[LINK] "Tweet, Screech, Hey: Animal Sounds Hint at Evolution of Language"

[rfmcdonald]

Lizzie Wade's ScienceNOW article, carried at Wired Science, reports on an ingenious comparative study of animal communications methods that has insights on the birth of human language.

One reason that human language is so unique is that it has two layers, says Shigeru Miyagawa, a linguist at the Massachusetts Institute of Technology (MIT) in Cambridge. First, there are the words we use, which Miyagawa calls the lexical structure. “Mango,” “Amanda,” and “eat” are all components of the lexical structure. The rules governing how we put those words together make up the second layer, which Miyagawa calls the expression structure. Take these three sentences: “Amanda eats the mango,” “Eat the mango, Amanda,” and “Did Amanda eat the mango?” Their lexical structure — the words they use — is essentially identical. What gives the sentences different meanings is the variation in their expression structure, or the different ways those words fit together.

The more Miyagawa studied the distinction between lexical structure and expression structure, “the more I started to think, ‘Gee, these two systems are really fundamentally different,’ ” he says. “They almost seem like two different systems that just happen to be put together,” perhaps through evolution.

One preliminary test of his hypothesis, Miyagawa knew, would be to show that the two systems exist separately in nature. So he started studying the many ways that animals communicate, looking for examples of lexical or expressive structures.

Vervet monkeys (Chlorocebus pygerythrus), for example, use different alarm calls to refer to different types of predators, such as snakes and leopards. “That’s already headed in the direction of the way in which we use different words,” says co-author Robert Berwick, a computational linguist at MIT. Still, the monkeys always use the calls in the same context—to warn others about predators that currently pose a threat. They can’t arrange the calls in new patterns to talk about predators they saw yesterday, predators they expect to see tomorrow, or the abstract idea of “predator.” Vervet monkeys have a lexical structure but no expressive structure.

Songbirds, on the other hand, appear to communicate using only an expressive structure. Nightingales (Luscinia megarhynchos), for example, can manipulate the patterns of their songs to form up to 200 different melodies. But no individual note has any particular meaning, the way the words in a sentence do. What’s more, an entire song always communicates the same message—the bird’s identity, location, and sexual availability. Birdsong is what Berwick calls a “holistic signal.” A song’s structure may vary, but it “means the same thing every time,” he says.

Because we see independent lexical and expressive structures in animals, it’s possible that human language could have evolved through the combination of similar preexisting systems, Miyagawa, Berwick, and University of Tokyo biopsychologist and co-author Kazuo Okanoya report online this month in Frontiers in Psychology.