Monkey Business

Remember the great tumult during the Sixties about talking dolphins? Because a dolphin brain is larger than ours, could it be that porpoises are potentially as bright as we are, maybe more so? John C. Lilly seriously tried teaching English to these clever little whales and for a time actually believed he had taught dolphins to mimic human speech. Like the black races of Africa, Lilly once said, porpoises are on the brink of becoming Westernized, a revolution with unpredictable consequences. “If dolphins come to understand our cold war,” he warned, “we don’t know how they will proceed to operate.”

After Lilly became convinced that several of his Florida porpoises had committed suicide, he abandoned his watery research to wander off into the jungles of parapsychology and Eastern mysticism. He reported fantastic encounters with extraterrestrial intelligence. He told reporters that dolphins were using ESP to “infiltrate” human minds. Eventually it became clear to almost everybody, as it had been all along to “establishment” biologists, that Lilly’s research was hopelessly flawed, and that the whale mind, though wondrous and unique, is not much more so, if at all, than the mind of a pig or an elephant. The lovable dolphins swam away from the press and television, leaving in their wakes a batch of careless books and TV documentaries, and surely the worst movie (The Day of the Dolphin) ever directed by Mike Nichols.

As the dolphin flap faded, a new media enthusiasm began to gather momentum. At the University of Nevada, Allen and Beatrice Gardner succeeded in teaching ASL (American Sign Language) to an infant female chimpanzee named Washoe. For the first time in history, it was loudly proclaimed, a lower primate had mastered a language in which it could talk to humans.

“Talk” and “language” are, of course, fuzzy terms with wide spectrums of meaning. A bluejay “talks” to other birds when it warns them of a cat. A cat “talks” when it asks to be fed by rubbing against your calf. Dogs communicate by barking, growling, whimpering, wagging their tails, and leaving symbolic messages on fire hydrants. Even so, the world was astounded by Washoe’s ability to understand hundreds of sign gestures, especially by her ability to combine signs in ways that suggested a rudimentary grasp of grammar.

The best-known instance of Washoe inventing a phrase was when her teacher, Roger Fouts, had taken her out in a rowboat and a swan glided by. Fouts signed “What’s that?” Washoe, knowing the signs for water and bird, responded with “water bird.” There were many other two-word combinations mastered by Washoe: Washoe sorry, Roger tickle, you drink, and so on.

Other researchers soon were teaching other visual languages to young chimps. In California David Premack symbolized words with plastic tiles of different shapes and colors. His star pupil, Sarah, became almost as famous as Washoe. Like Washoe, Sarah seemed to create significant phrases. David’s wife, Ann, wrote a book called Why Chimps Can Read.

In Georgia, Duane Rumbaugh tried a new tack. He had a computer built with a console of keys bearing patterns that represented words. A chimpanzee named Lana was taught to speak in this computer language of “Yerkish,” named for the Yerkes Primate Center in Atlanta where Rumbaugh did his work. Lana, too, apparently combined signs in meaningful sequences. She called a cucumber a green banana. She called an orange an orange apple.

The achievements of Washoe, Sarah, and Lana have now been surpassed, so it is claimed, by the fabulous linguistic feats of Koko, a female gorilla trained since 1972 by a psychologist, Francine (“Penny”) Patterson, in Stanford. It is not hard to understand why Penny—young, pretty, with long blond hair—has received such enormous publicity. What could be more dramatic than color photographs of Beauty and the Beast, heads together, raptly chattering to one another? Patterson wrote a cover story (the cover photo of Koko was snapped by Koko) for National Geographic (October 1978) titled “Conversations with a Gorilla.” The pair graced the cover of The New York Times Magazine (June 12, 1977). In Koko, A Talking Gorilla, a stirring film documentary that opened last December in Manhattan, Koko does a fine job of acting like a gorilla, but otherwise the film is mostly flimflam.

There is another reason for Penny’s growing fame. Her claims for ape intelligence far exceed those of any other trainer. For one thing, Koko loves to make up rhymes: Squash wash, do blue, bear hair, and so on. (She has learned English vocalizations by hearing Penny repeat them, and by using a typewriter speech synthesizer designed by the Stanford mathematician Patrick Suppes.) Once Koko made up the poem: Flower pink, fruit stink—fruit pink stink. Here is a sampling of Koko’s skill in inventing clever metaphors: Elephant baby (for a Pinocchio doll), eye hat (mask), finger bracelet (ring), white tiger (zebra), fake mouth (nose).

