• October 1, 2014

Angry Words

Will one researcher's discovery deep in the Amazon destroy the foundation of modern linguistics?

review-art-5831

Illustration by Steve Brodner for The Chronicle Review

Enlarge Image
close review-art-5831

Illustration by Steve Brodner for The Chronicle Review

A Christian missionary sets out to convert a remote Amazonian tribe. He lives with them for years in primitive conditions, learns their extremely difficult language, risks his life battling malaria, giant anacondas, and sometimes the tribe itself. In a plot twist, instead of converting them he loses his faith, morphing from an evangelist trying to translate the Bible into an academic determined to understand the people he's come to respect and love.

Along the way, the former missionary discovers that the language these people speak doesn't follow one of the fundamental tenets of linguistics, a finding that would seem to turn the field on its head, undermine basic assumptions about how children learn to communicate, and dethrone the discipline's long-reigning king, who also happens to be among the most well-known and influential intellectuals of the 20th century.

It feels like a movie, and it may in fact turn into one—there's a script and producers on board. It's already a documentary that will air in May on the Smithsonian Channel. A play is in the works in London. And the man who lived the story, Daniel Everett, has written two books about it. His 2008 memoir Don't Sleep, There Are Snakes, is filled with Joseph Conrad-esque drama. The new book, Language: The Cultural Tool, which is lighter on jungle anecdotes, instead takes square aim at Noam Chomsky, who has remained the pre-eminent figure in linguistics since the 1960s, thanks to the brilliance of his ideas and the force of his personality.

But before any Hollywood premiere, it's worth asking whether Everett actually has it right. Answering that question is not straightforward, in part because it hinges on a bit of grammar that no one except linguists ever thinks about. It's also made tricky by the fact that Everett is the foremost expert on this language, called Pirahã, and one of only a handful of outsiders who can speak it, making it tough for others to weigh in and leading his critics to wonder aloud if he has somehow rigged the results.

More than any of that, though, his claim is difficult to verify because linguistics is populated by a deeply factionalized group of scholars who can't agree on what they're arguing about and who tend to dismiss their opponents as morons or frauds or both. Such divisions exist, to varying degrees, in all disciplines, but linguists seem uncommonly hostile. The word "brutal" comes up again and again, as do "spiteful," "ridiculous," and "childish."

With that in mind, why should anyone care about the answer? Because it might hold the key to understanding what separates us from the rest of the animals.

Imagine a linguist from Mars lands on Earth to survey the planet's languages (presumably after obtaining the necessary interplanetary funding). The alien would reasonably conclude that the languages of the world are mostly similar with interesting but relatively minor variations.

As science-fiction premises go it's rather dull, but it roughly illustrates Chomsky's view of linguistics, known as Universal Grammar, which has dominated the field for a half-century. Chomsky is fond of this hypothetical and has used it repeatedly for decades, including in a 1971 discussion with Michel Foucault, during which he added that "this Martian would, if he were rational, conclude that the structure of the knowledge that is acquired in the case of language is basically internal to the human mind."

In his new book, Everett, now dean of arts and sciences at Bentley University, writes about hearing Chomsky bring up the Martian in a lecture he gave in the early 1990s. Everett noticed a group of graduate students in the back row laughing and exchanging money. After the talk, Everett asked them what was so funny, and they told him they had taken bets on precisely when Chomsky would once again cite the opinion of the linguist from Mars.

The somewhat unkind implication is that the distinguished scholar had become so predictable that his audiences had to search for ways to amuse themselves. Another Chomsky nugget is the way he responds when asked to give a definition of Universal Grammar. He will sometimes say that Universal Grammar is whatever made it possible for his granddaughter to learn to talk but left the world's supply of kittens and rocks speechless—a less-than-precise answer. Say "kittens and rocks" to a cluster of linguists and eyes are likely to roll.

Chomsky's detractors have said that Universal Grammar is whatever he needs it to be at that moment. By keeping it mysterious, they contend, he is able to dodge criticism and avoid those who are gunning for him. It's hard to murder a phantom.

Everett's book is an attempt to deliver, if not a fatal blow, then at least a solid right cross to Universal Grammar. He believes that the structure of language doesn't spring from the mind but is instead largely formed by culture, and he points to the Amazonian tribe he studied for 30 years as evidence. It's not that Everett thinks our brains don't play a role—they obviously do. But he argues that just because we are capable of language does not mean it is necessarily prewired. As he writes in his book: "The discovery that humans are better at building human houses than porpoises tells us nothing about whether the architecture of human houses is innate."

