Koko the Impostor: Ape sign language was a bunch of babbling nonsense
- Several great apes have been taught to use sign language. But do they actually understand it the way humans do?
- There are many fascinating and downright bizarre anecdotes in the field.
- But one prominent ape language researcher came to conclusions that effectively ended the entire field.
There is something miraculous about animals that communicate with human beings. A cat’s purr of affection or a dog’s frantic joy when you come home from work are clear examples of communication. Yet these are the animals’ languages, not ours. A creature that communicates in our language begins to feel like another person.
Great apes seem to lack the physical development of the larynx, or the neural capability, to learn human speech. Scientists debate the matter. However, with well-evolved articulating hands and arms, gorillas, chimpanzees, orangutans, and bonobos can master complex gestures. Researchers have harnessed this capability to teach them various forms of sign language. Occasionally non-signing forms of communication like pointing at pictograms also have been used.
Stories of ape sign language can feel shockingly human. Washoe was the first signing ape. When the chimp’s handler disclosed that her baby had died, Washoe reportedly signed back cry. The bonobo Kanzi learned to point to various symbols that represented about 350 words. The Koko project released a video of the namesake gorilla delivering a message about climate change. The animals appear to form thoughts and express them in one of our languages to meaningfully convey their ideas to us. Among these experiments, the story of one researcher and his chimpanzee stands out.
The strange tale of Nim Chimpsky
That researcher is Herbert Terrace, a professor of psychology at Columbia University. Nim Chimpsky — the name is a pun referring to the prominent linguist Noam Chomsky, then known for his groundbreaking research on linguistics — was his personal research study subject. Nim was raised like a human baby in a Manhattan apartment. His foster mother, Terrace’s student Stephanie LaFarge, taught him ASL. She also breast fed(!) the chimp and supposedly taught him to smoke weed(!!). One of Nim’s handlers reported that the animal requested the substance. Nim was also taught more than 100 signs.
Terrace rode the results of Project Nim to academic stardom in the 1970s. At the end of the study, he wrote a 1979 article in the prestigious journal Science. This paper became the seminal work in the field — and likely the source of its complete undoing.
Terrace carefully reviewed video footage of human-ape interactions. Specific frames and traced images from them are demonstrated in the paper. He noticed that the researchers prompted the apes by displaying signs to them, in English grammatical order, before recording the same signs repeated back by the ape. The animal was essentially mimicking the human’s behavior. The ape was aping it.
What about the heart-warming stories of human-ape understanding? Human handlers interacted with the apes for thousands of hours, and occasionally the human interpretation of a string of signs would stand out as interesting. But, this makes the interesting sign combinations look more like generous interpretations of anecdotes that were cherry-picked, or fed to the ape by human handlers, and not a conscious thinking pattern.
What’s more, the meaning of the signs was very easy to over-interpret. Is water bird the intellectual combination of two concepts to indicate a waterfowl? Or is it just rote repetition that a lake and a bird are nearby, combined with generous and wishful human interpretation? Studies in the field generally focused on picking unusual instances out of thousands of hours footage, rather than systematically studying whether apes expressed meaningful ideas. When Terrace did this, he found that interesting sentences began to look like drops in the ocean.
Most of that footage demonstrated the apes producing word salads that contained signs for food or affection they desired. Usually these sentences were very short, and in no sense grammatical. Terrace noted that nearly all Nim’s sentences were two or three words long; extended sentences were very rare. The general pattern was: Nim or me followed by eat, play, tickle, banana, grape, or the like. Human children begin with short sentences. But they rapidly develop the ability to form longer sentences, conveying meaningful thoughts, asking questions, and expressing new ideas. Nim never did any of these things.
Nim once formed a sixteen-word sentence: give orange me give eat orange me eat orange give me eat orange give me you. If that sounds to you more like the nonsense babbling of a parrot, or what your dog might say to you if he saw that you had an orange, and much less like the thoughts of a child, you can see the problem.
This situation was amusingly summed up by another famous researcher. Here is Noam Chomsky, on Nim Chimpsky:
“The ape was no dope. If he wanted a banana, he’d produce a sequence of irrelevant signs and throw in the sign for banana randomly, figuring that he’d brainwashed the experimenters sufficiently so that they’d think he was saying ‘give me a banana.’ And he was able to pick out subtle motions by which the experimenters indicated what they’d hope he’d do. Final result? Exactly what any sane biologist would have assumed: zero.”
Chomsky adds in a final dig:
“Then comes the sad part. Chimps can get pretty violent as they get older, so they were going to send him to chimp heaven. But the experimenters had fallen in love with him, and tried hard to save him. He was finally sent off to some sort of chimp farm, where he presumably died peacefully — signing the Lord’s Prayer in his last moments.”
Likewise, Terrace ultimately concluded that: “[Nim] was unable to use words conversationally, let alone form sentences.”
Koko the Impostor had a nipple fetish
Similar giant flaws appear to run deep through the stories of most signing apes. Former handlers have given accounts of concerning issues in studies of several animals. Let’s focus on the example of Koko the gorilla. Koko’s global warming message was obviously pieced together from a great many pieces of different footage, and no one seriously believes that a gorilla understands anthropogenic climate change. This brilliant lecture hilariously details many further problems with the study of Koko. No actual data was published of Koko’s signs. Instead, years of apparently random utterances were sifted through, and dubiously interpreted, to find heart-warming stories. The rest of the gibberish signing was ignored.
A transcript of a text message session with Koko demonstrates this issue. Read on their own, Koko’s words don’t make much sense. However, her interpreter comes up with explanations for them. Here’s an example, quoted from the text session:
(Handler): Koko, do you like to talk to people?
(Koko): Fine nipple.
(Handler): Yes, that was her answer. “Nipple” rhymes with “people,” OK? She doesn’t sign people per se, so she may be trying to do a “sounds like…” but she indicated it was “fine.”
Is that a fair interpretation? Did the gorilla understand and use rhyming in the English spoken language to play a clever conversational word game? The entire transcript is rife with apparently meaningless responses that may or may not be interpretable. The transcript also broaches the topic of Koko’s apparent nipple fetish, which eventually caused legal problems.
A dead language
Researchers familiar with the field often offer such statements as: “I do not believe that there has ever been an example anywhere of a nonhuman expressing an opinion, or asking a question. Not ever.” Another: “It would be wonderful if animals could say things about the world, as opposed to just signaling a direct emotional state or need. But they just don’t.”
Perhaps the harshest critic, respected semiotician and linguist Thomas Sebeok, concluded: “In my opinion, the alleged language experiments with apes divide into three groups: one, outright fraud; two, self-deception; three, those conducted by Terrace.”
However, to say the field is completely dead is not quite correct. One of the signing apes (Kanzi, now 41) still lives in captivity in an Iowa sanctuary. The researcher working with Kanzi has published on ape language as recently as the past decade. However, there are few, if any, other signing apes still alive, and it appears that none have been trained in decades.
The study of ape sign language is a perfect example of the art and the seductive nature of studies that appeal to our emotions, imaginations, and beliefs. The idea that intelligent animals are like people, that they can converse with us if only we teach them language, is enthralling. Who wouldn’t want this to be true? Often science shows us that reality is more amazing than we possibly imagined. Other times, it’s just wishful thinking.