Deconstructing Chomsky - Re-writing the Innate Rules of Grammar

Noam Chomsky, a rookie professor at MIT, published a ground-breaking book called Syntactic Structures, which set out a theory of Generative Grammar. He suggested that a Universal...
01 October 2009


Don't Sleep, There Are Snakes - Dan Everett's new book about his experiences of living alongside the Amazonian Pirahã tribe and what they taught him about linguistics.


There's change in the air. A movement afoot. The faint whiff of revolution, although many may not have noticed. That's because the change is taking place in Linguistics, and it
Don'll take time for word to spread.

This wasn't how it was in the 50s. Back then, Noam Chomsky, a rookie prof at MIT, published a ground-breaking book called Syntactic Structures, which both propelled Linguistics into the mass media and set the agenda for a fierce debate in linguistic theory, a debate which raged for the next twenty years in what were melodramatically dubbed 'The Linguistics Wars'. Chomsky set forth a theory of Generative Grammar which viewed language as an abstract and idealized mathematical system based on two key components: a Universal Grammar (UG) of basic linguistic principles and a Transformational Grammar of rules responsible for putting sentences together. Word got out, and for a moment Linguistics was sexy. Well, it was noticed, at least.

The most eye-catching of Chomsky's ideas was that language should be seen as an innate part of our make up, an 'organ' which comes to us hard wired and ready to go. All human beings start off with the same pre-adaptation and aptitude for language - it was argued - and all that remained was to switch flexible parameters to the appropriate settings according to the language a child learns first. For example, an English speaking child would set the 'word order parameter' to subject-verb-object ('Jack fills the pail') whereas a Japanese speaking child would have it set to subject-object-verb ('Jack the pail fills').

The main attraction of the UG theory was its very neat and minimalistic way of handling a concept as complex as language. But, there was an underlying weakness which would eventually be its undoing: the theory was a modern day example of a Just So story. It was a statement of anatomical fact with little more in the way of supporting evidence than that 'it could be so.' And yet it was accepted by contemporaries almost without question and in a semi-evangelical way, which would stifle necessary debate for many years.

The Chomskyan conception of language dominated Linguistics in the 1950s. Then, like 'New Labour', cracks appeared in the armour of Generative Grammar, the dream slowly faded, and the hope of a linguistic Brave New World was never fulfilled. As with anything placed up on a pedestal, critics began to chip away at the foundations and so it was only a matter of time before the whole thing came crashing down.

Some think the Chomskyan programme has fallen already; some don't. Such is the nature of academics, never agreeing on anything. At one point, things got so bad that 'war' was declared. In the 1960s and 1970s, some of Chomsky's former students decided the Chomskyan paradigm could be applied not only to the study of sentence structure, but also to the study of meaning, resulting in the theory of Generative Semantics. Chomsky railed at the impudence of these mutineers and banished them to another academic camp. Knowledge of grammar, in his view, could only be based on rules about sentence structure - on syntax. It was generally agreed that Chomsky won this battle, but the opposition rallied like a Medusa and re-emerged in different forms.

While veterans of the Generative Linguistics tradition have not let the original theory stand still, the concept of UG remains central to the framework, based on the nativist 'poverty of the stimulus' argument that there is simply not enough there in the environment where children grow up for them to learn something as complex as language. The opposition say that this is just something Chomsky decided must be true and they have been trying to empirically disprove the notion ever since. In this way, Linguistics has, like so many modern disciplines, become focused on the division of labour between nature and nurture. And that's not a division that is easily resolved.

The nativist conception of UG was revised and revitalised by Steven Pinker in his popular books, and dubbed by him The Language Instinct. Pinker's most outspoken critic, Geoffrey Sampson, responded with his own book on the subject - the first edition (the exciting and explanatory Educating Eve: The Language Instinct Debate) with a much better title than the second (the just plain explanatory The Language Instinct Debate). Sampson argues that children are very good at learning languages not because they are born with UG, but because they are very good at learning anything. What's innate in this 'nurturist' view (which is not as paradoxical as it first sounds: it is not purely nurturist, in the same way that the nativist position, by conceding that children have to hear language in their environment in order to learn it, is not purely nativist) are general cognitive abilities, not knowledge of grammatical structure per se.

So what can you believe in, if not UG? After all, if non-linguists are likely to know anything about Linguistics, it's that Chomsky said we are born with UG, and this sounds pretty convincing coming from such a hero of anti-establishment politics and all. Well, if the nature of our linguistic knowledge isn't UG, then you can believe that it's an open question, and an empirical matter awaiting further proof. Sorry, we linguists don't currently have an answer more definitive than that. But, if you realise that we are statistical machines, capable of swift and sophisticated processing and storage of language input, then you'll quickly realise that we don't necessarily have to have UG to learn language, because the 'poverty of stimulus' argument doesn't necessarily hold. At the very least, syntax doesn't have to be so important any more; abstract rules of sentence structure can be complemented with concrete knowledge of meanings, probabilities, and experiences.

The new standard bearer for the nurturists might be Dan EverettDan Everett, who has not been shy in taking on the nativists in the popular media. Everett's anti-UG book
Don't Sleep, There Are Snakes: Life and Language in the Amazonian Jungle,
published by Profile Books, which recounts his lifetime spent living with the Amazonian Pirahã tribe, has finally got the nature vs. nurture debate back in the public eye - well, in print and on the radio at least. The fact that the book is, moreover, a fascinating account of life in the rainforest means that the central linguistic message is carried sweetly in an attractive bundle of mass appeal. It has already been BBC Radio 4's Book of the Week, and has achieved impressive sales. What's more, it's still #1 in the category Travel & Holiday > Countries & Regions > Central & South America > Brazil > Amazon. Well, someone has to be!

So, maybe, just maybe, the message is getting out at last: we may not have an innate Universal Grammar after all. Sorry if you liked the idea of it. But, what you do have might be very much more remarkable: the ability to analyse, memorise, learn, and track all the language you're experiencing all the time. Impressive, right? If only we had a catchy name for it...

Further reading

Noam Chomsky (1998), On Language, The New Press.

Daniel Everett (2008),
Don't Sleep, There Are Snakes: Life and Language in the Amazonian Jungle, Profile Books.

Randy Allen Harris (1995), The Linguistics Wars, Oxford University Press.

Steven Pinker (2003), The Language Instinct: The New Science of Language and Mind, Penguin.

Geoffrey Sampson (2005), The Language Instinct Debate, Continuum.


Add a comment