The language Everett has focused on, Pirahã, is spoken by just a few hundred members of a hunter-gatherer tribe in a remote part of Brazil. Everett got to know the Pirahã in the late 1970s as an American missionary. With his wife and kids, he lived among them for months at a time, learning their language from scratch. He would point to objects and ask their names. He would transcribe words that sounded identical to his ears but had completely different meanings. His progress was maddeningly slow, and he had to deal with the many challenges of jungle living. His story of taking his family, by boat, to get treatment for severe malaria is an epic in itself.

His initial goal was to translate the Bible. He got his Ph.D. in linguistics along the way and, in 1984, spent a year studying at the Massachusetts Institute of Technology in an office near Chomsky's. He was a true-blue Chomskyan then, so much so that his kids grew up thinking Chomsky was more saint than professor. "All they ever heard about was how great Chomsky was," he says. He was a linguist with a dual focus: studying the Pirahã language and trying to save the Pirahã from hell. The second part, he found, was tough because the Pirahã are rooted in the present. They don't discuss the future or the distant past. They don't have a belief in gods or an afterlife. And they have a strong cultural resistance to the influence of outsiders, dubbing all non-Pirahã "crooked heads." They responded to Everett's evangelism with indifference or ridicule.

As he puts it now, the Pirahã weren't lost, and therefore they had no interest in being saved. They are a happy people. Living in the present has been an excellent strategy, and their lack of faith in the divine has not hindered them. Everett came to convert them, but over many years found that his own belief in God had melted away.

So did his belief in Chomsky, albeit for different reasons. The Pirahã language is remarkable in many respects. Entire conversations can be whistled, making it easier to communicate in the jungle while hunting. Also, the Pirahã don't use numbers. They have words for amounts, like a lot or a little, but nothing for five or one hundred. Most significantly, for Everett's argument, he says their language lacks what linguists call "recursion"—that is, the Pirahã don't embed phrases in other phrases. They instead speak only in short, simple sentences.

In a recursive language, additional phrases and clauses can be inserted in a sentence, complicating the meaning, in theory indefinitely. For most of us, the lack of recursion in a little-known Brazilian language may not seem terribly interesting. But when Everett published a paper with that finding in 2005, the news created a stir. There were magazine articles and TV appearances. Fellow linguists weighed in, if only in some cases to scoff. Everett had put himself and the Pirahã on the map.

His paper might have received a shrug if Chomsky had not recently co-written a paper, published in 2002, that said (or seemed to say) that recursion was the single most important feature of human language. "In particular, animal communication systems lack the rich expressive and open-ended power of human language (based on humans' capacity for recursion)," the authors wrote. Elsewhere in the paper, the authors wrote that the faculty of human language "at minimum" contains recursion. They also deemed it the "only uniquely human component of the faculty of language."

In other words, Chomsky had finally issued what seemed like a concrete, definitive statement about what made human language unique, exposing a possible vulnerability. Before Everett's paper was published, there had already been back and forth between Chomsky and the authors of a response to the 2002 paper, Ray Jackendoff and Steven Pinker. In the wake of that public disagreement, Everett's paper had extra punch.

It's been said that if you want to make a name for yourself in modern linguistics, you have to either align yourself with Chomsky or seek to destroy him. Either you are desirous of his approval or his downfall. With his 2005 paper, Everett opted for the latter course.

Because the pace of academic debate is just this side of glacial, it wasn't until June 2009 that the next major chapter in the saga was written. Three scholars who are generally allies of Chomsky published a lengthy paper in the journal Language dissecting Everett's claims one by one. What he considered unique features of Pirahã weren't unique. What he considered "gaps" in the language weren't gaps. They argued this in part by comparing Everett's recent paper to work he published in the 1980s, calling it, slightly snidely, his earlier "rich material." Everett wasn't arguing with Chomsky, they claimed; he was arguing with himself. Young Everett thought Pirahã had recursion. Old Everett did not.

Everett's defense was, in so many words, to agree. Yes, his earlier work was contradictory, but that's because he was still under Chomsky's sway when he wrote it. It's natural, he argued, even when doing basic field work, cataloging the words of a language and the stories of a people, to be biased by your theoretical assumptions. Everett was a Chomskyan through and through, so much so that he had written the MSN Encarta encyclopedia entry on him. But now, after more years with the Pirahã, the scales had fallen from his eyes, and he saw the language on its own terms rather than those he was trying to impose on it.