A reporter asked Koko who she liked best, Penny or her assistant. According to Penny, Koko looked back and forth, then diplomatically signed, “Bad question.” On another occasion Penny asked, “What are you afraid of?” Koko: “Afraid alligator.” Koko had never seen a live alligator. Penny thinks this shows how researchers can learn new facts about apes now that they can ask them questions.

According to Eugene Linden (who wrote a popular book about talking apes), in a wildly laudatory article (“Talk to the Animals,” Omni, January 1980), when Koko was asked where you go when you die she signed, “Comfortable hole bye.” Once when Penny became exasperated by the number of toys Koko had broken Penny muttered, “Why can’t you be normal like any other kid?” Koko, says Linden, signed “Gorilla.”

From the beginning large numbers of experts on animal behavior have been deeply skeptical of these extraordinary claims, but their animadversions appeared only in technical journals. Now the secret is out. Two books have been published, one popular, one technical, that give a strong case for the view that apes do not comprehend sign sequences in any way essentially different from a dog’s understanding of such commands as “Sit up and shake hands” or “Go get the newspaper.”

Nowhere on the jacket of Nim or in the book’s advertising does the publisher so much as hint that the book severely criticizes practically all earlier work with talking apes. Even the author, Herbert Terrace, a psychologist at Columbia University, plays down his doubts at the start of his book, though there is a reason. When he began training Nim Chimpsky, a baby male chimp named in honor of Noam Chomsky, he had high hopes of confirming the earlier findings. His book is an informal narrative, with marvelous photographs, about four years that he and his many assistants spent in teaching ASL to Nim. Not until Chapter 13, after Nim has returned to his birthplace in Oklahoma, does Terrace see the light.

Terrace’s complete disenchantment did not descend until he began to study his own extensive videotapes. Here are some of the things he learned:

Nim rarely initiated signing. Ninety percent of his signing was in response to gestures by teachers.

Half of Nim’s signs imitated part or all of what a teacher had just signed. In many cases his teachers were astonished to see how often they had unconsciously started a sign that Nim had noticed.

If Nim wanted something he first grabbed, signing only when the grab failed. He never initiated signs except when expecting such rewards as food, hugs, and tickling.

Most of Nim’s phrases were random combinations of signs, usually involving me, hug, and Nim—signs that fitted with almost all other signs, and which he had learned were likely to elicit favorable reactions.

Unlike children when they start to talk, Nim constantly interrupted teachers. He never learned the two-way nature of conversation. Researchers have attributed such interruptions to an ape’s eagerness to talk.

Nim’s mistakes were more often the confusing of signs similar in form rather than similar in meaning.

When Nim began to extend sentences beyond two or three words he simply added a string of nonsense words, usually repeating earlier signs. For example: “Give orange me give eat orange me eat orange give me eat orange give me you.” This in contrast to the longer utterances of children which expand the sense of shorter ones.

Nim never signed to another chimpanzee who knew ASL unless a teacher was present to coax him.

Nim Chimpsky finally convinced Terrace that Noam Chomsky, the most distinguished of skeptical linguistics experts, was right. Although apes have a remarkable memory that enables them to master hundreds of visual signs, Terrace believes there is no evidence yet that they understand any kind of syntax. Of course this may be true also of very young children, but children quickly go on to form sentences that require a firm grasp of the rules of form. When an ape learns to put a few signs together there is no reason, says Terrace, to suppose it is doing anything essentially different from a pigeon that has been taught to obtain food by pecking four differently colored buttons in a specific order regardless of how the buttons are arranged.

When Terrace examined the videotapes of other researchers he found the same disturbing features. In many cases of film released for public viewing and for fund raising, episodes had been edited so that initial promptings were not seen. A Nova documentary called The First Signs of Washoe consistently followed this practice. Uncut versions of the same episodes showed that every one of Washoe’s multi-sign statements came after similar signs by teachers.

“Can an Ape Create a Sentence?” is the title of Terrace’s report in Science (November 23, 1979). His reluctant answer is no. “Apes can learn many isolated symbols (as can dogs, horses, and other nonhuman species), but they show no unequivocal evidence of mastering the conversational, semantic, or syntactic organization of language.” Of the earlier researchers only Rumbaugh so far seems impressed by Terrace’s analysis. His own work, he told The New York Times (October 21, 1979) has been pushing him toward similar views.

