Pepperberg’s reason for teaching Alex words is not to teach him language, but to understand Alex’s cognitive abilities. It is possible to question Alex and discover that he understands the class concepts of color and shape, for example.
There is no reason to think that Alex, a pet-store baby, is a particularly talented bird. The difference between him and other members of his species is the way he was trained. African greys are renowned talkers, but they seldom use words and phrases with comprehension. Often they associate a phrase with a situation, saying “Good morning” when someone comes in or muttering “Bad bird, bad bird” when yelled at, but this is limited comprehension. Geier, a talented parrot, said “Auf Wiedersehen” when people left, but despite hours of tutelage, couldn’t learn to ask for water when he was thirsty.
Pepperberg’s method of training Alex, the model-rival method, is based on the work of Dietmar Todt. The animal sees a model and a rival performing the task to be taught. Rewards are related to the task. This is distinct from more usual methods of teaching, in which the reward is whatever the animal likes, and is not related to the task. The usual method for teaching a parrot the word “green” might be to hold up something green, say “green, green” and give him a peanut if he says something like “green.” But this doesn’t work. At best you teach the bird that “green” means “give me a peanut.”
Under the model-rival technique, there are two people in addition to the animal being trained. One person holds up something green, and asks the other to say “green,” or to say what color it is. If they say something close to “green,” they get the green object.
Meanwhile the animal is watching this procedure, perhaps thinking, “I want that green key, I bet I could do that, I can say ‘green’ better than that.” The task that the animal is supposed to perform has been modeled, and perhaps the animal’s rivalry has been aroused. This method of training works well with many animals, including people.
One of Pepperberg’s twists on the Todt method is that models take turns being teacher and pupil, showing that communication is a two-way street and not just a series of tasks that one individual forces another to perform. Since parrots are rebellious types, this probably helps.
It took nine months to teach Alex to use the words “same” and “different” with comprehension. But now if you want to understand the things that Alex considers similar—can he tell the difference between two colors, two birds, two people?—you only need ask. It is now possible to ask Alex “What same?” and have him answer “color” to tell you that a piece of wood and a key are both blue. Assuming that he is in the mood for your silly questions.
Pepperberg does not consider Alex’s tasks easy and natural for a parrot. She calls them “exceptional learning” and says that for that reason they require intensive training. To confirm that the model-rival method is what has worked with Alex (rather than her having happened to purchase a genius), Pepperberg tried training two young parrots with other methods. But Kyaaro and Alo did not learn words from audiotapes. They did not learn from videotapes, even videotapes featuring Alex. Even when people sat next to them and commented on the videotapes: nothing. So don’t get one of those recordings to play to your parakeet unless you enjoy hearing it yourself, because the bird won’t learn from it. Also, teaching yourself a foreign language from tapes? A real long shot.*
Alex finds “What color?” a boring question. He never asks for things by color, although he sometimes asks for things by shape. The first time he ever said “What color?” was one day when he was looking at his mirror image. He asked six times, and the student working with him gave him such answers as “That’s gray; you’re a Grey parrot.” That was enough. The researchers hastily dyed some objects gray and quizzed Alex to check: yes, he said they were gray.
Alex babbles secretly to himself. “He does a lot of things in private that he doesn’t do in public. You might get a lot of strange things in the evening.” In these surreptitious monologues he practices things that he will not say publicly until he gets them right.
Since Alex speaks and understands English words, you or I could discuss seeds, nuts, or the color of objects with him if he were in the mood. But we couldn’t chat about much else. “Alex’s input is mostly limited to what I deem necessary for experiments, which primarily require object or category label,” writes Pepperberg. He learns words faster when they describe things he actually wants and when he hears them in a social setting. Given how social parrots are, one wonders if he would learn more if he somehow had the chance to discuss what is really interesting to a parrot—social matters. Gossip. His environment is socially impoverished, but one can imagine that more socially fortunate parrots might wish to discuss such matters as who is cute, who flies funny, who is overrated, who has a crush on whom, and who bit whom and why.
Sign language
Meanwhile, Allen and Beatrix Gardner had had the brilliant notion of teaching chimpanzees words in a human language that didn’t tax their vocal tracts: American Sign Language. They began in 1966 with Washoe, a 10-month-old chimpanzee, raising her first in their home, and then in a trailer, and using only sign language in her company. Washoe acquired dozens of signs, which she principally used to request things. She combined signs in strings like “you me hide” to suggest a game of hide and seek. She made some word combinations like “open eat drink” to refer to the refrigerator. When Washoe moved to another site, the Gardners started a second project with several younger chimpanzees, and this time several native ASL speakers were among their caretakers. Eventually, in the care of Roger and Deborah Fouts, Washoe and the other signing chimps came to live together at Central Washington University. By this time Washoe knew at least 240 signs. When she adopted the 10-month-old Loulis, the researchers decided not to sign around the chimpanzees so that any signs Loulis learned would have to come from his mother or the other chimpanzees. He began learning at once, and Washoe was seen instructing him as she had been instructed. By the time humans resumed signing around the chimpanzees, Loulis had learned 55 signs.
