Uriel Cohen-Priva

September 4, 2012  |  Contact: David Orenstein |  401-863-1862
Uriel Cohen-Priva Assistant Professor of Cognitive, Linguistic, and Psychological Sciences
Languages and dialects are constantly changing, droppin’ letters when the context is clear. Linguist Uriel Cohen-Priva studies the evolution of language in a quantitative and testable way.

Listen closely to English speakers around the world and if your ear has been trained by linguist Uriel Cohen-Priva, you’ll notice that the t sound is missing from a lot of words where it is written. In a Cockney accent, for example, “butter” sounds something like “buh-uh.” In another place it sounds like “budder.”

“English tends to do very bad things to t,” said Cohen-Priva, who has joined Brown’s faculty in the Department of Cognitive, Linguistics, and Psychological Sciences.

Pronunciation mysteries like the missing t are more than just matters for Eliza Doolittle and Henry Higgins to sort through. They are examples of how languages evolve. Using sophisticated, interdisciplinary methods, Cohen-Priva has taken on the huge question of why sounds come and go from languages. It’s a core question in linguistics that spans the entire earth across the history — and future — of spoken language.

Cohen-Priva, who earlier this year earned a Ph.D. doing this research at Stanford University, knows that the answer lies beyond human physiology or perception. Those are universal. What varies across time and space and culture, he notes, is context and information. The sounds that tend to drop out are the ones that are redundant because people can infer them from context, even when they aren’t spoken. “Would you fancy jam or buh-uh on your crumpet, then?”

“The point is that across the English language when t is encountered it’s usually recoverable from context,” he said. “Usually you don’t need to hear the t to know the t is coming.”

In a similar vein, the s sound has been making an exodus from Romance languages. French has fewer of them than it used to. Some Spanish dialects are headed that way, too, Cohen-Priva says.

Anecdotes and observations are one thing. Science is another. Cohen-Priva has worked hard to develop a quantitative and testable approach to his research by using information theory to analyze the evolution of a variety of languages, including Arabic. The work involves scoring the informational value of each sound in every context in each language. Otherwise, even if he could make a convincing anecdotal case about why should t or s sounds suffer, he’d still be unable to say why not k or p sounds.

At Brown, Cohen-Priva hopes to expand his research to tackle another question closely related to his thesis: How does the articulatory and perceptual effort of communication affect the frequencies of sounds cross-linguistically?

The observation that sounds such as t, and k are more frequent than sounds such as d and g was first made by George Kingsley Zipf, a Harvard professor, in the 1930s and 1940s.

Cohen-Priva thinks that in this case, a combination of physiology and informational value comes into play. Sounds that are most common should enter many languages if they are easy to say, but only survive if they are not too easily implied by context.

For example, as a Germanic language, Old English was rife with g sounds. But g requires a lot of effort to say and, at least in English, often seemed expendable in context. In many contexts y replaced g. Dæg became “day.” Meanwhile, in other contexts, g has re-entered English more recently as in the word “egg” from Old Norse. In those cases, the g sound has become deeply entrenched and is hardly ever omitted.

Brown struck Cohen-Priva as the ideal place to pursue this kind of research because it is known for freely blending disciplines. His new department, for example, is not just about linguistics but also psychology and cognition.

“I see linguistics as a field of interactions with many other fields,” he said. “Brown is a place where such interactions are welcome and expected.”

Cohen-Priva is also excited to teach. One idea he has is to bring new research methods, such as easy online tools for recruiting and conducting experiments, into the undergraduate classroom. Students can harness Amazon Mechanical Turk to devise and run their own real experiments with people anywhere in the world.

There are a lot of questions to ask, after all, and a vast array of constantly evolving languages in which to ask them.