How We Know What Isn't So (15 page)

The metaphor also applies to how our beliefs fit together. We carefully choose furniture and works of art that do not clash, just as we try to avoid the dissonance produced by incompatible beliefs. If, over time, we find that our decor does not make a single, coherent statement, we might hold a garage sale and start anew. A similar phenomenon is observed when one undergoes an ideological conversion (such as joining a cult) and many of one’s earlier convictions are discarded to make room for new beliefs.

For the purposes of this chapter, however, the most telling analogue 87 between beliefs and possessions involves the tension between desire and constraint. We are tempted to buy as many of the best things in life that we can. As much of today’s world makes clear, the thirst for material possessions is hard to quench. But few of us can afford everything we desire. We have a budget, and some things are just too expensive. So we do without.

The same can be said of our beliefs. There are things we are sorely tempted to believe; to do so would be tremendously gratifying. To simply acquire many of these comforting beliefs, however, would extract too high a price in rationality and cognitive consistency. So not all are acquired, at least not as is. But if we could just view them from a slightly more flattering perspective, if we could just take the evidence in a little here and let it out a little there—if we could get them on sale!—we just might buy them.

*
Indeed, this would make a particularly good item for a demonstration of the “Lake Wobegon effect”: Asked to assess how objective or unbiased he or she is, the average person would no doubt rate him or herself above average.

*
Although this self-serving juggling of criteria can be attributed—as it is above—to the motive to see ourselves in a favorable light, it is important to note that this phenomenon can be explained in purely cognitive terms as well. In particular, people may use their own strengths as the basis of what constitutes success in a given domain because, after a lifetime of basing their actions on what they do well, those elements at which they excel simply come to mind more readily and thus figure more prominently in their assessments. These two rival explanations are not mutually exclusive, of course, and the most important point is that both processes result in people believing what they would prefer to believe.

Among the most widely known studies in the history of psychology is the conditioning of “Little Albert.”
¹
As most students of psychology have been told many times, every time the nine-month-old Albert came near a white rat, Watson and Raynor made a frighteningly loud noise behind Albert’s head by banging a metal bar with a hammer. Albert subsequently exhibited a strong fear of the rat even when it was no longer paired with the sound, a fear that did not readily diminish over time. Albert also exhibited a milder, but still pronounced, fear of a number of objects that had many of the same features as the rat, such as a rabbit, a white glove, cotton balls, and a white beard. The results of this experiment are often presented as evidence of how people can develop phobias of seemingly harmless objects, and of how our acquired fears can generalize to other, similar entities.

Although the story of Little Albert serves as a convenient vehicle for communicating some important ideas about the acquisition and modification of human emotional behavior, it suffers from a very serious flaw: Many of the events that are often described in secondhand accounts of this story never occurred.
²
The experimenters did indeed manage to make Albert afraid of the rat by pairing its presence with the loud noise seven times at the beginning of the experiment, a fear that remained strong five days later during a follow-up test. At that time, Albert also exhibited a strong fear of a rabbit, a dog, and a sealskin coat, a less pronounced “negative reaction” to a Santa Claus mask and to Dr. Watson’s hair, a mild response to the cotton balls, and a very favorable reaction to a set of wooden blocks and to the hair of Watson’s assistants.

After another five days, however, Albert showed such a slight reaction to the rat that the experimenters decided “to freshen the reaction” to it by presenting it with the loud noise once again; something they also did for the first time with the rabbit and the dog (thereby making them useless as stimuli in subsequent tests of generalization). Finally, when tested after another 31 days, Albert exhibited fear when touching the rat, the rabbit, the dog, the sealskin coat, and the Santa Claus mask. However, Albert also
initiated
contact with the very same rabbit and coat. After this final set of tests, Albert’s mother removed him from the hospital in which the study was conducted, and he was no longer available for subsequent assessment.

The actual details of Watson and Raynor’s study make it clear that Albert’s fear of the rat was not so intense, nor did it generalize as readily to other entities, as is often claimed in textbook accounts of this landmark study in the history of psychology. Eysenck, for example, claimed that “Albert developed a phobia for white rats and indeed for all furry animals.”
³
However, the contention that Albert developed a rat phobia is hard to reconcile with his mild reaction to the rat during the second test period, a reaction described by the experimenters as: “Fell over to the left side, got up on all fours and started to crawl away. On this occasion there was no crying, but strange to say, as he started away he began to gurgle and coo, even while leaning far over to the left side to avoid the rat.” His reported fear of “all furry animals” has also been exaggerated, given that his reaction to such animals was assessed only with respect to the rabbit and the dog (and even they, recall, were directly paired with the noise during the second test session). Indeed, the range of entities to which Albert’s fear reportedly generalized is the most frequently misrepresented result of the study. Different texts have made Albert afraid of a cat,
⁴
a white glove,
⁵
the fur neckpiece or fur coat of Albert’s mother,
⁶
and even a teddy bear.
⁷
Finally, in what may be the most intriguing distortion, a number of texts have re-written the ending of the tale, claiming that Albert’s fear had been eliminated by a “re-conditioning” procedure, sometimes described in detail, at the end of the experiment.
⁸

Why has the story of Little Albert been so frequently distorted, and why has it been distorted in the precise way that it has? There is little doubt that many of these distortions were introduced because they make the tale of Little Albert into a “good story.” There are numerous aspects of what constitutes a good story, several of which are illustrated by the accounts of Albert’s experiences at the hands of Watson and Raynor. These accounts tell a simple, coherent tale of how phobias can be acquired, a tale with a tidy (even happy) ending. This chapter discusses these and other elements of what constitutes a good narrative.

