The Structure of Scientific Revolutions

The hullabaloo surrounding the publication of my most recent book, The Language Myth, in September 2014, has taken even me by surprise. The book is a rebuttal of an earlier, very famous book, The Language Instinct, written by Steven Pinker, published over 20 years ago. Pinker’s book advocated the perspective on language and the mind developed by American linguist Noam Chomsky. And it did so very accessibly, presenting Chomsky’s complex and arcane theoretical arguments in a way that someone with almost no background in linguistic theory could make sense of. But in the intervening years, the debate, and the science have both moved on. Today, to objective eyes, Chomsky’s celebrated claim—that we are born with a genetically-programmed blueprint for grammar, which enables a human infant to effortlessly learn their mother-tongue, any of the world’s 7,000 or so languages—looks to be on increasingly shaky ground. In The Language Myth, I review the state of the art and, in light of this, provide a damning assessment: Chomsky’s entire edifice is a myth.

I was, of course, expecting a negative reaction from Chomsky’s disciples. But, given the practice in normal science, I was expecting a debate, and in particular, a challenge, with argument and evidence, to the substantive criticisms I levelled against the Chomskyan enterprise. This, it seemed to me, would have been a profitable way to take the field forwards. But things, at least, amongst some of Chomsky’s die-hard followers, haven’t panned out like that. In various fora, the reaction has sometimes appeared to border on the hysterical. For instance, in various public on-line venues I have been described as an “idiot” and a “quack”, and The Language Myth has been repeatedly derided as “junk” and “junk science”; there has been a public debunking exercise led by a “panel” of experts, embroidered with scorn and disdain, conducted in venues ranging from Facebook to Reddit. One prominent, and influential Chomskyan supporter, Professor Norbert Hornstein, has devoted, by now, an extraordinary number of posts to what amounts to a slim volume aimed at the general public, in his Faculty of Language Blogspot. He has repeatedly railed against The Language Myth, even issuing a call to (intellectual) arms: “criticize this in all venues, especially where non-linguists gather. Consider it part of your linguistic public service” (December 2014).

This reaction amongst some Chomskyans is, on the face of it, surprising. The publisher, Cambridge University Press took extreme care to ensure an exhaustive single-blind peer review process, with five world-leading experts providing the publisher with their expert advice at an early stage before a book contract could be contemplated. The pre-final draft of the book was also read by a sixth renowned expert who gave the book careful scrutiny prior to publication. And upon publication, experts have lined up to endorse the book including the universally respected Professor David Crystal, presumably with no theoretical axe to grind. Crystal provides the following assessment: “A much-needed, comprehensive critique of universal grammar. Vyvyan Evans builds a compelling case that will be difficult to refute.”

Moreover, since publication the book has received high praise and broad acclaim from reviews in the popular print media, not least for its clear and careful handling of the issues. For instance, a review in Times Higher Education, by a Chomskyan-trained linguist has this to say: “a comprehensive presentation of the case that human verbal communication emerges from use...The general reader can read this book from cover to cover and learn a great deal about language that challenges the established traditions. Equally, the more experienced reader will benefit from the alternative perspective it offers, and from the comprehensive reference lists to support the arguments that Evans makes."

So, what might have provoked the hue and cry amongst the Chomsky faithful? After all, the range of arguments presented in The Language Myth against the Chomskyan enterprise is hardly new. Several decades of findings from a range of distinct disciplinary perspectives, and from a wide array of high-profile and well-respected experts proffer many of the same criticisms that I make. Indeed, today, the overwhelming majority of the world’s linguists would not endorse the Chomskyan thesis: that our grammatical capacity is an innate biological endowment. And as Christina Behme and I observe in our reply (available here) to one particularly partisan critique of The Language Myth:

