Linguophiles: Help me settle this debate on the Chompskyian model of linguistics.

September 28, 2006 4:35 PM

Linguophiles: Help me settle this debate on the Chompskyian model of linguistics.

I was talking to a friend recently about "The Language Instinct," by Steven Pinker, which I've been reading recently. He presents many arguments which are fairly thorough as to why the Sapir-Whorf hypothesis is wrong and why the systems of language and grammar tap into something innate in our brains which children especially find malleable.

Essentially, my friend's claim was that modern linguists reject the Chompskyian model and it's been largely disproven. One of his major points was that without physical neurological proof, positing any model is useless, but I think this is basically antithetical to the progress of all sciences.

After a bit of hubbub, we concluded that neither of us knows enough about the other's perspective and the arguments already in the public discourse for each.

Basically--is this true? Say what you want about Chompsky's politics, but I find it hard to believe that this model was simply "discarded." It is a lot more logical and thorough, and accounts for a lot more than a Sapir-Whorfian POV.

If there are large bodies of work opposed to this model, which are the most well-known and thorough?

posted by Lockeownzj00 to writing & language (20 comments total)

Chomsky's claim is that "grammar" is innate and autonomous. It is this combination of claims for which there is no evidence. No phenomona claimed to be part of "universal grammar" have ever panned out.

Chomsky claims it is pointless to study meaning, and while he may discuss 'semantics' what he really is talking is the scope of quantifiers.

Any linguist that deals with meaning rejects Chomsky's framework.
posted by MonkeySaltedNuts at 4:57 PM on September 28, 2006

Essentially, my friend's claim was that modern linguists reject the Chompskyian model and it's been largely disproven. One of his major points was that without physical neurological proof, positing any model is useless, but I think this is basically antithetical to the progress of all sciences.

Well, very few linguists do any neurological work at all, and even fewer have more than a basic acquaintance with the literature. Hell, a lot of psychologists working at the borders of psychology and linguistics don't even do neurolinguistic work (they rely on experimental methods that are more about people's response to stimuli, as opposed to brain imaging, etc.) So I think your friend is a bit off base here. Positing a model in absence of direct neurolinguistic evidence is far from useless -- it is making a falsifiable claim. There may be a few linguists (or psychologists) who hold your friend's point of view (perhaps s/he took a class from one?) but they are not common.

Also, we are decades (if not further) from having any real understanding of the neurology underlying the human language faculty, so you basically couldn't work on the problem if you were only willing to consider direct neurological evidence. Most work like this that I know of is done by taking a linguistic claim, and doing brain imaging in some way or another while the subject is performing the relevant language task, and seeing if you get some kind of response that correlates with the task. That is, it involves testing linguistic claims that were formed in absence of any neurological proof.

As to whether the Chomskyan model has been "disproven", this is a little harder to approach. First of all, generative grammar has not been "disproven" and many people are still doing it (there are also linguists using non-generative models that achieve very similar results). Chomsky has proposed several models (with lesser and greater degrees of sketchiness) of syntax. Most mainstream syntacticians are working to some degree within the most recent one, which is known as "minimalism". So it hasn't been disproven in the sense that many syntacticians are using a chomskyan model. Of course there are many different versions and minimalism seems to be shaped as much by other practitioners in the field as by Chomsky. Furthermore, minimalism directly inherits most of the work in earlier Chomskyan models of syntax. Particular facts about these theories and analyses framed in them have been argued against, but the overall picture has been by no means abandoned. There are many other theories of syntax, (Lexical Functional Grammar, Head-driven Phrase Structure Grammar, Tree Adjoining Grammar(s), etc.) which have their adherents, but all of these are in some sense directly descended from Chomsky's work in the 50s and 60s, even if the practitioners might vehemently disparage current theory.

As far as I know, basically no linguists believe anything remotely close to a Sapir-Whorf point of view. (By the way, it may be worth pointing out that this is not in direct opposition to any theory Chomsky has come up with, so far as I know; the arguments for and against each are independent and it is probably possible if unlikely to be a Whorfian minimalist.)
posted by advil at 5:01 PM on September 28, 2006 [1 favorite]

Any linguist that deals with meaning rejects Chomsky's framework.

