Why language is not everything that Noam Chomsky said it is

Few scientific disagreements lead to public controversy. But there are times when the subject or the participants in a debate so capture the public imagination that otherwise dry, technical matters of discord among researchers erupt into the media, eliciting a wide array of opinions from experts and non-experts. Getting the public interested is good for science if it leads to deeper thinking about things that are of importance to understanding our species. And there is a controversy of just this type bubbling away for many years now in linguistics. Most recently, the disagreements in the field have pulled the American author Tom Wolfe into the fray, with a new book, The Kingdom of Speech, and a cover story in Harper’s Magazine on the topic. This has changed the debate a bit, engaging many more people than ever before, but now it’s centred around Wolfe, Noam Chomsky – and me.

As background to understanding what’s at stake in this controversy, we need a grasp of Chomsky’s important theoretical proposals regarding human language acquisition. From the late 1950s to the present, the American linguist has inspired thousands of articles and hundreds of books exploring what is referred to as the ‘language-acquisition device’. This is the idea that there is something unique in human biology dedicated to language. The principal evidence marshalled on behalf of this idea is that children know more than they could possibly have learned from the evidence available to them. This difference between what they know and the linguistic examples they are exposed to is known as the ‘poverty of the stimulus’. There are plenty of examples (and counterexamples) of this in the literature. Cognitive scientists are divided as to whether there is any such poverty, some even referring to it deprecatingly as the ‘poverty of imagination’ (on the part of Chomskyans). While that debate rages, however, its focal point has come to be how much, if any, of human grammar is innate.

That is where my work comes in. In 2005, I published a paper in the journal Current Anthropology, arguing that Pirahã – an Amazonian language unrelated to any living language – lacked several kinds of words and grammatical constructions that many researchers would have expected to find in all languages. I made it clear that this absence was not due to any inherent cognitive limitation on the part of its speakers, but due to cultural values, one in particular that I termed the ‘immediacy of experience principle’.

Although I realised that what I had written might be controversial, I was unprepared for the sheer number of academic papers, books and ad hominem attacks on me that have raged now for more than a decade provoked by that article. According to the different extremes in this debate, I am either an irrelevant, mistaken charlatan (Chomsky, in the Brazilian newspaper Folha de São Paulo, February 2009) or an ‘instant folk hero’ who knocked all the wind from Chomsky’s work on universal grammar as no one before had ever done (Wolfe, in Harper’s Magazine, August 2016).

It is likely that initially my work was picked up by the media because it was considered a problem for the core ideas of a man whom The New York Times sensationalistically labelled ‘the most important intellectual alive’, ie Chomsky. Looking back now, it is astounding that the point that has so inflamed my academic critics was my claim that the Pirahãs lacked subordinate clauses. You would have thought I had spit on a crucifix during church. I had, of course, expected someone to point out flaws in my reasoning or to give clear examples of data that I had missed or to conduct field research to test my claims. This is the norm in academic debates. But in the first round of criticisms, lasting for five years, what came my way instead was mainly name-calling.

My 2005 paper on recursion did not appear in a vacuum, of course. In 2002, Chomsky, Marc Hauser and Tecumseh Fitch published a paper in the journal Science on ‘The Faculty of Language’. They distinguished two related faculties, the broad faculty of language (FLB) and the narrow faculty of language (FLN). As they said:

FLB includes a sensory-motor system, a conceptual-intentional system, and the computational mechanisms for recursion, providing the capacity to generate an infinite range of expressions from a finite set of elements. We hypothesise that FLN only includes recursion and is the only uniquely human component of the faculty of language.  

Further, they suggested that ‘FLN may have evolved for reasons other than language, hence comparative studies might look for evidence of such computations outside of the domain of communication (for example, number, navigation, and social relations).’ Thus, most people took my argument that Pirahã lacks recursion as a pristine counterexample to their recursion proposal.

On the other hand, my actual conclusion in that paper was not primarily about recursion, but about the connection between culture and grammar. In fact, the word ‘recursion’ hardly appears in that paper. I addressed the lack of quantifier words, such as ‘all’, ‘each’, and ‘every’; the lack of colour words; the lack of numbers; the lack of creation myths; the lack of religion; the fact that the people have remained monolingual after centuries of contact; that the Pirahãs possessed the simplest kinship ever documented; and so on. What I concluded was almost innocuous:

For advocates of universal grammar the arguments here present a challenge – defending an autonomous linguistic module that can be affected in many of its core components by the culture in which it ‘grows’. If the form or absence of things such as recursion, sound structure, word structure, quantification, numerals, number, and so on is tightly constrained by a specific culture, as I have argued, then the case for an autonomous, biologically determined module of language is seriously weakened.  

