Strangers to Ourselves Read online

Oddly, these two selves appear to be relatively independent. There is increasing evidence that people's constructed self bears little correspondence to their nonconscious self. One consequence of this fact is that the two personalities predict different kinds of behavior. The adaptive unconscious is more likely to influence people's uncontrolled, implicit responses, whereas the constructed self is more likely to influence people's deliberative, explicit responses. For example, the quick, spontaneous decision of whether to argue with a coworker is likely to be under the control of one's nonconscious needs for power and affiliation. A more thoughtful decision about whether to invite a coworker over for dinner is more likely to be under the control of one's conscious, selfattributed motives.

  Because people cannot directly observe their nonconscious dispositions, they must try to infer them indirectly, by, for example, being good observers of their own behavior (e.g., how often they argue with their coworkers). How important is this kind of insight? It doesn't have to be perfect, because some positive illusions are beneficial. However, it is to people's benefit to make generally accurate inferences about the nature of their adaptive unconscious.

  The Personality of the Adaptive Unconscious

  There is considerable evidence that the adaptive unconscious has a stable, characteristic way of responding to the environment, thereby meeting Allport's definition of personality.' In Jonathan Miller's words, "Human beings owe a surprisingly large proportion of their cognitive and behavioral capacities to the existence of an `automatic self' of which they have no conscious knowledge and over which they have little voluntary control."'

  NONCONSCIOUS "IF-THEN" JUDGMENTS

  As we have seen, Walter Mischel and his colleagues argued that people possess a unique set of cognitive and affective variables that determine how they react to the social world. They describe five components of this "personality mediating system" that guide people's behavior: encodings (people's construals of themselves, others, and situations); expectancies about themselves and the social world; affect and emotions; goals and values; and competencies and self-regulatory plans. In short, they argue, people have distinctive "if-then" rules that determine how they respond in a particular situation; for example, "If I feel that I'm being ignored, I then get angry and aggressive."

  Each of the five components of Mischel's cognitive-affective system are signatures of the adaptive unconscious, such as chronic encodings of a situation. Consider, for example, how these encodings might be measured. One way would be simply to ask people to report their construals. To measure the distinctive ways in which people respond when they perceive that someone is not paying attention to them, we could construct a questionnaire that asked questions like this:

  Suppose you notice that your boss has not paid much attention to you over the past couple of weeks. Which of the following best reflects how you would interpret this lack of attention?

  (a) He/she has great confidence in my abilities.

  (b) He/she has lost faith in my abilities.

  (c) He/she has been quite busy-it has nothing to do with me.

  People's answers to questions like this might well reveal interesting things about their conscious belief system. Their answers might say little, however, about how their adaptive unconscious would interpret actual situations in which they think they are being ignored. Recall that a fundamental property of the adaptive unconscious is that people have no access to the ways in which it selects, interprets, and evaluates information. Thus, asking people to report their nonconscious reactions is fruitless; people may not know how they are likely to react.

  Alternatively, we could observe people's behavior very closely and try to deduce the "if-then" patterns of their adaptive unconscious. Though by no means easy, this approach bypasses the conscious explanatory system and may get directly at nonconscious encodings. This is the approach that Mischel and his colleagues have adopted. In one study, they systematically observed children in a residential camp for many hours, carefully noting the ways in which they behaved in a variety of situations. They were able to find "distinctive behavioral signatures" that permitted them to infer the children's "if-then" patterns of construal. For example, they observed how verbally aggressive the children were in five situations: when approached by a peer, when teased by a peer, when praised by an adult, when warned by an adult, and when punished by an adult. Some children were found to be very aggressive when warned by an adult, but relatively unaggressive in the other situations. Others were found to be very aggressive when a peer approached them, but relatively unaggressive in the other situations. Each of these children's "behavioral signatures" was stable over time; they seemed to reflect characteristic ways in which they interpreted the different situations."'

  Although this result might seem pretty straightforward-even obvious-it contrasts strongly with the way in which most personality psychologists study individual differences. Trait theorists would give the boys a standardized questionnaire and classify each on the trait of aggressiveness. The assumption would be that each boy possesses a certain level of aggressiveness that would allow predictions of their behavior, regardless of the nature of the situation. But clearly the trait approach would not be very useful here, because it does not take into account the fact that (1) the boys' aggressiveness would depend on how they interpret the situation (e.g., how threatening they found it); (2) not everyone interprets a situation in the same way; (3) their interpretations are stable over time; and (4) the interpretations are made by the adaptive unconscious. By taking each of these points into account we can predict the boys' behavior pretty well-better than if we had given them a questionnaire and assigned them a value on a single trait dimension.

