The Amazonian Language That May Lack a Universal Feature

Recursion is the trick where one phrase can be embedded inside another, indefinitely: "the cat that the dog that the boy chased bit ran." Noam Chomsky and colleagues argued in a 2002 paper that recursion is the one feature unique to human language and present in every human language. Take it away and you don't have language as we know it.

Daniel Everett went to live with the Pirahã, a small group along the Maici River in the Brazilian Amazon, as a missionary in 1977 and ended up a linguist. After almost three decades of fieldwork he published a paper in Current Anthropology in 2005 arguing that Pirahã has no recursion. To say something complex, a Pirahã speaker chains separate sentences together rather than nesting one inside another. He linked the absence to a culture that, he argued, prized immediacy: no creation myths, no fiction, no number system, no fixed color terms.

The response was loud. Chomsky publicly called Everett a charlatan. Other linguists ran new fieldwork and found counterexamples; some of Everett's own former colleagues disputed the recursion claim. Defenders of universal grammar offered a softer line: maybe the Pirahã can form recursive structures and culturally choose not to. Everett's reply was that linguistic theory shouldn't get to dismiss inconvenient field data by waving at unobservable mental machinery.

The debate has not closed. The June 2009 issue of Language, the journal of the Linguistic Society of America, ran a nearly 100-page exchange between Everett and his critics. What survives in any case is the demonstration that one fluent outsider, willing to live in a village for thirty years, can still threaten a research program built on data from couches in Cambridge.