EVER HEARD OF - GUT FEELINGS IN ACTION - Gut Feelings: The Intelligence of the Unconscious - Gerd Gigerenzer

Gut Feelings: The Intelligence of the Unconscious - Gerd Gigerenzer (2007)


A new scientific truth does not triumph by convincing its opponents and making them see the light, but rather because its opponents eventually die, and a new generation grows up that is familiar with it.

—Max Planck


A good name is better than riches.


The doorbell rang. The host rushed to the door to welcome the first guests arriving at the dinner party. He opened the door and turned to his wife. “May I introduce you to my new colleagues, Debbie and Robert.” Then he turned to his guests. “And I would like to introduce you to my dear wife, umm, ah, umm…” Panic spread over his face until his wife helped him out. “Joanne,” she said politely.

If a name sits too long on the tip of one’s tongue, time can become painfully endless, particularly if the person in question is intimately related. But it could be worse. Not being able to recall a name occurs more often the older one gets, particularly if one has a Y chromosome. Yet if the husband no longer recognized his wife’s name or her face, his gaffe would be in a different ballpark. He would be considered a clinical case and possibly end up in a neurological institution. Recognition memory is more reliable than recall memory both at the beginning and at the end of our lives and is also more fundamental; it is difficult, for instance, to recall individual information about a person whom one does not recognize.1 Recognition memory is an evolved capacity that the recognition heuristic makes use of. We have already had a glimpse at recognition; let us take a closer look at it now.

No single rule of thumb can guide someone through an entire life. But the following story of a guy named Reese explores how mere recognition profoundly shapes our intuitions and emotions in everyday life.

Names, Names, Names

Reese was born in Spokane, a town in the state of Washington, where he spent his youth. He was recently nominated for a prize for outstanding emergency unit patient care and invited to fly to London. When he travels outside of the country, he envies people who can say that they are from New York, or from some other city that doesn’t induce the question “Where?” He went to the Intercollegiate Center for Nursing Education in Spokane, near the Coeur d’Alene mineral fields. When he tells this, he is used to getting more blank stares and adds, “gold and silver mines,” whereupon his listeners at least say, “Ah, yes.” He rarely fails to add that Spokane was the place where, according to legend, Butch Cassidy and the Sun-dance Kid, the infamous bank and train robbers and foremost members of the Wild Bunch, died in obscurity. That usually gets him the response: “I’ve seen that movie. It’s cool. But weren’t they shot in Bolivia?” Whatever the truth is, at least these are names people recognize and they are linked to his hometown, which makes him feel an iota more important.

On the plane to London, Reese was seated next to a British woman in a Chanel suit who asked him what he did for a living. Reese explained that he works at the Westminster Clinic and hoped that she had heard of its reputation. In the course of their conversation he told her that he invested in stocks in order to pay for his children’s college education, but that he would never buy any he hadn’t heard of. He feels strongly about proper investing, because he wants his child to get a good education rather then ending up, as he did, at a no-name college. His three-year-old daughter already recognizes Mickey Mouse and Ronald McDonald. She loves to watch Disney movies and eat a Big Mac. Her face lights up when she hears of Madonna and Michael Jackson, even though she knows nothing about music. The little girl often begs him to buy toys advertised on TV and is afraid of people she has not seen before. The night before the trip, she got sick, and Reese drove her to a doctor they know instead of to a much closer clinic.

After he had arrived in London, he learned that the dinner invitation was black-tie. He did not own a tuxedo, nor did he recall a place to get one. When he looked up “London tailors” on the Internet, he recognized “Savile Row” and rushed there to buy a dinner jacket. At the dinner reception, he found himself standing in a large ballroom, uneasily scanning people dressed in black and white, looking in vain for a familiar face. He was relieved to see the woman from the plane. Reese was runner-up to the prize and had a good time in London until his last day, when one of his bags was grabbed by a thief. Asked for a description of the bag snatcher, he could not give reliable details, but later at the police station he recognized the man in one of the photos. Reese was glad to get back home to Spokane and be with his family. New faces and locations make him tense, but familiar ones give him a feeling of ease, even of intimacy.