  SCANNING PATTERNS: CHRONIC ACCESSIBILITY

  One rule the adaptive unconscious uses to judge information is accessibility, or how "energized" a category or construct is. Consider two people, Charlotte and Simon. For Charlotte, the category of "intelligence" is more accessible than the category of "friendliness," whereas for Simon it is the reverse. This means that when Simon and Charlotte meet a new coworker, Marsha, Simon is more likely to notice and remember how friendly she was, whereas Charlotte is more likely to notice and remember how intelligent she was. George Kelly referred to these accessible categories as "scanning patterns" that guide our construals of our social environments.I I

  A number of experiments have shown that these scanning patterns allow people to pick up information from their social environments quickly and efficiently. In one study, people were presented with twentyfour sentences about another person very rapidly-one sentence every two seconds. Imagine you were a participant; just as you read one sentence about the person, such as "admitted his blunder," another one appears on the screen, such as "stole from his friend's wallet." You might well experience information overload, finding it difficult to keep track of all the information and figuring out what the person is like.

  Unless, that is, you have a nonconscious scanning pattern that helps you organize the information. On an earlier test, some of the participants were found to have a chronically accessible category of "honesty." That is, honesty was one of the first traits on which they judged other people. For the remainder of the participants, honesty was not chronically accessible; it was not a trait that they used regularly to judge other people. In the experiment, the people for whom honesty was chronically accessible found it easiest to read the sentences and form an impression of the person, because many of the sentences had to do with honesty and these people were more prepared to process the sentences. In contrast, the people for whom honesty was not chronically accessible were more likely to experience information overload, finding it difficult to form an impression and showing poorer recall of the sentences. It is as if we have our antennae up for certain kinds of information about other people, depending on which categories are accessible to us. And this happens quickly, with no conscious awareness.'2

  But how does a category (such as honesty) become chronically accessib
le to someone in the first place? George Kelly noted that people develop constructs to make sense out of and predict their environments. As a result of their background and learning history, people develop regular, idiosyncratic ways of construing the world. The construct of honesty might be useful for one person, and the construct of friendliness more useful to another. A specific type of construct-our stored representations of significant others-is especially likely to become chronically accessible and applied when we meet new people.

  TRANSFERENCE: SEEING THE OLD IN THE NEW

  In Janet Malcolm's book Psychoanalysis: The Impossible Profession, one analyst poses the following question to another: "What would you call an interpersonal relationship where ... the persons within that relationship don't see each other for what they objectively are but, rather, view each other in terms of their infantile needs and their infantile conflicts?" The analyst replies, "I'd call that life.""

  Freud's discovery of transference-the way in which we superimpose infantile feelings toward our parents onto new relationships-has been called his "most original and radical discovery." 14 Freud focused primarily on the way in which unconscious sexual and aggressive drives, such as the Oedipus conflict, are played out in a person's relationship with the analyst. Harry Stack Sullivan and Melanie Klein took a broader view of transference, discussing how past relationship can influence people's perceptions of any new person they meet.

  The social/personality psychologist Susan Andersen argues that transference is best understood not in psychoanalytic terms, but as part of the nonconscious, social information-processing system-namely, the adaptive unconscious. Much as in our discussion in Chapter 1 of why the unconscious exists, Andersen suggests that there is no need to assume that transference is rooted in unconscious motivation, whereby people seek to cover up anxiety-provoking thoughts and feelings (e.g., "I love him because he is like my father"). Instead, transference may be part of the normal functioning of everyday life that is best understood in terms of modern research on social cognition. She argues that our mental representations of other people are stored in memory like any other chronic category. Because representations of relationships with significant others are self-relevant and frequently brought to mind, they become chronically accessible and are often used to interpret and evaluate new people we meet. In short, just as the construct of "honesty" or "kindness" can be activated and applied to a new person, so can the construct of a specific person such as "my mother" or "Uncle Henry."

  In a typical study, Andersen first asks people to name a significant other and to answer questions about what this person is like. Then, in what is ostensibly a different study, people receive descriptions of people they have never met. Andersen rigs it so that one of these new people shares characteristics of participants' significant others. For example, if you were in this study, you would be given information about several people, one of whom turns out to share some characteristics of an important person in your life (e.g., your mother).

  Andersen and her colleagues have found that people react quite differently to the new acquaintance who is like their significant other. They are more likely to remember things about this person and to evaluate them similarly to their significant other. For example, if you have fond feelings for your mother, you would have a positive reaction to the new acquaintance who is like her. If you have negative feelings toward your mother, you would dislike the new acquaintance.

