The Gap

by Thomas Suddendorf

  Extensive research on false-belief tasks has found that this developmental pattern is quasi-universal across cultures. By age three and a half children tend to pass these tasks. The capacity develops slightly earlier in children who have older siblings from whom to learn and in children who do better on language tasks. This suggests that the social communicative environment in which children grow up influences the development of theory of mind. In line with this, it has been found that deaf children of hearing parents who only start to learn sign language late are delayed in passing false-belief tasks, whereas deaf children born to deaf parents, who are exposed to sign language from the start, develop the capacity to pass these tasks like their hearing peers.

  Developmental psychologists have identified a whole battery of links between the emergence of false-belief understanding and different concepts. One important finding has been that at about the same time that children first become able to attribute false beliefs to others, they also begin to attribute them to themselves. That is, when asked what they thought was inside a candy box before they were shown that there were, say, pencils and not candies inside, younger children tend to state that they thought all along there were pencils inside. They do this in spite of having been excited about the prospect of candy just a minute ago.

  Young children do not seem to grasp how they come to know what they know. If you put an object in a bag and ask toddlers how they would find out the object’s color—by putting their hand in it or by peeking inside—they choose arbitrarily. Very young children do not seem to appreciate how perception and knowledge are linked. This is why they happily “show” you their new toy when you are on the phone with them (with video phones, of course, this example may not work anymore). In studies in which children were either shown or told new information such as the content of a drawer, children later failed to accurately explain how they know what they know. When you play hide and seek with them, you have to accept that they may want to go back to the same hiding place repeatedly. They do not yet appreciate that the aim is to get out of mind, not just out of sight.

  The developmental psychologist John Flavell showed that children around this age also fail to distinguish between the appearance of an object and reality. We generally take for granted that when milk is poured into a blue glass, it does not change the milk’s color but only our perception of it. However, young children have difficulty appreciating that something can look like one thing and still be another. They find it hard to consider two contradicting views of the same thing. Here the child needs to integrate two opposing thoughts about the milk—“is blue” and “is white”—and may need to tag these thoughts as “looks like” and “is truly.”

  Being able to concurrently consider contradicting interpretations of the same object or event is required in a lot of mind reading. Without it there can be no lies. You may say something that is not true—we all make mistakes—but to actually lie, you need to know that what you say is not true, and, furthermore, you need to want someone else to believe that it is true. In other words, to lie is to knowingly implant a false belief. Without mind reading there can be no pedagogy, for deliberate transfer of knowledge requires some understanding of what is or is not known by the pupil to devise a way in which future knowledge can be acquired. In short, theory of mind is essential for normal human cultural and social interaction.

  ALTHOUGH RESEARCHERS HAVE LONG FOCUSED on three- and four-year-olds, there are both earlier and later developments in this arena. For a while, developmental psychologists were nearly obsessed with the classic false-belief task as the watershed. The fact that passing such tasks demonstrates mind reading was understandably enticing. However, failure on such tasks does not necessarily entail absence of a capacity for mind reading per se. Nor does passing the task mean that a child has achieved all there is to achieve.

  For example, even five-year-old children have problems with increased embedding, such as I expect that you think that everyone knows that. In fact, even adults can typically only juggle five or six such nested ideas before we lose the plot. I think that you suspect that I intend you to believe this. Joey, in the TV sitcom Friends, for instance, struggles desperately to follow when Phoebe, on discovering that Chandler and Monica know that she and Rachel know that they are dating, succinctly states: “They don’t know that we know they know we know!” She warns Joey not to tell the others, and he despairs: “I couldn’t even if I wanted to.” As Robin Dunbar pointed out, only exceptionally clever writers like Shakespeare make our minds stretch with five or six levels of nesting, as he, for instance, intends us to believe that Iago wants Othello to suppose that Desdemona loves Cassio, who actually loves Bianca.

  Children have plenty to learn about embedding and the numerous intricacies of the mind. For example, it takes them some time to get to properly appreciate faux pas. Consider, for instance, a scenario in which, say, Jane visits Frank and then accidentally breaks his bowl. The bowl is a present that Jane gave Frank some years ago. Now imagine that when Jane apologizes for the accident, Frank, not remembering how he got it in the first place, tries to put her at ease by saying, “Oh, no worries. I never liked the bowl anyway.” Young children do not understand such a violation of etiquette. To appreciate it, one has to combine several different types of mental states: Frank did not remember where he got the bowl from, he did not have an intention to hurt Jane, but hurt she is. It can all be rather complicated, as I suspect you know. Indeed, our learning about the sensibilities and workings of other people’s minds might never stop. Just consider the misunderstandings characteristic of interactions between romantic partners or between different cultures. Many adults keep trying to learn more about minds. We may meditate, read self-help books, go to workshops, or study psychology.

  Sometimes minds seem to be awfully complex and unpredictable, and we may even despair at not understanding one another. Just as we get our time travels wrong, we regularly misjudge what is on others’ minds. And sadly we often feel misunderstood. At other times, we seem to see right through someone else’s thinking and “read them like an open book,” as the saying goes. We have meetings of minds and sometimes feel we thoroughly understand each other. We fall in love with each others’ minds.

