The brains of birds are not birdbrains

Holy cow! Birds got grammar!
A new study says birds can learn language rules thought exclusive to humans
By ANDREW CHUNG

The next time you're tempted to call someone a "bird brain," you might want to think twice.

It turns out that at least one bird, the European starling, can learn a kind of grammar that, in the past, language theorists thought separated humans from animals.

And why not? Other animals exhibit behaviours once considered the exclusive domain of humans. Take the use of tools — chimps and gorillas use sticks when they dig for termites.

But the idea that a bird can learn grammar rules, as starlings did in laboratory experiments, is startling.

The starling, with its heavily speckled body and arrowhead beak, is an intellectual heavyweight in the bird world.

Mozart bought one, quite impulsively, after he heard it in a pet shop singing his just-finished piano concerto in G-major. Starlings use a vast array of sounds in their songs, from rattles to warbles to screeches, and can mimic both other birds and foreign sounds, including human words.

So the species was well suited for the tests psychologist Timothy Gentner, of the University of California at San Diego, wanted to put it through.

Linguists, most notably famous MIT theorist Noam Chomsky, have held for years that it is only humans who are capable of doing a grammatical trick that goes beyond stringing a series of words together — that is, placing words in sub-clauses within a sentence to give it an entirely new meaning, and to theoretically do this endlessly.

It's called recursive centre-embedding, or, more simply, recursion. For instance, while you could say, "The dog chased the cat with grey fur," you could also say, "The dog, which just bit the little boy, chased the cat with grey fur."

You could go even further and say, "The dog, which just bit the little boy, and which has only three legs, chased the cat with grey fur."

Two years ago, Harvard researchers tried, and failed, to demonstrate the use of recursion in cotton-top tamarin monkeys.

Then Gentner and his co-researchers tried it out on starlings. But instead of testing whether the creatures instinctively knew the grammar rules, they sought to teach them — in their language, of course.

Starlings sing long songs based on recurring individual sounds, or "acoustical units." For the experiment, the researchers recorded and used eight different rattles and eight different warbles from a single male to create thousands of new songs.

In the non-recursive song, a warble was always followed by one of the eight rattles: As the researchers themselves put it — warble-rattle-warble-rattle.

The recursive pattern took a warble, followed that with a different warble, followed by a rattle and then another rattle. In other words: warble-warble-rattle-rattle.

After thousands of trials, nine out of 11 starlings were successfully trained in sound-isolation cages to press a button if they heard the recursive song pattern and not to press it if they heard the simple pattern. They were rewarded with food if they got the answer right and punished with temporary darkness if they got the answer wrong.

The research is published in the current edition of the journal Nature.

"It invigorates the debate about what makes human language unique," Gentner said. "Previously we thought that any non-human animal lacked this basic machinery."

The finding tells us a little more about ourselves, he also said.

"We're not saying they're capable of learning human language. But they can learn a type of patterning rule that is widespread in human languages. So we've got something in common with these little birds — part of the way humans see the world is shared."

The researchers did a series of control experiments to ensure the birds weren't just rote memorizing the patterns. The birds were also able to learn longer strings of warbles and rattles and, Gentner said, recognize novel patterns: "They could say, `Yes, that's a valid pattern, or `No, that one doesn't match either of the rules I have learned."

Unlike in humans, however, the birds' abilities have nothing to do with differences in meaning, or semantics

"Patterns by themselves aren't that interesting," Gentner said. "What we (humans) use them for is to code up meaning in very subtle ways."

In the example cited earlier, the dog chases the cat, and not the other way around. "The pattern tells you how to figure out who did what to whom. Nothing about the task we did had any semantic content at all. When we changed the patterns within a particular rule ... there's no change in the meaning of those patterns. So it might be that starlings lack the interface between these sorts of pattern rules and a much broader and richer semantic capacity."

Still, says New York University psychologist Gary Marcus in a Nature commentary on the study, "It's tempting to summarize the differences between humans and other species in a concise phrase, but most posited differences have turned out to be overstated."

And so what of that old insult, "bird brain"?

Says Gentner, "It's not quite so pejorative now, is it."