Chomsky

TalkLe Salon des Amateurs de la Langue

Join LibraryThing to post.

Chomsky

This topic is currently marked as "dormant"—the last message is more than 90 days old. You can revive it by posting a reply.

1tomcatMurr
Edited: Dec 14, 2009, 3:04 am

I have taken the liberty of copying a conversation from le salon and posting it here:

Geneg
I took a couple of language courses in college, mostly oriented toward teachers, and have been interested in language as a subject since. I don't know how one can separate the nature of speech from it's biology. I look at my dog, perched on the edge of consciousness, and wonder if he had a mouth instead of a snout if he could speak. I've tried to train him to say "throat" by gently pressing his throat between my thumb and fingers while saying throat, in a non-threatening voice, but he just doesn't seem to get it. I think it's because of the physical structure of his entire voice apparatus. So, yes, physical, too. What were some of Chomsky's insights into speech? Wasn't he all about the neuronal pathways and structures, as he put it, the deep grammar underlying human speech?

Booksfallaprt
Yeah, Chomsky's major deal was the theoretical focus on the hypothetical deep grammatical structures in the brain, underlying the manifestations (in any given language) that are functions of the vocal apparatus, situational choices, and ultimately nurture as opposed to nature. So if I say chien instead of dog, or "There is a cat in the house" as opposed to "a cat is in the house", the surface form of the utterances is different, but the deep structure is the same--and before getting to the level of a real dog or cat; we're still just talking about sound patterns conveying abstract meaning. To his credit, Chomsky didn't dismiss the study of variation, social and historical language use, and all the elements of language without anything directly to say about deep structure (although many of his followers did, leading to a dramatically named period in the 60s through 80s known as the "linguistics wars"), but he did certainly assign it a subsidiary position to the grad project of defining everything that language can be, as deduced from existing tokens and the attempt to apply rule-based systems (generative grammar, "minimalism", the post-Chomskyan optimality theory) to an idea of language as essentially a cognitive process.

And that's been a lot of the difficulty with Chomskyan generative ling--once you define the problem as "what are the possible forms of language" and identify the answer as existing in the structure of the brain, linguistics becomes a matter for neuroscience (for those aspects of language whose embedding in the brain can be empirically identified, e.g. the processing of speech inputs in the auditory and visual cortexes, or the effects of brain damage on speech in e.g. stroke) or for social-science fieldwork (how language varies geographically and socially; the preservation of dying languages) or for the humanities (rhetoric, discourse analysis--see the abovementioned Searle and speech-act theory) and not much room is left for the Chomskyan inferential-logical model (especially not the "only thing that really matters" space that was carved out for it in North American schools in the sixties).

And that's to say nothing of how the Chomskyan bias slights articulation and acoustics--production and perception are now being realized to take place much more on the level of reflex and body movement, and not on the level of underlying abstraction--and it's become unclear that it's really appropriate to posit separate centres for language, as opposed to semi-undifferentiated centres in the brain working in conjunction with muscle memory-type kinesthetic processing at the local (e.g. oral) level), at all. Language could just be a matter of reacting to stimuli, like feeling temperature or taking an advil, and the rest is all abstraction imposed by humans from outside. So interesting!

Geneg
One thing that strikes me is the ubiquity of the phrase as a component of language. Reminds me of a buffer. As if the brain buffers phrases while fitting them into some deeper narrative rendering their order immaterial. Of course languages tend to regularize phrase presentation. But there doesn't really seem to be a necessity for standardization.

Booksfallapart
Yeah, it's always interesting to consider whether things don't appear because they can't or just coincidentally. No language, I believe, has a verb-object-subject phrase order. But aaaaalmost no languages have, say, clicks, and yet they do exist, so is it possible for VOS order to exist and just be militated against in some way? And what would that way be? By extension, like you say, why phrases at all? Why parts of speech? Could there be a language with only verbs?

2tomcatMurr
Edited: Dec 14, 2009, 3:10 am

Gene, you are right about the ubiquity of the phrase in spoken language. just what people mean by 'phrase', however, is a source of bitter contention.

Computational linguistics has made it pretty clear that strings of words are in fact the main building blocks of language; and the main area of contention with Chomsky's notion of generative grammar. Very few linguists now support the notion of generative grammar.

My own position is that it seems likely that for highly inflected languages (e.g.Finnish, German) generative grammar plays a greater role than for less inflected ones (e.g. English, Mandarin Chinese).

