the students learn about themselves from the fact that they, too, agreed to collect the signatures?

  The real reason most people agreed was that the experimenter was quite persuasive and made it hard to say no, as evidenced by the fact that all students did agree to the onerous request. Instead of saying, "I did it because the guy twisted my arm," though, people mistakenly inferred that their behavior was a reflection of a strong attitude-stronger than they had actually felt before. In other words, they engaged in selffabrication. It is quite common for people to overlook situational influences on their actions and infer that they acted on the basis of their internal states-so common that this phenomenon is called the fundamental attribution error.'

  In most studies of the fundamental attribution error, the situational influences are rather subtle (such as an experimenter applying pressure to convince us to get signatures for a good cause) and easy to miss. What if the situational constraints or incentives are obvious? In these cases people correctly recognize that their behavior was caused by situational demands and thereby stop themselves from fabricating internal states. If our supervisor asks us to buy Girl Scout Cookies from his daughter's troop, and not-so-subtly implies that our next raise is contingent on saying yes, we will probably attribute our purchase of ten boxes of Thin Mints to his arm-twisting, rather than to the idea that the Girl Scouts are now our favorite charity and Thin Mints our favorite snack.

  But if situational influences are too strong, people make a different kind of self-fabrication error: they overattribute their actions to the situation, and underestimate how much they wanted to perform the behavior. Suppose that Bill has always loved to play the guitar and spends hours practicing. What happens when he also has a strong situational reason for the same activity, such as playing at a wedding for an enormous fee? It might seem as if Bill would enjoy the performance all the more, because he now has two reasons for doing it: the money he is earning and his love of guitar playing.

  Many studies show that in situations like this, people in fact overattribute their behavior to the situation and underestimate their intrinsic interest in the activity. The more Bill plays professionally, the less he is likely to enjoy playing the guitar, because he infers that he is "doing it for the money," not because he loves it. This is another form of selffabrication: because of a strong situational incentive or demand, people underestimate the magnitude of their internal interest in the activity.'

  A final example of self-fabrication is the case in which people's behavior might plausibly result from more than one internal state. Consider the case in which people find that their bodies are revved up; their heart is beating rapidly and they are short of breath. The way in which they interpret this arousal will determine the emotion they experience. If someone has just pointed a gun at them and said, "Give me your wallet," they will correctly interpret their arousal as a sign of fear. Often, however, there is more than one explanation for our arousal. Perhaps we are on a first date with a very attractive man or woman, and just narrowly avoided an automobile accident. How much of our arousal is due to fear over almost losing our lives versus attraction to our date?

  Again, if people were perfect at knowing the causes of their responses (in this case their arousal), there wouldn't be any problem. They could say, "Well, 61 percent of my arousal came from the nearly hitting that truck, and 39 percent is due to attraction to my date," and go on their way. Instead, people often make mistakes about the source of their arousal and end up with a self-fabricated feeling. People might underestimate how much their arousal resulted from the close call with the truck, for example, and assume that they are even more attracted to their date than they thought they were.

  To the extent that such self-fabrication is common, the self-perception process, whereby people observe their behavior to infer their feelings, is not such a good path to self-knowledge. People misunderstand why they are responding the way they are and wrongly infer that they don't like playing the guitar as much as they believed they did or that they are more in love than they thought.

  Sometimes, however, we have feelings of which we are not fully aware, and the self-perc
eption process has the potential to reveal these feelings. Consider the example from the previous chapter, in which people have prejudiced feelings toward members of a minority group, but convince themselves that they actually are not prejudiced at all. Or consider Henry Higgins, who couldn't see through his elaborate smoke screen of himself as a cultured, fair-minded English gentleman that he was in fact a coarse, misogynous brute. In cases such as these, people might be wise to become better observers of their own behavior. If an employer notices that he keeps passing over qualified African Americans, finding excuses to hire less qualified whites, he should begin to question his own level of prejudice. Henry Higgins would have been well advised to pay more attention to how he treated Eliza and Mrs. Pearce. Perhaps my friend Susan would have realized sooner than she did that she did not love Stephen, if she had paid more attention to the fact that she often found excuses not to see him on weekends.

  WHO DOES THE INFERRING?

  But there is a complication, namely the question of which part of the mind engages in the self-perception process, inferring how we feel from what we do. In the examples above, I assumed that people consciously make the effort to observe their behavior to figure out how they really feel. Although we can surely do this, the fact is that our adaptive unconscious might also be drawing inferences from our behavior, without our knowing it. In fact one of the major roles of the adaptive unconscious is to draw inferences about the nature of ourselves and the social world.

  We encountered an example of such nonconscious inference in a study by Stanley Schachter and Ladd Wheeler, in which people watched a comedy film after receiving an injection (see Chapter 6). Those who unknowingly received epinephrine (adrenalin) became physiologically aroused, and thus found themselves with an elevated heart rate and sweaty palms while watching the film. They attributed their arousal at least in part to the assumption that the film was very funny, as shown by the fact that they smiled and laughed a lot more during the film than people who had not received epinephrine. These inferences seem to have been made nonconsciously, however, because when asked how funny the film was, people in the epinephrine condition did not report that it was any funnier than other participants did. Instead, they relied on their conscious theories about how much they liked the kind of comedy they saw, such as the person who said, "I just couldn't understand why I was laughing during the movie. Usually, I hate Jack Carson and this kind of nonsense and that's the way I checked the scales." In other words, the adaptive unconscious inferred from people's level of arousal that the film was funny, causing people to laugh and smile more, whereas people's conscious selves drew different conclusions.

