Are We Smart Enough to Know How Smart Animals Are

by Frans de Waal

  What irony: we have traveled all the way from the absence of language as an argument against thought in other species to the position that the manifest thinking by nonlinguistic creatures argues against the importance of language. While I won’t complain about this turn of events, it owes a great debt to language studies on animals such as Alex: not so much because these studies demonstrated language per se but because they helped expose animal thought in a format that we easily relate to. We see a sharp-looking bird, who replies when spoken to, pronouncing object names with great accuracy. He faces a tray full of objects, some made of wool, some of wood, some of plastic, representing all colors of the rainbow. He is invited to feel every object with his beak and tongue, and then, after they have all been returned to the tray, he is asked what the two-cornered blue object is made of. By correctly answering “wool,” he combines his knowledge of color, shape, and material with his memory of what this particular item felt like. Or he sees two keys, one made of green plastic, the other of metal, and is asked “what is different?” He says “color.” Asked “which color bigger,” he answers “green.”8

  Anyone watching Alex perform, as I did in the early stages of his career, is blown away. Obviously, skeptics tried to ascribe his skills to rote learning, but since the stimuli changed all the time as did the questions asked, it is hard to see how he could have performed at this level based on stock answers. He would have needed a gigantic memory to handle all possibilities, so much so that it is in fact simpler to assume, as Irene did, that he had acquired a few basic concepts and was capable of mentally combining them. Furthermore, he didn’t need Irene’s presence to answer, nor did he even need to see the actual items. In the absence of any corn, he might be asked what color corn is and would say “yellow.” Particularly impressive was Alex’s ability to distinguish “same” from “different,” which required him to compare objects on a variety of dimensions. All these capacities—labeling, comparing, and judging color, shape, and material—were assumed to require language at the time that Alex began his training. It was an aggravating struggle for Irene to convince the world of his skills, especially since skepticism with respect to birds ran so much deeper than it ever was for our close relatives, the primates. After years of persistence and solid data, however, she had the satisfaction of seeing Alex turn into a celebrity. Upon his death in 2007, he was honored with obituaries in both The New York Times and The Economist.

  In the meantime some of his relatives had begun to impress as well. Another African gray not only mimicked sounds but added accompanying body movements. He’d say “Ciao” while waving goodbye with a foot or wing, or say “Look at my tongue” while sticking out his tongue, just as his owner had shown him. It remained a puzzle how a bird was able to draw such parallels between the human body and its own.9 Then there was Figaro, a Goffin’s cockatoo who was seen breaking off large splinters from a wooden beam in order to rake in nuts placed outside his aviary. Before Figaro, there had been no reports of toolmaking parrots.10 It makes me wonder if Kohts ever conducted similar experiments on her cockatoo, macaw, and ara. Given her keen interest in tools and her six untranslated books, I wouldn’t be surprised to hear about it one day. There is obviously still much to discover, as also became clear from tests of Alex’s counting abilities.

  Alex’s talents were accidentally revealed while researchers were testing Griffin —a parrot named after Donald Griffin—who was staying in the same room with him. In order to see if Griffin could pair quantities with sounds, they would click twice, to which the right answer would be “two.” But when Griffin failed to answer and got two more clicks, Alex, from across the room, chimed in with “four.” And after two more clicks, Alex said “six,” while Griffin remained mute.11 Alex was familiar with numbers and could correctly answer the question “what number is green?” after having seen a tray with many objects, including several green ones. But now he was doing addition, and more than that: he was doing it without visual input. Again, adding up numbers was once thought to be language-dependent, but this claim had already begun to wobble a few years back when a chimpanzee succeeded at it.12

  Irene set out to test Alex’s capacities more systematically by placing a few differently sized items (such as pasta pieces) under a cup. She’d lift the cup up for a few seconds in front of Alex, then put it down again. After this she would do the same for a second cup, then a third. The number of items under each cup was small, and sometimes there were none. After this, with only the three cups visible, Alex would be asked “how many total?” Out of ten tests, Alex mentioned the correct total eight times. The two that he missed, he got right the second time he heard the question.13 And all this in his head, because he couldn’t see the actual items.

  Unfortunately, this study was broken off by Alex’s unexpected death. But by then this diminutive mathematical genius in a grey suit had given us ample evidence that there is more knocking around in a bird’s skull than anyone had suspected. Irene concluded that “for far too long, animals in general, and birds in particular, have been denigrated and treated merely as creatures of instinct rather than as sentient beings.”14

  Red Herring

  At times, Alex’s talking made perfect linguistic sense. For example, once when Irene was fuming about a meeting in her department and walked to the lab with angry steps, Alex told her “Calm down!” No doubt the same expression had in the past been aimed at Alex’s own excitable self. Other famous cases include Koko, the sign-language gorilla spontaneously combining the signs for “white” and “tiger” upon seeing a zebra, and Washoe, the chimpanzee pioneer of this entire field, labeling a swan a “water bird.”