Speaking of Apes, edited by linguist-semiotician Thomas A. Sebeok and anthropologist Donna Jean Umiker-Sebeok, is a much needed anthology of important articles on both sides of the intensifying controversy over the capacities of apes for language. It is impossible to discuss such a wide variety of papers so I will concentrate mainly on the long introductory article, “Questioning Apes,” by the Sebeoks. Both are at Indiana University’s Research Center for Language and Semiotic Studies, of which Thomas Sebeok is chairman. Their introduction is the most powerful indictment in print of the early work on talking apes.^*

Psychologists have a term, “experimenter effect,” that covers all the insidious ways a researcher’s strong convictions can unwittingly distort data. The Sebeoks first remind us of obvious ways that scientists in any field can be unconsciously motivated to get positive results. The stronger the results the faster their career advances, and the more likely will their work attract funding. Assistants are strongly motivated to please an employer who pays their salary, and success often advances their own careers. If the work is controversial there is a tendency for research teams to form a cluster of insiders deeply suspicious of outsiders. They become, as the Sebeoks put it, a “dedicated group of enthusiastic workers, one that constitutes a tightly knit social community with a solid core of shared beliefs and goals in opposition to outside visitors…. In fact, it is difficult to imagine a skeptic being taken in as a member of such a ‘team.’ ”

Within this frame the Sebeoks see a variety of curious ways in which talking-ape results are easily twisted in the direction of belief. Consider, for example, the “Clever Hans effect.” The term comes from a classic 1907 study by Oskar Pfungst, a German psychologist, of a famous performing horse of the day who could answer difficult questions, including arithmetical problems, by pawing the ground. In most cases of such performing animals (there have also been “learned” dogs, pigs, and even geese) a trainer tells the animal when to stop by secret cueing, such as a slight sniff, but in the case of Hans, Pfungst was able to prove by ingenious tests that the horse had learned to respond to subliminal cueing on the part of spectators.

Talking-ape researchers have tried to exclude the Clever Hans effect, but the Sebeoks show convincingly that the effect is omnipresent. There is no evidence, they maintain, that successful teachers had any training in controlling unconscious facial movements, breathing rhythms, bodily tensions and relaxations, and so on. Some reactions, such as eye-pupil size, are probably uncontrollable. Pfungst reported his inability to avoid cueing Hans no matter how hard he tried.

Talking apes seldom perform well for strangers. Believers explain this by an ape’s emotional attachment to certain teachers, but it is as readily explained by assuming that over the years apes develop a special sensitivity to unconscious reactions peculiar to a loved human and which they naturally fail to perceive if someone new tries to talk to them. Could it be, the Sebeoks ask, that the best trainers are those most expressive in unconscious cueing? A study of unedited films shows that ape teachers are, in the authors’ words, “anything…but stone faced.” Even uncropped still photos reveal obvious cueing. The Sebeoks cite some horrendous instances in the photos illustrating Patterson’s National Geographic article, and in Mrs. Premack’s book.

Concerning the famous Washoe-swan incident, both Terrace and the Sebeoks point out what should have been obvious at once. Washoe may simply have signed “water,” then noticed the bird and signed “bird.” It is unlikely that Fouts could have concealed his elation. Washoe, observing this social reward, would henceforth associate the double sign with a swan.

There is no solid evidence that an ape has ever invented a composite sign by understanding its parts. In the course of several years an ape will put together signs in thousands of random ways. It would be surprising if it did not frequently hit on happy combinations that would elicit an immediate Clever Hans response. No teacher has bothered to record all the nonsense combinations produced by an ape, but every lucky hit is sure to be reinforced by cues of approval, and to go into a teacher’s records, reports, books, and lectures.

Even when an ape has memorized a sign it often makes errors in reproducing it. When this happens, the Sebeoks point out, ape teachers have a battery of excuses. Instead of a mistake it becomes a joke or a lie or an insult. Patterson is especially prone toward this kind of subjective evaluation. She asks Koko to sign drink. Koko touches her ear. Koko is joking. She asks Koko to put a toy under a bag. Koko raises it to the ceiling. Koko is teasing. She asks Koko what rhymes with sweet. Koko makes the sign for red, a gesture similar to the one for sweet. Koko is making a gestural pun. She asks Koko to smile. Koko frowns. Koko is displaying a “grasp of opposites.” Penny points to a photograph of Koko and asks, “Who gorilla?” Koko signs “Bird.” Koko is being “bratty.”

The National Geographic article reproduces a crayon picture by Koko that is captioned “Representational Art.” Its black squiggles, says Penny, are spiders. An orange scrawl is Koko’s drinking glass. A similar tendency to overhumanize ape behavior, though less blatant, infects all the earlier work. It is little different from the firm belief of sentimental pet owners that a beloved cat, or even a parrot, understands almost everything you say to it.