“The Chomskyan paradigm (especially his Minimalist Program) has faced serious and sustained opposition for decades. Developmental psychologists (e.g. Tomasello, 2003; MacWhinney, 2005) and linguists (e.g. Pullum & Scholz 2002; Sampson, 2002) have questioned the cogency of poverty of the stimulus arguments, field linguists have demonstrated that not all human languages share language universals (e.g. Evans & Levinson, 2009), syntacticians have shown inadequacies in data interpretation (e.g. Postal, 2004, 2014; Jackendoff 2011), experts on social cognition have shown how language structure is shaped by language use (e.g. Enfield & Levinson, 2006; Everett, 2012; Tomasello, 2008), computational modellers have simulated aspects of language acquisition previously claimed ‘unlearnable’ without the help of UG (e.g. Christiansen & Chater 1999, Clark & Lappin 2011, MacWhinney 2010), cognitive linguists have developed successful alternatives to the Chomskyan paradigm, (e.g. Croft & Cruse 2004; Evans & Green 2006; Geeraerts & Cuyckens 2007), and evolutionary theorists have provided persuasive arguments against the minimalist version of Chomskyan UG (e.g. Arbib 2008, Jackendoff & Pinker 2005, Tomasello 2008, Hurford 2011, Lieberman 2013).”

That all said, I think that what is new, however, is the timing. The weight of evidence has now accumulated to such a level that the tide may, now, have decisively turned against the Chomskyan perspective. Two reviews of The Language Myth, in fact, capture this very idea. A review of The Language Myth in The New Scientist has this to say: “Is the way we think about language on the cusp of a revolution? After reading The Language Myth, it certainly looks as if a major shift is in progress, one that will open people's minds to liberating new ways of thinking about language.” And in a review that appeared on the internet forum World Wide Words, a similar idea was presented as follows: “The Language Myth is a wide-ranging polemical dismissal of the received wisdom of many linguists. It’s worth reading also as a classic case study of an orthodoxy undergoing what Thomas Kuhn called a paradigm shift."

The Language Myth has appeared at a time when, for at least some, linguistics appears to be in the throes of a paradigm shift. When I began my postgraduate academic training, in the mid-1990s, the Chomskyan orthodoxy was still de rigeur. University-level classes in the core areas of theoretical linguistics deployed textbooks that inevitably took a Chomskyan perspective. But today, not only is it no longer de rigeur to use materials of this sort, at many English-speaking institutions it is no longer viewed as being heretical to eschew such materials—it once was! While there will always be differences arising from individual experience, and regional variation, my sense is that the theoretical perspective that infuses the discipline of language science is far more plural and diffuse than at any point in my academic career. And I’m sure, for Chomsky’s followers who may have glimpsed this brave new world from behind the safety of the turrets of minimalist linguistics at MIT, and the other Chomskyan redoubts, this is likely to be an increasingly terrifying prospect.

Lessons from Kuhn’s The Structure of Scientific Revolutions
Prior to the publication of Thomas Kuhn’s 1962 book, The Structure of Scientific Revolutions, it was widely assumed that science involved a process of steady progress. Accepted facts were added to, and knowledge accumulated steadily, over time. But Kuhn showed that science could also develop in abrupt ways, where conceptual continuity can be punctuated by periods of scientific revolution. Kuhn compellingly argued that the discovery of “anomalies” can lead to the emergence of a new paradigm. And new paradigms have a habit of scrutinising the data in new ways, asking different sorts of questions, and from different perspectives. And the result is that the rules of the game are changed in the process.

The Chomskyan enterprise was successful for a long period of time because there was no hard data to challenge its underlying assumptions. One of the most basic questions Chomsky asked was this: how do children acquire their mother tongue? Back in the 1950s and 1960s there were no good answers to this question, and little to go on. Children seem to be very good at learning language, and master it way before formal schooling begins. Hence, the idea that children must be born with a biological toolkit that enables them to acquire language was, on the face of it, highly appealing. As the eminent Chomskyan linguist, Neil Smith, put it as recently as 2005 (p. 21):

“One of Chomsky's achievements is to have demonstrated that, despite the easily observable richness in the world's languages, there is really only one human language: that the complex richness and bewildering array of different languages surrounding us are all variations on a single theme, most of whose properties are innately given."

The idea that all of the world’s languages share a common set of structural similarities, provided by our shared, biological endowment, gives credence to the thesis known as Universal Grammar. But as the search for universals has proceeded, it has increasingly become clear to all—except Chomsky’s die-hard disciples—that there is a huge, gaping hole in the enterprise. The search for "universals" keeps turning up nothing; nothing beyond the banal that is. There appears to be little in the way of absolute universals, common to all languages. And the problem is that, according to Universal Grammar, there should be.

