« Brainboard 3 | Gene Expression Front Page | Just a bit less scared.... »
March 18, 2003

Universal Grammar [part I]

Marc Miyake of the blog Amritas (on our permalinks) had an article up on FrontPage yesterday blasting Chomsky's Universal Grammar. His thesis is that UG is just another intellectual virus of the post-modern left. In it he quotes another of our favorite bloggers, Joanne Jacobs, who relates her school-day dissatisfaction with old Noam as well:

Structural linguistics was required for a degree in English at Stanford. I put it off till my last semester; finally I had to take the class. It consisted of uncritical worship of Noam Chomsky. I kept disrupting class by asking questions: Why do we believe this is true? Just because Chomsky says so? How do we know he's right? Why is this class required?

So how about it, is UG the new Freudism, just another high falutin academic fad, or is the majority linguistic paradigm really deserving of its special regard within university hallways? Marc is adamant that it is the former:

Prior to Chomsky, linguists engaged in a lot of data collection to understand the diversity of human language. I'm vehemently anti-PC, but in this case, I think the word 'diversity' is justified. There's a lot out there, and someone's got to catalog it.

However, Chomsky rejected this approach. He wanted to look into something deeper' (academese for 'pretentious and nonexistent'). So he invented something called 'universal grammar' which is somehow programmed into us at birth.

Unfortunately this is all highly misleading. Tailored for a political audience of a certain persuasion[1], Marc would like to convey the story in a way that makes Chomsky out to be the left-wing baddie (that the audience is already prepared to think of him as) who subverts decent American science. In Marc's moralized version the 'good' linguists were going about their proper duties collecting all kinds of data, while the 'bad' linguist decides to self-indulgently taint the field with PoMo gibberish (something FrontPage readers already outraged by Chomsky's radical politics will be eager to accept). But what Frontpage readers won't know is that the role Chomsky played was exactly the opposite- it was the gibberish he helped to contradict. Despite his poor reputation among modern conservatives, it should be made clear that Noam Chomsky was among the first to raise a serious challenge to the hegemony of the Boas school of thinking which denied the concept of a human nature and taught the infinite malleability of man.

What Marc doesn't tell you is that all the other linguists, like so many at the time, were radical environmentalists (Skinnerists) who believed in the infinite variety of human language- that there was no form that human language couldn't deviate from. Chomsky wasn't being pretentious when he volunteered to "look deeper", he was looking for the answer to pertinent question: Is language infinitely changeable, or does human language have fundamental commonality? Chomsky is a scientific hero b/c he dared to stand up, at a time when almost no one else dared, and question if an essential part of being human was innate.[2] Iíll let Steve Pinker summarize (far better than I could) the theoretical strength of Chomsky's theory of language:

Language is a human instinct. All societies have complex language, and everywhere the languages use the same kinds of grammatical machinery like nouns, verbs, auxiliaries, and agreement. All normal children develop language without conscious effort or formal lessons, and by the age of three they speak in fluent grammatical sentences, outperforming the most sophisticated computers. Brain damage or congenital conditions can make a person a linguistic savant while severely retarded, or unable to speak normally despite high intelligence. All this has led many scientists, beginning with the linguist Noam Chomsky in the late 1950's, to conclude that there are specialized circuits in the human brain, and perhaps specialized genes, that create the gift of articulate speech.[3]

Acquiring language is not a normal mental problem. Everyone sees small children pick up language so effortlessly that they might not realize how impossible a problem it is. No one would expect a three-year old to master calculus, it couldnít be done! But when a three year old learns grammar he is doing something much more complicated, and w/o effort. Children must learn the rules of their native grammar (whatever it might be) through a limited set of sample sentences that they hear. This limited information is mathematically insufficient for determining grammatical principles (Linguists working on a language for decades cannot figure out what a small child does in a few short years). This is what is called "The paradox of language acquisition". The normal computational function of the brain is insufficient to explain how children are able to do this- UG is the best current solution. Marc isnít much interested in providing an alternative in his article though. He continues:

There are innumerable problems with [UG]. For starters:

Where did this 'universal grammar' come from, and how did it end up becoming part of our biology? Not many Chomskyans are interested in evolutionary biology. 'Universal grammar' simply IS. (I myself suspect that there may be a universal grammar sans scare quotes, but I doubt that it has much in common with Chomskyan 'universal grammar'.)