In fact, linguists of the early 20th century nearly universally believed this and would have yawned at my conclusion. But it dramatically contradicts Chomsky’s conception of language. Although Chomsky refers to ‘language’ in his writings, he means exclusively a recursive grammatical system. Thus, his claim that ‘language’ derives from a narrow faculty of language that is populated only by recursion is a circular claim, because he is simply telling us how he has defined language for years. If there were a language that chose not to use recursion, it would at the very least be curious and at most would mean that Chomsky’s entire conception of language/grammar is wrong.

The tactic used against most of Chomsky’s critics is to ignore them. There is nothing wrong with this and it is standard in academics. I certainly don’t take the time to react to all my critics. There is positive work to be done. In this case, however, it was much harder to ignore my work because of the publicity that ensued. Chomsky made a clear claim – recursion is fundamental to having language. And my paper did in fact present a counterexample. Recursion cannot be fundamental to language if there are languages without it, even just one language without it.

The publicity began from the outset when the University of Chicago Press, the publisher of Current Anthropology, issued a press release about my 2005 article, at the request of the editor of the journal. This is common practice when editors think an article is noteworthy. Usually such press releases are ignored entirely. But in my case, several newspapers and radio programmes picked up the story and interviewed me about my work. A couple of magazines carried articles in which the results were briefly discussed. More prominent magazines even did feature stories, such as the April 2007 issue of The New Yorker magazine and, later, the aforementioned cover article by Wolfe in Harper’s Magazine.

The press’ attention to the work elicited a huge critical response (alongside many positive reviews and compliments) from academics and non-academics alike. As I look back a decade later, the cycle of criticisms against my work is interesting, reminiscent of the claim, attributed to Arthur Schopenhauer, that ‘All truth passes through three stages. First, it is ridiculed. Second, it is violently opposed. Third, it is accepted as being self-evident.’ The criticisms went from articles purporting to show that I was wrong to statements such as ‘He is a charlatan’ and ‘the work is irrelevant’, pretty much following Schopenhauer’s stages of acceptance. (While it isn’t clear that Schopenhauer said exactly this, he said similar things. Moreover, I am not claiming that my work is in any way ‘truth’. I could indeed be wrong.)

And now Wolfe, known for his take-downs of what he considers to be vacuous pomposity – such as his Radical Chic (1970) and From Bauhaus to Our House (1981) – has jumped into this controversy with both guns blazing. In his book The Kingdom of Speech (2016), he nearly drowns Chomsky in sarcastic criticism (not entirely undeserved, despite the horror Wolfe’s prose provoked in some quarters). Unsurprisingly, Wolfe’s targets and many others who respect Chomsky are firing back. And because I emerge as one of two heroes in The Kingdom of Speech (the other being the 19th-century naturalist Alfred Russel Wallace), I have come under even more fire from linguists and non-linguists. Some, like the attack on me by a pseudonymous writer (whose prose is quite good, I might add!) in the online 3:AM Magazine, are so wrong in content and vicious in tone that they don’t merit a reply.

But, despite the emotions involved, there seems to be a whiff of something important about all this. The average person recognises that language is central to what it means to be human. Every healthy human talks. Even those who yawn at the details of the debate between Chomsky and me on the role of recursion in human language are drawn to Wolfe’s book because they recognise that language is important.

There is another reason why people seem to have an emotional stake in this otherwise obscure academic dust-up. Thousands know Chomsky not because of his linguistic work but because of his political writings. For some, his intellectual authority on politics is believed to derive ultimately from his brilliance in inventing a theory in linguistics that is intellectually on a par with Albert Einstein’s theory of relativity. These folks are put off by criticism of his linguistics, I believe, because this could undermine in some way his ‘genius’ label, so important to their idolisation of him as a political figure. My criticisms of Chomsky are seen by some as analogous to those of some junior physicist at an obscure college saying that he or she had falsified Einstein’s theories.