  I am prepared to interpret this as a hint of deeper knowledge, but only after I see more evidence than we have today. It is good to keep in mind that these animals produce hundreds of signs every day and have been studied for decades. We’d need to know more about the ratio between hits and misses among the thousands of utterances recorded. How are these fortuitous combinations different from, say, Paul the octopus (nicknamed Pulpo Paul) who rose to fame after a string of correct predictions during the 2010 World Cup? In the same way that no one assumes that Paul knew much about soccer—he was just a lucky mollusk—we need to compare striking animal utterances with the probability of them coming about by chance. It is hard to evaluate linguistic skills if we never get to see the raw data, such as unedited videotapes, and hear only cherry-picked interpretations by loving caretakers. It also doesn’t help that whenever apes produce wrong answers, their interpreters assume that they have a sense of humor, exclaiming “Oh, stop kidding around!” or “You funny gorilla!”15

  Upon the death of Robin Williams, in 2014, when the whole world was grieving one of the world’s funniest men, Koko was said to be mourning, too. It sounded plausible, especially since the Gorilla Foundation, in California, called Williams one of her “closest friends.” The problem is that the two of them had met just once, thirteen years before, and that the only evidence of Koko’s “somber” reaction was a photo of her sitting with her head down and eyes closed, which was hard to distinguish from a dozing ape. I found the grieving claim to be a huge stretch, not because I doubt that apes have feelings or can grieve, but because it is nearly impossible to gauge an animal’s reaction to an event it has not witnessed. While it is entirely possible that Koko’s mood was affected by the people around her, this is not the same as grasping what had happened to a member of our species whom she barely knew.

  All responses to death and loss thus far observed in apes concern individuals who were truly close (such as mother and offspring, or lifelong friends) and whose corpses the apes were able to see and touch. Mourning triggered by the mere mention of someone’s death requires a level of imagination and understanding of mortality that most of us don’t assume. It is precisely because of such inflated claims that the whole field of talking apes has fallen into ill repute over the years, and why no new projects of its kind are being initiated. Those that still
do exist tend to resort to feel-good stories and publicity stunts to raise funds. There is too much of this going around, and too little hard-nosed science.

  You won’t often hear me say something like this, but I consider us the only linguistic species. We honestly have no evidence for symbolic communication, equally rich and multifunctional as ours, outside our species. It seems to be our own magic well, something we are exceptionally good at. Other species are very capable of communicating inner processes, such as emotions and intentions, or coordinating actions and plans by means of nonverbal signals, but their communication is neither symbolized nor endlessly flexible like language. For one thing, it is almost entirely restricted to the here and now. A chimpanzee may detect another’s emotions in reaction to a particular ongoing situation, but cannot communicate even the simplest information about events displaced in space and time. If I have a black eye, I can explain to you how yesterday I walked into a bar with drunken people … and so on. A chimpanzee has no way, after the fact, to explain how an injury came about. Possibly, if his assailant happens to walk by and he barks and screams at him, others will be able to deduce the connection between his behavior and the injury—apes are smart enough to put cause and effect together—but this would work only in the other’s presence. If his assailant never walks by, there will be no such information transfer.

  Countless theories have attempted to identify the benefits that language bestows upon our species and to explain why language may have arisen. In fact, an entire biennial international conference is devoted to exactly this topic, where speakers present more speculations and evolutionary scenarios than you can imagine.16 I myself take the rather simple view that the first and foremost advantage of language is to transmit information that transcends the here and now. There is great survival value in communication about things that are absent or events that have happened or are about to happen. You can let others know that there is a lion over the hill, or that your neighbors have picked up weapons. This is just one idea out of many, though, and it is true that modern languages are far too complex and elaborate for this limited purpose. They are sophisticated enough to express thoughts and feelings, convey knowledge, develop philosophies, and write poetry and fiction. What an incredibly rich capacity it is: one that seems entirely our own.

  But as with so many larger human phenomena, once we break it down into smaller pieces, some of these pieces can be found elsewhere. It is a procedure I have applied myself in my popular books about primate politics, culture, even morality.17 Critical pieces such as power alliances (politics) and the spreading of habits (culture), as well as empathy and fairness (morality), are detectable outside our species. The same holds for capacities underlying language. Honeybees, for example, accurately signal distant nectar locations to the hive, and monkeys may utter calls in predictable sequences that resemble rudimentary syntax. The most intriguing parallel is perhaps referential signaling. Vervet monkeys on the plains of Kenya have distinct alarm calls for a leopard, eagle, or snake. These predator-specific calls constitute a life-saving communication system, because different dangers demand different responses. For example, the right response to a snake alarm is to stand upright in the tall grass and look around, which would be suicidal in case a leopard lurks in the grass.18 Instead of having special calls, some other monkey species combine the same calls in different ways under different circumstances.19

  After the primate studies, the usual rippling has added birds to the list of referential signalers. Great tits, for example, have a unique call for snakes, which pose a grave threat as they slither into nests to swallow the young.20 But whereas these kinds of studies have helped raise the profile of animal communication, some serious doubts have been raised, too, and language parallels have been called a “red herring.”21 Animal calls do not necessarily mean what we think they mean: a critical part of how they function is how listeners interpret them.22 On top of this, it is good to keep in mind that most animals do not learn their calls the way humans learn words. They are simply born with them. However sophisticated natural animal communication may be, it lacks the symbolic quality and open-ended syntax that lends human language its infinite versatility.