This amounts to a Kuhnian “anomaly”—there are others, which I detail in The Language Myth. But anomalies such as this have led to many researchers beginning to drift away from the Chomskyan enterprise, becoming agnostic, or even shifting intellectual allegiance. And this includes those that were once amongst the most ardent of Chomsky’s followers.

In a previous post in this blog, The Shapeshifting Malleability of Universals in Universal Grammar, I observed that the quest for universals in the Chomskyan enterprise has, over the years, been steadily downsized. Today, all that remains is the claim for a highly abstract operation, variously termed ‘merge’ (or ‘recursion’—which may or may not be the same thing as ‘merge’). One problem is that various experts disagree as to how to define the terms, a point that Christina Behme and I, make, in our response article. And, even recursion may not be universal.

Daniel Everett, a linguist anthropologist, with over 30 years experience conducting field research, first argued, back in 2005, that at least one language, Pirahã—on which he is the world’s leading academic authority—fails to exhibit recursion, which I also discuss in The Shapeshifting Malleability of Universals in Universal Grammar. But this claim has been rebutted vehemently by Chomskyan linguists in an exchange of papers, subsequently published in the journal Language. For the Chomskyan enterprise, the real story—and a damning indictment on the failure of the enterprise— is that after over 50 years of looking, no true universals have been found. And the last vestige of a once promising project—the claim that what is universal is a highly abstract operation—is all that’s left. And if Everett is correct, even this claim appears to be on shaky ground too.

The embarrassment has been further compounded by the move, since the 1990s, to reduce the status of the enterprise, by none other than Chomsky himself. In the early days, Chomsky self-confidently dubbed his proposals “theories”: the Standard Theory, the Extended Standard Theory, the Revised Extended Standard Theory, and so on. But by the mid-1990s, Chomsky’s approach had been reconceived, not as a theory, but as a ‘program’. And this addressed the paucity of support for the specific theories, which were in continual need of revision, as new and contradictory findings came to light: the perspective was programmatic, and this after 50 years of intensive effort.

Anomalies and crisis
One of the observations that Kuhn makes is that on their own, findings of fact that militate against a theory, won’t falsify it. This follows as scientific paradigms are more than mere epistemological entities, consisting of findings based on reason and supported by facts. A scientific paradigm is an ideological organism, given sustenance by people, of flesh and blood, whose careers and even livelihoods are invested in the paradigm.

In the case of the Chomskyan enterprise, many independent observers perceive Chomsky’s circle to behave much like a cult, dazzled and bamboozled, in equal measure by The Great Man. Whatever the reason for, and the psychology of, Chomsky’s appeal to those captivated in his orbit, his followers often do appear to behave like devout disciples. They exhibit sometimes quite startling degrees of devotion, poring over his pronouncements and writings, curating, zealously defending, and interpreting his works and even mythologising his contribution in a way that is both remarkable and unusual in science, and is certainly without precedent in linguistics.

When confronted with an anomaly, one that doesn’t fit with the predictions made, Kuhn observed that this can lead to a crisis for the paradigm. And one response to crisis is that a paradigm’s defenders will “devise numerous articulations and ad hoc modifications of their theory in order to eliminate any apparent conflict.” (Kuhn, 1962: pp 78). Chomsky himself is quite explicit on this strategy. At various points in his writings, Chomsky has advocated what he dubs a Galilean approach to linguistic enquiry, an issue I also addressed in my earlier post The Shape-Shifting Malleability of Universals in Universal Grammar.

Writing in his 2002 book, On Nature and Language, Chomsky claimed that: “[Galileo] dismissed a lot of data; he was willing to say: “Look, if the data refute the theory, the data are probably wrong.” And the data that he threw out were not minor”. (Chomsky 2002: 98). He continues saying that “the Galilean style . . . is the recognition that . . . it often makes good sense to disregard phenomena and search for principles” (Ibid.: 99), by “discarding recalcitrant phenomena,” (Ibid: 102). And in 2009, in his opening remarks to a volume edited by Piattelli-Palmarini and colleagues, Chomsky explains, in describing his “scientific” approach: “You just see that some ideas simply look right, and then you sort of put aside the data that refute them” (Ibid.: 36).

