Are We Smart Enough to Know How Smart Animals Are

by Frans de Waal

  Animal research in this area began perhaps with the uncertainty response noticed by Tolman in the 1920s. His rats seemed to hesitate before a difficult task as reflected in their ‘‘lookings or runnings back and forth.”36 This was most remarkable, since at the time animals were thought to simply respond to stimuli. Absent an inner life, why be in turmoil about a decision? Decades later the American psychologist David Smith gave a bottlenose dolphin the task to tell the difference between high and low tones. The dolphin was an eighteen-year-old male named Natua, in a pool at the Dolphin Research Center in Florida. As in Tolman’s rats, Natua’s level of confidence was quite manifest. He swam at different speeds toward the response, depending on how easy or hard it was to tell both tones apart. When they were very different, the dolphin arrived with such speed that his bow wave threatened to soak the electronics of the apparatus. They had to be covered with plastic. If the tones were similar, though, Natua slowed down, waggled his head, and wavered between the two paddles that he needed to touch in order to indicate a high or low sound. He didn’t know which one to pick. Smith decided to make a study of Natua’s uncertainty, mindful of Tolman’s suggestion that it might reflect consciousness. The investigator created a way for the animal to opt out. A third paddle was added, which Natua could touch if he wanted a fresh trial with an easier distinction. The tougher the choice, the more Natua went for the third paddle, apparently realizing when he had trouble coming up with the right answer. Thus the field of animal metacognition was born.37

  Investigators have essentially followed two approaches. One is to explore the uncertainty response, as in the dolphin study, while the other is to see if animals realize when they need more information. The first approach has been successful with rats and macaques. Robert Hampton, now a colleague at Emory University, gave monkeys a memory task on a touchscreen. They would first see one particular image, say a pink flower, then face a delay before being presented with several pictures, including the pink flower. The delay varied in length. Before each test, the monkeys had the choice to either take it or decline it. If they took the test and correctly touched the pink flower, they gained a peanut. But if they declined, they only got a monkey pellet, a boring everyday food. The longer the delay, the more the monkeys declined taking the test despite its better reward. They seemed to realize that their memory had faded. Occasionally, they were forced to take a trial without a chance of escape. In those cases they fared rather poorly. In other words, they opted out for a reason, doing so when they couldn’t count on their memory.38 A similar test with rats gave similar results: the rats performed best on tests that they had deliberately chosen to take.39 In other words, both macaques and rats volunteer for tests only when they feel confident, suggesting that they know their own knowledge.

  A rhesus macaque knows that food has been hidden in one of four tubes, but he has no idea which one. He is not allowed to try every tube and will get only one pick. By bending down to first peek into the tubes, he demonstrates that he knows he doesn’t know, which is a sign of metacognition.

  The second approach concerns information seeking. For example, jays placed at peepholes were given an opportunity to watch food—waxworms—being hidden before they were allowed to enter the area to find it. They could look through one peephole to see an experimenter put a waxworm in one of four open cups, or they could look through another to see another experimenter with three covered cups plus one open one. In the second case, it was obvious where the worm would end up. Before entering the area to find the worm, the birds spent more time watching the first experimenter. They seemed to realize that this was the information they needed most.40

  In monkeys and apes, the same sort of test has been done by having them watch an experimenter hide food in one of several horizontal pipes. Obviously, the primates remembered where he had put the food and confidently selected the correct pipe. If the food hiding had taken place in secret, however, they were not sure which pipe to pick. They peeked into the pipes, bending down to get a good look, before selecting one. They realized that they needed more information to succeed.41

  As a result of these studies, some animals are now believed to track their own knowledge and to realize when it is deficient. It all fits Tolman’s insistence that animals are active processors of the cues around them, with beliefs, expectations, perhaps even consciousness. This viewpoint being on the rise, I asked my colleague Rob Hampton about the state of affairs in this field. The two of us have offices on the same floor of Emory’s psychology department. While sitting in mine, we first watched the video of Lisala carrying her huge rock. Like a real scientist, Rob immediately began to imagine how to turn this situation into a controlled experiment by varying the locations of the nuts and the tools, even though for me the beauty of the whole sequence was Lisala’s spontaneity. We had nothing to do with it. Rob was impressed.

  I asked him if his work on metacognition had been inspired by the dolphin study, but he rather saw this as a case of convergent interests. The dolphin study did come out first, but it wasn’t about memory, which was Rob’s focus. He was inspired by the ideas of Alastair Inman, a postdoc in Sara Shettleworth’s Toronto lab, where Rob worked at the time. Alastair wondered about the cost of memorizing things. What is the price of holding information in mind? He set up an experiment on pigeon memory that was similar to the metacognition test for monkeys that Rob developed.42

  When I asked what he thought of people who draw a sharp line between humans and other animals, such as Endel Tulving’s shifting definitions, Rob exclaimed: “Tulving! He loves to do that. He has done a great service to the animal research community.” Tulving says those things, Rob believes, because he thinks it’s fun to set a high bar. He knows that others will go after it, so he pushes them to come up with clever experiments. In his first monkey paper, Rob thanked Tulving for his “incitement.” Meeting the senior scientist not long thereafter at a conference, Tulving told Rob, “I have seen what you wrote, thank you!”