Recognition is the ability to tell the novel from the previously experienced, or the old from the new. Recognition and recall carve out our world into three states of memory. When visitors enter my office, they will be one of three kinds: those whose faces I do not recognize, those whom I recognize but don’t recall anything else about (this is called tartling in Scotland), and those whom I both recognize and recall something about. Note that recognition memory is not perfect; I may have a wrong feeling of déjà vu, or not remember that I have already encountered a person. However, such errors need not be dysfunctional, because, as we will see shortly, forgetting can actually benefit the recognition heuristic.

The capacity for recognition is adapted to the structure of the environment. Herring gulls recognize their hatched chicks in order to rescue them from danger. Their nests are on the ground, making it easy for the chicks to stroll away and be killed by a neighbor. But they do not recognize their own eggs and are happy to sit on those of other gulls, or on a wooden dummy provided by an experimenter.2 Outside of tricky experiments, they do not seem to need this capacity for recognizing eggs because their eggs don’t roll far enough away to reach the next adult’s nest. Lack of recognition can be exploited in the natural world as well. European cuckoos take advantage of other birds’ inability to recognize their own eggs and offspring and lay their eggs in others’ nests. The host birds seem to have wired into their brains the rule of thumb “feed any small bird sitting in your nest.” In this particular bird environment, where nests are separated and chicks cannot move between nests, individual recognition is not necessary for taking care of the chicks.

In contrast, humans have an extraordinarily large capacity for recognizing faces, voices, and pictures. As we wander through a stream of sights, sounds, tastes, odors, and tactile impressions, some novel and some previously experienced, we have little trouble telling the two apart. In a remarkable experiment, participants were shown 10,000 pictures for five seconds each. Two days later, they correctly identified 8,300 of them.3 No computer program to date can perform face recognition as well as a human child can. Why is this? As mentioned in chapter 4, humans are among the few species whose unrelated members exchange favors, such as trading goods, engaging in social contracts, or forming organizations. If we were not able to recognize faces, voices, or names, we would not be able to tell whom we’d encountered previously, and as a consequence, not recall who treated us fairly and who cheated. Hence, social contracts of reciprocity—“I share my food with you today, and you return the favor tomorrow”—could not be reinforced.

Recognition memory often remains when other types of memory, such as recall, become impaired. Elderly people suffering memory loss and patients suffering certain kinds of brain damage have problems saying what they know about an object or where they have encountered it. But they often know (or can act in a way that proves) that they have encountered the object before. Such was the case with R., a fifty-four-year-old policeman who developed such severe amnesia that he had difficulty identifying people he knew, even his wife and mother. One might be tempted to say he had lost his capacity for recognition. Yet in a test in which he was shown pairs of photographs consisting of one famous and one unfamous person, he could point to the famous persons as accurately as healthy people could.4 It was his ability to recall anything about the persons he recognized that was impaired. Because recognition continues to operate even when everything else breaks down, I view it as a primordial psychological mechanism.

As the $1 million question in chapter 1 illustrates, the purpose of the recognition heuristic is not to recognize objects, but to make inferences about something else. Here we investigate it in more detail.


The recognition heuristic is a simple tool from the adaptive toolbox that guides intuitive judgments, both inferences and personal choices. A judgment is called an inference when a single, clear-cut criterion exists, such as whether the Dow Jones will go up this week or whether or not a given player will win Wimbledon. Inferences can be right or wrong, and they can earn or lose a fortune. When no single, easily verifiable criterion exists, a judgment is called a personal choice—choosing a dress, a lifestyle, or a partner. Personal choices are more a matter of taste than of being objectively right and wrong, although the line between the two can get blurred.

Consider personal choices first. A business professor told me that he relies on brand-name recognition when purchasing a stereo system. He does not waste time consulting specialized magazines to learn about the avalanche of stereos featured on the market. Rather, he only considers brand names he has heard of, such as Sony. His rule of thumb is

When you buy a stereo, choose a brand you recognize and the second-least expensive model.

Brand-name recognition narrows down choice and the principle of the second-lowest price is added to make the final decision. The rationale is that if one has heard of a company, it is likely because its products are good. The professor’s justification for the additional step is that the quality of stereo technology has reached a level at which he is no longer able to hear the difference. His price principle arises from avoiding the cheapest and potentially least reliable model that companies manufacture for the low-price market. This rule saves time and likely protects him from being taken in.

