HOW INTUITION WORKS - UNCONSCIOUS INTELLIGENCE - Gut Feelings: The Intelligence of the Unconscious - Gerd Gigerenzer

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

Part 1. UNCONSCIOUS INTELLIGENCE

It is a profoundly erroneous truism, repeated by all copy-books and by eminent people when they are making speeches, that we should cultivate the habit of thinking of what we are doing. The precise opposite is the case. Civilization advances by extending the number of important operations which we can perform without thinking about them.

—Alfred North Whitehead1

3. HOW INTUITION WORKS

Charles Darwin thought of the hive bee’s art of cell making as “the most wonderful of all known instincts.”2 He thought that this instinct evolved from numerous successive and slight modifications of simpler instincts. I believe that the evolution of cognition can be understood in a similar way, based on an adaptive toolbox of “instincts,” which I call rules of thumb, or heuristics. Much of intuitive behavior, from perceiving to believing to deceiving, can be described in the form of these simple mechanisms that are adapted to the world we inhabit. They help us master the primary challenge for human intelligence: to go beyond the information given.3 Let us begin with how our eyes and brains make unconscious bets.

BRAINS MAKE THINGS UP

King Henry VIII is known as having been a self-centered and forever mistrustful ruler who went through six marriages, with two of his wives joining the large tally of eminent persons he executed for alleged treason. As the story goes, his favorite dinner enjoyment was to close one eye and behead his guests. Would you like to give it a try? Close your right eye and stare at the smiling face located on the top right-hand side of Figure 3-1. Hold the book about ten inches away from your face, then move the book slowly toward you and away again, keeping your left eye focused on the smiling face. At some point, the sad face on the left will disappear, as if beheaded. Why does our brain act like a guillotine? The region in which the face disappears corresponds to the “blind spot” in the retina of the human eye. The eye acts like a camera, with a lens that directs light rays so that a picture of the world is created on the retina. The sheet of photoreceptors on the retina is much like a sheet of film at the back of the camera. But unlike film, there is a hole in it, through which the optical nerves exit the retina to transport the information to the brain. Because the hole has no photoreceptors, objects that would be processed in this region cannot be seen. When you look around with one eye closed, you might therefore expect to see a blank slate corresponding to this blind spot. In fact, you won’t notice a thing. Our brains “fill in” the blank slate with a good guess. In Figure 3-1 (top), the best guess is “white” because the surrounding field is white. That guess makes the sad face disappear. In the same way, Henry VIII “beheaded” his guests by centering the image of their heads in his open eye’s blind spot.

image

Figure 3-1: Seeing is betting. Close your right eye and stare at the smiling face in the top panel. Move the page closer to you while still staring; at some point the sad face on the left will disappear. Repeat the procedure with the bottom panel. At some point, your mind will repair the broken fork on the left. This creative process illustrates that the nature of perception is unconscious betting, not a veridical picture of what is out there.

Now try something with your brain more constructive than beheading. Close your right eye and stare at the smiling face at the lower part of Figure 3-1, then move the book slowly toward you and away again. You will see that the broken fork on the left is miraculously repaired. The brain again makes its best guess based on the surrounding information: an elongated object crosses the blind spot from one side and continues on the other, thus it is likely that it exists in between. As in the case of the beheaded guest, these intelligent inferences are unconscious. Our brain cannot help but draw inferences about the world. Without them, we would see details but no structures.

Evolution could have created a better design in which the optical nerves exited from the backside rather than the surface of the retina. And in fact it has, but not for us. An octopus has no blind spot. The cells that carry information to the brain are located in the outer portions of the retina, so that the optic nerves do not have to cross through the retina. But even if evolution had favored us instead of the octopus, the general point remains, as the next section will illustrate. A good perceptual system has to go beyond the information given; it has to “invent” things. Your brain sees more than what your eye sees. Intelligence means making bets, taking risks.

I believe that intuitive judgments work in the same way as these perceptual bets. When given insufficient information, the brain makes things up based on assumptions about the world. The difference is that intuition is more flexible than perception. Let us first see how exactly these perceptual inferences work.

UNCONSCIOUS INFERENCES

To understand in more detail how our brains “go beyond the information given,” consider the dots on the left-hand side of Figure 3-2. They appear concave; that is, they recede into the surface like small dents. The dots on the right-hand side, however, appear convex; that is, they project up from the surface, extending toward the observer. When you turn the book upside down, the concave dots will change into convex dots, and vice versa. Why do we see the dots this way or the other?