The claim that, "Not many Chomskyans are interested in evolutionary biology." is an utter falsehood. Evolutionary biologists_love_the idea of universal grammar. Evolutionary biologists swear by universal grammar. Tooby and Cosmides essentially developed their evolutionary psychological model off of Chomsky's idea. Instead of one undifferentiated mass, scientists began, largely thanks to Chomsky, to imagine the brain as a sort of Swiss army knife, with many specialized sub-systems adapted for unique problems; and in many ways this view of the brain has found powerful evidentiary support. Marc confuses viable questions as damning critiques. It isnít necessary to know any of the things Marc asks before UG can be useful as a theoretical solution (although they would certainly enrich the theory). In fact many evolutionary biologists are searching for those answers, and with progressive success. Theoretical Biologist Martin A. Nowak (Princeton) is one important scientist who is looking into the role of language in human evolution (and the role of evolution in human language). The beginning of the answer to Marcís questions can be found in Nowakís 2001 paper in the journal Science titled the Evolution of Universal Grammar.

[Part II later. . .]

[1] Marc got all sorts of props for this article over at the FreeRepublic, which is world-famous for its concern with issues relating to neurological theories of syntax :P

[2] Similarly, around the same time, it was widely assumed that facial expressions could vary arbitrarily and w/o limit until psychologist Paul Ekman demonstrated that facial emotions had continuity across all cultures. Donít ask me how, but somehow itís "programmed into us at birth".

[3] Seven years after Pinker penned that introduction in The New Republic the first of such language genes was found.

Posted by Jason Malloy at 02:17 AM

one of the freepers says that they aren't too impressed by linguists. you know what impressed me about lingustics? indo-europeanologists figured out what an indo-european language would look like extrapolating from latin, greek, sanskrit, etc. the problem of course was that the oldest indo-european language recorded was classical era. wait! they found hittite! recorded circa 1600 or so-even before linear B greek (which was a bit later in chronology both historically and linguistically as far as decipherment by ventres)-and guess what, hittite looked how it should have looked according to linguistic theories. it predicted something! that impresses me.

now, linguists, tell me, is this an academic myth meant to impress ppl steeped in the natural sciences that worship at the altar of predictive power?

and you know, many linguists become programmers, how many PoMo intellectuals do? wait, don't answer that :)

Posted by: razib at March 18, 2003 02:49 AM

Figured i'd do people a favor and break the critique up into two parts. I haven't read Chomsky*, its pretty heavy stuff, but I can tell why the majority of linguists find him useful. Next post I'll deal with things like "transformations" and "deep structures" which aren't as PoMo as Marc insists.

*Pinker is to Chomsky, what Dawkins is to Hamilton. Thank God there's people to transmit this stuff down the IQ chain.

Posted by: Jason Malloy at March 18, 2003 03:18 AM

well-i think pinker duz more original research that dawkins. dawkins is a popular book machine-but he holds a chair all about popularizing science....

Posted by: razib at March 18, 2003 03:59 AM

Yeah but, is Marc wrong? Isn't the big knock on UG and Chomskyans their focus on the English language? Don't the rules of UG conveniently mirror the syntax rules of English? Marc is saying that the UG theory itself as explained by Chommy is wrong--not that there isn't something like it. I mean, is Tooby and Cosmides' idea anything like Chommy's?

Posted by: Justin Slotman at March 18, 2003 05:13 AM

My point being: you don't have to be a Chomskyan to believe there's a universal grammar something-or-other. There you go.

