March 02, 2004

Monkeys saying things again -- NOT

Andrew Carstairs-McCarthy in the latest Science mentions the view that although animals are capable of "knowing how..., e.g., how to get food", they are not capable of attaining propositional knowledge, or "knowing that", since "only humans have language, by means of which propositions can be entertained or expressed". However, he then argues against this view, citing the work of Herb Terrace in defense of the propositional competence of at least some non-human primates. He says:

Terrace's research with macaques (humble monkeys, not apes) casts doubt on the claim that only humans have declarative knowledge. To obtain a food reward, macaques can quickly learn to punch a sequence of five or more symbols on a keyboard--and do it consistently right, even when the keyboard configuration is randomly shuffled. Thus, the macaques learn not just a sequence of manual movements (analogous to learning a passage on the piano) but an abstract sequence of symbols, whose application involves different motor commands on each occasion.

I say this is not propositional knowledge. This is a complex form of knowing how to get food, as Carstairs-McCarthy almost admits ("To obtain a food reward", he notes). I've said this before and I'll now say it again: contra all the stupid stories people tell and are prepared to believe about communication with apes and parrots and dolphins and other worthy but stubbornly uncommunicative beasts, I see not a flicker of evidence that any animal has ever expressed a proposition (and they may well, therefore, never have understood one, either).

I found something on the web that's relevant to this, something that you may not have seen. It concerns an inside report about Kanzi, the allegedly language-competent bonobo.

Steve Jones, of Perth (Western Australia), posted this message in an archive of discussion about evolution on the American Scientific Affiliation, an organization devoted to "science in a Christian perspective", in which he says that a friend of his on another list (a closed list, apparently, so he felt he should not give the friend's name) had posted the following about Kanzi. Keep in mind while you read the quote below that the experiments with Kanzi are widely regarded as perhaps the most successful experiments on communication ever done with any animal.

I am amused by the Ape Story, mostly because I have met Kanzi! My Philosophy of Mind professor ... was a thorough naturalist, and thought it his responsibility to let us all know about the mental capabilities of our nearest relatives. So, we took a field trip to Rumbaugh's laboratory to see Kanzi, the famed bonobo, and his sister Panbonisha.

I was distinctly unimpressed. My class had been told about Kanzi's ability to understand complex commands, but he refused to perform or obey when we were present. The Rumbaughs had a huge electronic board with hundreds of symbols on it; whenever a symbol was pushed, the board would electronically pronounce the word associated with the symbol. This is how the bonobos are supposedly able to communicate as well as a three- year-old human. Again, Kanzi refused to push any of the symbols; his sister Panbonisha did push some of the symbols repeatedly, but it was difficult to tell if she was really communicating or just having fun making noise. For example, Panbonisha pushed a button repeatedly that said, "Chase." Of course, the trainers were happy to offer extensive commentary and interpretation: "See, she's trying to say that you [one of the humans] should chase him [another human]. She loves the game of chase." All of the alleged communication consisted of the ape pushing a button, and the trainers giving elaborate exegesis thereupon.

My personal opinion is that the Rumbaughs are possibly guilty of a little wishful thinking. And as for the assertion that Kanzi has the language abilities of a 3-year-old, I could read the newspaper at 3. Kanzi's nowhere close.

Comment from me would be almost superfluous. Except to say that there is a similar anecdote about a different chimpanzee reported in Joel Wallman's book Aping Language about a native ASL signer who was once employed, along with some other graduate students, to hang out with a famous allegedly signing chimp, and keep notes on what she supposedly said. It turned out he got the reputation of being the slacker on the project, because he never wrote down very much in his notebook. He said there wasn't much to write. The non-native signers were saying "Ooh, look, isn't that the sign for water, she must be thirsty" and wrote down that the chimp had asked for water, and this native signer just wasn't seeing that anything in his native language had been uttered at all. (I have since learned that the native signer in question, whom Wallman does not name, was Ted Supalla, now a Professor of Brain & Cognitive Sciences, Linguistics, and American Sign Language at the University of Rochester.) Much the same story as reported above for Kanzi, in other words. As for an actual declarative sentence, a statement that something was true or false? Fuhgeddaboutit.

I've said before and I'll say it again (what I tell you three times is true): I do not believe that there has ever been an example anywhere of a non-human expressing an opinion, or asking a question. Not ever. It would be wonderful if animals communicated propositionally -- i.e., could say things about the world, as opposed to just signalling a direct emotional state or need. But they just don't.

Posted by Geoffrey K. Pullum at March 2, 2004 06:42 PM