This is simply false -- most people working in mainstream semantics assume some version of Chomsky's syntax (I'm one of them). It is not necessary to make this assumption (you just need some coherent view of the syntax), but if you want to communicate with other linguists it is practically a requirement. It certainly does not hinder the study of semantics or of pragmatics.

At one point Chomsky did claim it was pointless to study meaning, but this claims have been muted significantly since the mid-80s. You'll also notice (if you were to look) that MIT produces quite a few semanticists, and employs some of the more prominent ones in the field.
posted by advil at 5:06 PM on September 28, 2006

Most mainstream syntacticians are working to some degree within the most recent one, which is known as "minimalism". So it hasn't been disproven in the sense that many syntacticians are using a chomskyan model.

"Mainstream syntacticians are using it" is very far from equaling "it is proven." Mainstream astronomers used the Ptolemaic theory for centuries, and mainstream doctors bled people for even longer. In this case, there are plenty of linguists who think Chomsky led the field off on a wild-goose chase into the woods from which it's still trying to find its way back. (And listening to people from MIT talk about Chomsky is like listening to people from the Vatican talk about the Pope; it can be entertaining, but you're not going to get much perspective.)
posted by languagehat at 5:40 PM on September 28, 2006

Chomsky's claim is that "grammar" is innate and autonomous. It is this combination of claims for which there is no evidence. No phenomona claimed to be part of "universal grammar" have ever panned out.

This isn't a sufficient argument. I can come up with numerous real-world examples which really can only point to the Chompskyian model.

First of all, generative grammar has not been "disproven" and many people are still doing it

Just a few notes: one, generative grammar, while I certainly put stock in it, has flaws, and is a technicality of the overarching theory. I think the important thing to take away is that like Einstein's Theory of Relativity, the model can be largely correct and yet still retain some currently insurmountable flaws. Obviously, the method/model with the best and most rational explanation should be favored.

By the way, it may be worth pointing out that this is not in direct opposition to any theory Chomsky has come up with, so far as I know

While they are not diametrically opposed, the Sapir-Whorfian model, if it were true, renders a variety of situations that do actually occur impossible, situations which the Chompskyian perspective explains--ie, workers migrating to a central location and forming an essentially grammarless pidgin; somehow the next generation supplants the 'dialect' with grammar and consistency by some innate, unconscious understanding of communication, and form a veritable language--creole.

I think the existence of creole languages alone should confirm the Chompskyian model.

As is probably apparent, I have a fair amount of confidence in it, but I want to see this alleged 'counter-evidence.'

Languagehat, I'd like to hear more of your opinion, as this level of depth is exactly what I sought in starting the thread. Forget about semantics, for a second--what about the 'innate language' concept is so repellant? What do you think of Pinker?
posted by Lockeownzj00 at 6:00 PM on September 28, 2006

I'm not a linguist but I've done coursework in the field, so I'll just give a general answer:

The way you pose the question leaves room for misunderstanding. Chomsky has made many different claims in linguistics, some of which are now discredited and some of which form a general assumption among many people in the field.

I think most working linguists accept that there is some innate language-acquisition ability in humans, that there's a critical period for language learning, and that there are some pathways our languages will always go down. That is, when children are acquiring their first language, they are not just trying out every possible "theory" (about the syntactic rules and the semantics of the language) that would account for the inputs they have gotten -- they are only trying out a much narrower range of such "theories".

But I think there is controversy over how much, and exactly what, is innate, and how it all works. Particular claims Chomsky makes about this topic are much less widely believed.
posted by LobsterMitten at 6:09 PM on September 28, 2006

Also there's no "p" in Chomsky or Chomskyan/ian.
posted by LobsterMitten at 6:10 PM on September 28, 2006

And since you ask about academics' opinions of Pinker: The working philosophers of language and linguists that I know regard his work as poor. The Blank Slate especially, but even The Language Instinct is regarded as being sloppy and distorting the real picture of how things stand in the field, to fit the conclusions Pinker wants. I don't have enough expertise to make these assessments more specific, but I know that there is generally derision when his name comes up. (And it's not just because he's a popularizer -- the same attitude does not crop up against Steven J. Gould for example.)
posted by LobsterMitten at 6:15 PM on September 28, 2006

(And listening to people from MIT talk about Chomsky is like listening to people from the Vatican talk about the Pope; it can be entertaining, but you're not going to get much perspective.)