Another reason for negative reactions to criticism of Chomsky is what I refer to occasionally as ‘Ivy-league bias’. Imitation is a stronger force in cultures than innovation. Everything goes more easily if we imitate rather than innovate – so we buy our clothes at the same department stores and eat out at the same chain restaurants. And when we imitate people – wearing their jerseys, singing their music, repeating their ideas – we are doing what most cultures do, copying people with prestige and status. For example, reporters rarely innovate when covering science, that is, they rarely come to their own opinion about difficult material. Rather, they establish a set of ‘go-to’ experts to cite. And those sets are populated largely by Ivy-league professors. There is nothing wrong with that. I simply point out that it is common. It saves one from the excruciating work of original thought.

But Chomsky is no Einstein. And linguistics is not physics. Unlike Einstein, for example, Chomsky has been forced to retract at one time or another just about every major proposal he has made up to his current research, which he calls ‘Minimalism’. Concepts that helped make him famous, such as ‘deep structure’ and ‘surface structure’, were thrown out years ago. And unlike physics, there is no significant mathematics or clear way to disprove Chomsky’s broader claims – part of the reason for the current controversy.

Over time, universal grammar has been reduced from a rich set of supposedly innate principles to whatever it is about human biology that makes human language possible (by which definition, as I have said numerous times, the physical brain itself is all there is to ‘universal grammar’). And once he got down to the narrow faculty of language, supposedly the one thing that makes human language possible and unavailable to other creatures, Chomsky claimed it was nothing more nor less than recursion – the ability to put one thing inside another of the same type. Then Pirahã came along.

Recursion is common in English and many other languages. For example, put the noun ‘truck’ and the noun ‘driver’ inside a single noun, and you get ‘truck-driver’. Put a sentence inside another sentence and you get ‘John said that he did not do it,’ where ‘he did not do it’ is a sentence inside the larger sentence, ‘John said that…’ Or, much more fun, ‘Oysters that oysters eat themselves eat oysters,’ which can also come out as ‘Oysters oysters eat eat oysters.’

The Kingdom of Speech has been the object of scathing reviews on blogs, in comments, and in newspaper articles by linguists, anthropologists, psychologists, journalists and evolutionary biologists – often people with a belief that Chomsky’s views on linguistics and politics are proven or so well-established that challenges from outside science – the unwashed – are somehow sacrilegious. I have no dog in this fight. Wolfe’s book is not my book. Wolfe interviewed me for hours by phone over a period of several months. His conclusions are his own.

But Wolfe can take care of himself. He doesn’t need my defence. Nevertheless, my reaction to most of the reviews of Wolfe’s work is that they ‘doth protest too much, methinks’. Yet some reviewers apparently feel that, to trash Wolfe’s book, they must trash my work. That gets my attention. It is disappointing to see that every criticism of my work in the reviews of Wolfe’s book, from Chomsky’s to all the others I have read, miss the mark widely.

Like anyone, I could be wrong. Pirahã might one day be proven to have recursion. But no one has done anything remotely close to that yet, and none have successfully rebutted my 2005 analysis. There was a critical discussion of my work in the journal Language, as I mention directly, where I and my critics went back and forth, but it is fair to say that neither side was more convinced after the exchange than before. Some of the reviews and discussions of Wolfe’s book have even remarked that the debate is irrelevant. That is also wrong. The debate is crucial to our understanding of human language and evolution, and so far the evidence supports my view, not Chomsky’s.

When he was asked to comment on the documentary about my work, The Grammar of Happiness (2012), Chomsky asserted: ‘There is no question that the language is built on a recursive process.’ With this remark, Chomsky introduces a step not anticipated by Schopenhauer – simple assertion, without investigation: ‘I insist that I am correct and that my opponent is wrong, by definition.’

Other linguists did conduct the honest and painstaking work of trying to show that my analysis was incorrect. Andrew Nevins at University College London, David Pesetsky at the Massachusetts Institute of Technology and Cilene Rodrigues now at the Pontifical Catholic University of Rio de Janeiro have criticised me many times, vociferously, in the media and in academic journals but, at least in the first five years, never said the work was irrelevant. Everything Wolfe says about their actions in The Kingdom of Speech is right-on.

On the other hand, these linguists did take my work seriously, publishing a long, strongly critical article in the journal Language. I replied in the same issue, by invitation of the journal’s editor. Their article basically claimed that my 1983 PhD dissertation provides evidence that there is recursion in the language. My response to the critical article was and is simple – new data changed my mind and convinced me that Pirahã did not have recursion. That is really all there was to our exchange, though it took an unprecedented 87 pages in the flagship journal of the Linguistic Society of America. No one said in that exchange that Pirahã was irrelevant. Because no one thought so. That is a later idea.