  One reason mind reading is such an exceedingly complicated affair is that people frequently say one thing but mean quite another. I refer here not only to lies but to devices such as irony and metaphor. You may, for instance, say, “That’s just wonderful; it’s exactly what I needed,” on hearing news about a strike at the airport—as you are packing your bags. In irony we may state the opposite of what we mean, and if we do this in a harsh tone, it turns into sarcasm. We may exaggerate or understate, make fun through satire or parody, make vague suggestions, or employ double entendres. Not only actors can pretend to have one mental state while in fact being in quite a different one. People can cry and yet not actually be sad, and vice versa. We have some control over what we want to reveal and what we want to conceal about our minds. We can manage the impressions we give. Strategic games unfold as we try to influence each other’s minds. It is the stuff of our lives—or at least of our soap operas.

  While there is much left to learn once children pass false-belief tasks, failing these tasks need not mean younger children are clueless about minds. Indeed, as we have seen, newborns already have a special interest in social stimuli and over the first few months at least appear to develop some appreciation of the minds of others. They form expectations about the goals and intentions of animate objects—even of cartoons on computer screens. Joint activities, shared attention, gaze following, and declarative pointing suggest that minds increasingly matter to them. In the second year they show some competence at appropriately interpreting emotions, desires, and intentions. They copy what someone else intends to do, even if that person fails to achieve the goal. They get upset if you do not join in and attend what they attend to. “Look at me, look at me,” my two-year-old daughter is yelling in the background, as
if to illustrate the point.

  In fact, even toddlers may have some understanding of false beliefs. When examining spontaneous responses, such as looking direction, toddlers and young children seem to have an expectation of where someone who holds a false belief about the location of an object will look first. When you ask them, however, they will insist that this person will search for the target where it actually is, not where it is mistakenly believed to be. This suggests that they have some problem inhibiting their own knowledge of the facts to reasoning about others’. But they may already develop a basic capacity for mind reading a lot earlier than previously thought. There continues to be debate about the nature of this early understanding. The psychologists Ian Apperly and Stephen Butterfill, for instance, argue that humans have two different systems that track belief: one that is implicit and evident in toddlers, and one that is explicit and develops later in preschoolers.

  Researchers are increasingly finding evidence that theory of mind develops gradually through social scaffolding and not with the big bang once imaged. The developmental psychologists Henry Wellman, Candi Peterson, and colleagues, for instance, have documented how mind reading typically progresses from an early understanding that individuals can have diverse desires to an understanding of diverse beliefs. Children pass standard false-belief tasks before they appreciate that someone can feel one thing and yet display a different emotion.

  Do other animals show any of these capacities? Do they consider the minds around them? Do they stare deeply into each other’s eyes and link their minds? Since Premack and Woodruff’s paper, extensive efforts have been made to find out.

  IF YOU LOOK STRAIGHT AT a rhesus monkey you might well be attacked. For primates, staring into another’s eyes is typically a threatening gesture. Therefore, primates largely avoid eye contact, and face-to-face interactions are surprisingly unusual. Even chimpanzees look into each others’ eyes only on rare occasions.5 When you stare into the eyes of chimpanzees, one thing you will notice is that their eyes do not show any white. Human eyes physically differ from those of other primates in that our sclera is white and more of it is exposed in relation to the eye outline. Human eyes signal gaze direction. We advertise where we look, and we read where others are looking. The eyes of other primates, if anything, seem to camouflage gaze direction. They do not roll their eyes to express disdain or shed tears to express their sorrow.

  We use the eyes extensively in mind reading without words. For instance, soccer players often rely on this when taking penalty kicks. In my youth I could reasonably reliably convert the shot by simply peeking briefly at one corner of the goal, running up, and then slotting the ball in the other corner. This technique relies not on the accuracy or velocity of the shot but almost entirely on fooling the keeper. Eventually clever keepers began to pick up on this simplest of tricks and tried to thwart me by jumping in the opposite corner to the one I looked at. Some even moved closer to the other post and hence “offered” one side. The battle became increasingly more difficult, as I had to size up the keeper’s ability to read my intentions and do the opposite of what I thought he thought. This is “theory of mind” in action.

  FIGURE 6.1.

  Ockie and me. Chimpanzees on the odd occasion do make eye contact.

  Following Premack and Woodruff’s paper, field studies added enthusiasm for the possibility that other primates read minds and demonstrate this in their actions. The work of primatologists like Jane Goodall suggested that primate societies are much more complex than had been previously assumed. The notion that social pressures drove the evolution of intelligence gathered momentum, and it was fitting to suppose that primates may have evolved some mind-reading capacities to understand and control the behavior of others. For instance, reports of apparent tactical deceptions in primate societies suggested that they can reason about each others’ desires and beliefs—and that they may be able to deliberately implant false beliefs.6 In a classic example, a baboon was observed mating with a female behind a rock in a position such that the alpha male of the group could see his head but not what he was doing. Was the baboon intending to mislead the alpha male into thinking he was doing something else? As we have seen, leaner explanations can also account for such observations. For instance, the baboon may have been punished in the past by the alpha male when mating in the open but not when mating behind rocks. So he may mate behind rocks without necessarily reasoning about what the alpha male can or cannot see and know. Nevertheless, early laboratory experiments supported the richer interpretation of mind-reading in at least our closest living relatives. In the early 1990s the comparative psychologist Daniel Povinelli reported studies that suggested chimpanzees, though not monkeys, might be able to reason about others’ knowledge and take their perspective.

