The Gap

by Thomas Suddendorf

  Alas, animal field studies have produced hardly any evidence for pretend play in the wild. You might rightly wonder how one could know whether, say, a wolf bolting into the distance is pretending to chase a rabbit or is simply running. Indeed, it is exceedingly difficult to distinguish imaginary play from other types of play in the absence of props such as the dolls and toys that human children use (which in itself may hint at differences). There is one published report of purported pretend play with something resembling a toy from a chimpanzee the researchers called Kakama. While traveling with his pregnant mother, Kakama carried a log for several hours that he seemed, according to the observing researchers, to treat suspiciously like a baby. This behavior included making a nest and placing the log inside it. When two field assistants, who did not know of these events, observed Kakama some months later engaging in similar behaviors, they collected the (new) log and labeled it as his “toy baby.” Although this is suggestive, there may be reasons other than pretend play that could account for the observed behavior in this anecdote.

  When there are many convergent anecdotes, researchers may become more confident in a richer interpretation. Andrew Whiten and Dick Byrne systematically collected and categorized many reports of what they called tactical deception in primates. These are cases of apparent social manipulation, which suggest that primates at times pretend in order to get an advantage or avoid punishment. For example, a lower-ranking male chimpanzee, when almost caught by the alpha male in a sexual act with a female, quickly covered his erect penis with his hands, as if to hide the evidence. Other reports suggested not just concealing but active misleading. For example, a baboon being chased suddenly stopped, jumped up onto his hind legs, and peered intently into the distance, as if he had spotted a predator. The pursuant baboon then also halted and stared in the same direction, giving up the chase. Was the pursued animal pretending that there was danger, to bring the chase to a halt? Such examples certainly make it seem as though the animal acted as if something is the case when it knows that it is not. In Whiten and Byrne’s survey, complex tactical deception was particularly frequent in great apes. Only gorillas and chimpanzees, for instance, produced cases in which an individual spotted a half-hidden favorite food item but then ignored it until a competitor disappeared from the scene and it was safe to secure the treat. There were even accounts of counter deception, where a competitor would not completely leave the scene but return quickly to snatch the food from the would-be deceiver. It is clear that primates’ maneuvering in their social worlds requires some smarts, so it is tempting to conclude that great ape minds are capable of pretending.

  A killjoy, however, would remind us that these are all merely anecdotes. We simply cannot be sure what is driving any of these overt behaviors. Perhaps the apparent counterdeception was mere coincidence. The ape came back and was lucky to stumble across the other ape as he was attempting to retrieve a choice food item. Similarly, the baboon that was chased might have actually thought, mistakenly, that there was a predator in the distance. Or it might have been in a similar situation in the past and subsequently learned that stopping suddenly and staring in the distance could lead to the end of an annoying chase. Neither of these options requires that the primate pretended that something she knows is not true is actually the case. As with Clever Hans, behavior can look sophisticated and yet be driven by simpler means.

  There are some famous reports of pretense from great apes held in human captivity. Researchers raising young great apes have offered rich interpretations of their charges playing with dolls or toy animals as if they were real things. For instance, Sue Savage-Rumbaugh has written that chimpanzees bathe dolls and sometimes make a toy animal “bite” others. The gorilla Koko has chased people around the room with a rubber snake. To cite a personal anecdote, I once brought a small rubber crocodile to the two chimpanzees, Ockie and Cassie, we sometimes test at Rockhampton zoo. They were intrigued, and when I pretended that the crocodile was trying to bite my co-investigator Emma Collier-Baker, they quickly jumped to her side and appeared to play along—even to the extent of reassuring her when the “danger” had passed. At least, that is how I remember the event, although frankly I cannot be sure whether any pretense took place from the chimpanzees’ perspective.

  The most often-cited case comes from the chimpanzee Viki, who is said to have repeatedly pulled an imaginary toy. Sometimes Viki acted as if the pretend cord had tangled around an obstacle and then placed one fist above the other, seemingly strained backwards repeatedly until, with a little jerk, she walked on as if she had loosened the stuck imaginary toy. She would then continue on her way, trailing the pretend pull toy behind. At least this is what it looked like to the human observer. This anecdote is perhaps stronger behavioral evidence because the ape seems to follow through with the logical implications of the imaginary events. Children do this a lot. For example, if an empty glass filled with imaginary juice is spilled, they then proceed to pretend to clean up the resulting imaginary mess on the floor. I know of only one other incident, reported at a conference, of a chimpanzee possibly following through with such pretend implications, in a case involving play with what appeared to be imaginary building blocks.

  Human children, of course, use words to elaborate and share their pretend play. The gorilla Koko, having been trained in human sign language (more on that in the next chapter), has been documented to use appropriate signs in her apparent pretend games. In response to being given an ape doll, for instance, she hugged it and then signed “drink.” She then made the doll sign “drink” by taking its thumb to its mouth. Again, though, there remain obvious difficulties with interpretation. Did the gorilla intend to make the doll sign “drink,” or did she merely put its thumb into its mouth?

