What constitutes proof in ape language studies? How would you go about proving that an ape has acquired language? The truth is that nobody knows. Many people will tell you what doesn't constitute proof, but the scientific community has not agreed on a standard, that if met, would definitively prove that an ape has acquired language.

Writing on this topic, Sue Savage-Rumbaugh, Stuart Shanker and Talbot Taylor had this to say: "...[I]f at the end of extensive experimental research, a reputable research team proposes a definitive answer [to one of the key questions in ape language research] ... there are no agreed methods by which the scientific community can determine whether that conclusion is or is not well-founded." (Apes Language and the Human Mind 1998. 142.)

The problem of proving that an ape has acquired language is not really different in theory from the problem of proving that a human being has acquired language. And in fact, some human beings face a similar fight for the recognition of their own independent communication. Those are the people who use assisted communication.

As long as they use their articulatory apparatus to speak out loud, most people enjoy the benefit of the presumption that they are in fact speaking, and not just making noises that sound like speech. People with normal speech don't have to prove anything. But for those with neural damage that prevents them from having normal control over their movements, the path to recognition that they do understand language and are able to use it is a difficult one. The controversy surrounding facilitated communication still rages.

In this hub I will explain the problem and offer some unconventional examples that I believe ought to be considered as proof, both in the case of humans and of apes.

In the sixties and early seventies there was a lot of research into non-humans and language ability. Many of the subjects of these experiments were apes, although research was also conducted with other animals, such as parrots and dolphins.

One of the most successful projects of this period was the work of Beatrice and Allen Gardner with a female chimpanzee by the name of Washoe. Washoe was brought up in a human home, using a method called cross-fostering. She was taught American Sign Language and acquired an impressive vocabulary that she could use in context to express her wishes, desires and opinions.

Herbert Terrace attempted a similar experiment with a male chimpanzee whom he named Nim Chimpsky, after the famous linguist, Noam Chomsky. Like the Gardners, it was Terrace's intention to raise Nim in a human home and to teach him sign language. However, Nim's early experiences were very different from Washoe's. For one thing, Terrace did not volunteer to adopt Nim himself. He was an unmarried man, and he did not feel up to the challenge of raising a baby chimpanzee by himself. Instead, he asked a former research assistant of his to adopt Nim into her home. However, conflicts concerning parenting methodology and other matters arose, and Nim was abandoned by his adoptive mother at an early age and brought up by a whole string of different assistants, none of whom served as the central figure in Nim's life. Deprived of a stable family life, Nim experienced attachment issues very similar to those that human children deal with when they are denied a strong relationship with a reliable parent.

Terrace was also dealing with serious funding issues. Bringing up a chimpanzee in your own home does not cost more than bringing up a human child, assuming you are willing to be a stay-at-home parent and supervise all day long. However, Terrace was not bringing up Nim in his home. He needed a special place for Nim to stay which was not anybody's home, and he had to pay people to stay with Nim, and it was very expensive. Terrace also believed, as a scientist, that every waking moment of Nim's life had to be documented, either on film or in copious research notes, and this was very expensive, too. Eventually, Terrace gave up because he was unable to find sufficient funding, and the return on his investment in Project Nim seemed too low compared to the trouble that Nim was getting into.

Nim was sent back from New York to the place of his birth in Oklahoma, and he was forced to deal with very harsh conditions and to face the unpleasant realization that as a chimpanzee in a human society he had no rights. At one point, he was sold into medical research, and although many campaigned for his release, incuding Terrace, his remaining life was pretty bleak. If you are interested in learning more, I recommend the biography written by Elizabeth Hess.

(Washoe also ended up in the same facility from which she was taken as an infant, but she was lucky to have Roger Fouts, an assistant and student of the Gardners', with her, and he saw to it that her experience was not as stressful as Nim's nor nearly as lonely.)

About the same time that Nim returned to Oklahoma, Terrace began to review the tapes of Nim signing. He came to the conclusion that Nim was not using language spontaneously to express himself. It seemed to him that Nim was signing to please his trainers. Terrace published his conclusions, and then what happened was this: all researchers working in animal communication were invited to a conference which retrospectively is known as The Clever Hans Conference. There, not only was Nim Chimpsky denounced, but it was suggested that Washoe, and all other animals engaged in language experiments, were also not actually using language. They were all compared to Clever Hans, a German horse who had used cues from humans in order to answer math questions correctly. In 1907 it had been discovered that Clever Hans could answer questions correctly only when the human with him knew the answer and Hans could see the human. Now it was being suggested that the same sort of cuing was going on in all animal language experiments.