The signing chimpanzees of Central Washington sign to each other when there are no people around, using it as a small part of their communicative repertoire. One example of this integration of sign with other modes of communication is given in a journal article: Moja was about to get some fruit juice when Tatu approached. Moja communicated her resentment of the intrusion by the classic chimpanzee technique of screaming, whereupon Tatu signed, “Smile.” Moja turned away and ignored Tatu, a further and eloquent form of communication.
When the chimpanzees sign to each other, it is usually for “reassurance, social interaction, and play.” They urge other chimps to hurry, and they ask to be chased. This contrasts with the pedagogical conversations humans insist on having, all “What this?” and “What color that?” When Moja was in estrus, she did a lot of signing, mostly urging a certain male chimpanzee to “hug.” (At other times chimps requested hugs in a more chaste spirit.)
Inspired by the Gardners’ work with Washoe, psychologist Penny Patterson wanted to see if a gorilla could learn to sign. Raised by Patterson after a year at the San Francisco Zoo, Koko is now said to have the largest vocabulary of any signing ape. Her first signed word was “food,” quickly followed by “drink,” “more,” “out,” “dog,” “comegimme,” “up,” “toothbrush,” and “that.” She understands a great deal of spoken English, is amused by rhymes, and will sign words that rhyme if spoken aloud. Like certain other signing apes, and most people, Koko dislikes being drilled on vocabulary and becomes recalcitrant if it goes on for any length of time.
Chantek
Chantek is an orangutan raised until the age of nine years old by H. Lyn Miles, who considers him her cross-fostered son. He was born in 1977 at the Yerkes Regional Primate Research Center and was loaned to Miles, an anthropology professor at the University of Tennessee, when he was nine months old. She and her colleagues taught him a modified form of sign language. Miles didn’t try to tea
ch him long lists of signs to use on quizzes, preferring to focus on signs that interested him.
The goal of Project Chantek was not to show that Chantek could or could not acquire language, but to examine his cognitive and communicative development. Miles charted Chantek’s progress through the stages outlined by Piaget for human children. Chantek “was immersed in a human cultural environment and learned the rules for behavior and interaction, a process anthropologists call enculturation,” writes Miles. They took walks, they fed the squirrels, they played with blocks, they went out for fast food, they painted, they made shadow puppets on the wall at bedtime, they talked to the nice policeman on the horse.
At first he was taught signs by molding his hands. “Food-eat” was the first thing he signed. Later he learned by imitation. When he only knew 20 to 35 words, Chantek invented the idea of signing with his feet instead of his hands.
Over seven years, it took Chantek an average of 53 days to master signs for food, 69 days for beverages, 117 days for actions (give, listen, kiss), and 565 days for pronouns. It took just 19 days to add “give” to his vocabulary, but it took 915 to add “shoe.” Chantek didn’t wear shoes, but he was happy to accept gifts, especially gifts of food. He learned the word “up” early and quickly, initially as a request to be picked up. But it took him over a year to learn the word “down,” because he never wanted to be put down.
He combined signs into phrases, showing word order preferences. At the age of four he called contact lens solution “eye drink.” “Cheese meat bread” of course means a cheeseburger. Charmingly, “noisy dog” or “noisy sky dog” means a helicopter. He invented a self-explanatory sign that means a View-master, and another sign for balloon. Chantek was fond of deception: of claiming an urgent need to use the bathroom when he merely wished to play with the taps; of claiming to have eaten his vegetables when he had actually concealed them behind the toilet; of popping an eraser in his mouth, signing “food-eat,” and displaying his open and apparently empty mouth, when in fact he had wickedly hidden it in his cheek for later use.
“When Chantek saw his first video of other orangutans he called them orange dogs and he wanted no part of them,” says Miles. (This is not so different from the attitude of wild orangutans toward strangers, it must be pointed out.) “He calls himself an ‘orangutan person’ and he now lives with other ‘orangutans’ but originally they were orange dogs.”
When he was nine, Chantek was reclaimed by the Yerkes Center. At first Miles was able to continue some work with him, but then a decision was made to treat him as an ape, not as a child, and Miles was not allowed to see him for seven years. When they finally met again, Chantek signed, “Mommy give ice cream.”
The signs Chantek remembered after his long silence were objects of human culture, Miles reports. Perhaps he could handle other concepts with normal orangutan communication, or perhaps he tried to express his apely wishes through sign and wore them out with failure.
Biruté Galdikas witnessed Gary Shapiro’s endeavors to teach sign language to free-ranging orangutans being rehabilitated at Camp Leakey, her Borneo headquarters, for life in the wild. He taught 30 signs to his best pupil, Princess, although there were others she “refused to learn.” These efforts were also witnessed by Galdikas’s toddler child Binti Brindamour. “Naturally and effortlessly, Binti between age two and three learned more signs than Princess at age five or six learned through rigorously structured, formal lessons.” While no one doubts that the child has more language aptitude than the orangutan, it is possible that this difference was accentuated by their different learning situations. In short, Princess may have been annoyed by her rigorous lesson plan, while Binti was, inadvertently, getting the benefits of a model-rival training regime.