Now consider inferences. The recognition heuristic can make accurate inferences when there is a substantial correlation between recognition and what one wants to know. For simplicity, I assume that the correlation is positive. Here is the recognition heuristic for inferences about two alternatives:

If you recognize one object but not the other, then infer that the recognized object has a higher value.

Whether the correlation is positive or negative can be learned from experience. Substantial correlations between name recognition and quality exist in competitive situations, such as the value of colleges, companies, or sports teams. When this correlation exists, ignorance is informative; the fact that you haven’t heard of a college, firm, or team tells you something about it. An easy way to measure the degree to which your ignorance is informative is your recognition validity. Let’s take the sixteen Gentlemen’s tennis matches in the third round of Wimbledon 2003:

Neither an expert who recognizes all the names in a set nor someone totally ignorant of the players can use the recognition heuristic to infer the winner. Only if you are partially ignorant, that is, you have heard of some but not all of the players can this heuristic guide your intuitions. To determine your personal recognition validity, mark all the names of tennis players you have heard of. Now take the pairs in which you have heard of one player but not the other. Count all cases in which the player you recognized won the game (the winner is always listed first in the left-hand column, and second in the right-hand column). Divide this number by the number of all pairs in which you recognized one but not both, and you get your recognition validity for this set. For example, if you recognize only Roddick, Federer, Schuettler, Agassi, and Novak, the recognition heuristic predicts the outcomes right in 4 out of 5 times; that is, the validity is 80 percent. Relying on this rule of thumb, you would get only the Novak-Popp match wrong, although you know almost nothing. You can use your recognition most effectively if you have heard of half of the players, not more and not less. If your recognition validity is above 50 percent, there is wisdom in your ignorance; you do better than chance.

If you want to test it in other cases, here’s the formula for all pairs in which one alternative is recognized and the other is not, in a class such as cities, companies, or sports teams:

Recognition validity = number of correct inferences divided by number of correct plus incorrect inferences.

The recognition heuristic, just like every rule of thumb, does not always lead to the correct answer. As a consequence, the recognition validity is typically smaller than 100 percent.

Your recognition does not have the same validity for all problems, but depends on the class of objects (such as this year’s Wimbledon Gentlemen’s Singles contestants) and the kind of inference to be made (such as who will win). Consider potentially fatal diseases and infections. When people inferred which of two diseases, such as asthma or tularemia, is more frequent, a study reported recognition validities around 60 percent.5 That is, the recognized disease was the more widespread one in 60 percent of the cases. This is better than chance, but not as good as for predicting the winners of Wimbledon tennis matches, which has been reported to be around 70 percent. When one wants to infer which of two foreign cities has a larger population, the validity is even higher, around 80 percent. In each case there exists a “mediator” that brings the names of diseases, players, or cities to the attention of the general public. These mechanisms include newspapers, radio, television, and word of mouth.


Figure 7-1: How the recognition heuristic works. Impact of quality: high-quality objects are mentioned more often in the media than low-quality objects. Impact of publicity: those that are mentioned more often are recognized more often. Recognition validity: those that are recognized more often are hence usually higher quality (based on Goldstein and Gigerenzer, 2002).

Figure 7-1 illustrates how the recognition heuristic works. On the right-hand side we see partially unknowledgeable people; that is, they have limited name recognition. On the left is what they are trying to infer (quality), such as who will win a sports game, which city is larger, or which product is better. On the top, there is a mediator in the environment, such as newspapers. The quality of a player or a product may be reflected by how often they are mentioned in the press. If so, the impact of quality is high. A manufacturer of running shoes, for example, could decide to produce high-quality shoes and trust that the product quality will lead to a higher profile in the media. In turn, the more often a name occurs in the news, the more likely it is that a person will have heard of the name, regardless of its actual quality. The manufacturer could then instead settle for a mediocre product, invest directly in publicity, and bet that people will buy the product because they have heard of it. Here the impact of publicity is an influential factor. This means shortcutting the triangle, as indeed many advertisers do. Measuring the impact of quality and publicity, we can predict in which situations relying on name recognition is informative or misleading.

So much for theory. But do people rely on the recognition heuristic in the real world? Let’s begin with soccer matches.