  There is not much we can do to control the nonconscious inferences we draw about ourselves. The best approach is to try to perform the selfperception process consciously as well. In this way, people's conscious self-narratives are likely to match better the changes that are occurring nonconsciously, such as what films they find funny, whom they like, and in what situations they are most comfortable. I don't mean to suggest that people should become so self-vigilant that they constantly question the accuracy of their self stories. When faced with an important decision, however-such as whether to marry or have children-people might he wise to be good observers of themselves, and engage less in the kind of fruitless introspection discussed in Chapter 8.

  Remember Mike from the previous chapter, who believes he is shy but seems anything but? It would seem to be to his advantage to pay more attention to his own behavior, to see that he often acts in a quite extraverted manner. By so doing he might realize that his theory about his shyness is out of date, and revise it to fit his adaptive unconscious. People's conscious stories about themselves are often too negative or limiting, and it is to their benefit to revise these stories to better match their nonconscious traits, abilities, and feelings. More often, perhaps, people's conscious theories are too positive. Whereas it can be useful for us to maintain positive illusions about ourselves, if we are to grow and change for the better, we need to recognize that we may be more prejudiced than we thought or, like Henry Higgins, less kind.

  Do Good, Be Good

  If people do have an overly positive view of themselves, they may not want to revise their conscious stories downward to match their more negative unconscious states. Rather, people would be better off changing their nonconscious states to match their more positive self-stories. People who hold nonprejudiced, egalitarian attitudes toward other social groups at a conscious level, but possess more prejudiced attitudes at an automatic, nonconscious level, do not want to revise their conscious narratives to match their nonconscious states. They prefer to do the reverse, changing their nonconscious, prejudiced attitudes to match their conscious, egalitarian ones. Similarly, if Henry Higgins were to recognize that he had an inflated view of himself, he would presumably want to change for the better.

  But how? It is not easy to know what our nonconscious states are, much less to change them. Aristotle suggested that "We acquire [virtues] by first having put them into action ... we become just by the practice of just actions, self-controlled by exercising self-control, and courageous by performing acts of courage." William James offered similar advice: "Seize the very first possible opportunity to act on every resolution you make, and on every emotional prompting you may experience in the direction of habits you aspire to gain"' In other words, the first step to changing our nonconscious inclinations is to change our behavior. People who are concerned that they might be prejudiced at a nonconscious level could try their best to act in nonprejudiced ways as often as possible. Doing so can lead to change at the automatic level in two ways. First, it provides the opportunity for people to infer from their behavior, nonconsciously, that they are nonprejudiced people, according to the self-perception process discussed earlier. That is, it provides the adaptive unconscious with new "data" from which to infer attitudes and feelings.

  Second, as suggested by William James, the more frequently people perform a behavior, the more habitual and automatic it becomes, requiring little effort or conscious attention. One of the most enduring lessons of social psychology is that behavior change often precedes changes in attitudes and feelings. Changing our behavior to match our conscious conceptions of ourselves is thus a good way to bring about changes in the adaptive unconscious.

  But why stop with trying to mold the adaptive unconscious to our conscious conceptions of ourselves? Sometimes people are dissatisfied with both their conscious and nonconscious feelings or traits in a particular area. Their goal is not one of self-knowledge, whereby they bring their conscious narratives more into line with their nonconscious states, but self-improvement, whereby they change both. Perhaps the "alter your behavior first" strategy can work here as well, bringing about desired changes in people's conscious narratives and their adaptive unconscious. In short, if we want to become a better person, we should follow a "do good, be good" strategy. By acting in ways that are helpful and caring toward others, we will come to view ourselves as more helpful and caring people.

  Now, I know that this sounds simplistic. People do not transform themselves into saints by doing one kind act. People who no longer love their partners cannot make themselves fall in love again simply by acting as if it were so. An extremely shy person cannot suddenly become the life of the party by deciding to chat with a few strangers. I think we underestimate, however, how much we can change feelings and traits by changing our behavior.

  I have always considered myself to be a bit introverted, for example, and I think this conscious conception is true, in the sense that my nonconscious dispositions and inclinations are on the shy side. I have often wished I were more comfortable in large groups, and a few years ago I decided that the answer was simply to act in a more extraverted way whenever possible. I made more of an effort to chat with people, such as someone at a party whom I had never met, rather than talking only with my friends or sticking close to the buffet table. The more I did so, the more comfortable I became in
such situations. I will never be like my wife, who can chat with anyone, anytime, with great ease and charm. But I think I have become more extraverted as a result of my little experiment.

  Some of this change is simply due to practice, of course. The more I have made the effort to chat with people, the better I have become at small talk. This practice effect has also been true with my teaching; when I first taught large lecture courses to hundreds of people I was a basket case, but little by little I learned how to do it without my teeth chattering. After years of teaching, large lecture courses are now my favorites. My stand-up skills will never earn me a spot as the guest host of The Tonight Show, but they're better than they used to be.

  Deliberately changing our behavior is beneficial beyond the practice it gives in a new way of acting. It also promotes a new self-definition. The more I find myself chatting comfortably with a new acquaintance at a party, or gathering steam in a lecture to a large class, the more my view of myself changes. This can happen at both a nonconscious and a conscious level. My adaptive unconscious is more likely to draw the inference that I am an outgoing person, and this inference has become a part of my conscious self-narrative as well. The more my self-definition changes, the easier it becomes to act in yet more extraverted ways automatically, rather than having to force myself to make the effort. Automatic selves produce automatic actions. The new "extraverted" Tim takes the controls and steers me in directions I never would have gone before, like chatting amiably with the person sitting next to me on a plane, rather than keeping my nose in a book.