  What are we to conclude? Can captive great apes really be said to engage in pretend play, or are they merely mimicking human behavior they see around them? A romantic perspective might point out that we would probably be content to accept many of the reported anecdotes as evidence for pretend play had we seen them in a child. A killjoy perspective would question what part of the alleged pretense is in the mind of the ape and what is in the mind of the human observer. Clearly, then, evidence of pretense, whether in play or in deception, remains tenuous.

  A neutral observer might highlight that neither extreme view should overlook inconvenient facts. On the one hand, although there are reports of pretend play in great apes, the list is relatively short, and even if the reports are accepted at face value, great apes do not show anywhere near the sophistication or amount of pretend play that human children do. Children often put their heads together in pretense, creating one scenario after another. On the other hand, while members of all great ape species have been repeatedly recorded performing apparent pretend activities, monkeys and other animals, even if reared in similarly intimate human contexts, have not. As it stands, the current data merely raise the possibility that our closest animal relatives have some ability to entertain alternative worlds in their minds. To unequivocally establish that great apes can think about things other than what they directly perceive, we need stronger evidence. We need carefully controlled experiments.

  IN RECENT YEARS EXPERIMENTAL APPROACHES for animals have increasingly been adapted from human developmental psychology. Most children acquire their sophisticated mental capacities in a predictable manner, and developmental psychologists have created nonverbal tests to identify the steps involved. Performance on these tests often predicts later capacities and performance on other measures (including on verbal tasks that cannot be adopted to test animals). The emergence of pretend play tends to coincide with the development of various other capacities, such as reasoning about hidden objects and recognizing one’s image in mirrors. Systematic experimentation has bolstered the case for imagination in great apes with research from these domains.

  The Swiss child psychologist Jean Piaget, who produced what is still the most comprehensive and influential theory of how human minds develo
p, constructed a progressive series of simple search tasks to measure children’s growing capacity to reason about objects that they no longer directly perceive.

  For young infants, “out of sight” quickly turns to “out of mind.” A brief distraction is typically enough to stop them from searching for something they had previously pursued. Children seem to only gradually acquire an understanding of the fact that things exist independent of whether they can still perceive them or not. Even adult philosophers wonder whether a tree falling in the woods really does make a sound when no one is there to perceive it.6 Philosophers aside, adults generally trust that people and objects continue to exist regardless of whether someone senses them or not.

  Piaget called this “object permanence” and proposed developmental stages through which infants progressively pass over the first two years of life.7 At around twelve months of age infants can reliably find a target object if they witness it being hidden beneath one of several boxes. They remember where it is and so have achieved what Piaget called stage 5 object permanence. However, when the target is first put in a small container, and this container is then used to move it somewhere else (so-called invisible or hidden displacement), they are at a loss. Imagine putting a small toy in your hand, closing it, and then putting it in the pocket of your jacket. You then pull out your hand and show your toddler that the hand is now empty. It is obvious that, unless there is some trickery, the object must now be in the pocket. Yet young infants and many animal species do not know where to look for the object. As long as there are no other clues, such as the smell of the treat or someone pointing to the pocket, they are at a loss as to where the toy has gone. They either do not remember the past situation or cannot relate this memory to the current situation to infer the solution. Toddlers only pass such tasks from around age eighteen to twenty-four months. This is stage 6 object permanence and demonstrates reasoning about the movement of the target that could not be directly perceived.

  Because Piaget’s search tasks are nonverbal and relatively easy to adapt to the needs of different species, they have become among the most widely used comparative tests ever. All you need are some boxes and a cache of desirable treats. Many species have been studied over the years, and several species have demonstrated stage 5 object permanence. Species that have formally passed this test include cats, chimpanzees, dogs, dolphins, gorillas, magpies, orangutans, parrots, and various species of monkeys. However, the sixth stage, assessed with the invisible displacement task and careful controls, has only been passed by a select group. Monkeys typically fail it, whereas members of all the great ape genera have passed these tasks repeatedly.

  You may want to try out the invisible displacement task on your pets. In fact, domestic dogs were one of the few species that had been reported to have passed these tasks. However, be careful. When our research group had a closer look, we found that dogs had “cheated.” They can find the target only under specific conditions. For example, they consistently perform well when the container (the displacement device) is put next to the hiding place, but fail when this cue is not available. In other words, they would only pass if you, after having put your hand in one of, say, four pockets to hide a treat, placed your empty hand next to the pocket where the food is. This suggests that the dogs simply learned this association and relied on this cue in their search. When we set up the experiment so that rules such as “search next to the displacement device” did not lead to success, dogs performed randomly. When we gave these carefully controlled versions of the task to chimpanzees and twenty-four-month-old children, on the other hand, they continued to pass consistently.8

  FIGURE 3.1.

  Test apparatus for standard displacement task; the chimpanzee Ockie selecting a box.

