It’s a regular Georgia afternoon in the late 1970s, and two chimpanzees, named Sherman and Austin, are about to do something phenomenal.
They are separated into two rooms by a barrier with a single hole in the middle. In one room, Austin has just seen some bananas placed in a box, and locked. In the other, Sherman is sitting with an array of tools to choose from, each of which is appropriate for opening a different type of food box. As soon as Austin sees that the bananas have been placed in a lock-based food box, he rushes over to a keyboard and pushes the abstract symbols that produce the phrase get key. When he does, those same symbols light up in Sherman’s room, and he runs to his tool selection, picks out the key, and hands it through the opening in the wall to a waiting Austin, who uses the key to open the box and grab the bananas, ultimately handing half of those bananas back through the hole to Sherman.
Let’s think about what just happened there. Austin had to recognize the problem at hand, think of the correct tool to solve that problem, realize that the tool he needs is sitting with another chimpanzee, recognize that he has information the other chimpanzee doesn’t, remember the abstract symbols for the verb and noun that will communicate that information to the other chimpanzee, who himself has to recognize those symbols and respond appropriately to them, all culminating in Austin acknowledging that the bananas he received were the result of a division of labor which merits an equal sharing of the reward.
That one experiment shows off an amazing arsenal of communicative, organizational, linguistic, and abstract mental abilities all of which, two decades earlier, were believed to be the exclusive domain of humans. Prior to 1966, the idea of a primate using language with another primate to accomplish a task requiring division of labor would have been simply unthinkable, but that year a team of researchers, Beatrix and Allen Gardner, had the idea of teaching American Sign Language to a young chimpanzee named Washoe, and thus began a two decade exploration of the ape mind, carried out in the teeth of fierce academic scorn, which changed forever how we think about the nature of animals, and ourselves.
As central and foundational as the Gardners’ work was, however, the beginning of primate language research stretches back a bit further than them, to the late 1940s when Catherine and Keith Hayes raised a chimpanzee named Viki in the same manner as they would a human child. They found that she had an advanced capacity to perform perceptual tasks, but in attempting to train her to speak, they found she never moved beyond four basic words. Linguists the world around took comfort – as smart as they were, apes evidently did not have language.
And there the issue would have remained had Beatrix and Allen Gardner not had the stroke of insight to try once more, only this time using sign language instead of the spoken word as the means of inter-species communication. As we would later discover, because of their skull placement apes don’t have a windpipe configuration that would allow them to produce consonants, which do the heavy lifting when it comes to communicating orally. But, the Gardners wondered, do apes still produce the requisite mental machinery to associate abstract signs with actions and objects, and to both communicate their needs and understand the communications of others? If a chimp could be taught to sign, perhaps that question could be answered.
In 1966, they took in a ten month old chimpanzee named Washoe and proceeded to raise her as the Hayeses did, but with a crucial change: now, all communication would be done in sign only. They felt that attempting to communicate simultaneously via speech and sign would complicate their methods, and insisted on strict silence from anybody working with Washoe. As opposed to the operant conditioning approach of B.F. Skinner’s reigning behaviorist model, which focused on drill, repetition, and reward, the Gardners adopted a natural approach that let language unfold as a regular part of the daily routine. Once Washoe grasped the utility of signing for making her wishes known, she rapidly increased her vocabulary, learning 30 signs in her first 22 months on the way to an ultimate total of around 350 words that she could produce and understand, and use in simple combinations to make requests. She was able to pass double-blind tests of comprehension which established that she understood signs as abstract markers for general concepts.
When the Gardners published their results in a 1969 Science article, making the modest and substantiated claim that Washoe used her signs in a communicative and natural fashion that suggested linguistic capacities, the academic world erupted in outrage mixed with occasional enthusiasm. The dominant linguistic paradigm at the time was the Chomskyan theory of innate syntax – that human minds have a basic hard-wired structure for linguistic syntax which is unique to us as a species. For linguists of this tradition, apes producing language was a dangerous notion indeed, and the next two decades would see them throwing their collective institutional muscle into attacking any research that suggested that primates could meaningfully communicate.
At that point, working in primate language was an uphill battle, and yet the field attracted a series of talented researchers who showed through rigorous and ingenious testing that Washoe was no fluke, and that apes’ gift for language and hierarchical thinking runs deep.
Francine “Penny” Patterson attended a lecture by the Gardners in 1971 and was enthused by the idea of extending their work to other animals. She ultimately chose to work with gorillas, knowing full well their reputation as the least cooperative of the great apes. The general consensus was that gorillas were stubborn, slow-witted, and impossible to work with, and Patterson’s colleagues attempted to steer her clear from the choice. She pressed on, however, and in 1972 began working with an infant gorilla named Koko, a relationship which would continue for 46 years, until Koko’s death in 2018.