Founded in 1863, the Football Association (FA) is the ruling body for English football (soccer). It represents more than a million players belonging to tens of thousands of clubs and organizes national competitions. The FA Cup is the oldest soccer competition in the world and the major knockout tournament for English clubs. Teams are randomly paired, so that well-known clubs often compete against lesser-known clubs in lower divisions. Consider the following match from the third round of an FA Cup:

Manchester United plays Shrewsbury Town

Who will win? In a study, fifty-four British students and fifty Turkish students (living in Turkey) predicted the outcomes of this and thirty-one other FA Cup third-round matches.6 British participants had plenty of knowledge about the previous records and current conditions of the two teams and could ponder the pros and cons before deducing who would win. The Turkish participants had very little knowledge about (or interest in) English soccer teams, and many protested their ignorance during the testing. Nevertheless, the Turkish forecasters were nearly as accurate as the English ones (63 percent versus 66 percent correct). The reason for this strong performance was that the Turkish laypeople intuitively followed the recognition heuristic consistently in 95 percent (627 out of 662) of all cases. Recall that an expert who has heard of all teams cannot use the recognition heuristic. A person who has never heard of Shrewsbury Town but only of Manchester United can guess the answer faster by relying on partial ignorance.


Every year, millions of spectators watch the tennis matches at Wimbledon, one of the four annual grand slam tennis events, and the only one still played on natural grass. In the 2003 Gentlemen’s Singles Championship, 128 players competed. We have already seen the names of the 32 players who made it into the third round. The players were ranked by the Association of Tennis Professionals and by the Wimbledon experts’ seeding. For each of the 127 matches, one can predict that the player with the higher rank will win the game. In fact, the two ATP rankings predicted the winners correctly in 66 percent and 68 percent, respectively, of the matches. The experts did even slightly better. Their seeding predicted the outcome of 69 percent of the matches correctly.

How do ordinary people intuitively judge who wins? A study showed that in 90 percent of the cases where laypeople and amateur players had heard of one contestant but not of the other, they followed the recognition heuristic.7 The amateurs had heard of the names of about half of the players, whereas the laypeople had only heard of fourteen, on average. All players were ranked according to the number of ordinary people who recognized their names, and the prediction was made that the one with higher name recognition would win. I refer to this ranking as collective recognition. Would you bet money on the combined ignorance of people who had not even heard of half of the Wimbledon contestants?


Figure 7-2: How to predict the outcomes of the 2003 Wimbledon’s Gentlemen’s Singles matches. Benchmarks are (1) the ATP Champions Race, the official worldwide ranking of tennis players for the calendar year, (2) the ATP Entry Ranking, the official ranking for the last fifty-two weeks, and (3) the seeding, which represents the expert ranking by Wimbledon officials. The outcomes were also predicted by the collective recognition of laypeople who had heard of only a few players and amateur players who recognized only half of them. Collective recognition based on partially ignorant people predicted the actual outcomes as well as or better than the three official benchmarks did (Serwe and Frings, 2006).

The collective recognition of the laypeople predicted the outcomes of 66 percent of the matches correctly, which was as good as the number predicted by the ATP Entry Ranking. That of the amateur group predicted the outcomes of 72 percent of the matches correctly, which was better than each of the three official rankings (Figure 7-2). A study of Wimbledon 2005 replicated this striking result. Both studies demonstrate that collective wisdom can emerge from individual ignorance and indicate that there is a beneficial degree of ignorance, where less knowledge is more. But they do not tell us when and why.


Let me begin with the odd story of how we discovered, or stumbled across, the less-is-more effect. We were testing a completely different theory for which we needed two sets of questions, one easy, the other difficult. For the easy set we chose one hundred questions—such as “Which city has more inhabitants, Munich or Dortmund?”—which were randomly drawn from information about the seventy-five largest German cities.8 We asked students from the University of Salzburg, where I was teaching then, who knew lots about German cities. We thought a hundred similar questions from the seventy-five largest American cities would provide the difficult set, but when we saw the results, we could not believe our eyes (we had not yet read chapter 1 of this book). The students’ answers were slightly more often correct for American cities, not German ones! I did not understand how people could answer questions equally well about something that they knew less about.