These days, though, that is exactly what my critics claim, namely, that even if I am right, my analysis of Pirahã is irrelevant to the important enterprise of understanding the nature of human language. They assert that if the Pirahã can learn a recursive language, such as Portuguese, the national language of Brazil, then they indeed can ‘do’ recursion and so the fact that their language might lack it has nothing to say about recursion as a prerequisite for human language. As it turns out, this is a poorly reasoned argument though it is frequently mistaken as ironclad. The reasoning is mistaken on several counts. Before stating them, I should perhaps remind those who have followed the debate that I have written about the Pirahãs’ ability to use recursion from the very beginning of my research. The Pirahã ability to think recursively is not a new discovery.

Chomsky and others have even made the false claim that many Pirahãs speak Portuguese, strengthening the case that they can use recursion. Again, they use this to claim that it is therefore the case that the Pirahãs do have recursive abilities and so the hypothesis of the narrow faculty of language is safe and sound after all. But, as I said, this doesn’t follow. For example, it is extremely telling that, of the Pirahãs who speak any Portuguese at all, no Pirahã has ever learned Portuguese subsequent to learning to speak Pirahã. The only Pirahãs who have learned Portuguese well have learned it by being raised outside the village, with Portuguese as their native language. Jeanette Sakel of the University of the West of England in Bristol has studied exactly this. She has visited the Pirahãs with me and conducted her own research, publishing several articles in refereed journals on the Pirahãs’ use of Portuguese.

Pirahãs who speak Pirahã natively and are culturally Pirahãs speak very little Portuguese, if any. And, according to Sakel’s work, when they do speak it is non-recursive Portuguese. This is a (fascinating) fact about the contact between Pirahã and Portuguese cultures and the basic values of the Pirahãs. It is not a question of intelligence. They are not stupid or backwards or genetically isolated weirdos. It is the connection between their culture and grammar that brings this about.

Ultimately, this dispute is about whether Pirahã is a ‘counterexample’ to Chomsky’s work or whether it is an ‘exception’. Superficially, counterexamples and exceptions are identical. Both are facts that don’t fit a prediction. Whether you set them aside and proceed with your theory (as with exceptions) or take them seriously enough to revise or abandon your theory (as with counterexamples) is a cultural decision, reflecting your values and knowledge structures, as well as your social roles. It is what seems most elegant to you or fits the facts you are most interested in, according to your value system. The Chomskyan view is that Pirahã is not only not a counterexample, but an (irrelevant) exception. The new Chomskyan view that ‘recursion is just a state of mind’ won’t make Pirahã irrelevant. Thinking recursively is not the same as having a recursive grammar.

Why won’t the fact that the Pirahã think recursively save the Hauser, Chomsky, Fitch (2002) idea that the narrow faculty of language is recursion? Although Chomsky and company dwell on it at length, they miss the point. The real lesson is that if recursion is the narrow faculty of language, but doesn’t actually have to be manifested in a given language, then likely more languages than Pirahã (such as perhaps the Indonesian language Riau) could lack recursion. And by this reasoning we derive the astonishing claim that, although recursion would be the characteristic that makes human language possible, it need not actually be found in any language.

Such reasoning removes any empirical connection between the narrow faculty of language hypothesis and actual linguistic data. This problem arises because Chomsky confuses the possession of recursive thought with the locus and primary function of recursive ability. I have made the case many times that the crucial question is not whether humans have the ability to think recursively, but where that ability resides. And I have concluded that recursion facilitates thought and helps language increase the rate at which information can be transmitted. It is easier to communicate, for example, the following complex information in the recursive sentence (i) than in the non-recursive sentence (ii):

(i) John said that Mary thinks that John falsely believes that the moon is made of green cheese.

(ii) John spoke. Mary thinks. John is wrong. John believes. The moon is green cheese.

Recursion is not the biological basis for language. It is an enhancement of human thought.

Ironically, although my work is claimed to be irrelevant, the same crowd still works hard to prove that I am wrong. Nevertheless, I have answered every single objection that has been raised. Yet none of my responses are mentioned by the critics. When people criticise my work, they cite only my critics, not my responses, so far as I can tell.

Pirahã could in fact be irrelevant to theories of language for either of two reasons. First, it could be irrelevant because Chomsky’s theory does not say that a language actually has to use recursion for recursion to be the basis or ‘narrow faculty’ of language. Second, and related to the former, is the idea that the genotype of universal grammar can be and is manifested in a range of phenotypes.