It is possible, of course, that apes do have a feeble talent for creating meaningful composite signs, but by the principle of Occam’s razor, the Sebeoks insist, should we not accept simpler explanations first? So far there is no reason to suppose that Koko’s remarkable utterances are anything more than responses to unwitting cueing on Penny’s part, or to Penny sifting out from thousands of nonsense combinations those that make sense to her, not to Koko. An objective evaluation of a phrase like “bad question” cannot be made without a videotape of the scene to make sure the details are correctly recalled, and without knowledge of how many of the ape’s unlearned and spontaneous two-word combinations are nonsense. Otherwise we have nothing more than a collection of anecdotes.

Some researchers, especially Premack, have tried “double blind” tests to rule out Clever Hans effects, and whenever these controls were tight the ape’s ability dropped almost to chance. Much is made of the slight deviations from chance, but the Sebeoks list numerous ways in which bias could have slipped into these efforts to exclude it. We are not told what controls were placed on photographers. Reports often fail to note the presence of others who happened to be around but were deemed too irrelevant to mention. One-way windows eliminate visual cues but not sound cues. Details are sparse about randomizing procedures and the rules followed in scoring.

There are religious beliefs—in the West notably those of the Catholic Church and conservative Protestantism—that make it necessary to assume that human beings have an immortal soul denied to the beasts that perish. Mortimer J. Adler wrote a book a few years ago called The Difference of Man and the Difference It Makes in which he enlarges on Thomist arguments that the ability to understand syntax is one of the main ways a human mind differs from the mind of a beast. At the time Adler wrote his book the dolphin language flap was in full swing, and Adler made much of the fact that if we ever succeed in conversing with a whale his thesis will be undermined. If the book has a new edition you can be sure Adler will underplay porpoises and concentrate his dialectical fire on apes.

It is good to understand that this sort of metaphysical objection to talking apes, reinforced by Revelation, is not behind the views of Chomsky, Terrace, the Sebeoks, or any other major critic. What they are saying is much simpler. Contemporary humans and apes are terminal branches on the tree of evolution. Transitional types that flourished over the millennia during which human beings acquired the ability to talk are no longer available for study. Chomsky believes that evolution gave to humans, as it did not give to any living lower primates, a capacity for language that is deeply interlocked with the inherited structure of their brains.

Little is gained by quibbling over the meaning of “language.” As Chomsky says in his contribution to the Sebeok anthology, this is a conceptual not a scientific question. If you define flying, he writes, as rising into the air without the aid of special equipment and landing some distance away, then human broad-jumpers can fly about thirty feet. Chickens do slightly better—about three hundred feet.

Suppose, Chomsky continues, we label the four colors pecked by pigeons with four words: Please-give-me-food. “Do we want to say that pigeons have the capacity for language, in a rudimentary way? This is much like the question whether humans can fly, almost as well as chickens though not as well as Canada geese. The question is not clear or interesting enough to deserve an answer.”

The central empirical question can be simply put. Do apes have the ability to link visual signs in ways that justify saying they are using syntax? Yes, say most of the researchers and many outsiders. Jane H. Hill, an American anthropologist, closes her contribution to Speaking of Apes by writing: “It is unlikely that any of us will in our lifetime see again a scientific breakthrough as profound in its implications as the moment when Washoe…raised her hand and signed ‘come-gimme’ to a comprehending human.”

No, say some researchers and a growing number of outsiders. If no, Chomsky concludes, then a study of ape signing can be expected to cast as little light on human language or conversely as a study of human jumping can cast light on the mechanism of bird flight or conversely. One can teach two pigeons to bat a ball, writes Ms. Hill, quoting a familiar aphorism, but is it ping-pong? She thinks it unjust to apply this skepticism to talking-ape research. Chomsky holds the opposite opinion.

No one can rule out the hope that as talking-ape research continues, under better controls, it may turn out that apes do have a dim awareness of syntax. If so, then the researchers will have made a point even though it may not be a big one. At the moment, however, the situation seems little different from that which confronted biologists a century ago. Here is how Darwin summed it up in a section on language in The Descent of Man:

As the voice was used more and more, the vocal organs would have been strengthened and perfected through the principle of the inherited effects of use; and this would have reacted on the power of speech. But the relation between the continued use of language and the development of the brain has no doubt been far more important. The mental powers in some early progenitor of man must have been more highly developed than in any existing ape, before even the most imperfect form of speech could have come into use….