Washoe had established that apes could understand the concepts of signs representing a class of objects. What Patterson and Koko would go on to prove is that apes are just as proficient with understanding spoken English as they are with American Sign Language. Unlike Washoe, whose team of researchers spoke only in signs, Koko’s team used simultaneous sign and spoken communication. The technique paid off, as in 1974, at just three years of age, Koko demonstrated under controlled conditions the ability to identify objects and situations described to her either orally or via signs with an accuracy just under that of mentally handicapped children, and by the end of her life she had demonstrated comprehension of over 2000 words. She also showed an ability to understand and respond appropriately to questions (a skill not rigorously tested with Washoe), with answers that were grammatically correct 83% of the time, and factually correct 84% of the time, a result even more remarkable when you factor in Gorilla Stubbornness.
That stubbornness manifested itself in three particularly fascinating ways: jokes, deceptions, and insults. Koko was a master at putting together pejorative terms to coin fresh insults which she used with abandon when frustrated (of which dirty toilet devil is probably my favorite and which I fully intend to work into my rotation), at attempting to throw the blame for broken objects onto other people or gorillas, and at pushing her linguistic abilities to absurd comedic extremes. Once, while handling a large white sheet one of her human companions asked Koko what color the sheet was. “Red,” answered Koko. Pressed by the companion, she insisted on her answer and then held up a tiny speck of red lint that had been on the sheet before bursting into a massive grin.
Francine Patterson worked with Koko and, beginning in 1976, with a male gorilla named Michael who developed an unheard of facility for using “Wh” type words in questions, for a half century, collecting not only invaluable data about gorillas’ ability to understand and produce a plethora of signs in single and multi word combinations, but also a treasure trove of anecdotal evidence that gave us our first shadowy look at the mental life of our near relatives, including their thoughts upon death, birth, fairness, and guilt. It was stunningly important work which had the great misfortune of being done in the very teeth of the Great Linguistic Counter-Attack which all but buried primate studies in the late 1970s.
Things had been bad enough with the legion of Chomskyan warriors insisting, based upon their definition of Language As Syntax, that any results obtained with apes were unclassifiable as language, but the worst blow fell in 1979 when Herbert Terrace announced the results of his attempt to scientifically reproduce the work of the Gardners with a chimp named Nim. As a reproduction of the Gardners’ methods it was a mockery – whereas Washoe had been raised by a stable roster of companions fluent in ASL who could teach signs naturally as opportunities arose, Nim was placed in a household where nobody knew ASL, was taught signs in a sterile lab environment devoid of natural context, and was shuttled between different caretakers as she proved unmanageable in her original home, ultimately ending her days tragically in a primate research facility after Terrace had finished with her. Not surprisingly, Nim foundered under these inconsistent and sterile circumstances, and only learned a small fraction of the signs Washoe mastered in a comparable time, and used them with far less spontaneity or creativity.
…working in primate language was an uphill battle, and yet the field attracted a series of talented researchers who showed through rigorous and ingenious testing that apes’ gift for language and hierarchical thinking runs deep.
Rather than taking the obvious lesson – that his experimental approach was gravely flawed – Terrace concluded that Nim’s underperformance was a clear sign that apes couldn’t learn language, that they were only copying what their instructors fed them, and he said as much in a massively influential 1979 article. The linguists, seeing a primate language researcher disavowing his own field, seized upon the article, and held it up as conclusive evidence that the entire endeavor was a misguided, self-deluding waste of time.
Terrace’s article hit the moment that the third great name in the history of primate language, Sue Savage-Rumbaugh, was completing a series of remarkable experiments that demonstrated beyond a doubt the advanced problem solving and communicative skill of primates. In 1976 she began working with Sherman and Austin, the pair of chimpanzees who collaborated to win some bananas in our opening paragraphs, at the Language Research Center in Georgia, where researchers employed a keyboard featuring an array of abstract symbols to communicate with primates. After the chimps had been trained as to the meanings of a set of symbols, she set about designing experiments to show that they were capable of using these communicatively in creative problem-solving scenarios. At first, she would place a food item in an opaque food container with one of the chimpanzees watching, then walk to the room where the other was waiting. Sherman and Austin would only get the food within when the waiting chimpanzee used the symbol corresponding to that item, which meant that the chimpanzee who had seen it had to communicate that information through the symbolic keyboard.
Their extreme facility with this, and the later tool exchange scenario, showed not only that apes had the capacity to imagine the incomplete knowledge states of others (which we had assumed only humans could do), but could harness symbols to communicate the necessary piece of information to another chimpanzee.