  For Rob, the big question in relation to consciousness is why we actually need it. What is it good for? After all, there are lots of things we can do unconsciously. For example, amnesic patients are able to learn without knowing what they have learned. They may learn to make inverse drawings guided by a mirror. They acquire the hand-eye coordination at about the same rate as any other person, but every time you test them, they’ll tell you that they’ve never done it before. It is all new to them. In their behavior, though, it is obvious that they have experience with the task and have acquired the required skill.

  While consciousness has evolved at least once, it is unclear why and under which conditions. Rob considers it such a messy word that he is reluctant to use it. He adds, “Anyone who thinks they have solved the problem of consciousness hasn’t been thinking about it carefully enough.”

  Consciousness

  When in 2012 a group of prominent scientists came out with The Cambridge Declaration on Consciousness, I was skeptical.43 The media described it as asserting once and for all that nonhuman animals are conscious beings. Like most scientists studying animal behavior, I really don’t know what to say to his. Given how ill-defined consciousness is, it is not something we can affirm by majority vote or by people saying “Of course, they are conscious—I can see it in their eyes.” Subjective feelings won’t get us there. Science goes by hard evidence.

  But in reading the actual declaration, I calmed down, because it is a reasonable document. It doesn’t actually claim animal consciousness, whatever that is. It only says that given the similarities in behavior and nervous systems between humans and other large-brained species, there is no reason to cling to the notion that only humans are conscious. As the document puts it, “The weight of evidence indicates that humans are not unique in possessing the neurological substrates that generate consciousness.” I can live with that. As you can see from this chapter, there is sound evidence that mental processes associated with consciousness in humans, such as how we relate to the
past and future, occur in other species as well. Strictly speaking, this doesn’t prove consciousness, but science is increasingly favoring continuity over discontinuity. This is certainly true for comparisons between humans and other primates, but extends to other mammals and birds, especially since bird brains turn out to resemble those of mammals more than previously thought. All vertebrate brains are homologous.

  Although we cannot directly measure consciousness, other species show evidence of having precisely those capacities traditionally viewed as its indicators. To maintain that they possess these capacities in the absence of consciousness introduces an unnecessary dichotomy. It suggests that they do what we do but in fundamentally different ways. From an evolutionary standpoint, this sounds illogical. And logic is one of those other capacities we pride ourselves on.

  8 OF MIRRORS AND JARS

  Pepsi was the star of a recent study on Asian elephants. The adolescent bull passed a mirror test conducted by Joshua Plotnik by carefully touching a large white X that had been painted on the left-hand side of his forehead. He never paid attention to the X that had been put with invisible paint on the right-hand side; nor did he touch the white one until he walked up to the mirror in the middle of a meadow. The next day we reversed the sides of the visible and invisible markings, and Pepsi again specifically felt the white X. He rubbed off some of the paint with the tip of his trunk and brought it to his mouth, tasting it. Since he could know its location only via his reflection, he must have connected his mirror image with himself. As if to make the point that the mark test isn’t the only way to do so, Pepsi took one step back at the end of testing to open his mouth wide. With the mirror’s help, he peered deeply inside. This move, also common in apes, makes perfect sense given that one never gets to see one’s own tongue and teeth without a mirror.1

  Years later Pepsi towered over me as a nearly adult male. He was very gentle, though, lifting me up and putting me down on the orders of his mahout. Revisiting Thailand to see the camp in the Golden Triangle, where the Think Elephants International Foundation conducts its research, I met Josh’s team of enthusiastic young assistants. Every day they invite a couple of elephants to their experiments. With a mahout sitting high up on their neck, the colossal animals lumber to the testing site on the jungle’s edge. After the mahout gets off to squat in the background, the elephant performs a few simple tasks. She feels an object with her trunk, after which she is asked to pick a matching one from among several; or she stretches her trunk to smell the difference between two buckets depending on what the students put into them.2

  A marked Asian elephant in front of a mirror. The mark test requires an individual to connect her reflection with her own body, resulting in inspection of the mark. Only a handful of species pass this test spontaneously.

  Everyone knows that elephants are smart, but there is an enormous scarcity of data similar to those for primates, corvids, dogs, rats, dolphins, and so on. All we have for the elephant is spontaneous behavior, which doesn’t allow for the precision and controls that science desires. Discrimination tasks like the ones I witnessed are an excellent starting point. But even if the pachyderm mind may be the next frontier in evolutionary cognition, it is a most challenging one given that the elephant is probably the only land animal never to be seen alive on a university campus or in a conventional lab. While science’s preference for easy-to-keep species is understandable, it has its limits. It has given us a small-brain perspective on animal cognition, one that we have had trouble shedding.