More to the point, however, this chapter also examines how this need or desire to tell a good story can distort the accuracy of information we receive secondhand, and thus bias some of the most important information upon which we base our beliefs. Much of what we know in today’s world comes not from direct experience, but from what we read and what others tell us. An ever-higher percentage of our beliefs rest on a foundation of evidence that we have not collected ourselves. Therefore, by shedding light on the ways in which secondhand information can be misleading, we can better understand a common source of questionable and erroneous beliefs.

To understand what constitutes a good story, it is necessary to examine the needs of the speaker and listener, and the goals they try to achieve in their interaction.
*
Because communication or conversation is a reciprocal process, it is not surprising that many of the needs and goals of the speaker and listener are complementary. This is well illustrated by one of the most basic goals of communication, to ensure that the act of communication is “justified.” For the speaker, this means, among other things, that his or her message should be worthy of the listener’s attention; for the listener, it means that the interaction must in some way be worthwhile. To satisfy this basic goal, it is necessary that certain preconditions be met. The message should be understandable (i.e., not assume too much knowledge on the part of the listener), and yet not be laden with too many needless details (i.e., not assume too
little
knowledge on the part of the listener).
⁹

Sharpening and Leveling
. For the purpose of understanding the formation of erroneous beliefs, it is important to note that satisfying even these very basic enabling conditions can introduce distortion in what is communicated. Classic studies by psychologists F. C. Bartlett
¹⁰
and Gordon Allport and Leo Postman
¹¹
demonstrate that when people are given a message to relay to someone else, they rarely convey the message verbatim. The limits of human memory and the implicit demand that the listener not be burdened with too many details constrain the amount and kind of information that is transmitted. What the speaker construes to be the gist of the message is emphasized or “sharpened,” whereas details thought to be less essential are de-emphasized or “leveled.” Secondhand accounts often become simpler and “cleaner” stories that are not encumbered by minor inconsistencies or ambiguous details.

The case of Little Albert is a good example. Albert did develop some fear of the rat, and his fear did generalize somewhat to other entities. However, the evidence for both the extent of his fear and the amount of generalization was rather inconsistent and hard to interpret. Because these inconsistencies interfered with the main story about classically conditioned anxiety, many authors managed to set them aside. Watson’s original report mentioned that Albert’s fear had to be “freshened” after a few days, and that the loud noise was directly paired with the rabbit and dog as well. Nevertheless, in subsequent accounts by Watson himself and by other authors, these details were leveled out of the story.

A particularly interesting consequence of the processes of sharpening and leveling concerns our impressions of people we only know about secondhand.
¹²
Everyday experience seems to tell us that we often develop exaggerated or extreme impressions of people we have heard or read about but never met. The most telling evidence in this regard is that when we finally meet someone we have been led to believe is, say, unusually charismatic and compelling, or uncommonly wicked and detestable, we are often “disappointed.” The person often seems less worthy of positive or negative regard than we had been led to expect. Sharpening and leveling can help explain this phenomenon.

When someone tells us about another person and his or her actions, the account we receive tends to be organized around the person rather than the context in which the actions took place. The person, after all, generally constitutes the core of the story. Information about the person and the action tends to be sharpened, whereas information about the surrounding context and various mitigating circumstances tends to be leveled. There are a couple of reasons for this disparity in emphasis. First, with all else being equal, people tend to think of actions and actors as going together: An actor’s dispositions are considered to be more of a preeminent cause of behavior than the dictates of the surrounding context.
¹³
It may thus seem more natural, typically, to construct an account of a person’s actions that makes greater reference to the type of person involved than to the nature of the existing circumstances. Second, it is probably
easier
to construct such accounts: People and their actions can often be described in the same terminology; situations and actions usually cannot.
¹⁴
There are compulsive people and compulsive actions, but there is really no such thing as a compulsive situation—although situations do vary in how much they call for compulsive behavior.

Because of this asymmetry in what is transmitted via secondhand accounts, our impressions of people we have heard about but never met may be relatively unaffected by how their actions may have been elicited or constrained by various situational determinants. Their behavior may thus seem to be more a product of underlying personal dispositions, leading us to form more extreme impressions of such people than we would have if we had witnessed their actions firsthand.

A series of recent studies provides support for these ideas.
¹⁵
In one set of experiments, a group of “first generation” subjects watched a videotape of a “target” person describing two events from his or her past. The subjects then rated the target person on a variety of trait dimensions, and provided a tape-recorded account of what they had seen. Subsequently, a group of “second generation” subjects listened to these secondhand accounts and then made the same trait ratings. As predicted, second generation subjects made more extreme ratings of the target than did their first generation counterparts. Furthermore, an analysis of the accounts provided by the first generation subjects indicated that they did indeed underemphasize the situational determinants of the target person’s actions. An event that the target person regretted, for example, tended to be described as a bad deed rather than as a likely product of difficult circumstances. Aspects of the target person’s dispositions were sharpened, whereas features of the surrounding context were leveled.