  Indeed, chimpanzees can find objects even if you invisibly displace them more than once. Imagine you put the treat in your hand and your closed hand then goes in and out of two of several available pockets before emerging empty. The target can thus logically be expected to be in either of the visited locations, but not in the pockets you did not put your hand in. In a formal test of this sort, subjects are given a second choice if they select one of the visited hiding places and it happens to be empty. When Emma Collier-Baker and I gave such a task to chimpanzees, they selected as their second choice the logical alternative hiding place rather than one of the other options, passing all of Piaget’s object permanence tasks (this final one is called 6b).

  To confirm that the task really measures what we intend it to measure, we also tested children with these same object permanence tasks, and, as expected, they solve it by age two. Emma also has tested orangutans and gorillas, and these data suggest that they too are able to pass all these tasks. Like human two-year-olds, but unlike other primates, great apes have demonstrated that they can think about things they did not perceive. These results are in line with several other problem-solving experiments that have been done by the comparative psychologist Josep Call and others in recent years. For instance, great apes can work out the location of hidden food from clues about where it is not.

  EVEN WHEN WE HAVE ESTABLISHED that an animal can reliably solve certain problems, there can be disagreement about what that behavior entails in terms of mental capacities. Take research on mirror self-recognition. Adult humans commonly spend considerable time looking in mirrors and fussing with their appearance. Just think of the multibillion-dollar cosmetic industry. Although a lot of animal species adjust their posture and appearance to suit different situations, such as puffing up to increase their size to predators, it is not clear if they are actually aware of what they look like. When our pets see themselves in mirrors, be they cats, dogs, fish, lizards, or birds, they do not stop and use their reflections to groom themselves. Instead, even when they can learn to use a mirror to find or avoid things, they seem confused about their own reflection, often treating it as if it were another individual or ignoring it altogether. Cats sometimes check behind the mirror, just to be sure. Even monkeys do not seem to get it. Great apes, on the other hand, can show interest in their image and may use reflective surfaces to investigate parts of their body that are otherwise out of sight, such as facial skin aberrations or their nether regions.

  Darwin briefly described the reactions of apes and children to their mirror image, but it was not until 1970, when Gordon Gallup developed an objective test for mirror self-recognition, that research moved from collections of anecdotes to systematic experimentation. Gallup anaesthetized chimpanzees and placed an odorless red mark on their face. Upon recovery he presented them with a mirror and found that the chimpanzees used it to inspect the unusual marks with their hands. He concluded that they must hence recognize their image. This experiment has been replicated many times since, though typically using surreptitious marking rather than anaesthetization. Orangutans and gorillas have also passed the task repeatedly.9 Versions of this simple test have been widely used with a variety of other animals as well as with children.

  This is a great little game to play with toddlers. You can wipe rouge on children’s faces under the pretense of cleaning, or you can surreptitiously place a large sticker in their hair in the process of patting their head. Wait a little to ensure that the child has not noticed the mark directly and then present a mirror. Though they are typically interested in their reflections, you will find that toddlers younger than about fifteen months do not touch the mark on their face. They will be surprised when you take the sticker out of their hair even if they have studied the sticker’s reflection extensively in the mirror. Only from the middle of the second year onwards do toddlers investigate their own head upon seeing their reflection. By twenty-four months close to all children retrieve the mark instantly.10 They know they are seeing themselves and may also start to talk about that fact. A recent direct comparison of human and chimpanzee infants found that they develop the capacity to pass Gallup’s task at comparable ages.

  FIGURE 3.2.

  The chimpanzee Cassie looking at his m
irror image.

  If you try to do the same task with your pets, however, you will find that they do not catch on. They do not use reflections to investigate changes to their appearance; so, like many other species that have been tested, they fail the mirror mark test. Monkeys such as baboons, capuchins, and macaques all fail the task even after hours of exposure. Yet there have been occasional headline-grabbing claims that animals other than the great apes recognize their mirror image. In the first instance, the influential behaviorist B. F. Skinner (famous for his work on the effects of reward and punishment he studied using the boxes that bear his name) and his colleagues conditioned pigeons to peck on their body in response to mirrors. Pigeons displayed this self-pecking only after extensive reinforcement over hundreds of trials11 and do not spontaneously use mirrors to explore their bodies.

  Much media coverage followed the announcement that two dolphins demonstrated mirror self-recognition. Note, however, that because they lack hands with which to touch the mark, one cannot give dolphins the standard test. It’s true that the dolphins in this study spent more time at the mirror when marked than when not visibly marked. Yet does this behavior demonstrate mirror-guided self-exploration? It would not be too surprising if dolphins were able to recognize themselves; after all, they have big brains and probably see their own reflection more often than any other mammals do, as they frequently jump out of the water. However, the current evidence for dolphin mirror self-recognition is not watertight and at minimum requires replication. Replication is also a problem for the final two claims. Two magpies and one elephant were recently reported to have passed the task. In another study, however, all elephants failed. So far we only have strong evidence, replicated in different laboratories, from experiments with members of the great ape genera. They can examine themselves in mirrors like humans do.12