Campus & Community

Researchers debate origin of language:

7 min read

Do monkeys have anything interesting to say?

Psychologist Marc Hauser and his colleagues have come up with new ideas about why animals can communicate but not talk. (Staff photo by Stephanie Mitchell)

If chickens could talk, would they have anything interesting to say? Most scholars think not. But Marc Hauser, a Harvard professor of psychology, disagrees with them.

“Actually,” he says, “chickens have interesting things to say, but no way to say them.” And the same goes for monkeys, chimpanzees, and dolphins. “If you study animals in the wild like I do,” Hauser continues, “you see a richness of social relationships, kinships, strategies, and coalitions.”

Birds sing, chimps grunt, and whales whistle, but those sounds fall far short of expressing the richness of their experiences. Their lack of language goes to the question of why humans have it but no other animals do.

That question in turn leads to two major theories of the origin of language. One is the idea that language arises from bird song, dolphin whistles, monkey hoots, and other precursors that extend back through hundreds of millions of years of evolution. The other theory maintains that language is a uniquely human adaptation, or series of adaptations, with no precursors among other species.

Hauser and his colleagues have come up with a third idea, which includes elements of both a long evolutionary history and a recent adaptation. “It’s a hypothesis that hasn’t been put on the table before,” he says, “and it opens the door to a whole new set of research questions about communication and language.”

Hauser, Massachusetts Institute of Technology linguist Noam Chomsky, and Harvard psychologist W. Tecumseh Fitch argue that the difference between animal communication and human language is recursion, the ability to take discrete elements, like words or numbers, and recombine them in a way that creates an infinite variety of expression. Animals can’t do this, so, although they have interesting experiences, even thoughts, they have no way to string elements together to create an unlimited range of descriptions and statements.

Monkeys count

Recursion might not have evolved specifically for the purpose of allowing humans to warn, converse with, or bore each other, the researchers write in the Nov. 22 issue of Science. It could have come from adaptations that help animals navigate or handle numbers. Tamarin monkeys, for example, can count up to four precisely and distinguish larger numbers in the sense that they know 12 objects are more than eight objects. Hauser believes that the human ability to count to higher numbers precisely came only after we evolved language and developed words like “twenty-nine” and “one thousand two hundred forty.”

Natural selection is the great sculptor of evolution that favors the persistence of features, like feathers and hands, which give those who possess them greater survival potential than those who do not. Such selection could have acted on the ability to handle numbers and to get from place to place in a way that produced an unanticipated result – language.

In other words, a capability to do the internal computations involved in navigating between feeding and breeding grounds, or understanding that 12 is greater than four, could have led to a byproduct now enjoyed by only one species.

If things did, indeed, happen this way, it doesn’t mean language arose as a quick leap in evolution. The most probable time for its origin was after the ancestors of apes and humans went their separate evolutionary ways, about 6 million years ago. “That’s plenty of time for language to have evolved gradually,” Hauser comments.

Before that time, other nonverbal communication systems evolved. Take the alarm system used by vervet monkeys in East Africa, for instance. Vervets sound different warnings for different threats. Leopards, eagles, snakes, and baboons all elicit specific calls, because each enemy hunts in a different way. A leopard warning sends vervets running into trees. An eagle alarm has them hide under bushes. A snake signal causes the monkeys to stand on their hind legs. A baboon call sends them scurrying to treed areas.

This system allows vervets to respond to a threat without seeing what’s going on. It’s like someone yelling “fire” in a theater. You don’t need to see smoke or flames to start you toward the exit.

However, such brief warnings are not as helpful to survival as words. They don’t contain useful information, such as “There’s one leopard coming from the right and another behind you.”

The same holds true for other grunts, shrieks, whistles, and songs. The information they provide doesn’t go beyond “I’m the dominant animal in this territory,” “Here’s food,” or “I’m a female/male looking for a mate.” Hauser, who has studied chimps in the wild, notes that they produce specific calls when they find food, but the sounds do not reveal whether it’s bananas or pomegranates.

Dolphins whistle during contact with humans and each other, and they navigate with the help of echoes from clicking sounds. But their sound repertory falls far short of a language. The same is true, as far as we know, for the songs of whales.

Bird song is rich in recombinations of discrete notes, but the melodies don’t lead to the type of recursion involved in language. “No matter how long and complicated,” Hauser points out, “the songs contain only one meaning: ‘I’m a male of this species in this territory, and I’m calling to attract a mate.’”

Artistic birds

Although their theory limits recursion to humans, Hauser, Chomsky, and Fitch encourage their colleagues to look for it in other species. Certain male bowerbirds chew berries to make a colored juice with which they decorate their nests to attract females. Some scholars would brand such tool use as recursive. Hauser insists that “it’s artistic but not recursive.”

It’s another example of animals that boast some of the predecessors of language without the valuable recursive property that makes human conversation and literature possible. Hauser and Fitch demonstrated this forcefully in laboratory experiments with tamarin monkeys. They wanted to determine whether the simians have the capability to learn a statistical rule involving groups of paired consonants and vowels, sounds like sa, me, hi, mo, gu. The rule allows these sounds to be combined in an infinite variety of expressions because it is based on recursion. If the monkeys learned the rule, they would pay less attention to sound pairs that are consistent with it than to sounds that violate the rule. While the clever tamarins can do less taxing forms of number and sound discrimination, they fail to learn this statistical rule.

Hauser and Fitch did the same experiment with Harvard students, and, as you might suspect, the students paid more attention to violations of the pairing rule. “They did this easily without consciously knowing what the rule is,” Hauser points out.

Other tests reveal that tamarins and human infants share an ability to distinguish between the sounds of two different languages, like Dutch and Japanese. Both species can also determine when one word ends and another begins in a stream of adult speech. To Hauser, this means that such perceptual, but not recursive, abilities did not evolve for the purpose of learning language; they existed before speech evolved. The result, Hauser says, “suggests that the recursive property of language may not have evolved until after 6 million years ago, although sensory and motor aspects of language expression were in place well before that time.”

The question however, is still open, the researchers admit. To prove their new theory conclusively requires more experiments to be done, especially with human infants and other species, including chimpanzees, dolphins, and even chickens.