15 August 2001 18:30 GMT
by Bea Perks, BioMedNet News

Don't underestimate the complexity of language evolution, warned a British computer modeler today. Simon Kirby, research fellow at the University of Edinburgh's Language Evolution and Computation Research Unit, welcomes recent publications by theoretical biologists on language evolution, but fears that their mathematical models oversimplify the problem. Today, he called for a greater multidisciplinary approach to tackle what he brands "the most complex problem in science."
Kirby is both "pleased and frustrated" by research from theoretical biologists on the subject of language evolution. "It under-represents what's gone on before," he told BioMedNet News. Their main contribution, he says, has been "waking people up" to research that's been going on for years.

Biologists submit papers to prominent journals like Nature and Science, whereas linguists stick to journals prominent in their own field, such as Language, he suggests. "It's only recently that people [in the field] have thought we really have to talk to a much wider audience," he said.

The theoretical biologist and language evolution expert Martin Nowak, head of the Program in Theoretical Biology Institute for Advanced Study at Princeton University, has published on language evolution in both Science and Nature in the past year. Last month he wrote an Opinion article in the journal Trends in Cognitive Sciences, summarizing his latest work. The idea of a biologist writing for cognitive scientists impresses Kirby: "He's trying to do what we should have been doing all along!"

Nevertheless, this and a related article in the September issue of Trends in Ecology & Evolution, by the mathematical biologist Michael Stumpf, at the University of Oxford's Zoology Department, ignore contributions from the field of linguistics, says Kirby.

Both articles review Nowak's application of evolutionary biology to the understanding of language evolution. The approach is "solidly set in the framework of evolutionary game theory," writes Stumpf. "Individuals (the players) that fare best in one generation will produce relatively more offspring in the next ... Nowak et al. connect fitness to linguistic capacity."

Syntax and grammar, in particular, offer an evolutionary advantage, suggests Nowak. A non-syntactic language might have separate words for a lion running, a lion sleeping, a zebra running and a zebra sleeping. Whereas a syntactic language might have nouns (lion, zebra) and verbs (run, sleep), notes Stumpf. With just those four options (lion runs, lion sleeps, zebra runs, and zebra sleeps), grammar has no advantage. But with a growing list of verbs and nouns, grammar will become advantageous - just one extra verb could apply to many nouns, cutting the cost of learning more words.

But it's not that simple says Kirby, "the unique thing about language is that its not like any other evolutionary system."

Kirby and his team have devised a computational model that he claims takes into account more of language's true complexity. Language, he says, evolves on three levels: Individuals learn words, syntax evolves within populations, and a linguists have recently uncovered a third system - an intermediate between the two. "We're only just beginning to see how that works," he said.

During the cycle of learning and production of behavior "the languages themselves seem to evolve," said Kirby, who likens the system to Adam Smith's definition of an economy operating under "an invisible hand." Language develops thanks to the combined effects of many individuals, says Kirby, an aspect he says is not taken into account by the mathematical models of Nowak. Kirby is preparing a paper on his latest findings.

But not all cognitive scientists question the biologists and their mathematical models. "I think highly of Nowak's work," Stephen Pinker told BioMedNet News today. Pinker, professor in the Department of Brain and Cognitive Sciences at the Massachusetts Institute of Technology, is renowned for his work on linguistic behavior and the evolution of language. Writing in a commentary on Nowak's paper last year in Nature, he noted: "Language, like sex, aggression, and cooperation, is a game it takes two to play ... Game theory can provide the external criteria for utility enjoyed by the rest of evolutionary biology."