I'm not sure who (if anyone in this thread) this was aimed at, but I'm not and have never been associated with MIT. For that matter, I'm not primarily a syntactician, and don't have a stake in this particular debate (I was attempting to provide a relatively neutral assessment of the field's take on the matter). If anything, my personal opinions would tend more toward the contrary as well.

"Mainstream syntacticians are using it" is very far from equaling "it is proven."

I didn't attempt to claim that it (for purposes if this paragraph, let us fix "it" as, say, some version of the minimalist program) was proven, but rather that it has not been disproven, and is for practical purposes accepted by the majority of the field, and I do stand by this statement. Offering a pointer to counter-arguments might be more productive than analogizing Chomsky's theories to bleeding or heliocentrism.
posted by advil at 6:47 PM on September 28, 2006

Also, I think LobsterMitten put some of the things I wanted to say but was not very coherent about in a very clear way.
posted by advil at 6:50 PM on September 28, 2006

Mr ownzj00, if you're going to be engaging in lengthy discussions about the theories of Mr. Chomsky, could you perhaps learn to spell his name correctly?

Yeah--I donno what got into me. For some reason I kept instinctively putting that p there, for no real reason.

I'd like to hear more about Pinker. I can see where the argument's coming from, but so far, I can't find anything in The Language Instinct that seems to be entirely skewed. Then again, I'm only about 80 pages in.
posted by Lockeownzj00 at 7:05 PM on September 28, 2006

As an autodidact and tyro in this entire subject, what I mainly know is that the real linguists I've talked to about Chomsky tend to be very abusive -- but only when other linguists aren't listening, because expressing strong anti-Chomsky feelings is not good for the career.

As to Pinker, I haven't read "The Language Instinct" but I have read other books by him and my opinion is that he is entirely too in love with the theory of evolution and the basic principle of genetic predisposition.

Especially when it comes to our brains, he assumes that a lot of higher functions are directly controlled by the genes. He completely ignores the fact that the brain is to a significant extent self-organizing. And he ignores the fact that objects as complex as the brain which develop fractally can exhibit emergent properties which are not, strictly speaking, under detailed genetic control.

He engages heavily in what Stephen J. Gould described as "Just So" stories: he finds some behavior, shows how it enhances survival, and from that concludes that it is a direct result of evolution and thus must be directly controlled by the genes.

But he never actually shows cause and effect. He never proves any of it. He engages constantly in a monumental case of Post Hoc fallacy: if a behavior enhances survival THEN it must be the result of evolutionary selection. And that argument doesn't follow.

The problem is that he pretty much totally discounts the idea that favorable behaviour could be learned and not be instinctive. And when it comes to behavior of creatures like us, that's a major blind spot in his speculations -- which is pretty much all they are.

Pinker is a good writer, but take his work with a big grain of salt.
posted by Steven C. Den Beste at 7:48 PM on September 28, 2006

Lockeownzj00 and advil (to some extent) are confusing Chomsky's theories (of innate autonomus syntax) with generative grammer in general.

General generative grammar seeks a formal (maybe algebraic) system to model language. Most any linguist that tries to elaborate the 'rules of language' who doesn't resort to vague handwaving and wishful thinking is working in "Generative Grammar", though some of the 'formal' systems can be more or less formal and precise.

Chomsky has long tried to define 'grammar' to mean syntax. And 'innate language' to mean that there is a seperate 'organ' in the brain called the 'syntax/grammar module'.

Yes, some parts of language may be innate. But Chomsky hasn't explored this large question, he has just by fiat declared that syntax is the only innate part and has come up with no evidence.