Neither universal grammar (the innate biological capacity of all grammars, according to Chomsky) nor the narrow faculty of language predict that all languages will look alike. Not all languages have click sounds, as we find in Bantu languages, for example. Not all languages have tones, as we find in Mandarin. Not all languages have interdental consonants – the ‘th’ sounds in English words such as ‘throw’ and ‘then’. Many linguists who accept the idea of universal grammar or even the narrow faculty of language (and if the relationship between those two is unclear to you, it is because it is unclear to everyone) believe that it is a ‘toolbox’. Each language can pick and choose the features it wants from the innate, universal toolbox. The idea that the toolbox exists is not threatened just because the entire toolbox is not used in every language. After all, no mechanic would say that the tools she owns but doesn’t use are any less real than the tools she does use.

But the toolbox is not the appropriate analogy for recursion and language. This is because no tool, no feature of language such as tones, clicks and so on has been claimed to be the ‘only uniquely human component of the faculty of language’ as recursion has been. Recursion is not a tool to be selected but the very engine of language itself, according to Chomsky. It is utterly unlike the other features. It is foundational. All others are derivative, according to the hypothesis of the narrow faculty of language.

However, as I noted, Chomsky still seems to want to cover all the bases – I am wrong, I am a charlatan, and I am irrelevant. In an interview in La Voce di New York in October 2016, Chomsky asserts that ‘careful scholarship’ has refuted my claims. But this is false, however much critics might wish to believe it. I have answered successfully every single counteranalysis that has been raised in the professional literature. In fact, in a recent paper in the journal PLOS ONE, several cognitive scientists from the University of Rochester and MIT explored a number of natural Pirahã stories collected by a missionary who preceded me (so as not to be ‘tainted’ by any ‘confirmation bias’ I might bring to the task); their translations were subsequently checked by that same missionary (Steve Sheldon, of Wycliffe Bible Translators). They concluded: ‘Our analysis has failed to find strong support for syntactically embedded structures in Pirahã.’ Of course, this doesn’t prove anything. But it certainly suggests that an unambiguous determination that I am wrong about recursion has not yet emerged from the data.

If recursion is just a component of human intelligence generally, even if it were the fact of human intelligence, then it is available to be used or not used by human languages. It isn’t the narrow faculty of language. It is the way humans think. And we have no strong evidence that we are the only creatures that can think recursively. Even Hauser, Fitch and Chomsky comment that other animals could have it, just not in language. The human brain is bigger and more complex than other brains, and I have written in many places that it is the greater abilities and computational power of the human brain that most likely underwrite human language, without any need to appeal to the concept of a ‘faculty of language’ either narrow or broad.

In short, the question is not whether humans can think recursively. The question is whether this ability is linked specifically to language or instead to human cognitive accomplishments more generally (it could be connected to both, but that is less likely given what we know about the organisation of the brain).

If I am correct, then I have shown that the sentential grammars of human languages don’t need to be constructed recursively. People might all think recursively but lack recursion in their grammars. What I have shown is that for the very reason that the Pirahãs can think recursively, then if their language lacks recursion, recursion is not fundamental to human language but is rather a component of human cognition more generally. To claim otherwise, again, is to claim that all languages of the world can lack recursion but recursion is still alone the narrow faculty of language. And that is empirically vacuous gibberish. If there is anything innate and specific to the human capacity for language, the Pirahã data shows that recursion is not part of it.

The question remains though as to why linguists and others get so emotional about this debate. I believe that there are two reasons. First, many people have enormous respect for Chomsky. And in my opinion he has earned every bit of that respect. I admire him tremendously, having spent the first 25 years of my career working in his theory. Second, however, if Chomsky is wrong, many research careers built on his theory are on shaky ground. One linguist asked me if the fact that so many very smart people were attracted to Chomsky’s theory didn’t indicate that it was right. Not at all. There have been many brilliant theologians. But I still don’t believe there is a god to theologise about. Intelligence is no guarantee of being right, and science is not a democracy.

Recursion is not the basis of human language. One language shows that. Language does not seem to be innate. There seems to be no narrow faculty of language nor any universal grammar. Language is ancient and emerges from general human intelligence, the need to build communities and cultures.

If the Schopenhauer axiom is right, the third round of criticisms of my work will be ‘We knew that all along’. This reminds me of the quote attributed to William James: ‘A great many people think they are thinking when they are really rearranging their prejudices.’

This article first appeared on Aeon.