The answer is again that the eye does not have sufficient information to know for certain what is out there. But our brains are not paralyzed by uncertainty. The brain makes a “bet” based on the structure of the environment, or what the brain assumes the structure to be. Assuming a three-dimensional world, it uses the shaded parts of the dots to guess in what direction of the third dimension they extend. In order to make a good guess, it assumes that

image

Figure 3-2: Unconscious inferences. The mind automatically infers that the dots in the left-hand picture are curved inward, that is, away from the observer, and those in the right-hand picture are curved outward, that is, toward the observer. If you turn the book around, the inward dots will pop out and vice versa.

1. light comes from above, and

2. there is only one source of light.

These two structures are characteristic in human (and mammalian) history, where sun and moon were the only sources of light. The first regularity also holds approximately true for artificial light today, which is typically placed above us—although exceptions, such as car lights, exist. The brain goes beyond the little information it has and relies on a simple rule of thumb adapted to these assumed structures:


If the shade is in the upper part, then the dots recede into the surface; if the shade is in the lower part, then the dots project up from the surface.


Consider the dots on the right. They are bright in the upper part, and shaded in the lower part. Thus, the brain’s unconscious inference is that the dots extend toward the observer, with light hitting the upper part and less light hitting the lower part. In contrast, the dots on the left side are shaded in the upper part and bright in the lower part; for the same reasons the brain bets that they must be curved inward. These assumptions, however, are generally not conscious, which is why the great German physiologist Hermann von Helmholtz spoke of unconscious inferences.4 Unconscious inferences weave together data from the senses using prior knowledge about the world. There is a debate whether they are learned individually, as von Helmholtz and the Vienna psychologist Egon Brunswik argued, or are acquired by evolutionary learning, as the Stanford psychologist Roger Shepard and others have maintained.

These unconscious perceptual inferences are strong enough to act upon, but unlike other intuitive judgments, they are not flexible. They are triggered by external stimuli in an automatic way. An automatic process cannot be changed by insight or information external to the process. Even now, when we understand how the intuitive perception works, we cannot change what we see. We continue to see the concave dots suddenly pop out of the surface when we turn the book upside down.

Humans would not be called Homo sapiens if all inferences were like reflexes. As we have seen, other rules of thumb have all the advantages of perceptual bets—such as being fast, frugal, and adapted to their environment—but their use is not fully automatic. Although typically unconscious in nature, they can be subjected to conscious intervention. Consider how children infer the intentions of others.

WHAT DOES CHARLIE WANT?

From an early age, we intuitively have a feeling of what others want, what they desire, and what they think of us. But how do we arrive at these feelings? Let’s show a child a schematic drawing of a face (“Charlie”) surrounded by a seductive selection of chocolate bars (Figure 3-3).5 We then say, “This is my friend Charlie. Charlie wants one of these treats. Which one does Charlie want?” How could a child possibly know? Yet almost all children immediately point to the same treat, the Milky Way. In contrast, children with autism tend to fail on this task. Some pick one, some another, and many egotistically pick their own favorite treat. Why do non-autistic children have a clear intuition about what Charlie wants, whereas autistic children do not? The answer is that the non-autistic children automatically engage in “mind reading.” Mind readers can work with only minimal clues. They notice, but perhaps not consciously, that Charlie’s eyes are pointed to the Milky Way, and so infer that this is the one he wants. However—and this is crucial—when asked what Charlie is looking at, children with autism answer correctly. What they do not seem to do as well as other children is to make the spontaneous inference from looking to wanting:


If a person looks at one alternative (longer than at others), it is likely the one the person desires.


In non-autistic children, this mind-reading heuristic is effortless and automatic. It is part of their folk psychology. The ability to infer intentions from a gaze seems to be localized in the superior temporal sulcus of our brain.6 In children with autism this instinct appears to be impaired. They don’t seem to understand how the minds of others work. In the words of Temple Grandin, an autistic woman who holds a PhD in animal science, much of the time she feels “like an anthropologist on Mars.”7

image

Figure 3-3: Which one does Charlie want?

As with the unconscious inferences in perception, this simple rule for inferring desire from gaze may well be anchored in our genes and does not need much learning. However, unlike perceptual rules, the inference from gaze to desire is not automatic. If I have reason to assume that Charlie wants to deceive me, I can change my impression that he prefers the Milky Way. I might conclude that he is only looking at the Milky Way in order to influence me to take it, so that he can easily get the Snickers he in fact desires. Here we have a candidate for a rule of thumb that might be genetically coded and unconscious but nevertheless brought under voluntary control. In fact, autistic people sometimes use this voluntary control when trying to understand the secrets of mind reading. Grandin reported that, like a cognitive scientist, she tries to discover the rules that ordinary people use unconsciously and are unable to tell her. Then she uses a rule consciously, as if it were the grammar of a foreign language.