David Pesetsky, a linguistics professor at MIT and one of the authors of the critical Language paper, thinks Everett was trying to gin up a "Star Wars-level battle between himself and the forces of Universal Grammar," presumably with Everett as Luke Skywalker and Chomsky as Darth Vader.

Contradicting Everett meant getting into the weeds of the Pirahã language, a language that Everett knew intimately and his critics did not. "Most people took the attitude that this wasn't worth taking on," Pesetsky says. "There's a junior-high-school corridor, two kids are having a fight, and everyone else stands back." Everett wrote a lengthy reply that Pesetsky and his co-authors found unsatisfying and evasive. "The response could have been 'Yeah, we need to do this more carefully,'" says Pesetsky. "But he's had seven years to do it more carefully and he hasn't."

Critics haven't just accused Everett of inaccurate analysis. He's the sole authority on a language that he says changes everything. If he wanted to, they suggest, he could lie about his findings without getting caught. Some were willing to declare him essentially a fraud. That's what one of the authors of the 2009 paper, Andrew Nevins, now at University College London, seems to believe. When I requested an interview with Nevins, his reply read, "I may be being glib, but it seems you've already analyzed this kind of case!" Below his message was a link to an article I had written about a Dutch social psychologist who had admitted to fabricating results, including creating data from studies that were never conducted. In another e-mail, after declining to expand on his apparent accusation, Nevins wrote that the "world does not need another article about Dan Everett."

In 2007, Everett heard reports of a letter signed by Cilene Rodrigues, who is Brazilian, and who co-wrote the paper with Pesetsky and Nevins, that accuses him of racism. According to Everett, he got a call from a source informing him that Rodrigues, an honorary research fellow at University College London, had sent a letter to the organization in Brazil that grants permission for researchers to visit indigenous groups like the Pirahã. He then discovered that the organization, called FUNAI, the National Indian Foundation, would no longer grant him permission to visit the Pirahã, whom he had known for most of his adult life and who remain the focus of his research.

He still hasn't been able to return. Rodrigues would not respond directly to questions about whether she had signed such a letter, nor would Nevins. Rodrigues forwarded an e-mail from another linguist who has worked in Brazil, which speculates that Everett was denied access to the Pirahã because he did not obtain the proper permits and flouted the law, accusations Everett calls "completely false" and "amazingly nasty lies."

Whatever the reason for his being blocked, the question remains: Is Everett's work racist? The accusation goes that because Everett says that the Pirahã do not have recursion, and that all human languages supposedly have recursion, Everett is asserting that the Pirahã are less than human. Part of this claim is based on an online summary, written by a former graduate student of Everett's, that quotes traders in Brazil saying the Pirahã "talk like chickens and act like monkeys," something Everett himself never said and condemns. The issue is sensitive because the Pirahã, who eschew the trappings of modern civilization and live the way their forebears lived for thousands of years, are regularly denigrated by their neighbors in the region as less than human. The fact that Everett is American, not Brazilian, lends the charge added symbolic weight.

When you read Everett's two books about the Pirahã, it is nearly impossible to think that he believes they are inferior. In fact, he goes to great lengths not to condescend and offers defenses of practices that outsiders would probably find repugnant. In one instance he describes, a Pirahã woman died, leaving behind a baby that the rest of the tribe thought was too sick to live. Everett cared for the infant. One day, while he was away, members of the tribe killed the baby, telling him that it was in pain and wanted to die. He cried, but didn't condemn, instead defending in the book their seemingly cruel logic.

Likewise, the Pirahã's aversion to learning agriculture, or preserving meat, or the fact that they show no interest in producing artwork, is portrayed by Everett not as a shortcoming but as evidence of the Pirahã's insistence on living in the present. Their nonhierarchical social system seems to Everett fair and sensible. He is critical of his own earlier attempts to convert the Pirahã to Christianity as a sort of "colonialism of the mind." If anything, Everett is more open to a charge of romanticizing the Pirahã culture.

Other critics are more measured but equally suspicious. Mark Baker, a linguist at Rutgers University at New Brunswick, who considers himself part of Chomsky's camp, mentions Everett's "vested motive" in saying that the Pirahã don't have recursion. "We always have to be a little careful when we have one person who has researched a language that isn't accessible to other people," Baker says. He is dubious of Everett's claims. "I can't believe it's true as described," he says.