Given the pared down version of Universal Grammar that exists today, amounting primarily to merge/recursion, Daniel Everett’s assertion—that Pirahã doesn’t exhibit recursion—presents a problem. Of course, Everett could be mistaken. But if he were correct, this leaves the Chomskyan enterprise in tatters, right? Wrong! As Kuhn observed, recalcitrant data—and the lack of Pirahã recursion, a parade example of recalcitrant data—can be circumnavigated by an ad hoc move. And Chomsky, writing with colleagues Fitch and Hauser in 2005 makes exactly such a move: ‘the putative absence of obvious recursion in one of [the human] languages . . . does not affect the argument that recursion is part of the human language faculty [because] . . . our language faculty provides us with a toolkit for building languages, but not all languages use all the tools”. This constitutes a Kuhnian ad hoc move to avoid a potentially problematic issue for the Chomskyan paradigm.

Another move, and one that is even more disingenuous, is to go beyond the recalcitrant data, and attack the opponent directly. In a 2009 Brazilian newspaper article, reporting on an interview, provided to the newspaper by Chomsky, Chomsky was quoted as describing Everett as a “charlatan”. Here, not only is the veracity of the “finding” being attacked; in addition, the credibility, motivations, and research ethics of the opponent are also called into question.

Fear and loathing in the Academy
This brings us back to the response by some Chomskyans to The Language Myth. Rather than focusing on the substantive criticism, and the wide literature surveyed in the book, which ostensibly points to deep theoretical and empirical problems for the Chomskyan perspective, the response, thus far has been to avoid these. The anomalies are too many and too pervasive, presumably, to be countered. Rather, in response to crisis, the approach has been to attack the book itself, and, in so doing, the opponent. In one published review of The Language Myth, Professor David Adger (Lingua 2015), seeks to characterise The Language Myth as providing a wholesale misrepresentation, and misunderstanding of the Chomskyan enterprise, based on out-of-date sources. This strategy, in the face of crisis is summarised by Adger’s ‘review’ of The Language Myth which he posted on the Amazon website: “An embarrassing hodge-podge of misunderstandings and misrepresentation backed up by out of context quotations and out of date discussions.”

One illustration of Adger’s modus operandi, faced with devastating criticism of his idol’s position, is his attempt to undermine the use of the term ‘language instinct’ in The Language Myth. In his Lingua review article, he suggests that the term is employed, in The Language Myth, as some kind of technical term. Adger’s point is to try and show that this amounts to an instance of both misunderstanding and misrepresentation: that The Language Myth shouldn’t be taken seriously as it is no more than a caricature by an opponent who doesn’t really grasp the issues.

But as Christina Behme and I point out in our response article to Adger, (available here), this is patently not the case—Adger is playing fast and very loose. Moreover, the term ‘instinct’ is widely deployed, as an informal means by Chomsky’s followers, to describe the language faculty. Steven Pinker names his book ‘The Language Instinct’, and Adger’s own research was reported as relating to a language ‘instinct’ in a review piece in The New Scientist, a fact which he drew attention to, circulating a tweet amongst his followers celebrating the review of his work on the language ‘instinct’. Not only does this kind of behaviour illustrate double standards, what it also shows is exactly those sorts of defensive behaviours that Kuhn observed defenders of a paradigm under threat might invoke; in response to crisis engendered by a paradigm shift the strategy is to adopt a bunker mentality, and attack the opponent, as a misguided attempt at defence. The Language Myth is a high-profile book that underlines the shift away from a former orthodoxy. And it therefore constitutes a salient and important target for defenders of the Chomskyan enterprise, as they attempt to respond to the scientific crisis they were in prior to its publication.

A more egregious example of this strategy arises in the blogging of Professor Hornstein. In a previous post, Hornstein has railed against the publisher of The Language Myth, for allowing the “junk” that is The Language Myth to be published: In a post published in early 2015, Hornstein says: “The scandal of [Evans’] published work goes beyond the work itself. The bigger scandal is that Cambridge University Press (Yes, CUP, the CUP!!) published this junk. … CUP has embarrassed itself with this book and it owes Generative Grammar an apology.”