  Following these initial findings, Povinelli set up his own chimpanzee research center at the University of Louisiana and examined their understanding of seeing, pointing, intention, and knowledge. To general surprise, he failed to find additional support for chimpanzee theory of mind. Instead, he discovered numerous reasons for lean interpretations of their behavior. In scores of studies, he obtained consistently negative results from his group of young chimpanzees. For example, they would beg from someone who was wearing a bucket over her head just as often as from someone who could see what was going on. When one trainer saw where food was hidden and another could not see this hiding—because they had left the room, looked away, or wore a blindfold or a bucket—the apes were equally likely to follow the advice of the ignorant as of the knowledgeable human. Chimpanzees that had been trained on a cooperative task, such as pulling a box with two ropes, ignore that a naïve chimpanzee lacks the relevant knowledge to get the task done and do not teach them. Povinelli championed the killjoy cause and concluded that chimpanzees only reason about behavior, not about minds.

  Povinelli’s research suggests that theory of mind is uniquely human. Great ape behavior may be driven by more basic calculations. Indeed, we might be misguided if we think our own behaviors are caused by theory of mind simply because we explain them in such terms. Perhaps we often merely reinterpret behavior in mental terms. For instance, to keep with the theme of soccer that I kicked this section off with, when an offensive player tries to dribble past a defensive player, a common trick is to make the opponent commit to one side and to delay one’s own final move in order to capitalize on that commitment. When we explain these actions afterwards, we may say that we wanted to fake out the defender, misleading him into thinking we were going to go the other way. Yet it is not clear such thoughts drive our actions. We have time before a penalty to plot deception and counterdeception. In the heat of the action, however, we can’t stand there and formulate explicit notions of each other’s intent. We may automatically go through all the motions and only afterwards interpret them in mental terms.

  Povinelli and colleagues argue that ape behaviors that look as though they might be driven by mind reading, such as apparent deception, empathy, or grudging, may not be. When we see chimpanzees engage in a chase of rapid side-to-side movements, we may be misinterpreting their behavior in terms of faking a turn to fool the pursuer into thinking a certain way; the apes may merely be concerned with actions.7 Only humans, they claim, have evolved a capacity to reinterpret behavior in mental terms.

  There is an alternative explanation to this proposal, however. Andrew Whiten and I have suggested that in some cases the rapid actions of deception and counterdeceptions are only automatic because we have initially engaged in a lot of practice involving explicit cognitive considerations. Human skill development is replete with examples in which behaviors that had once been governed by slow conscious processes become automatic with training. Just think of the complexities of driving a car. Initially you need to carefully think about what your arms and legs are supposed to be doing to control the vehicle. With experience you can focus on having a conversation or listening to the radio without attending to how your autopilot, as it were, is dri
ving the car. Similarly, strategies that allow one to get around defenders in soccer may become automatized only after extensive training. It requires a lot of practice to become a good player. Perhaps theory of mind also develops like other skill acquisition, from effortful, explicit, and controlled to fast, automatic, noneffortful processing that comes with practice. Initial mental simulations can give way to quicker shortcuts. For example, we may tend to quite automatically follow the gaze of others not, as Povinelli suggests, because we have some low-level mechanism for this but because we had sufficient experience and practice with these situations.

  In spite of Povinelli’s considerable output, other laboratories—particularly the Max Planck Institute for Evolutionary Anthropology in Leipzig—gradually flooded the field with findings supporting richer interpretations of animal behavior. The comparative psychologists Michael Tomasello and Josep Call and their colleagues showed, for example, that gaze-following abilities in apes were more sophisticated than previously thought. Even dogs and monkeys appear to be able to follow gaze under some circumstances. Chimpanzees can project somebody else’s line of sight geometrically beyond their own immediate visual field and do so even around barriers. They move in ways that allow them to see what another is looking at—even if the target is behind an obstacle. This suggests that chimpanzees may interpret gaze in terms of what the other person might see. They sometimes even glance back and forth as if to check what in fact the experimenter is finding so interesting.

  In collaboration with Brian Hare, the Leipzig group produced some ingenious studies supporting the possibility that chimpanzees reason about minds. In one set of experiments chimpanzees had to compete with another, more dominant chimpanzee for food. The chimpanzees were found to preferentially head toward seizure of whichever of two food items was visually screened from the dominant competitor. This suggests that they do understand something about what another can see. When the opaque screens were made transparent, the preference for the “concealed” food vanished, presumably because they recognized that it no longer blocked the view of the competitor.