Talking Dogs

/a>; also mentioned at Normblog):

She [the German researcher] and her colleagues conducted a series of experiments on the 9-year-old border collie after his owners claimed he knew the names of about 200 objects — his collection of toys, balls and stuffed animals.

In the first experiment, the researchers put 10 of Rico’s toys in one room and Rico and his owner in another. The investigators then instructed the owner to order Rico to fetch two randomly selected items. As Rico ran into the other room and began searching for the items, he could not have picked up any hints from his owner because the owner was out of sight.

In 40 tests, Rico got it right 37 times, demonstrating he had a vocabulary comparable to dolphins, apes, sea lions and parrots that have undergone extensive training.

Notice the phrasing: the dog, apparently, has a “vocabulary”.

The findings are the latest evidence that animals are capable of more complex communication than had been thought, and that dogs, in particular, are especially astute at comprehending their human companions.

Rico’s talents are prompting some scientists to go so far as to speculate that dogs might even have the capacity to speak — like Mister Ed, the talking horse in the 1960s television show — if only they had the necessary anatomy.

“If Rico had a human vocal tract, one would presume that he should be able to say the names of the items as well, or at least try to do so,” said Sue Savage-Rumbaugh, who studies animal communication and intelligence at Georgia State University. “It also raises the issue of whether Rico and/or other dogs or other mammals might already be trying to say words, but have great difficulty being understood.”

Oh bollocks. What’s happening here is surely not that remarkable. Dogs have excellent hearing: they can discriminate between different words. But let’s not get carried away: they don’t “understand” the words: they respond to them as commands. If for example the dog’s owner had trained the dog differently, it would go and fetch a ball, say, on the command “fetch”, a book on the command “bingo”, and a teddy bear on the command “vomit”. There’s no vocabulary, no understanding of the word itself, merely an understanding of what’s required of it.

The article continues:

But scientists have since determined that many species have complex vocalizations that enable them to communicate sophisticated information among themselves. And researchers have been able to teach many creatures to recognize human symbols and language. A bonobo ape named Kanzi, for example, can understand dozens of symbols. Alex, an African gray parrot, can identify a wide array of objects, shapes, colors and materials. And recent studies have shown that dogs are remarkably adept at reading human gestures.

But again the conclusions – helped by the carefree use of that word “understand” – are simply not warranted by the evidence. Sue Savage-Rumbaugh has spent years trying to get bonobo chimps to communicate through language (in the great tradition of Sir Arthur Streeb-Greebling’s attempts to teach ravens to fly underwater), but her conclusions are always marred by her determination to present the results in the best light possible. Kanzi can combine symbols (pictures) in complex ways to get what it wants, certainly, but that can (and should, by Occam’s razor) be explained by straightforward learning, as encountered in simpler form with those staples of the experimental psychology lab, rats and pigeons, rather than asserting, as Savage-Rumbaugh constantly does, that if only her poor beasts had human vocal tracts they’d be chattering away in English.