And in his most recent post, Hornstein splutters with rage, having discovered that Language, the flagship journal of the Linguistic Society of America, proposes making an “event”, in his words, out of the publication of The Language Myth. Hornstein describes this in the following terms:

“It has come to my attention that Language is considering making an "event" of Vyvyan Evan's junk book The Language Myth. What do I mean by an "event"? Well, and here I quote: "because of the potentially controversial nature of the book, Language is planning a new type of review, in which we target the book for commentary papers by 4-5 individuals who have different academic perspectives." These reviews are to be about 1500 words. So, Language is going to make a BIG DEAL (6-7500 words of criticism plus a reaction by Evans, I would assume) out of this book pretending that there is something there, pretending that it is "controversial" in the sense that that the book raises many interesting issues that people of good faith can understand in different ways and that debating would enlightening.”

Hornstein is incandescent because, in his words: “the book is not controversial. It is junk. Pure, unadulterated, complete junk. Reading it will make you dumber.” The frenzied invective used is less calculated than Adger’s more measured strategy. With Adger, there is a clear attempt to present the book as, at best, naive, at worst a deliberate hatchet job: either way, as Adger would have it, The Language Myth is a misrepresentation, and hence completely wrong up front, and hence not worth bothering with. For Hornstein, who is seeing red, the book will actually make you dumber if you have the misfortune to read it. Of course, the fact that Cambridge University Press did deem it worthy of publication, and Language have decided to make an "event" out of its publication, call into question the perspectives of both Adger and Hornstein, and the other disciples of Chomsky. But overall, this is an illustration of a clear and deliberate strategy, a response to crisis brought on by a shifting paradigm, for which The Language Myth is a high profile emblem.

Radical fundamentalism in language science
One way of viewing the hullabaloo surrounding the publication of The Language Myth is to see Chomskyan linguistics as more than merely another paradigm whose longevity is under real threat. It is tempting to see it as a form of scientific fundamentalism. In a previous post I have pointed out that subscribing to Chomsky’s position, that there is a biologically instantiated language faculty, amounts to a leap of faith; and given Chomsky’s cult-like status amongst some of his more ardent followers, daring to criticise the word of The Great Man amounts to an act of heresy. Arguably, on those grounds, The Language Myth can be legitimately described as “junk” and an intellectual holy war can be legitimately declared against it and its author.

But some of the narrative that has been used in the context of criticising the book does also suggest deep fear and loathing, which, I suspect, is a consequence of insecurity about the status and the future of the Chomskyan enterprise. History will self-evidently be the arbiter of the value, if any, of Chomsky’s eventual contribution, and that of his followers. While I, and many others, believe his ideas to be wrong-headed, no would presumably deny his supreme importance in the history of ideas, and especially in the development of 20th century cognitive science. But in the meantime, there remains a fossilised paradigm: Kuhn argued that paradigms that are subject to revolution can remain fossilised until the champion dies.

In the final analysis, my assessment is that the Chomskyans are, more than ever, out of touch. They peddle a perspective which is unfalsifiable, and borders on pseudoscience, a point I’ve made in previous posts in this blog. And there is no compelling evidence to support the claim for a biologically endowed language faculty along the Chomskyan lines. Indeed, the evidence overwhelmingly points in the opposite direction—read The Language Myth for a state of the art review--I assure you, it won't, in fact, make you dumber.

Chomskyans can also appear to be deluded. In his most recent blog post, Professor Hornstein claims that two thirds of linguists are Chomskyan, making the explicit, and frankly, preposterous claim that by making an “event” out of the publication of The Language Myth, Language must somehow hate two thirds of all linguists. The delusion is that while once two thirds of American-based linguists may have broadly assumed a Chomskyan perspective, in the heyday of the orthodoxy, today, the Chomskyan paradigm is no longer the orthodoxy. For many linguists, Chomskyan linguistics is not relevant to their research. And in time, its relevance will almost entirely diminish. But until that point, some die-hards will doubtless continue to rage against the dying of the light.

Read Norbert Hornstein’s extraordinary outburst, in his blog post, here.

Read the Behme and Evans response article to David Adger here.

See Behme's 2014 critique of Chomsky: 'A Galilean Science of Language', here.