Salzburg has excellent restaurants. That night my research group had dinner at one of them to mourn the failed experiment. We tried in vain to make sense of this bewildering result. Finally, the insight dawned. If the students knew sufficiently little, that is, had not even heard of many of the American cities, they might intuitively rely on their ignorance as information. When it came to German cities, they were not able to do this. By exploiting the wisdom in missing knowledge, they scored equally well with the American cities. It has been said that researchers are like sleepwalkers whose creative intuition guides them to intellectual destinations they could never clearly see beforehand. Instead, I had been like a sleepwalker who had failed to understand the creative hunches of the intuitive mind. Discovery arose by serendipity: failing to do one thing and yet achieving another, more interesting one.

But how exactly does the less-is-more effect arise? Consider three American brothers who apply to a new school in Idaho. The principal tests each of them on their general knowledge, starting with geography. He names two European countries, Spain and Portugal, and asks which one has more inhabitants. The youngest brother goes first. He has not even heard of Europe, let alone the countries, and just guesses. The principal tests him on other pairs of countries, but the little brother performs at chance level and fails. Now it’s the second brother’s turn. Unlike his little brother, he occasionally watches TV news and has heard of half of the European countries. Even though he thinks he is guessing because he knows nothing specific about the countries he recognizes, he gets two-thirds of the questions correct and passes. Finally, the oldest brother is tested. He has heard of the names of all the countries, although he also knows nothing specific beyond their names. Surprisingly enough, he does worse than the second brother.

How can that happen? The youngest brother who has heard of none of the countries cannot use the recognition heuristic and performs at chance level (Figure 7-3). The oldest brother who has heard of all the countries cannot use it either and also performs at chance level (50 percent). Only the middle brother who has heard of some but not all of the countries can use the recognition heuristic; since he has heard of half of them, he can use the heuristic most often and does best. He scores 65 percent correct. Why? The recognition validity is 80 percent, a typical value. Half of the time the middle brother has to guess and half of the time he can use the heuristic. Guessing results in 25 percent correct (half of the half ) and using the heuristic results in 40 percent correct (80 percent of the other half ). Together, that makes 65 percent—much better than chance, even though he knows nothing about population sizes. The line in the figure that connects the three brothers shows how they would have performed at intermediate levels of name recognition. On the right side of this curve, a less-is-more effect is visible: the brother who recognizes all the countries performs less well.


Figure 7-3: The less-is-more effect. Shown are three brothers who are asked which of two countries has more inhabitants. They know nothing about the countries, except that the middle brother has heard of the names of half of the countries, and the older has heard of all. The youngest brother and the oldest brother cannot rely on name recognition because they have heard of none and all, respectively, and perform at chance level. Only the middle brother can rely on name recognition, which improves his performance without knowing any facts.

Now consider three sisters who apply to the same school. The two older ones have some knowledge, such as that Germany is the country with the largest population in Europe. That extra knowledge may get them 60 percent correct (that is, 10 percentage points more than chance) when they recognize both countries.9 The principal puts them to the same test. Like the youngest brother, the youngest sister has not heard of any European country and guesses at chance level, but the oldest, who has heard of all, now gets 60 percent correct. Neither of them can use the recognition heuristic. The middle sibling, who has heard of half of the countries, is again better than the oldest, as the curve in Figure 7-4 shows.

Does that mean that partial ignorance is always better? The upper curve in Figure 7-4 shows a situation where the less-is-more effect disappears. This is when the validity of one’s knowledge matches or exceeds that of recognition. In the upper curve, both have a value of 80 percent. This means that someone knows enough to get 80 percent of the questions right when she recognizes both countries and 80 percent when she recognizes only one.10 Here, less is no longer more.


Figure 7-4: The less-is-more effect when people know something. The oldest sister has heard of all the countries and knows some facts and gets 60 percent of the questions right. The middle sister has not heard of half of the countries, and therefore can rely on name recognition and gets more answers right than her older sister. This less-is-more effect only disappears when the validity of knowledge is the same as that of recognition.

Daniel Goldstein and I have shown that a less-is-more effect can emerge in different situations. First, it can occur between two groups of people, when a more knowledgeable group makes worse inferences than a less knowledgeable group—take the performance of the American and German students on the question of whether Detroit or Milwaukee is larger (chapter 1). Second, a less-is-more effect can occur between domains, that is, when the same group of people achieve higher accuracy in a domain in which they know little than in a domain in which they know a lot. For instance, when American students were tested on the largest American cities (such as New York versus Chicago) and on the largest German cities (such as Cologne versus Frankfurt), they scored a median 71 percent correct on their own cities but slightly higher on the less familiar German ones, with 73 percent correct.11 This effect was obtained even though many Americans already knew the three largest U.S. cities in order and did not have to make any inferences. Third, a less-is-more effect can occur during knowledge acquisition, that is, when an individual’s performance first increases but then decreases again. All these are expressions of the same general principle, which help to understand why tennis amateurs were able to make better predictions than the official rankings of the ATP and Wimbledon experts.