People know that sentences have parts as they do that chairs have legs. The question of how much of human 'general cognitive ablilites' are used in all aspects of language cannot be addressed within a framework that is trying to enshrine the 'only real true part of language'.
posted by MonkeySaltedNuts at 7:58 PM on September 28, 2006

LobsterMitten: And since you ask about academics' opinions of Pinker: The working philosophers of language and linguists that I know regard his work as poor. The Blank Slate especially, but even The Language Instinct is regarded as being sloppy and distorting the real picture of how things stand in the field, to fit the conclusions Pinker wants.

Well, coming from an educational/developmental psych perspective, I couldn't read The Blank Slate after the hacked straw man of Watson and Skinner. Among Pinker's many sins, he rips Watson's "12 infants" out of historical context as an argument against an Ev. Psych. that wouldn't exist for 50 years, rather than a Social Darwinism and Eugenics movement that had become entrenched in academca in Europe and America during Watson's lifetime.

As a former biologist turned psychologist, I was rather surprised to see such a clumsy treatment of the nature v. nurture argument. That is, the number of people who accept Pinker's straw man of a blank slate hypothesis appears to be quite thin. Most will agree that individual differences in ability and temperment exist, but tend to focus on the fact that human behavior appears to be amazingly plastic in response to environmental situations. Pinker's basic problem is that even with the many limitations of behaviorism, it explains a heck of a lot of human and animal behavior.

And at least as far as LobsterMitten points out, from the developmental psych lit there is little doubt that something really amazing and almost magical is going on in infancy involving the acquisition of grammar and a fair amount of vocabulary. One hypothesis in cognitive psychology is a limited number of cognitive processes build the richness of human cognition. And ironically, there is possibly also room to support a very limited Sapir-Whorf variation in that those cognitive processes appear to be strengthened by early use, and weakened by early disuse.

The problem is that depending on how you slice the data, it's possible show influence from both nature and nurture. The basic question is not either/or but by what mechanisms do each contribute to the complex phenomena that is behavior?
posted by KirkJobSluder at 8:02 PM on September 28, 2006

Lockeownzj00 and advil (to some extent) are confusing Chomsky's theories (of innate autonomus syntax) with generative grammer in general.

Um, generative grammar was started by Chomsky in the 50s, and so is quite plausibly referred to as one of Chomsky's theories, especially from the outside. It simply wasn't clear which thing (all of generative grammar, some particular theory, just the part about innateness, etc.) the OP meant by "Chomsky's theories", so I tried to answer for several cases. And like I said, Chomsky no longer tries to restrict grammar to just syntax. For that matter, it's wrong to say he always has -- one of the pioneering works in English phonology, which is in the generative tradition, was co-authored by Chomsky. (Chomsky and Halle 1968)
posted by advil at 11:02 PM on September 28, 2006

advil, Pinker was at MIT until recently; maybe that's what languagehat was referring to.

And about the "p" typo -- it's very tempting, isn't it? There's also a famous chimpazee who was taught sign language, called Nim Chimpsky, which makes the error that much more fun to make.

Here's a page that has links to a bunch of reviews of The Language Instinct by linguists when it came out, and replies by Pinker to some of the things they say.

As I recall, The Language Instinct gives a reasonable very general outline, and it's fun to read -- or was when I read it very, very long ago. But it's sloppy about specifics and glosses over what are in fact active debates. So some of the assertions of Chomsky's that are rejected by most academics, you will not hear anything at all about in that book.

(I haven't read The Blank Slate but have heard it was a travesty of a crap book -- so, wouldn't recommend that for a follow-up!)
posted by LobsterMitten at 11:21 PM on September 28, 2006

I can not speak with authority on Chomsky's most current view of syntax but as a psychologist studying language processing I can speak with some authority on Pinker. The Language Instinct is a wonderfully written and entertaining book with lovely examples. But a lot of the science is just plain wrong. As far as I can tell, Pinker is just a clever, entertaining, not quite accurate science writer who has done little original research of note but who second-hand promulgates the notion that
evolutionary explanations underlie everything.
posted by bluesky43 at 5:20 AM on September 29, 2006

I'm not sure who (if anyone in this thread) this was aimed at

No one, and certainly not you—I've always been glad to see you in threads and know that I don't have to be the standard-bearer for linguistics all by myself. You mentioned that "MIT produces quite a few semanticists, and employs some of the more prominent ones in the field," which inspired a little riff on MIT on my part. No offense intended (except to worshipers at the shrine of Chomsky).