Posted by: Justin Slotman at March 18, 2003 05:22 AM

Good post.

Posted by: Charlie Murtaugh at March 18, 2003 06:11 AM


Here is the conclusion to Marc's essay:

Is Chomsky a double fraud in both science and politics? I honestly don't know. I have never met him and don't want to - the urge to verbally attack him is too strong. Maybe he really believes what he says in one or both fields. But in any case, Chomsky is a troublemaker on two fronts. He is like Lenin and Lysenko rolled into one.

To compare Chomsky to Lysenko is not only wrong*, it is absolutely shameful. Chomsky completely revolutionized his field and is responsible for not only the current paradigm in linguistics, but the one of cognitive science in general. (and even proved useful to the technical fields!) That isn't worthy of suggestions of Fraud, that is worthy of one thing: respect.

And no, you can't believe in some sort of universal grammar or innate language acquisition device (or even foundational mental structures for language) without giving Chomsky a measure of proper credit. (Lysenko...feh)

As for the English thing that's part of Post II.

*Even if we were to accept Marc's critique there's still a huge difference between working off of a theory with little hard evidence to support it, and working off of a theory that actively contradicts well established scientific knowledge.

Posted by: Jason Malloy at March 18, 2003 07:57 AM

Yeah, I happen to think Chomsky is a great linguist. But, like many academics, he thinks that because he's brilliant in one field, he must be brilliant in many (politics, international relations). In reality, in political theory, he's about as intellectually well-developed as your average college student. I mean, really, he wrote paeans to Pol Pot.

He should stick to linguistics.


Posted by: David at March 18, 2003 08:42 AM

Well written. A convincing argument that Marc Miyake is engaged in partisan fuzzy thinking...
One of the interesting things about linguistics is that it reaches all the way from what might be considered soft, fuzzy science (e.g. anthropological/linguistic studies in the Amazon Basin), where PoMo influence is strong, to very hard and rigorous CS-related science (build me a parser for language X). Computational Linguistics draws a broad range of people for just that reason, I think.

Posted by: at March 18, 2003 08:43 AM

Has anyone here had a look at Geoffrey Sampson's Educating Eve? It's a thoroughoing (and to my non-expert ear) pretty convincing attack on "innatism" in linguistics. It specifically takes on both Pinker and Chomsky.

Posted by: Aaron Baker at March 18, 2003 08:47 AM

Gottfried Wilhelm Leibniz(for any Freepers out there:German philosopher, mathematician, and logician best known for having invented the differential and integral calculus, independently of Newton) said it best in the debates about the tabula rasa raging in the 17th century:
"There is nothing in the intellect that was
not first in the senses. Except the intellect itself."
Of course the brain is predisposed to learn language. To argue otherwise is analagous to creationism.

Posted by: martin at March 18, 2003 09:33 AM

Of course the brain is predisposed to learn language.

Not necessarily Martin. Going off of first principles alone, nothing's to say that consciousness itself is a mandatory result of higher brain function much less language ability.

But of course we do speak. And of course it is our biology that enables us to that. But we also water ski and our biology also enables us to do that....but we don't say we have an innate water-skiing function. A question might be 'is the brain specified for speech function'? (adaptation vs. by-product)

That parts of the brain have special roles in speech function is fairly clear, but that doesn't necessarily answer all of the questions, and "predisposed" can be a tricky word. (for instance are we "predisposed" to seek shade? admire sunsets?)

*Personally (obviously) I think the case for general adaptative language function overwhelming and for UG specifically, pretty compelling. I'm just saying that you don't necessarily get there from 1st principles.

