Are gorillas using sign language really communicating with humans?
What is the bottom line with Koko the gorilla's ability to learn sign language? I know she only communicates through her handler, who seems to engage in a great deal of subjective translation. I saw an excerpt in Harper's Magazine of a supposed Internet chat with Koko a few years ago that made me rather dubious that the gorilla was capable of any use of language. Nonetheless, there is a strong perception out there that Koko has learned to sign. What is the straight dope?
Don't be too hard on Koko, Fabian. If you judged strictly from Internet chat, you'd have to question the linguistic abilities of many humans. (D00dz!) Scientists have debated for years whether gorillas really understand language or are just, you know, aping us. The consensus among animal researchers seems to be that they understand at some level, but are less adept at using language themselves. When I read transcripts of Koko's alleged conversations I often think: Jeez, a trained monkey could do better.
A couple obvious problems present themselves when one looks into this talking-ape business. The first, as you suggest, is that interpretation of the gorilla's conversation, if such it be, is left to the handler, who generally sees any improbable concatenation of signs as deeply meaningful. During the 1998 on-line chat you saw bits of in Harper's (the whole thing is at www.koko.org/world/talk_aol.html) [Editor's Note: archived page], for example, Koko, without being prompted or questioned, made the sign for nipple, which Francine Patterson, her trainer, interpreted as a rhyme for "people." (Patterson further claimed that this was a reference to the chat session's audience.) Even if you buy the idea that gorillas, who cannot speak, grasp the concept of rhyme, this sounds like wishful thinking. Similar examples abound: "lips" is supposedly Koko's word for woman, "foot" her word for man. Koko made a lot of signs, and sometimes expressed desires or other thoughts, but nothing in the transcript suggests a sustained conversation, even of the simple sort you might have with a toddler.
That brings us to the second problem. What constitutes language use? In 1979 Herbert Terrace of Columbia University published a skeptical account of his efforts to teach American Sign Language to a chimpanzee named Nim Chimpsky. Nim accomplished the elementary linguistic task of connecting a sign to a meaning, and could be taught to string signs together to express simple thoughts such as "give orange me give eat." But in Terrace's view Nim could not form new ideas by linking signs in ways he hadn't been taught — he didn't grasp syntax, in other words, arguably the essence of language. (A dog, after all, may understand that bringing his leash to his owner is a sign that he wants to go out, but nobody sees that as evidence of language use.)
Terrace's work was a major blow to talking-ape proponents. But their case started looking stronger in 1990, when researcher Emily Sue Savage-Rumbaugh of Georgia State University presented evidence of language development in a bonobo chimp named Kanzi. One of the more telling complaints made about gorillas like Koko who communicated via sign language was that they often babbled, producing long, apparently meaningless strings of signs. Their handlers would then pluck a few lucky hits from the noise and declare that communication had occurred. Savage-Rumbaugh got around this problem by teaching Kanzi to point to printed symbols on a keyboard, a less ambiguous approach. She claimed that the ape demonstrated a rough grasp of grammar using this system. What's more, when presented with 653 sentences making requests using novel word combinations, Kanzi responded correctly 72 percent of the time — supposedly comparable to what a human child can do at two and a half years old.
Today, from what I can tell, scientific opinion is divided along disciplinary lines. Many researchers who work primarily with animals accept or at least are receptive to the idea that apes can be taught a rudimentary form of language. Linguists, on the other hand, dismiss the whole thing as nonsense. Personally I'm happy to concede that the boundary between animal and human communication isn't as sharply drawn as we once thought. Animals (not just primates — check out Alex the talking African gray parrot sometime) can use language in limited ways. They can respond to simple questions on a narrow range of subjects; they can express basic thoughts and desires. I'll even buy the possibility that some are capable of employing elementary syntax. However, all this strikes me as the equivalent of teaching a computer to beat people at chess — a neat trick, but not one that challenges fundamental notions about human vs nonhuman abilities. I've seen nothing to persuade me that animals can use language as we do, that is, as a primary tool with which to acquire and transmit knowledge. I won't say such a thing is impossible. But in light of the muddled state of the debate so far, the first task is to decide what would constitute a fair test.