Chomsky hasn't exactly risen above the fray. He told a Brazilian newspaper that Everett was a "charlatan." In the documentary about Everett, Chomsky raises the possibility, without saying he believes it, that Everett may have faked his results. Behind the scenes, he has been active as well. According to Pesetsky, Chomsky asked him to send an e-mail to David Papineau, a professor of philosophy at King's College London, who had written a positive, or at least not negative, review of Don't Sleep, There Are Snakes. The e-mail complained that Papineau had misunderstood recursion and was incorrectly siding with Everett. Papineau thought he had done nothing of the sort. "For people outside of linguistics, it's rather surprising to find this kind of protection of orthodoxy," Papineau says.

And what if the Pirahã don't have recursion? Rather than ferreting out flaws in Everett's work as Pesetsky did, Chomsky's preferred response is to say that it doesn't matter. In a lecture he gave last October at University College London, he referred to Everett's work without mentioning his name, talking about those who believed that "exceptions to the generalizations are considered lethal." He went on to say that a "rational reaction" to finding such exceptions "isn't to say 'Let's throw out the field.'" Universal Grammar permits such exceptions. There is no problem. As Pesetsky puts it: "There's nothing that says languages without subordinate clauses can't exist."

Except the 2002 paper on which Chomsky's name appears. Pesetsky and others have backed away from that paper, arguing not that it was incorrect, but that it was "written in an unfortunate way" and that the authors were "trying to make certain things comprehensible about linguistics to a larger public, but they didn't make it clear that they were simplifying." Some say that Chomsky signed his name to the paper but that it was actually written by Marc Hauser, the former professor of psychology at Harvard University, who resigned after Harvard officials found him guilty of eight counts of research misconduct. (For the record, no one has suggested the alleged misconduct affected his work with Chomsky.)

Chomsky declined to grant me an interview. Those close to him say he sees Everett as seizing on a few stray, perhaps underexplained, lines from that 2002 paper and distorting them for his own purposes. And the truth, Chomsky has made clear, should be apparent to any rational person.

Ted Gibson has heard that one before. When Gibson, a professor of cognitive sciences at MIT, gave a paper on the topic at a January meeting of the Linguistic Society of America, held in Portland, Ore., Pesetsky stood up at the end to ask a question. "His first comment was that Chomsky never said that. I went back and found the slide," he says. "Whenever I talk about this question in front of these people I have to put up the literal quote from Chomsky. Then I have to put it up again."

Geoffrey Pullum, a professor of linguistics at the University of Edinburgh, is also vexed at how Chomsky and company have, in his view, played rhetorical sleight-of-hand to make their case. "They have retreated to such an extreme degree that it says really nothing," he says. "If it has a sentence longer than three words then they're claiming they were right. If that's what they claim, then they weren't claiming anything." Pullum calls this move "grossly dishonest and deeply silly."

Everett has been arguing about this for seven years. He says Pirahã undermines Universal Grammar. The other side says it doesn't. In an effort to settle the dispute, Everett asked Gibson, who holds a joint appointment in linguistics at MIT, to look at the data and reach his own conclusions. He didn't provide Gibson with data he had collected himself because he knows his critics suspect those data have been cooked. Instead he provided him with sentences and stories collected by his missionary predecessor. That way, no one could object that it was biased.

In the documentary about Everett, handing over the data to Gibson is given tremendous narrative importance. Everett is the bearded, safari-hatted field researcher boating down a river in the middle of nowhere, talking and eating with the natives. Meanwhile, Gibson is the nerd hunched over his keyboard back in Cambridge, crunching the data, examining it with his research assistants, to determine whether Everett really has discovered something. If you watch the documentary, you get the sense that what Gibson has found confirms Everett's theory. And that's the story you get from Everett, too. In our first interview, he encouraged me to call Gibson. "The evidence supports what I'm saying," he told me, noting that he and Gibson had a few minor differences of interpretation.

But that's not what Gibson thinks. Some of what he found does support Everett. For example, he's confirmed that Pirahã lacks possessive recursion, phrases like "my brother's mother's house." Also, there appear to be no conjunctions like "and" or "or." In other instances, though, he's found evidence that seems to undercut Everett's claims—specifically, when it comes to noun phrases in sentences like "His mother, Itaha, spoke."

That is a simple sentence, but inserting the mother's name is a hallmark of recursion. Gibson's paper, on which Everett is a co-author, states, "We have provided suggestive evidence that Pirahã may have sentences with recursive structures."