It was a striking result, but represented only the beginning of Savage-Rumbaugh’s groundbreaking work. In 1980, she began working with Kanzi, a bonobo adopted by Matata, a female at the Language Research Center. Matata had not shown any marked ease with the keyboard, but one day when Matata was absent, Savage-Rumbaugh was astonished to see Kanzi march over to the keyboard and start spontaneously and accurately using its keys. Kanzi had never been directly trained with the keyboard, but had picked up its use and symbols simply by passively observing Matata attempting to learn them. Savage-Rumbaugh grasped the implication: perhaps the formalism of sign training sessions was a hindrance to symbol acquisition, and that spending time with language users in rich environments would lead Kanzi to pick up the rudiments of language more easily than repetitive conditioning exercises.
Utilizing to the full a nearby forest, she created a series of stations stocked with unique food and items and went out with Kanzi on expeditions that compelled him to use language to the utmost in order to work through the navigation of his favorite locations and activities. This laissez faire attitude towards word acquisition bore fruit as Kanzi showed a gift for word acquisition and for multi-word combinations but, most importantly, proved incredibly adept at understanding spoken English. In comprehension trials, he scored an astounding 95%.
These were the bitterest years of the linguistic war, as Savage-Rumbaugh’s reports on controlled double-blind tests of Kanzi’s comprehension were rejected by journal after journal as reviewers simply refused to believe her results in spite of her documented and rigorous methods. However, as with most theoretical academic fashions, a counter-attack to Chomsky’s hegemony was in the works. Rogue linguistic theorists began speculating that language, far from being the result of some special syntactical structures in the brain, was produced by a more generalized capacity to hierarchically order tasks, an ability evidenced by ancient man’s capacity to create stone tools. Under this model, a general capacity for creating ordered solutions to solve a given problem, rather than a specific syntactical organ, formed the substrate for eventual language acquisition.
This revived the utility of primate language studies for a growing sector of the scientific community, and when Kanzi demonstrated the ability to learn from a human the ability to make stone hand axes for herself which she then used to solve problems in her environment, a whole world of shared cognitive capacity opened and the graying boundary between species became less distinct still.
Washoe. Koko. Kanzi.
Gardner. Patterson. Savage-Rumbaugh.
Between them they had shown that apes can learn and use symbols, produce novel combinations of those symbols to describe new phenomena, apply personal word order rules to convey meaning, use words to joke and deceive, and communicate effectively with other primates to jointly solve novel problems. They can understand spoken English and reply via signs and symbols. They even use symbols when playing by themselves, and have been observed teaching other apes their own specialized versions of standard signs. Some linguists are, inevitably, still unsatisfied, and have played a half century game of moving the goalposts back, yard by yard, with each demonstrated primate capacity, but to those without a vested interest in its debunking, the trio’s accumulated evidence is compelling, and the small group of researchers who stuck to their methods and results in the midst of almost universal dismissal are earning at last some small measure of the recognition that is their due.
Beatrix Gardner died in 1995. Washoe’s language learning was taken over by Roger Fouts in the Seventies, where she was observed teaching language to other chimpanzees. Washoe died in 2007.
Nim died in 2000. After having finished his experiments, Herbert Terrace gave her to a research facility where she was kept caged and showed constant signs of distress. He visited her once ten years later for a photo opportunity and then never again.
Koko died June 19, 2018 after having become a world celebrity after images of her gingerly handling her pet kitten were published in a 1985 issue of National Geographic. Penny Patterson continues her work with Ndume, a gorilla she began working with in 1979.
Kanzi is still alive and well, and will tell you all about it, if you ask him.
Lead image via Wikimedia, creative commons license. Bonobos Kanzi and Panbanisha with Sue Savage-Rumbaugh. Photograph from William H. Calvin’s ALMOST US in Kindle format.
FURTHER READING: Each member of our featured trio wrote at least one accessible account of their remarkable research. Francine Patterson’s The Education of Koko (1981) is one of my favorite books, and has been since I found it on a shelf in the Hayward library two decades ago and happily ignored the world while I read it on a bench outside. Sue Savage-Rumbaugh’s Kanzi: The Ape at the Brink of the Human Mind (1994) covers both her Sherman-Austin experiments and those with Kanzi, as well as detailing the bitter twists of the late 70s and early 80s Syntax Wars. Meanwhile, for straight up Scientist Presenting Her Methods and Results realness, Teaching Sign Language to Chimpanzees (1989) by Beatrix and Allen Gardner contains a chapter where the Gardners describe their work with Washoe, and another describing their next generation of chimp studies. Meanwhile Catherine Hayes wrote a book in 1951 about her experience with Viki called The Ape in Our House. I’ve never gotten my hands on a copy, but I sure want to.
And for more awesome Women in Science comics, check out the archive and my books, Illustrated Women in Science – Volume 1, 2 and 3.