WHAT MAKES A GUT FEELING WORK?

Intuitive feelings seem mysterious and hard to explain—and most social scientists have steered clear of them. Even books that celebrate rapid judgments shy away from ever asking how a gut feeling arises. Rules of thumb provide the answer. They are typically unconscious but can be lifted to the conscious level. Most important, they are anchored both in the evolved brain and in the environment. By making use of both evolved capacities in our brain and environmental structures, rules of thumb and their product—gut feelings—can be highly successful. Let me walk you through this scheme.

· Gut feelings are what we experience. They appear quickly in consciousness, we do not fully understand why we have them, but we are prepared to act on them.

· Rules of thumb are responsible for producing gut feelings. For instance, the mind-reading heuristic tells us what others desire, the recognition heuristic produces a feeling of which product to trust, and the gaze heuristic generates an intuition of where to run.

· Evolved capacities are the construction material for rules of thumb. For example, the gaze heuristic takes advantage of the ability to track objects. It is easy for humans—in contrast to robots—to track a moving object against a busy background; at three months old, babies have already begun to hold their gaze on moving targets.8 Thus, the gaze heuristic is simple for humans but not for present-day robots.

· Environmental structures are the key to how well or poorly a rule of thumb works. For instance, the recognition heuristic takes advantage of situations where name recognition matches the quality of products or the size of cities. A gut feeling is not good or bad, rational or irrational per se. Its value is dependent on the context in which the rule of thumb is used.

image

Figure 3-4: How gut feelings work. A gut feeling rapidly appears in consciousness, based on unconscious rules of thumb. These are anchored in the evolved capacities of the brain and the environment.

Automatic rules, such as inferring depth from shade, and flexible rules, such as the recognition and gaze heuristics, all work according to this scheme. Yet there is an important difference. An automatic rule is adapted to our past environment without a present evaluation as to whether it is appropriate. It is simply triggered when a stimulus is present. Life has survived on this form of mindlessness from time immemorial. The flexible rules, in contrast, involve a quick evaluation of which one to use. If one doesn’t work, there are others to choose from. The phrase “intelligence of the unconscious” refers to this quick process of evaluation. Brain imaging indicates that it might be associated with the anterior frontomedian cortex (see chapter 7). Gut feelings may appear simplistic, but their underlying intelligence lies in selecting the right rule of thumb for the right situation.

TWO WAYS TO UNDERSTAND BEHAVIOR

Like other approaches in the social sciences, the science of intuition attempts to explain and predict human behavior. Otherwise, it is unlike many of the other approaches. Gut feelings and rules of thumb are not the same kinds of explanations as fixed character traits, preferences, and attitudes, the key difference being that, as mentioned, rules of thumb are anchored not just in the brain but also in the environment. Explaining behavior in this way is what we call an adaptive approach, which assumes that people’s behavior develops in a flexible way as they interact with their environment. Evolutionary psychology, for example, tries to understand present-day behavior by relating it to the past environment in which humans evolved.9 Brunswik once compared the mind and the environment to a married couple who have to come to terms with one another. I will use his analogy to sketch out the difference between mental (internal) and adaptive explanations.

To generalize, let me distinguish between two ways a husband and wife can interact: to be kind and try to make each other happy, or to be nasty and try to hurt each other. Consider two couples, the Concords and the Frictions, who are in many respects similar. Yet the Concords are kind, warm, caring, and get along very well, while the Frictions fight, yell, insult each other, and are on the verge of splitting up. How can we explain the difference?

According to a widespread account, every person has a set of beliefs and desires, and these are the causes of behavior. For instance, Mr. and Ms. Friction might have sadomasochistic impulses and derive pleasure from hurting each other, and they simply maximize that pleasure. Alternatively, the couple might not have such desires but might instead have failed to calculate how they should behave. The first is the rational account and the second the irrational one, both assuming that people rely on mental calculations equivalent to Franklin’s balance sheet. A third account is in terms of personality traits and attitudes, such as an overly aggressive temperament, or a dismissive attitude toward the other sex. Note that each of these explanations seeks the cause of a person’s behavior in the individual mind. Personality theories investigate traits, attitude theories study attitudes, and cognitive theories focus on probabilities and utilities, or beliefs and desires.