Lockeownzj00, my basic problem with Chomsky is that he took some interesting ideas (mostly those of his mentor Zellig Harris, although they developed those ideas in very different ways) and tried to turn them into the One True Orthodoxy, seeding linguistics departments throughout the country with disciples and trying to stifle opposing views. Linguistics was never like that before he came along, and it created a very unpleasant atmosphere. (To any Chomskyans who may be tempted to disagree: yeah, right, and the Inquisition wasn't as bad as it's made out to be.) Furthermore, his basic approach—that all languages are the same once you remove a few superficial layers of rules, so you might as well just study your own language and call it universal—did a lot to damage the great American tradition of fieldwork just at a time when it was needed more than ever because of the accelerated dying out of minor languages worldwide. And his "universals" have been shot down as quickly as he's proposed them, so his approach is not only counterproductive, it's wrong.

Take all that with a grain of salt, because I'm obviously biased and haven't been professionally involved with the field for over a quarter of a century, but that's where I'm coming from.
posted by languagehat at 5:59 AM on September 29, 2006

The Sapir-Whorf "hypothesis" is a canard; it doesn't exist except as a figure of very imprecise speech. There is no such "hypothesis" in the work of either Sapir or Whorf.

The assertion that grammatical structures, lexical resources, and phonological patterns influence thought and experience is true on its face. What do you think a phoneme is, chopped liver? It's a constraint on your ability to perceive sound contrasts imposed by the arbitrary conventions of the language community into which you happened to be born. Boas pointed this out in the 20s. Does that mean you cannot learn to perceive (as meaningful) sound contrasts that are comprised of allophonic or non-phonemic segements from your language, or sounds you cannot make yourself? Of course not. But it isn't easy, as the persistence of accents in adult second-language learners demonstrates.

Pinker's old book, which is very helpful in explaining some concepts, was also a direct attack on anthropological linguistics, a field about which he knows almost nothing. He presents a very narrow view of language and equates its narrowness with its scientificness. Saussure did the same thing, and so did Chomsky. It is perfectly scientific to view language as a domain of human social behavior, and to ask how the neurobiological basis for language, which of course represents a history of evolutionary adaptation, interacts with other cognitive and behavioral faculties and patterns and functions as a medium of social life and cultural evolution.

Pinker is vying for the celebrity science guru role. It requires him to invent enemies of his perspective. It's quite tiresome and unscientific of him.
posted by fourcheesemac at 7:47 AM on September 29, 2006

Well, I definitely have a different perspective after reading all this. I did not know about Pinker fascist state, heh. I found many things I still disagreed with, but alas, I fear my thread has died.

But what's with this?

who second-hand promulgates the notion that
evolutionary explanations underlie everything.

Yeah, I would say, evolution being the lifeblood of existence, and all of human creation evolving constantly pretty much does justice to this postulation.

And he ignores the fact that objects as complex as the brain which develop fractally can exhibit emergent properties which are not, strictly speaking, under detailed genetic control.

Again I say I have not read this book fully and I have not read any of his others, but so far he has made either no or little reference to genes. He has essentially alluded to chemical reactions and brain functions, of which I don't think anybody disagrees exist.

The Sapir-Whorf "hypothesis" is a canard; it doesn't exist except as a figure of very imprecise speech. There is no such "hypothesis" in the work of either Sapir or Whorf.

Hardly. Besides the fact that Whorf actually did make several arguments which amounted to a general, theoretical idea (hypothesis) in his work, even if he hadn't, that umbrella of ideas, if but for pragmatic purposes, can be and is referred to as "Sapir-Whorfian."
posted by Lockeownzj00 at 7:05 PM on September 29, 2006