If that turns out to be true, it would undermine the primary thesis of both of Everett's books about the Pirahã. Rather than the hero who spent years in the Amazon emerging with evidence that demolished the field's predominant theory, Everett would be the descriptive linguist who came back with a couple of books full of riveting anecdotes and cataloged a language that is remarkable, but hardly changes the game.

Everett only realized during the reporting of this article that Gibson disagreed with him so strongly. Until then, he had been saying that the results generally supported his theory. "I don't know why he says that," Gibson says. "Because it doesn't. He wrote that our work corroborates it. A better word would be falsified. Suggestive evidence is against it right now and not for it." Though, he points out, the verdict isn't final. "It looks like it is recursive," he says. "I wouldn't bet my life on it."

Another researcher, Ray Jackendoff, a linguist at Tufts University, was also provided the data and sees it slightly differently. "I think we decided there is some embedding but it is of limited depth," he says. "It's not recursive in the sense that you can have infinitely deep embedding." Remember that in Chomsky's paper, it was the idea that "open-ended" recursion was possible that separated human and animal communication. Whether the kind of limited recursion Gibson and Jackendoff have noted qualifies depends, like everything else in this debate, on the interpretation.

Everett thinks what Gibson has found is not recursion, but rather false starts, and he believes further research will back him up. "These are very short, extremely limited examples and they almost always are nouns clarifying other nouns," he says. "You almost never see anything but that in these cases." And he points out that there still doesn't seem to be any evidence of infinite recursion. Says Everett: "There simply is no way, even if what I claim to be false starts are recursive instead, to say, "'My mother, Susie, you know who I mean, you like her, is coming tonight.'"

The field has a history of theoretical disagreements that turn ugly. In the book The Linguistic Wars, published in 1995, Randy Allen Harris tells the story of another skirmish between Chomsky and a group of insurgent linguists called generative semanticists. Chomsky dismissed his opponents' arguments as absurd. His opponents accused him of altering his theories when confronted and of general arrogance. "Chomsky has the impressive rhetorical talent of offering ideas which are at once tentative and fully endorsed, of appearing to take the if out of his arguments while nevertheless keeping it safely around," writes Harris.

That rhetorical talent was on display in his lecture last October, in which he didn't just disagree with other linguists, but treated their arguments as ridiculous and a mortal danger to the field. The style seems to be reflected in his political activism. Watch his 1969 debate on Firing Line against William F. Buckley Jr., available on YouTube, and witness Chomsky tie his famous interlocutor in knots. It is a thorough, measured evisceration. Chomsky is willing to deploy those formidable skills in linguistic arguments as well.

Everett is far from the only current Chomsky challenger. Recently there's been a rise in so-called corpus linguistics, a data-driven method of evaluating a language, using computer software to analyze sentences and phrases. The method produces detailed information and, for scholars like Gibson, finally provides scientific rigor for a field he believes has been mired in never-ending theoretical disputes. That, along with the brain-scanning technology that linguists are increasingly making use of, may be able to help resolve questions about how much of the structure of language is innate and how much is shaped by culture.

But Chomsky has little use for that method. In his lecture, he deemed corpus linguistics nonscientific, comparing it to doing physics by describing the swirl of leaves on a windy day rather than performing experiments. This was "just statistical modeling," he said, evidence of a "kind of pathology in the cognitive sciences." Referring to brain scans, Chomsky joked that the only way to get a grant was to propose an fMRI.

As for Universal Grammar, some are already writing its obituary. Michael Tomasello, co-director of the Max Planck Institute for Evolutionary Anthropology, has stated flatly that "Universal Grammar is dead." Two linguists, Nicholas Evans and Stephen Levinson, published a paper in 2009 titled "The Myth of Language Universals," arguing that the "claims of Universal Grammar ... are either empirically false, unfalsifiable, or misleading in that they refer to tendencies rather than strict universals." Pullum has a similar take: "There is no Universal Grammar now, not if you take Chomsky seriously about the things he says."

Gibson puts it even more harshly. Just as Chomsky doesn't think corpus linguistics is science, Gibson doesn't think Universal Grammar is worthwhile. "The question is, 'What is it?' How much is built-in and what does it do? There are no details," he says. "It's crazy to say it's dead. It was never alive."

Such proclamations have been made before and Chomsky, now 83, has a history of outmaneuvering and outlasting his adversaries. Whether Everett will be yet another in a long line of would-be debunkers who turn into footnotes remains to be seen. "I probably do, despite my best intentions, hope that I turn out to be right," he says. "I know that it is not scientific. But I would be a hypocrite if I didn't admit it."

subscribe today

Get the insight you need for success in academe.