Common sense suggests that forgetting stands in the way of good judgment. Earlier in this book, however, we met the Russian mnemonist Shereshevsky, whose memory was so perfect that it was flooded with details, making it difficult for him to get the gist of a story. Psychologists have worked out in detail how forgetting can benefit the use of recognition.12 Consider once again the oldest sister who recognizes all countries (Figure 7-4). If we could move her to the left on the curve toward the middle sister, she would do better. By forgetting some of the countries, she’d be able to use the recognition heuristic more often. Too much forgetting would be detrimental: if she moved too far to the left toward the youngest sister, she would again do worse.

In other words, if the oldest sister can no longer correctly remember every country she has heard of, this loss works to her advantage. This effect is only obtained if her memory errors are systematic, not random; that is, she tends to forget the smaller countries. Figure 7-4 shows that what constitutes a beneficial degree of forgetting also depends on the amount of one’s knowledge: the more one knows, the less beneficial it is to forget. The oldest brother would gain more from forgetting than the oldest sister would. This would lead us to expect that people who know more about a subject forget less often. Did forgetting evolve to aid inferences by rules of thumb? We do not know. We are, however, beginning to understand that cognitive limits are not simply liabilities, but can enable good judgment.


Let’s look at the game show in chapter 1 again. This time the show master asks a group of three the $1 million question: “Which city has more inhabitants, Detroit or Milwaukee?” Again, none of them knows the answer for sure. If the members disagree on their best bet, one might expect that the majority determine the group decision. This is called the majority rule.13 In an experiment, the following conflict arose. Two group members had heard of both cities and each concluded independently that Milwaukee was larger. But the third group member had not heard of Milwaukee, only of Detroit, and concluded that the latter was larger. What was their consensus? Given that two members had at least some knowledge about both cities, one might expect that the majority got their way. Surprisingly, in more than half of all cases (59 percent), the group voted for the most ignorant person’s choice. This number rose to 76 percent when two members relied on mere recognition.14

It may seem odd that group members let their answers be dominated by the person who knew the least. But in fact one can prove that this is a successful intuition when the validity of recognition is larger than that of knowledge, which was the case for the participants in this experiment. Hence, the seemingly irrational decision to follow the most ignorant member increased the overall accuracy of the group. The study also showed a less-is-more effect in groups. When two groups had the same average recognition and knowledge, the group who recognized fewer cities typically had more correct answers. For instance, the members of one group recognized on average only 60 percent of the cities, and those in a second group 80 percent; but the first group got 83 out of 100 questions correct, whereas the second got only 75. Thus, group members seemed to intuitively trust the value of recognition, which can improve accuracy and lead to the counterintuitive less-is-more effect.

How conscious was the group decision to follow the wisdom in ignorance? The videotaped group discussions show that in a few cases, the least knowledgeable members articulated that a particular city must be smaller because they had not heard of it, and the others commented on that. Yet in most cases, the discussions lacked explicit verbalization or reasoning. What is noticeable, however, is that people who could rely on the recognition heuristic made snap decisions, which seemed to impress those who knew more and needed time to reflect.


If you read magazines or watch TV, you will have noticed that much advertisement is noninformative. The notorious Benetton campaign, for instance, only presented their brand name together with shocking images such as a corpse in a pool of blood or a dying AIDS patient. Why do firms invest in this type of advertisement? The answer is to increase brand-name recognition, important because of consumers’ reliance on the recognition heuristic. Oliviero Toscani, the designer and man behind the Benetton campaign, pointed out that the ads had pushed Benetton beyond Chanel into the top-five best-known brand names across the world, and that Benetton’s sales increased by a factor of ten.15 If people did not rely on brand-name recognition in consumer choice, noninformative advertisement would be ineffective and so obsolete.