  To what extent are people aware of this process? Andersen suggests that it occurs quickly and nonconsciously. It is not as if people say, "Hm, Sue is a lot like my mother, so I guess she is a warm, nurturing person." Instead, the adaptive unconscious selects, interprets, and evaluates new information very quickly and does so in terms of categories that are accessible-in this case, accessible representations of important people in our lives. In support of this interpretation, the transference process occurred in one study when the information indicating that a target person was like a significant other was presented subliminally. Even though people were unaware that the target person shared some characteristics with their significant other, they still "transferred" their feelings about their significant other onto the target person. This transference process, which occurs outside of awareness, appears to be an important source of individual differences in how people react to new acquaintances.''

  To the psychoanalytically inclined, Andersen's research is likely to seem quite consistent with what is already known about transference and object relations; and indeed, in some ways it is. To the more empirically inclined student of the adaptive unconscious, however, Andersen's work is novel in two respects. First, she has developed a new method to study transference systematically in controlled experiments. Second, she has shown that transference can be explained easily by modern theories of social cognition (e.g., ideas about how chronically accessibility constructs of all sorts, including those about significant others, influence people's judgments and behavior), with no need to introduce additional theoretical constructs such as repression, resistance, or the management of anxiety. It is part of the normal functioning of an adaptive unconscious, and not necessarily part of the emotional hijinks of a dynamic unconscious.

  WORKING MODELS OF ATTACHMENT

  Further evidence for the nonconscious influence of past relationships comes from research on attachment relationships. Initially this work focused on the internal working models of attachment that infants formed about their parents, which are measured by observing how an infant interacts with his or her parent and a stranger in a laboratory setting (in a procedure known as the Strange Situation). The parent is asked to leave the room and then return several times, and the infant's reactions to these separations and reunions are observed. On the basis of these reactions, the infant is classified as having a secure, avoidant, or anxious/ambivalent working model of attachment. Infants with secure attachment models are distressed when their parent leaves but seek comfort when he or she returns. They have parents who are sensitive to and responsive to their needs. Infants with avoidant attachment models typically have parents who have rebuffed their attempts to be intimate. In the laboratory session they typically show little distress when their parent leaves and they do not seek comfort from their parent when he or she returns. Infants with anxious/ambivalent attachment models typically have parents who alternate between unresponsiveness and excessive affection. They fear that others will not reciprocate their desire for intimacy and are preoccupied with their parent's availability in the laboratory session. A fourth attachment style has recently been identified called "disorganized." Infants with this style show contradictory reactions, such as crying when they are separated from their parent but ignoring their parent when he or she returns. Some researchers suggest that infants with this attachment style are more likely to have parents who are depressed or neglectful.

  These attachment styles are hypothesized to become internalized and to guide how people react to others, even outside the parent-child relationship. One study, for example, measured infants' attachment styles in their second year of life and then observed their behavior in a summer camp at ages ten and eleven. Compared to children with avoidant or anxious/ambivalent models of attachment, children with secure attachment styles as infants spent more time with peers at the summer camp, were more likely to develop friendships, and were more likely to evaluate other children in a positive light.16

  In recent years, researchers have looked at working models of attachment in adults. They assume that people's chronic way of viewing important past relationships (i.e., with their parents) colors their perception of behavior in current relationships, particularly with romantic partners. One way of measuring adult attachment models is to ask people to report their feelings about their romantic relationships, with the assumption that they can easily access and report these feelings. In one version, people are given descriptions of three adult attachment relationships and asked to choose the one that best applies to them. For example, if you chose th
e following statement, your adult romantic relationships would be classified as anxious/ambivalent: "I find that others are reluctant to get as close as I would like. I often worry that my partner doesn't really love me or won't stay with me, I want to merge completely with another person, and this desire sometimes scares people away."''

  A second method of measuring adult attachment is the Adult Attachment Interview (AAI), which involves a lengthy interview in which people are asked questions about their relationship with their parents. The interviewer pays attention not only to what people say but to how they say it and to their nonverbal reactions. Researchers who use this method assume that people are not fully aware of their working models of attachment; hence the need to infer what these models are from people's behavior during the interview. The AAI appears to be a valid measure of adult attachment, in that it also predicts interesting things, such as problem behaviors in adolescence (e.g., delinquency, drug use, school dropout, and teenage pregnancy) and the bonds people establish with their own children.

  So far the story is pretty straightforward: there are two ways of measuring adult models of attachment (self-report questionnaires and the AAI), and both seem to do a pretty good job, in that they predict interesting social behaviors. But here's the rub: the two measures do not correlate very well. If you came out as securely attached on one measure, you are likely to be classified as securely attached on the other at a level only slightly better than chance."

  One explanation for this lack of correspondence might be that the techniques are simply measuring different kinds of attachment. The AAI focuses on people's memories of their relationship to their parents, for example, whereas the self-report measures focus on people's conceptions of their current romantic relationships. Most researchers in this area, however, assume that memories of parental relationships and views of romantic relationships are influenced by the same internal models of attachment and thus should be related.