  Perhaps hand gestures offer a better parallel, since in the apes they are under voluntary control and often learned. Apes move and wave their hands all the time while communicating, and they have an impressive repertoire of specific gestures such as stretching out an open hand to beg for something, or moving a whole arm over another as a sign of dominance.23 We share this behavior with them and only them: monkeys have virtually no such gestures.24 The manual signals of apes are intentional, highly flexible, and used to refine the message of communication. When a chimp holds out his hand to a friend who is eating, he is asking for a share, but when the same chimp is under attack and holds out his hand to a bystander, he is asking for protection. He may even point out his opponent by making angry slapping gestures in his direction. But although gestures are more context-dependent than other signals and greatly enrich communication, comparisons with human language remain a stretch.

  Does this mean that all the attempts to find languagelike qualities in animal communication have been a waste of time, including training projects, such as those with Alex, Koko, Washoe, Kanzi, and others? After Terrace’s paper, linguists eager to rid their territory of hairy or feathered “intruders” made the fruitlessness of animal research their mantra. They were so contemptuous of it that, at a 1980 conference—the title of which contained the words Clever Hans—they called for an official ban on any and all attempts to teach animals language.25 This unsuccessful move was reminiscent of nineteenth-century anti-Darwinists for whom language was the one barrier between brute and man, including the Linguistic Society of Paris, which in 1866 forbade the study of language origins.26 Such measures reflect intellectual fear rather than curiosity. What are linguists afraid of? They had better pull their heads out of the sand, because no trait, not even our beloved linguistic ability, ever comes about de novo. Nothing evolves all of a sudden, without antecedents. Every new trait taps into existing structures and processes. Thus, Wernicke’s area, a part of the brain central to human speech, is recognizable in the great apes, in which it is enlarged on the left side, as it is in us.27 This obviously raises the question of what this particular brain region was doing in our ancestors before it was recruited for language. There are many such connections, including the FoxP2 gene that affects both human articulated speech and the fine motor control of birdsong.28 Science increasingly views human speech and birdsong as products of convergent evolution, given that songbirds and humans share at least fifty genes specifically related to vocal learning.29 No one serious about language evolution will ever be able to get around animal comparisons.

  In the meantime, language-inspired studies have dispelled the notion that natural animal communication is purely emotional. We now have a far better grasp of how communication is geared to an audience, provides information about the environment, and relies on interpretation by those receiving the signals. Even if the connection with human language remains contentious, our appreciation of animal communication has greatly benefited from this flurry of research. As for the handful of language-trained animals, they have proven invaluable at showing what their minds are capable of. Since these animals respond to requests and prompts in a way that we find easy to interpret, the results speak to the human imagination and have been instrumental in breaking open the field of animal cognition. When Alex hears a question about the items on his tray, he inspects them carefully and comments on the one that he was asked about. We have no trouble putting ourselves into his shoes, given that we understand both the question and his answer.

  I once asked Sue Savage-Rumbaugh, who worked with Kanzi, the bonobo who communicates by pressing symbols on a keyboard, “Would you say that you study language or intelligence, or is there no difference?” She replied:

  There is a difference because we have apes who have no linguistic a
bilities in the human sense, but who do quite well on cognitive tasks such as solving a maze problem. Language skills can help elaborate and refine cognitive skills, though, because you can tell an ape who is language-trained something that he does not know. This can put a cognitive task on a whole different plane. For example, we have a computer game in which apes put three puzzle pieces together to make different portraits. After having learned this, they get four pieces presented on the screen, and the fourth piece is from a different portrait. When we first did this with Kanzi, he would take the piece of a bunny face and put it together with a piece of my face. He kept trying, but of course it wouldn’t fit. Since he understands spoken language so well, I could say to him: “Kanzi, we’re not making the bunny, put Sue’s face together.” As soon as he heard this, he stopped making the bunny, and stuck to the pieces of my face. So, the instructions had an immediate effect.30

  Since Kanzi lived for years in Atlanta, I met him multiple times and was always impressed by how well he grasped spoken English. What struck me was not his self-produced utterances—which were rather basic, certainly below the level of a three-year-old child—but the way he reacted to those by the people around him. In one videotaped exchange, Sue asks him “Put the key in the refrigerator,” while she wears a welding mask to prevent Clever Hans Effects. Kanzi picks up a chain of keys, opens the fridge, and puts the keys into it. Asked to give his doggy a shot, he picks up a plastic syringe and injects it into his stuffed toy dog. Kanzi’s passive comprehension is greatly helped by his familiarity with a large number of items and words. This has been tested by playing spoken words to him through headphones while he sits at a table and selects a picture of the object that he hears being mentioned. But that he is excellent at word recognition still doesn’t explain why Kanzi appears to understand entire sentences.