I quote from Sue Savage-Rumbaugh and Roger Lewin: "For some time during the mid-1970s, Thomas Seobok , a linguist at Indiana University, had bee expressing strongly negative views on ape language research. And in May 1980 he organized a conference under the auspices of the New York Academy of Sciences, which made his position brutally clear. The conference was called 'The Clever Hans Phenomenon: Communications with Horses, Whales, Apes and People'. ... There was even a move, fortunately thwarted, to have the conference vote for a ban on the research. At a press conference at the end of the meeting, Sebeok expressed his views most stridently of all: 'In my opinion, the alleged language experiments with apes divide into three groups: one, outright fraud; two, self-deception; three, those conducted by Terrace.'" (Kanzi: The Ape at the Brink of the Human Mind,1994, p.50-51.)

Many researchers in animal language experiments lost their funding following the Clever Hans Conference. Others persisted. Irene Pepperberg found novel ways of funding her work with Alex the Parrot. Francine Patterson went on working with Koko the gorilla. Sue Savage-Rumbaugh continued to work with chimpanzees and bonobos.

While working with a parrot who speaks, as does Pepperberg, one is less obviously subject to the accusation of cuing, because clearly nobody is telling Alex which vowels and consonants to put together in what sequence in order to answer the questions that Irene Pepperberg poses. If the social contact with Pepperberg is giving Alex any sort of clues, it might be on the content of his answer, but not its linguistic form.

Parrots are lucky in that their articulatory apparatus, while quite different from ours, is capable of producing comprehensible speech. The other great apes are not able to produce speech that humans can comprehend. That is why alternative methods of communication, such as sign language or lexigrams, have to be used. The moment it's not audible, the researcher opens himself up to the accusation that not just the answer itself, but also its linguistic form, is being cued.

It's one thing to be able to determine that the green key is the bigger one. It's quite another to know how to say "green" in response to a question like "which color bigger?". Even if Alex could somehow tell that Irene thought the green key was bigger (maybe by the way she was looking at it), she can hardly be accused by her body language to have caused Alex to move his vocal tract in such a way as to produce the word "green." Also, she cannot have clued him into the fact that green was a color. So clearly he knows what the word "green" stands for, and he knows what the word "color" stands for. And he knows that the question "which color?" requires an answer that is a color. All of this is very big in and of itself! And really, in a study about linguistic ability, who cares if a parrot is good at estimating sizes?

I'm not saying that I doubt in the slightest that Alex knew that the green key was bigger. I'm convinced in the validity of the entire demonstration. But the point, for purposes of standards of proof, is that it doesn't matter if Alex gave the right answer without cuing. He's demonstrated his linguistic competence, either way!

With apes who use lexigrams, the whole process is considered more suspect, because if the researcher inadvertently looks at the right lexigram for the answer, this might be a cue for the subject, not just for picking out the right answer, but also for picking out the right word to stand for the answer. Because of this, researchers in ape language studies are expected to conduct double blind tests, where neither the human nor the subject know the question or the answer. Many ingenious methods have to be found to minimize cuing.

In the following clip, Sue Savage-Rumbaugh is wearing a welder's mask to minimize cues to Kanzi about what she expects him to do as she gives him novel requests to comply with. Kanzi has never heard these sentences before. This is evidence of comprehension of English syntax.

Notice how patient both Sue Savage-Rumbaugh and Kanzi have to be in order to go through the entire list of novel sentence for purposes of proof. Clearly, Kanzi understands what he is told to do. Equally clearly, after a few tries, this exercise gets really old for both of them. They have to keep going, because they have set themselves this artificial task for the purpose of winning points with the scientific community.

In the following clip, Kanzi is asked to point at lexigrams for the words pronounced by a researcher who is not in the room. Kanzi hears the request through headphones, so Sue, who is in the room with him, will not know what he is being asked, and therefore will be unable to clue him in to the right answer.

When I watch these clips, the first thing that strikes me is how patient Kanzi is. Yes, he knows his lexigrams, and he understands English. But what is truly amazing to me is that he'll sit still through these really unchallenging tests without protest!

Kanzi is a bonobo, and he's a unique individual, too. I really admire both him and Sue for their patience and perseverance.

My adopted son Bow is a common chimpanzee. While I believe he is just as smart as Kanzi, I have to admit that he's a lot more stubborn and much less cooperative.

Bow started spelling out words in the summer of 2007 at the age of five and a half years. Up to that point, we were using lexigrams in Hebrew and English, but we made little progress, because Bow pointed faster than we were able to see. Then Bow started taking our hands and using them as pointing devices, making sure we took in what he was pointing at before proceeding further. From pointing at words in standard spelling he went on the spell out words by pointing at letters. His achievement is remarkable. There is just one problem: proof.

How do we know that with the physical contact between us during the process of pointing, we are not inadvertently cuing Bow as to what to say and how to say it?

Bow won't spell when there is no one in the room. Bow will not talk to strangers. He refuses to answer questions that he thinks we already know the answer to. He uses language only to communicate, and I cannot bribe or cajole him to use it for anything else.