Posted by: Jason Malloy at March 18, 2003 11:51 AM

Jason-perhaps I'm running ahead of the evidence-but all data I see seem to indicate the brain is as predisposed to communicate as the stomach is to digest. True, human language might be to the brain what twinkies are to the stomach, but communicative intent is just as hardwired as is the secretion of digestive enzymes. Of course fully formed sentences have to be learned-but some type of mentalese is inherent in the neurons, i.e. the slate is not blank. I see this as occuring along an evolutionary continnuum from pure (i) genetic programming, to(ii)instinct molded by learning, to (iii) acquired art, e.g. (i) Sub-oscine songbirds develop normal song even with being reared in isolation or artificially deafened in their development or being reared in isolation, indicating their song is not learned but rather genetically programmed; (ii) all oscine songbirds will develop highly degraded song if they are deafened during development, and develop degraded but species-typical songs if reared in isolation. It is quite clear that human language is in the same category as the songs of oscine songbirds, and similar effects are noted in the unfortunate case studies we have. The burden is on anyone asserting language is an art acquired by a blank slate mind to explain the available data.
The posted article offers no alternative explanation, and it's clear why the author is having trouble securing a job. He's a crank without even the courtesy to offer a crankpot hypothesis.
If the mind only "thinks" in purely formed words-how do we learn language in the first place-or any new word for that matter?

Posted by: martin at March 18, 2003 01:55 PM

P.S. Jason - those questions are not addressed to you per se-but the author and perhaps Mr. Baker, supra. I'm not dogmatic about UG-it just appears to be the theory that best explains the observable facts.
Mr. Baker-I haven't read Sampson's book-but I read in a review he is a dualist, and asserts that language is learned by the mind-not the brain. Is that correct?

Posted by: at March 18, 2003 02:03 PM

dr. miyake taught at my alma mater for a few years-hey, dr. miyake, do you know jason mak-VP of APASU form circa 1996-2000 or something? i was kind of friends with him for a while. also, do you know carol lin??? she took a bunch of chinese classes i think.

Posted by: razib at March 18, 2003 02:09 PM

Good post.

Pinker's "Blank Slate" has a good description of the Romantic roots of Chomsky's political views.

Is this Atrios blog worth reading? What little I heard of it, it sounds pretty consistently dumb, but I may be wrong.

Posted by: Steve Sailer at March 18, 2003 05:27 PM

screeching left-wing version of instapundit. judge from that :)

Posted by: razib at March 18, 2003 05:44 PM

Hmm..praise from Razib, Godless, Murtaugh, and Sailer. Unless Chomsky himself drops in unexpected a la Kevin MacDonald*, I'd say it can only go down from here.

*so what ever happened with that anyway?!

Posted by: Jason Malloy at March 18, 2003 06:02 PM

KMD toles me he was 2 buzy & there were too many points that he'd replied to many times in other debates. he'll prolly swing by again though-he can't avoid a judeoversy :)

Posted by: razib at March 18, 2003 06:25 PM

Dog, separated at birth, 6 months isolation, dog barks! Just like dogs otherside of world! Wimpers when sad, growels when mad, sorta yelps when happy. Turns head from side to side when trying to discern sounds from bigger brained animals.

Posted by: reader at March 18, 2003 06:51 PM

a freeper saw the following message by razib here (the first reponse): "one of the freepers says that they aren't too impressed by linguists" and has posted a new message saying:

"Tell them the FReeper's name is AmishDude. And he's still not impressed. On the academic food chain, they're still herbavores."

Um, that's all. I'm just a messenger.

Posted by: robber_baron at March 18, 2003 11:45 PM

uh, i'm not a herbivore myself-i'm an omnivore....

Posted by: razib at March 19, 2003 04:09 AM

Sampson does plump for dualism near the end of his book--a view that, whatever its merits or demerits, didn't seem to me to have much bearing on his critique of Chomsky, Pinker, et al. The book really does appear to me to be an important contribution to exactly the debate being aired here, and I hope discomfort over dualism won't keep people from reading it.

Posted by: Aaron Baker at March 19, 2003 06:56 PM