Doug's Darkworld

Well, an acquaintance showed me some exciting video the other day. It was of Kanzi, a Bonobo that has learned to communicate with humans by pointing at symbols. Kanzi knows thousands of symbols, and videos of him are all over youtube. It’s pretty impressive stuff. Kanzi can give and understand a vast array of commands, and interacts with his handlers regularly using the symbols. To primatologist Sue Savage-Rumbaugh, who has been studying Kanzi all Kanzi’s life, Kanzi exhibits “advanced linguistic aptitude.” Kanzi has even been interviewed on TV, heck, how many people can say that? Is the Kanzi the breakthrough primatologists have been striving for for decades, have humans and an animal learned to communicate? Well, yes. The more important question though is this, is Kanzi the the holy grail of animal communication research, has Kanzi learned to speak?

Alas, while there is a lot to be learned from Kanzi research, don’t place any pre-orders for handy Bonobo house servants. Let’s start from the beginning. In 1969 a chimpanzee named Washoe rocked the scientific world, the first chimp to learn sign language. Washoe was a media sensation, and launched a whole raft of primate sign language research. People everywhere loved the idea that chimps could talk. Sure, their vocal cords can’t pronounce human words, but with sign language, that barrier was broken! Unfortunately, upon closer examination, Washoe well, washed out. Her handlers had been wildly optimistic about their interpretations of many of her hand movements. Even one of her most famous examples of “speech,” her making the signs for water and bird upon seeing a swan, isn’t particularly amazing. A swan is a bird, and it was on water, all it really showed was that Washoe knew the signs for water and bird. Science moved on, and while a few researchers went forward, other than in the popular perception, signing chimps were a dead end.

Then, along came Sue Savage-Rumbaugh and Kanzi. No messing around with ambiguous hand movements, by learning actual symbols, his communications were clear. Kanzi learned thousands of symbols, and could use them to signal his wants and even to some extent communicate his internal states. Watching the videos of him is pretty amazing, at least on a  superficial level. Kanzi can hear complex commands and act on them, surely that means he is using language similar to how humans do. Had Sue Savage-Rumbaugh done it, was Kanzi the first animal to speak with a human?

Alas, no. There’s a number of problems with the “Kanzi is speaking” scenario. The first is how he acquired language. When human babies start learning words, they almost immediately begin constructing sentences out of them. And as they learn more words, their sentences get longer and more complex. When Kanzi (or other “talking” chimps) start to learn words, they pretty much don’t make sentences out of them. And as they learn ever more words, their sentence construction remains at their initial very modest levels. Kanzi’s average sentence length is … 1.15 words. In other words, Kanzi for the most part uses exactly one symbol to express himself. And while Kanzi’s understanding of symbols might seem impressive, it’s more substance than real. Yes, Kanzi seemingly can understand commands involving several words, but that is not necessarily language. IE if one tells Kanzi to “put the doll in the bucket in the other room” all Kanzi really has to know is that he is expected to manipulate the doll, the bucket, and the room. That’s not language.

More accurately, Kanzi does not appear to understand grammar at all. Grammar is how words strung together modify each other, the essence of language. IE take these two sentences, “Man bites dog.” and “Dog bites man.” A human child can understand the clear distinction between these two sentences almost as soon as they start learning to speak. Kanzi can’t, when carefully tested with simple sentence pairs like this, his “understanding” doesn’t rise above chance levels. Despite learning language for decades, Kanzi is 26, he doesn’t understand grammar at all. As one primatologist puts it, no ape has ever asked a question or expressed an opinion.

Will humans ever communicate with animals? Not looking good, human’s facility with language most definitely is something that no animal, no matter how clever, has ever demonstrated. Is there a lesson here? Of course, I’m always illustrating some point or other. The main point being how people’s public perception of science is often at odds with reality. Most people one talks to about signing and symbol using chimps are absolutely convinced that indeed, these animals are “speaking.” I suspect this is a combination of wishful thinking; both on the part of the public, the media, and on the part of the very sincere researchers involved. Sadly, just because a handful of researchers and the public thinks that something is a scientific reality, doesn’t actually make it so.

Lastly, Kanzi is a curious example of borderline research. Nothing is ever black and white, the boundaries between science and nonsense aren’t as clear cut as many would believe. Talking apes aren’t pseudoscience, actual scientists are working in the field. And they sincerely believe they are onto something. I suspect the amount of research devoted to this will decline over time, that’s usually the case with unproductive lines of research. Still, all this talking chimp research has at least cleared up one thing: Chimps can’t be taught to talk.

(The above image is claimed as Fair Use under US copyright law. It’s not being used for profit and is central to illustrating the post. I got it from this fine site, which presumably holds the credit and copyright. And yes, I have been ill. I have returned and am blogging again. That’s good or bad depending on one’s perspective I suppose.)