When you ask Bow a question, he'll just as soon lie than tell you the right answer. So, how can we get any kind of double blind testing into place?

The answer is that at the moment, we can't. We are working on a computer interface for Bow, and we hope to make our demonstrations more objective by allowing the computer to sound out the words that Bow types even when we can't see him typing. But that is still in the works, and while we wait for all the kinks to be ironed out, I've discovered some old clips that I believe prove Bow the one spelling out the words, because he's saying things that the person who is with him cannot possibly know.

In the following picture, you see Delight Wang standing in our living room with my daughter Sword. Delight and her mother June Sun stayed with us for six months in 2003.

In the picture below, Delight is holding Bow while Sword looks on. When we spoke among ourselves we didn't call Delight by her English name. Her Chinese given name is 忻之. It's pronounced like this: [ʃɨ n ʐ ɨ] .

At the end of August 2007, I showed the pictures in our photo album to Eden Michaelov. Eden did not know Delight, and she did not know Delight's Chinese name. She took the sheets from the album containing the photo of Bow, Delight and Sword, and she asked Bow, at a time when I was not there, what the name of the little girl in the photo was.

If Bow had thought this was a test for purposes of proof, he probably wouldn't have answered. But he took Eden at her word. He knew that Eden didn't know, and he was willing to tell her.

Eden: I want to know something. I saw this picture. Who is this girl? What is she called? That's Bow, right? Who is holding Bow here? Who is holding him? This is Sword. Who is this?

Bow: (spells) shin nun.

Eden: I don't know. I don't know her. What is her name. Do you know her? This is Sword; who is this? Tell me!

Bow: (spells)  shin nun gimmel.

Eden: (Tries to sound it out.) Shanag?

Eden: What are you doing? Tell me! What is her name?

Bow: (spells) shin nun.

There were three consonants in the Chinese name. When we write in Hebrew, it is common to specify only the consonants. Bow chose the three Hebrew consonants that most nearly conformed to the Chinese pronunciation. Hebrew doesn't have a [ʐ ] sound, but the gimmel stands for [g] the closest available sound, and what native speakers normally use to render a [ʐ ].

What are the odds that someone selecting three letters out of the twenty-two letter Hebrew alphabet in sequence at random would have chosen this sequence? The odds are 1 in 10,648.

Eden didn't know the answer. No one else was there besides Bow. Is there any other explanation than that it was Bow who selected the letters?

The next example is from the following year. Our Intern in the summer of 2008 was Katie Thurston She stayed with us till the beginning of September.  Katie is English, and she had been a student at the University of Edinburgh. Not only did she not speak Hebrew, something that Bow held against her, but her pronunciation of English was not American, and hence unfamiliar. At first Bow mistrusted her. He told me she talked funny. Then he started making up stories about her. He said she was an English spy, and that her name was Ruthie. He made up all sorts of interesting stories about what happened when I went out and left him alone with Katie.

Bow had always given new interns a hard time, but now that he could use language , there was an added element to his hazing: he tried to stir up trouble by making up tales. Eventually, though, Bow came to respect and like Katie, and they developed a close enough relationship that he felt safe with her. He began to spell with her. She even tried to learn a little Hebrew from him, but he refused to teach her, claiming that she was stupid, because she did not know Hebrew.

Bow could have helped our cause had he been willing to teach Katie Hebrew. It would have been proof that the words were coming from him and not her, had he been willing to provide English to Hebrew translations. But Bow is not that cooperative. Instead, on one occasion at least, he tried to mystify Katie by telling her in Hebrew what she could understand only when spelled out in English.

In the clip from September 3, 3008, transcribed below, Bow told Katie that he was full in three ways:

1. He used the number 7, which when spelled out makes the word for "sated":

   שבע

2. He spelled out the word for "full": מלא
3. When she didn't understand the above, he spelled out in English: "I am full."
   

Katie had picked up a little Hebrew in her three months with Project Bow. She knew the words for "Mommy", "Auntie" and "no." But this did not make her a speaker of Hebrew, and she did not understand what Bow was telling her.

Katie did not know that if spelled out in letters the Hebrew word for seven was identical in spelling with the word for "sated". Katie did not know the Hebrew word for full. There were no other people present, besides Katie and Bow. Is there any better explanation than that Bow was the one who pointed at the letters?

Some have suggested that all the researchers who have worked with Bow and had him spell for them are self-deluded. But how could self-delusion have resulted in the spelling out in context of words that the researchers did not know? What are the odds?

Real language use does not consist in answering questions that everybody already knows the answers to, as in the case of most double blind experiments approved by the scientific community. Real language use is coming up with spontaneous utterances that convey new information that your interlocutor does not know.

I would like to suggest a new standard of proof: if there is no other logical explanation than that the communication must have come from the subject, then we must accept that the subject has mastered language!

(c) 2009  Aya Katz