Reproduced with permission of Punch Ltd.

Darwin saw parallels between the evolution of species and of languages

Charles Darwin spotted it. In The Descent of Man, he wrote: "The formation of different languages and of distinct species and the proofs that both have developed through a gradual process are clearly the same." He'd been struck by ideas that William Jones had advanced 50 years earlier, that the similarities between languages as disparate as Sanskrit, Latin, and Old Persian, suggest a common historical ancestry.

These foundations for an entirely new field of research were largely ignored for almost a century. Language wasn't recognized as a heritable trait, subject to processes of natural selection. Instead, studies of biological and language evolution followed different trajectories, evidence of which is still reflected in the departmental structure of most universities. The divide is finally being bridged, even if a common terminology has taken a...


Groundbreaking work by Noam Chomsky laid the foundations for this type of language research, of trying to understand how human language is possible at all. Why, for example, do children all over the world, regardless of their backgrounds, arrive at roughly the same set of rules governing the way they communicate? Natalia Komarova, at the University of California, Irvine, says that such questions are important.

"Every child faces an enormously complex task while mastering her native language. Without formal instruction, she has to figure out all the rules of underlying grammar." Komarova points out that this goal must be achieved using only a limited set of sentences, those actually heard, from the infinite number of possibilities. "In fact," she says, "if you're completely open-minded it's mathematically impossible to guess the correct rule."

Somehow, though, children do manage this impressive feat, overcoming what is known as the paradox of language acquisition. Chomsky's revolutionary insight was in postulating an innate universal grammar, which helps children navigate the infinite realm of choice. Komarova and others, such as Harvard University's Martin Nowak, are beginning to explore the evolutionary origins of the universal grammar by drawing on some of the tools that evolutionary biologists have been developing over the last few decades.

Nowak says that the complexity of the problem has hampered progress. "To study [language] from an evolutionary standpoint requires detailed knowledge of several fields." His own work lies at the center of a rapidly growing literature that uses evolutionary biological tools to yield insight into some of the basic questions on language evolution.12

Nowak points to limitations of current methodologies. "Most traditional linguistic models have assumed that one party is the teacher and the other is the student. However, that's not the case in real life," he says, referring to recent studies of a newly emerging sign language in a Nicaraguan deaf community, whose evolution through successive cohorts of children has been closely followed.3 "Children appear to learn from their peers; they arrive at a consensus by which they communicate effectively." This research supports Chomsky's theory of an innate grammar and provides valuable field data on how new languages evolve, but it greatly increases the complexity of the models theoreticians must now develop. Fortunately, a precedent already exists.


<p>HELLO, HELLO?</p>

Courtesy of Claude Bramble

Can linguistics shed light on animal communication, such as alarm calling in vervet monkeys? (shown above) If so, animal models might in turn help us undestand human language.

At its simplest level, language can be viewed as a sequence of signals between speaker and listener. This is where methods lifted from evolutionary biology come into play. The late John Maynard Smith championed the adoption of game theory,4 which revolutionized ethology during the 1970s and 1980s. Game theory is now applied to a diverse range of research questions, enabling the study of frequency dependence: population-dynamic situations in which an individual's best response depends on what others are doing.

Niyogi, who recently collaborated with Komarova and Nowak,1 says such a conceptual shift is essential. Traditional linguistic approaches can't really tackle more realistic scenarios, such as how a group of children might learn from a more fluid set of teachers, including both peers and a wider, age-structured group, he says. "Linguists are going to have to come to terms with the fact that we need a more sophisticated, population-based setting."

Another tantalizing possibility is to define the limits of the universal grammar. An overly intricate grammar would take too long to learn, as too many examples would be needed to test the validity of each linguistic rule. After all, humans have only a limited time for learning language before they reach adulthood. "This is really where linguistics meets evolutionary biology," says Komarova. "There is a selection pressure to make universal grammar smaller and easier to learn." An ultracompact grammar isn't necessarily the most efficient, though. Larger grammar pools increase flexibility, making it possible to express more complex ideas and thus facilitate innovation.

Cognitive scientist Gary Marcus, at New York University, is interested in the origins of the human mind.5 He emphasizes the complexity of the task ahead and the need for new, more flexible approaches. "It's easy to view language as the product of natural selection, but we shouldn't see it as a single trait shaped by a single selection pressure," he says. Instead, we need to see it as a mixture of cognitive mechanisms, some old and some new, each of which might be subject to numerous differing, even competing, selection pressures.


Simon Kirby, reader in the evolution of language and cognition at the University of Edinburgh, says that the cross-disciplinary approach has helped dismantle the dogma that all languages are equal. "One of the main contributions coming from the evolutionary biology side is that languages improve qualitatively," he says, "so a language can become more efficient at conveying information as it evolves.

The flow of ideas between disciplines isn't all one-way traffic. Theories of learning are vital for any attempt to understand human language. "When the story is finally told about how language has evolved, learning theory will play a central role," says Niyogi. He argues that both the linguistics and computer science communities, which together are responsible for the major developments in learning theory, share the view that its role hasn't been sufficiently incorporated into evolutionary models. "You can't simply collapse 50 years of language research into a single parameter, call it "P" and say it's analogous to cultural transmission."

Not all the findings from ethological game-theory studies necessarily map onto linguistics research. Language is a consensus between speakers and listeners. In other biological systems, mutualisms are often kept honest because both partners are continually attempting to exploit one other. But Harvard University's Steve Pinker says that the cooperative nature of language is the most likely factor driving its evolution.

"I don't think that language could have evolved primarily as a technique of manipulation," he says. Most animal signals are clear-cut examples of manipulation, in which signalers and receivers exert selective pressure to produce more effective means of exploitation and resistance, respectively. Pinker draws a distinction between these types of interactions, such as mating calls, and the complexity of language, which requires a huge investment by the listener simply to decode the information. Why, Pinker asks, would the listener bother to do this if only to be manipulated as a result?

Nevertheless, it's clear that many of the other central evolutionary concepts, such as kin selection, are vital to language. "Human interactions are not entirely antagonistic. There are kin, spouses, and reciprocators, where the relationship is largely (albeit not completely) nonzero-sum and positive," says Pinker, drawing attention to the fact that often both parties can benefit through successful communication. "I think the cooperative story has to be basically right. Lying and manipulation are a parasitic overlay."


Comparing animal signals, morphologic structures, genetic systems, and learning processes across species has paid dividends, according to Tecumseh Fitch at the University of St. Andrews' School of Psychology in Scotland. "Many historical linguistics methods can be successfully applied to nonhuman systems," says Fitch. "Bird and whale-song dialects show geographic variation and changes through time, just like human language does." This raises the possibility of using other animal species, what Niyogi calls a "language fruit fly," to better understand human speech.

Clearly these are exciting times for language research. And there's no shortage of raw materials, says Kirby. "We have an embarrassment of data. Language is going on all the time, all around us."

Interested in reading more?

Magaizne Cover

Become a Member of

Receive full access to digital editions of The Scientist, as well as TS Digest, feature stories, more than 35 years of archives, and much more!
Already a member?