The effect of brand-name recognition extends to food as well. In an experiment, participants had a choice between three jars of peanut butter.16 In a pretest, one brand had been rated as higher quality, and participants could identify the higher-quality product 59 percent of the time in a blind test (substantially higher than chance, which was 33 percent). With another group of participants, the scientists put labels on the jars. One was a well-known national brand that had been advertised heavily and which all participants recognized; the other two were brands they had never heard of before. Then the experimenters put the higher-quality peanut butter into one of the jars with the unfamiliar labels. Would the same percentage of participants still choose the best-tasting peanut butter? No. This time 73 percent chose the low-quality product with the recognized brand label, and only 20 percent chose the high-quality product (Figure 7-5). Name recognition was more influential than taste perception. In a second tasting test, the researchers put exactly the same peanut butter into three jars, labeling two with unfamiliar labels and one with a brand-name label. The result was nearly identical. In this case, 75 percent of the participants chose the jar with the recognized brand, even though its content was the same as that of the other two jars (Figure 7-6). Nor did marking one brand with a higher price than the other two have much effect. Taste and price mattered little compared with the influence of the recognition heuristic.


Figure 7-5: Brand names taste better. In a taste test, people were asked to compare high-quality peanut butter put in a no-name brand jar with low-quality peanut butter put in a jar from a nationally recognized brand (Hoyer and Brown, 1990).


Figure 7-6: Brand-name recognition motivates consumer choice. In another taste test, the same peanut butter was in three different jars, one of which was a nationally recognized brand (Hoyer and Brown, 1990).

Relying on brand-name recognition is reasonable when firms first increase product quality, and increased quality subsequently increases name recognition, by word of mouth or the media. This world is represented in Figure 7-1, with consumers on the right, product quality on the left, the media on the top, and a strong correlation between product quality and media presence. Noninformative advertisement, however, shortcuts this process. Firms spend huge sums of money to directly increase the recognition of their brand name in the media. The competition for space in consumers’ recognition memory can impede or conflict with any interest in improving the product itself. In this case, the correlation between quality and media presence might be nil.

When consumers can only tell the difference between competing products by looking at the label, brand-name recognition and reputation become a substitute for genuine product preferences. Many beer drinkers have a favorite personal brand and claim that it tastes better than others. They swear it has more aroma, more body, is less bitter, and has just the right carbonation. These preferences are what some consumer theories take for granted, and that make more choice desirable as it gives every consumer a good chance of finding a match. Yet blind taste tests have repeatedly shown that consumers were unable to detect their own preferred brand. Some three hundred randomly selected American beer drinkers (who drank beer at least three times a week) were given five national and regional brands of beer.17 The beer drinkers assigned “their” brands superior ratings over all competitors, as long as the label was on the bottle. When the label was taken off, and the test was “blind,” none of the groups favoring a certain brand rated it as superior!

If consumers can only tell the difference between competing brands by name, then there is little economic justification for the idea that more choice is always better. Firms that spend their money on buying space in your recognition memory already know this. Similarly, politicians advertising their names and faces rather than their programs, and colleges, wannabe celebrities, and even small nations operate on the principle that if we do not recognize them, we will not favor them. Taken to the extreme, being recognized becomes the goal in itself.


An effective use of the recognition heuristic depends on two processes, recognition and evaluation. The first asks, “Do I recognize this alternative?” and determines whether the heuristic can be applied. The second asks, “Should I rely on recognition?” and evaluates whether it should be applied to the situation at hand. For instance, most of us are hesitant to gather and eat unfamiliar mushrooms that we find while hiking in the woods. Yet when we find the same mushrooms on our plate in a good restaurant, we will probably not even think twice about gobbling them down. In the woods, we follow the recognition heuristic: if it’s unknown, it might be poisonous. In the restaurant we do not follow it, because in this environment, the unknown is usually safe. This evaluation process is not always conscious. People intuitively “know” when a lack of recognition indicates a lack of safety.