Nim Chimpsky
Then there was the historic Nim Chimpsky, famous as a failure and a phony at signing. In the first exciting heyday of teaching chimpanzees to sign, some grandiose claims were made for the apes’ abilities and the ones that they were undoubtedly about to show. In 1973 researcher Herbert Terrace got a young male chimpanzee and named him (sort of) after linguist Noam Chomsky. Nim was raised first in a family home, and then in a semi–laboratory setting. He was tutored in sign language by 60 students, most of whom had little familiarity with signing. These sessions were taped. Preparatory to writing a paper on Nim’s achievements, Terrace reviewed the tapes and was shocked at what he saw. He declared that Nim repeated himself wildly and meaninglessly in hope of getting food rewards, showed no sign of syntax, rarely offered signs spontaneously, interrupted constantly, was often just imitating whatever sign his tutor had just made, and showed no signs of turn-taking, as happens in conversations. Furthermore, he charged that all signing apes did the same.
Terrace’s reversal on whether apes could learn to sign meaningfully had a devastating effect on ape sign research, and quite a few apes were shipped back to the primate centers whence they had come, including Nim, not quite four years old, who was sent back to the Institute for Primate Studies, where he lived in a cage or on a monkey island with other chimpanzees.
When Nim was six (still preadolescent), some researchers at the institute tried to test Nim as he had been tested in Terrace’s study, administering a drill on naming objects. “The result was a series of threats from Nim, culminating in a bite.” Subsequently they videotaped some sessions with Nim to see what effect social setting had on his signs.
In conversational sessions, Nim was taken for a stroll to a grassy area, played with, and chatted with in a casual and indulgent manner. True, they did sometimes point to things and ask Nim what they were, but they tried not to be overbearing about it. Nor did they give Nim food rewards, though if he asked for someone’s hat or shoes he got them. (Unlike Chantek, Nim was interested in shoes.) There was a box of toys. Nim enjoyed this, didn’t bite, and displayed much more normal use of signs. During the training sample his interruption rate was 31 percent, but in conversation it was only 9 percent. He made four times as many spontaneous contributions (using signs that others hadn’t used), going from 14 percent to 60 percent.
In the same grassy setting, they tried to repeat a training-type session, complete with food rewards, but Nim was uncooperative. “The training session was terminated after eight minutes because of increasing irritation and volatility on the part of both participants.”
In short, the better you treated Nim, the better-spoken he was. One of the complaints about Nim was that he just fired off signs randomly in hopes of being rewarded, but since he was taught on a reward system, this seems like a reasonable strategy. (Nim later went to an animal sanctuary, where they vowed he would never be quizzed again.)
Artificial symbol systems
In the meantime, some researchers worked with apes and artificial symbol systems—magnet boards, computer screens, or computerized keyboards. Some used lexigrams, intricate geometric shapes that don’t resemble the objects they symbolize, displayed on a keyboard.
Two chimpanzees, Sherman and Austin, who began training with lexigrams at the ages of two and a half and one and a half, readily used them to ask for things. When they were later asked to name those same things, they couldn’t do it. They could press the lexigram for “apple” to request an apple, but if they were asked, using that lexigram, to get an apple, they didn’t understand. All the symbols had to be retrained.
With time, the chimpanzees began to use their lexigrams in a new way. “We had taught Sherman and Austin key elements of communication—request, naming, and comprehension. Once these were in place, other aspects of communication emerged spontaneously. The chimps began to pay close attention to each other’s communications; they engaged each other before delivering their message; they gestured to emphasize or clarify messages; they took turns.”
Kanzi
The same research team wanted to see how bonobos (once called pygmy chimpanzees) would do with lexigrams. An adult female bonobo, Matata was tutored for two years by Sue Savage-Rumbaugh in the mea
ning of 12 lexigrams. Matata was not an apt student. After two years her progress compared unfavorably to that of chimpanzees. She had learned lexigrams for banana, juice, raisin, apple, pecan, and orange, but could not generalize from the symbol for “banana” to a picture of a banana. She was patient with Savage-Rumbaugh’s insistence on teaching her, just as she was patient with her exuberant and disruptive child, Kanzi, who grabbed her food, somersaulted over her head while she was trying to press a lexigram key, and slapped the keyboard to get her attention.
At this discouraging time (funding was running out), when Kanzi was nearly two and a half, Matata was sent away for a few months for breeding. Kanzi was left behind. The first day she was gone, when Kanzi was not searching for his mother, he astounded the humans by using the keyboard with unsuspected proficiency. He used all 12 lexigrams in 120 separate utterances. He combined lexigrams. He used lexigrams not just to ask for food but to comment on it. He used them to describe what he planned to do next. “Kanzi had been keeping a secret.” Everything they had tried to teach Matata, Kanzi knew. He also knew things they had not even tried to teach her.* (Kanzi may be another inadvertent example of the efficacy of the model-rival method.)
Becoming a Tiger: The Education of an Animal Child Page 16