The Language Instinct

StevenPinker, The Language Instinct, Penguin Books 1994 - ISBN 0060976519

From the cover: Language is an instinct, as innate as building a web for a spider, or flying to geese. Pinker talks about every aspect of human language, its origin, its uniqueness to humans, the way children acquire it, its structure, ...


A must-read if you have kids younger than four. You'll never think about the brain or intelligence the same way again. Among other things, you'll never believe all those stupid uses of the WhorfianHypothesis any more.

This is one of five books that I buy in lots because I'm ever giving copies away. -- JimCoplien


My husband and I kept snatching this book out of one another's hands; if he left it unattended, he'd come back to find me nose-deep and reluctant to return it. A fascinating overview of what is known about human cognition, language, and language development. I especially enjoyed the chapters that discussed language defects, and what suggest about the innate components of language. If you believe in talking animals, you will not enjoy Pinker's review of the research to date.

A book which rings many of the same bells for me is TheMoralAnimal, a discussion of human ethics as they may have been influenced by natural selection. --BetsyHanesPerry


I was finally able to read a good portion of this recently. To be fair, he does not do a very good job of discussing how, as my Psycholinguistics professor at CarnegieMellonUniversity explained, Chomsky's view loses influence the further away you get from MIT. Many linguists use the same examples (and many he does not cite) to show language is more social/learning based, with few specialized structures (or even none at all [How is this possible in light of studies on creoles? --DavidPorter ]). My guess is that it is a GoldilocksSolution, where Pinker is close to one extreme.

While reading the book, I couldn't help but wonder, though, that if there really is a LanguageInstinct?, could there also be a ProgrammingInstinct?? Stronger (more advanced?) in some than in others? --DavidGauthier

Yes. Consider that modus ponens is an instinct.

No - simple logical rules aren't particularly inherent to the way people think. They have done studies with preliterate cultures, and members tend not to do syllogisms: "All animals in the arctic circle are white, polar bears live there. What color are they?" "Well, I've never seen one, but most of the bears I have seen are brown." People who were taught reading wavered between brown and white. And this should come as no surprise, considering how often we are duped by simple logical fallacies. -JoshuaGrosse

I want to see an ActualReference for this study. There are too many urban legends out there about some study proving this, that, or the other thing. Also, I would have to see the actual methods used. In particular, many languages have very specific ways of indicating that a statement is supposed to be hypothetical or about something that is specifically outside of one's personal experience, and if the interviewer did not phrase things exactly right, the interviewee could have easily misunderstood the question. I have seen a lot of research thoroughly botched this way and tend not to believe these things when I read them.

I believe the study concerning syllogisms is by Luria (1971), available at http://www.marxists.org/archive/luria/works/1979/culture.htm. "But on the basis of my words, 'in the north, where there is always snow, the bears are white,' can you gather what kind of bears there are in Novaya Zemlya?" - "If a man was sixty or eighty and had seen a white bear and had told about it, he could be believed, but I've never seen one and hence I can't say. That's my last word. Those who saw can tell, and those who didn't see can't say anything!"

As for TheLanguageInstinct itself, it's a good introduction to the field of linguistics but a poor representation of current thought in the field. (I did buy multiple copies to give to friends and family just so they could have an idea of what I do in grad school!) Chomsky's views are definitely overrepresented (see HeadDrivenPhraseStructureGrammar for one alternative), although I would say that most linguists, even those outside of Chomsky's camp, agree that something about language is innate, and it is one of the tasks of linguistic theory to determine just what that something is. I don't think many people outside of Chomsky's camp are emotionally attached to the idea, though. If it can be shown that human language is entirely learnable and that language acquisition is part of general cognitive development, well, that's interesting too! -- SteveConley

You misunderstand what is, and what is not, Chomsky-ian. HeadDrivenPhraseStructureGrammar is not Chomsky (aside from historical ways in which Chomsky has agreed with the rough idea of anything being something vaguely like "head driven"). You can identify Chomsky's camp easily: it is literally unintelligable to other linguists who are not Chomsky followers (citations available to back that, it's not an exaggerated flame).

Similarly Chomsky does not have a coherent theory (not coherent enough to be either true or false) of the innateness of language. This is essentially universally agreed by non-Chomskyites, and this was, IIRC, even discussed by Pinker, wasn't it? -- DougMerritt

Actually, you seem to have misunderstood Mr Conley's post above. He wasn't suggesting that HPSG has anything to do with Chomsky; in fact, it's completely different from "Minimalism". I like to think that Minimalism is on the wane, since it's a top-down theory (Chomsky had a BigIdea, and he and other interested researchers continually scrabble for more data to fit into it), and since other theories of syntax (such as HPSG) have taken a more rigorous approach from the outset, they will likely prove more viable.

Also, SteveConley is sugar-coating his answer a bit -- pretty much all linguists accept that language is innate, for some definition of "innate". However, the controversy of the innate-ness of language is very similar to the controversy surrounding evolution of species, with the caveat that even fewer people who rail against this idea have the information/motivation to really study the issues.

You're right that Chomsky's immense corpus of work on Transformational Grammar (and later incarnations) is pretty much opaque to outsiders, but that's a particular syntactic theory and isn't relevant to the issue of a language instinct. And on the subject of "coherent theories", no one has a completely unified theory of any aspect of language, but obviously it's a fallacy to assume that it must therefore be based on metaphysics.

Agreed on all counts (including that I misread Conley -- but that's ok, since it triggered this good clarification from you).

Part of what I meant was that, although technical research on grammar is naturally opaque to non-linguists, on the other hand, non-Chomskyian research is not typically opaque to other linguists, even if they are outsiders to the camp in question. Chomsky, however, is the special case: his more recent stuff (last 2 decades, say) is opaque even to other linguists who are outside the Chomsky camp.

One thing that confuses these issues is words like "innate" and "instinct". In modern biology there is a good understanding that old debates on "nature versus nurture", etc, were too polarized to meaningful, in a very strong sense. MuAnswer applies. In almost all cases, phenotypes are shaped by an interaction of environment and heritage. And each term is complex; environment may mean nurture, society, the environment of the womb, the environment the mitochondria provide the nucleus, the environment that nuclear metabolic RNA provide to the nominal genes on exon DNA, etc.

It is therefore appropriate to regard words like "instinct" as being loose and of having some poetic license. And in that sense, as you said, pretty much all linguists accept that language is innate, for some definition of "innate"., yes. -- DougMerritt


See also SeeingVoices.


CategoryBook CategoryNaturalLanguage

EditText of this page (last edited July 28, 2005) or FindPage with title or text search