The tendency to explain behavior internally without analyzing the environment is known as the “fundamental attribution error.” Social psychologists have studied this tendency in the general public, but the same error creeps into social scientists’ explanations as well. A person who takes financial risks with the stock market is not the same as one who takes social risks in dating or physical risks in mountain climbing. Few among us are risk seeking in all respects. As a student, I was fond of personality and attitude research but learned the hard way that it rarely predicts behavior well, and for good reason. The idea of fixed traits and preferences overlooks the adaptive nature of Homo sapiens. For the same reason, knowing the human genome does not mean understanding human behavior; the social environment also has a direct influence, possibly even on DNA’s production of growth hormones.10 As Brunswik observed, to understand the wife’s behavior, one needs to find out what her husband is doing, and vice versa.

Adaptive theories focus on the relation between the mind and the environment, rather than on the mind alone.11 Would that suggest a different story for the Concords and Frictions? Here we need to think about how rules of thumb interact with an environment’s structure. What underlies the spouses’ behavior? Consider a rule of thumb that has been dubbed tit for tat:


Be kind first, keep a memory of size one, and imitate your partner’s last behavior.


Assume that Ms. Concord, who unconsciously uses this rule of thumb, is solving a task with her husband for the very first time (caring for their first newborn, shopping for clothes together, or preparing dinner and doing the dishes). Ms. and Mr. Concord are kind to each other on the first occasion. The next time, she imitates his cooperative behavior, he imitates hers, and so on. The result can be a long harmonious relationship. The phrase “keep a memory of size one” means that only the last behavior (kind or nasty) is imitated and needs to be remembered. A relationship can grow if partners are willing to forget mistakes in the past, but not if one partner digs out the same old skeleton from the closet over and over again. In this case, forgetting means forgiving.

Most important, the same rule of thumb can lead to opposite behaviors, kind or nasty, depending on the social environment. If Ms. Concord had married someone with the maxim “Always be nasty to your wife, so that she knows who is the boss,” her behavior would be the inverse. Triggered by her husband’s nasty behavior, she would in turn react nastily to him. Behavior is not a mirror of a trait, but an adaptive reaction to one’s environment.

Tit for tat works well if the partner also relies on it and does not make errors. Assume that the Frictions also intuitively rely on tit for tat. They similarly started out as a caring couple, but Mr. Friction said something hostile in a burst of anger, and since then there has been no end to exchanging blows. Ms. Friction was hurt, so she responded with something similar. That made him hit back at the next opportunity, and on it went. Now, the original incident is long forgotten, but they have become trapped in an endless behavioral pattern. Mr. Friction feels that her last insult is to blame for his own, and she thinks the same of her own response. How can the Frictions stop the game, or not play it in the first place? They could rely on a more forgiving rule called tit for two tats.


Be kind first, then keep a memory of size two, and be nasty only if your partner did so twice; otherwise be kind.


Here, if he accidentally insults her, she gives him a second chance. Only if it happens twice in a row does she retaliate. Tit for two tats works better for couples in which one partner behaves unreliably without being intentionally malicious. Yet its leniency is vulnerable to exploitation. For instance, think of a man who gets drunk and beats his wife one night, but deeply regrets it the following day and is gentle and thoughtful. If her mind works by tit for two tats, she will remain kind to him. A shrewd man can consciously or unconsciously repeat the game for a long time to exploit her sense of forgiving. Switching to tit for tat would mean that he could no longer take advantage of her.

How well would these simple rules work in interactions with many partners rather than one, who may rely on different kinds of rules? In a widely publicized computer tournament, the American political scientist Robert Axelrod had fifteen strategies compete with each other, and the winnings were totaled up for all games. The winning strategy was tit for tat, despite its simplicity.12 In fact, the most complex strategy was the least successful. Axelrod then worked out that if someone had submitted tit for two tats to his tournament, it would have been the winner instead. It is good at avoiding rounds of mutual recrimination, as happened with the Frictions. Does this mean that tit for two tats is generally better than tit for tat? No. Just as in real life, there is no single best strategy—it depends on the games the other players play. When Axelrod announced a second tournament, the eminent evolutionary biologist John Maynard Smith submitted tit for two tats. But the saintly heuristic did not win. Confronted with nasty strategies that tried to exploit the softies, it ranked far down the list. Once again the winner was tit for tat. Its wisdom is in its building blocks. In general, cooperation pays, forgetting pays, and imitation pays. And, most important, the combination pays. If one followed the biblical tenet “Turn the other cheek,” exploitation would be a likely consequence.

Like the perceptual rules encountered in this chapter, tit for tat is based on evolved capacities, including those for imitation. These capacities are not the same as traits; rather, they are the stuff from which the rules of thumb are built. The next two chapters will address how they are anchored in the brain and in the environment, respectively.