The evaluation process is absent in automatic (reflex-like) rules of thumb. The recognition heuristic, by contrast, is flexible and can be consciously suppressed. How the evaluation process works is not yet known, but we have some clues. One aspect of this process seems to be whether reliable knowledge about what one wants to know can be retrieved. For instance, when members of Stanford University were asked whether Sausalito (a small town just north of the Golden Gate Bridge with only seventy-five hundred residents) or Heingjing (a made-up name that sounded like a Chinese city) had the larger population, most no longer relied on name recognition. They knew for sure that the town around the corner was small and guessed it must be Heingjing.18 The source of recognition is another aspect that seems to be evaluated. In the same study, people were asked whether Chernobyl or Heingjing was larger, and only a few decided on Chernobyl, known mainly for its nuclear power plant accident, which has nothing to do with its size. Not relying on name recognition in such cases is an adaptive and smart response—except in this study, of course, where the experimenter deceived people by using a nonexistent city.

Is There a Neural Correlate of the Evaluation Process?

I have argued that the recognition heuristic is used in a flexible way. That means that the mind evaluates whether it should be used in a given situation. This kind of evaluation is what I refer to as the intelligence of the unconscious. If such an evaluation process exists, it should be separate from the processes of recognition. Therefore we should find distinct neural activities in the brain when people decide whether to follow the heuristic. In a study using a brain imaging (functional magnetic resonance imaging, or fMRI) technique, we presented pairs of cities from Canada, England, France, Holland, Italy, Spain, and the United States to participants who were in a scanner.19 One group’s task was to infer which city in the pair had the larger population; the other group’s task was simply to indicate which cities they had heard of. Note that the first task involves an evaluation process whereas the latter only involves recognition memory.


Figure 7-7: A neural correlate of the evaluation process. When people decide whether to rely on the recognition heuristic, a specific neural activity can be observed in the anterior frontomedian cortex (afMC).

Did the brain scans indicate a neural correlate of this evaluation process? Such an activity needs to be very specific. That is, the particular brain activity should be observed when participants in the first group followed the recognition heuristic compared to when they did not, and should be absent when the participants in the second group simply indicated whether they recognized a city or not. Our study showed a specific activity in the anterior frontomedian cortex (Figure 7-7) of the first group but not the second. The location of the activity suggests that the process is not impulsive but a form of evaluation based on unconscious intelligence. There is still controversy about what this part of the brain actually does for us, though it has already been proposed that it serves evaluative functions, controls errors, and handles response conflicts. Its specific activity indicates that there is indeed a neural correlate to the evaluation process, that is, to the intelligence of the unconscious.

Where’s My Steak?

No concert hall can afford a program that consistently ignores recognition. People like to hear the same music over and over again, and La Traviata fills the house that a nameless opera may leave vacant. This taste for familiarity can conflict with a taste for variety. In 2003, the Berlin Philharmonic toured the United States with Sir Simon Rattle as the conductor. At the time, the Berlin orchestra was arguably the best in the world. In New York, they performed Debussy’s La Mer, which many did not know. In Rattle’s words:

During the New York performance of Debussy’s La Mer, half of the audience sat there with folded arms and disapproving faces. They were there because they wanted a Porterhouse steak; instead we served them an unfamiliar meal. The other half of the audience perked its ears, but the first half remained skeptical to the end. “Where’s my steak?” they asked themselves. It would be a mistake to give in and appease their culinary desires.20

Here, the audience mistrusted the musical taste of one of the best conductors in the world and went with the intuition “We don’t know it, so we don’t like it.” Simon Rattle, who fills concert halls based on his name alone, can afford to ignore the audience’s limited recognition of music, but less well known musicians and orchestras that strive for success rarely can.

There are more subtle ways to overcome novelty aversion. The eighteenth-century French economist and statesman Turgot was a reformer. As the story goes, he wanted to introduce potatoes to France, but peasants resisted the unknown food, until he came up with a trick. He ruled that only experimental state-run farms would be allowed to plant potatoes. Soon the peasants protested, clamoring for the privilege to grow them. Here, novelty aversion was trumped by a competitive social motive: “If another has it, I want it, too.” Thus, the impulse to prefer what one recognizes can be overridden in various ways. A conductor like Rattle can ignore name recognition, although not consistently, and a politician like Turgot overcame it by putting the aversion to novelty in conflict with envy.

Nonetheless, much of the time the gut feeling to “go with what you know” is a helpful guide in life. An effective use depends on two processes, recognition and evaluation. The first decides whether the simple rule can be applied, the second whether it should. People tend to follow the recognition heuristic intuitively when it is valid, and collective wisdom based on individual ignorance can even outperform experts.