Cornell Chronicle index page Table of Contents Front page of this issue

It's 'status' that decides whether a language survives, CU engineers say

By Bill Steele

The Tower of Babel might get built after all. While thousands of different languages are spoken around the world, 90 percent of them are dying and are expected to vanish in the next few decades.

But Cornell engineers have come up with a mathematical model that, for the first time, quantifies "language death" and may offer strategies for those who want to preserve an endangered language.
Professor Steven Strogatz, left, and engineering graduate student Daniel Abrams collaborated on a mathematical model that quantifies "language death." Robert Barker/University Photography

The key factor is status, according to Cornell graduate student Daniel Abrams and Steven Strogatz, Cornell professor of theoretical and applied mechanics, who described the model in the Aug. 21 issue of the journal Nature. Others, they say, have used mathematical modeling to study the evolution of grammar, syntax and other structural features; but they believe this is the first attempt to quantify competition between languages.

The status of a language is determined by the social and economic opportunities it offers its speakers, the researchers say in their paper. In Wales, for example, "Parents want their kids to speak English for the opportunity," Abrams explained. "If they only speak Welsh, they're not going to be able to move to London and get a good job."

The researchers tested the model against historical data on the decline of Welsh, Scots Gaelic and Quechua, among other endangered languages. When population figures are plugged in, the model produces a family of curves depending on the value assigned to status, and in each case one of the curves agreed with the published data on language decline. Most of the data came from published census figures, but in the case of Quechua, Abrams traveled through Peru, interviewing Catholic priests to find out when the last Mass had been celebrated in the old language.

Quechua, Abrams noted, is a language that would not be considered endangered by most measures. It was the common language of the Inca empire and still has some 10 million speakers scattered over Peru, Ecuador, Argentina, Bolivia, Colombia and Brazil. "It was once what English is today, the common language people used to interact even when they had their own local language, but it's now disappearing very quickly," being replaced by Spanish, Abrams said.

The model predicts that when languages are competing for speakers, the balance always is unstable. "Multilingual" societies, like Switzerland, really consist of mostly separate monolingual populations living side by side, the researchers said. Even bilingual individuals are not truly so, Abrams said. "People almost always have a mother tongue or speak one language better," he said.

"We don't take into account social structure or geographic distribution," he added. "The amazing thing is that it still works very well."

An example of a situation where the model doesn't work is in the persistence of Spanish in the United States, which he attributes to a constant resupply of native speakers.

The conclusion is that a language can be preserved by boosting its status. In Quebec, Abrams pointed out, 20 years ago French was disappearing, but the provincial government passed laws requiring its use in certain places, adopted immigration policies that favored French speakers and even ran an advertising campaign saying, in effect, "French is cool."

But in a lot of other places, Abrams said, English has a very high status and, "This is driving the disappearance of languages around the world."

The Nature paper is titled "Modeling the Dynamics of Language Death."

September 11, 2003

| Cornell Chronicle Front Page | | Table of Contents | | Cornell News Service Home Page |