  Elephants Listening

  Southeast Asians have a long-standing cultural relation with elephants. For thousands of years, these animals have carried out heavy forest work, transported royalty, and served in hunting and warfare. They have always remained wild, though. The species is not domesticated in the genetic sense, and free-ranging elephants still often sire the offspring of captive ones. Not surprisingly, elephants are less predictable than many domesticated animals. They can be hostile to people, occasionally killing a mahout or tourist, but many of them also form lifelong bonds with their caretakers. In one story, an elephant at the age of ten pulled her drowning mahout out of a lake after hearing his cries for help a kilometer away; in another, a fully grown bull would charge anyone who came close except the wife of the village elder, whom he would caress with his trunk. Young elephants grow so used to people that they learn how to fool them by stuffing a trunkful of grass into the wooden bells around their necks so as to muffle their sound. This way they can move about unnoticed.

  African elephants, in contrast, are rarely brought under human control. They live their own parallel lives, even though the massive ivory trade is now putting them in danger to the point that we face the dismaying prospect of permanently losing one of the world’s most beloved and charismatic animals. The elephant’s Umwelt being largely acoustic and olfactory, the protection of wild populations against poaching and conflict with humans requires methods that are not immediately obvious to our visual species. Studies focus on the extraordinary senses of these animals. One study, in arid Namibia, followed free-ranging elephants equipped with GPS collars. It discovered that these animals are aware of thunderstorms at enormous distances and adjust their travel routes to precipitation days before it actually arrives. How do they do this? Elephants can hear infrasound, which are sound waves far below the human hearing range. Also used in communication, these sounds travel over much longer distances than the ones we are able to discern.3 Is it possible that elephants can hear thunder and rainfall hundreds of miles away? It seems the only way to explain their behavior.

  But isn’t this just a matter of perception? Cognition and perception cannot be separated, though. They go hand in hand. As the father of cognitive psychology, Ulric Neisser, put it: “the world of experience is produced by the man who experiences it.”4 Since the late Neisser was a colleague of mine, I know that nonhuman minds were not his foremost interest, yet he refused to view animals as mere learning machines. The behaviorist program was ill suited to all species, he felt, not just ours. Instead, he emphasized perception and how it is turned into experience by picking and choosing what sensory input to pay attention to and how to process and organize it. Reality is a mental construct. This is what makes the elephant, the bat, the dolphin, the octopus, and the star-nosed mole so intriguing. They have senses that we either don’t have, or that we have in a much less developed form, making the way they relate to their environment impossible for us to fathom. They construct their own realities. We may attach less significance to these, simply because they are so alien, but they are obviously all-important to these animals. Even when they process information familiar to us, they may do so quite differently, such as when elephants tell human languages apart. This ability was first demonstrated in African elephants.

  In Amboseli National Park, in Kenya, the British ethologist Karen McComb studied elephant reactions to different human ethnic groups. The cattle-herding Maasai sometimes spear elephants in order to show their virility or to gain access to grazing grounds and water holes. Understandably, elephants flee the Maasai, who approach them in their characteristic red ochre robes, but they don’t avoid other people on foot.5 How do they recognize the Maasai? Instead of focusing on their color vision, McComb explored what is perhaps the elephant’s keenest sense: sound. She contrasted the Maasai with the Kamba people, who live in the same area but rarely interfere with elephants. From a concealed loudspeaker, McComb played human voices saying a single phrase in the language of either the Maasai or the Kamba: “Look, look over there, a group of elephants is coming.” It is hard to imagine that the precise words mattered, but the investigators compared the elephants’ reaction to the voices of adult men, adult women, and boys.

  Herds retreated and “bunched” together (forming a tight circle with calves in the middle) more often after playbacks of Maasai than of Kamba voices. Maasai male voices triggered more defensive reactions than those of Maasai women and boys. Even after the natural voic
es were acoustically transformed so as to make male voices sound more female, and vice versa, the outcome remained the same. The elephants were especially vigilant upon hearing the resynthesized voices of Maasai men. This was surprising because the pitch of these voices had now the opposite gender’s qualities. Possibly the elephants identified gender by other characteristics, such as the fact that female voices tend to be more melodious and “breathy” than those of males.6

  Experience played a role, because herds led by older matriarchs were more discriminating. The same difference was found in another study in which lion roaring was played from a speaker. Older matriarchs would charge the speaker, which is quite different from their hasty retreat from Maasai voices.7 Aggressive mobbing of men carrying spears is unlikely to pay off, yet driving off lions is something elephants are good at. Despite their size, these animals face other dangers, including very small ones, such as stinging bees. Elephants are vulnerable to stings around the eyes and up their trunks, and young elephants lack a thick enough skin to protect themselves against a mass attack. Elephants give deep rumbles as an alarm to both humans and bees, but the two sounds must differ because playbacks induce quite different responses. Upon hearing the bee-rumble from a speaker, for example, elephants flee with head-shaking movements that would knock insects away, a reaction not shown to the human-rumble.8