Chomsky's notion of deep structures are making a bit of a comeback, with the difference that the structures are now thought to be even deeper: perhaps only on the level of a verb-noun distinction only, where the difference between 'structure and 'category' becomes unclear.

3MeditationesMartini
Dec 19, 2009, 8:25 pm

I should note that my take is only my take, and that if you want to experience a bunch of smart people bashing Chomsky on a regular basis more eloquently than I possibly could, you should visit languagehat.com (I think Anna also recommended this site). I should also note that I like Noam's politics, if not his linguistics.

4Porius
Dec 21, 2009, 8:37 pm

too timid, too timid.

5MeditationesMartini
Dec 22, 2009, 12:40 am

heh

6bobmcconnaughey
Dec 29, 2009, 9:41 pm

As a total ignoramus in re any background in linguistics - my impression is that Chomsky's role in linguistics was rather akin to Freud's in psychology. It's not that he was right - it's that he made linguistics a discipline worth thinking about. Being "right" isn't nearly as important as stimulating thinking.

7Mr.Durick
Dec 29, 2009, 10:35 pm

One of the Psammetichuses caused an experiment to be done that established Phrygian as the natural language of human beings. That was reported in Herodotus, I believe.

But William Jones really got the ball rolling in linguistics. He was not the first to recognize similarities in Latin, Greek, and Sanskrit, but his speech engendered a movement.

Chomsky co-wrote The Sound Pattern of English which was pretty good. Structuralism had fallen into a period of tedium, and Chomsky's conjectures brought life back to the study of languages.

Robert

8MeditationesMartini
Dec 30, 2009, 4:57 pm

>7 Mr.Durick: Well, it depends on how broadly you define "linguistics", I suppose. Structuralism had fallen into a period of theoretical tedium, maybe, but there was still plenty of interesting descriptive work being done--work that was somewhat enervated by the advent of the Chomskyan theory-first model. And the whole linguistic turn in philosophy was moving into its poststructural phase contemporaneous with Chomsky's major work, and that is deserving of note. And there was Labov and the rise of sociolinguistics starting in the sixties. If you're referring to The Sound Pattern of English specifically, though, I do agree that it was interesting and novel.

9bjza
Dec 30, 2009, 6:32 pm

>6 bobmcconnaughey: Chomsky isn't quite a Freud. Linguistics was actually a very well-respected field at the time when Chomsky arrived on the scene. (Take for example the influence it payed on Lévi-Strauss.) If someone is interested in the history of that period, The Linguistics Wars and Linguistic Theory in America are both decent introductions. Newmeyer has generative bias, but even he notes that the idea that linguistics was a failed science or in decline would have been rejected by most pre-Chomskyan linguists.

>2 tomcatMurr: I wonder if you can you clarify what you meant by: "Computational linguistics has made it pretty clear that strings of words are in fact the main building blocks of language; and the main area of contention with Chomsky's notion of generative grammar."

Are you thinking of only statistical approaches?

11tomcatMurr
Edited: Dec 30, 2009, 10:38 pm

>9 bjza:.
Sure, I'll try. Forgive me if a lot of what follows is obvious: I don't know how much background knowledge you have. :)

Generative grammarians hold to the idea that production of language happens by selecting grammar rules and combining them with vocabulary; that every utterance is therefore minted anew by the selection and combination of vocabulary with grammar rules. They hold that the main building blocks of language are grammar rules. On a neuro-processing level, this means the rapid selection and application of grammar rules, combined with a monitor function for self-correction and clarification. (This is a gross oversimplification of a very complex model, of course.). This is reflected in the way languages are taught: learning grammar rules and then drilling them with substituted lexical items.

However, the presence of repeated strings of language in corpora implies that language production is rather the product of memory rather than application of grammar rules; that the role of memorised lexical items of varying length (from two word combinations - 'of course'- to whole phrases- 'do you know what I mean?') is more than just an adjunct to the application of grammar rules.

so this leaves us with two neuro-processing models: the model which holds that the brain produces language by the application of grammar rules (Chomsky's generative grammar), and a model which holds that the brain produces language by the selection and production of pre-memorized strings of lexical items. Of course the truth probably lies somewhere in the middle, and there is no reason why the brain cannot switch constantly between these two modes of language production.

My own hunch as a researcher in CL and teacher is that the second model is probably more likely.

I hope that's clear?

12bjza
Dec 31, 2009, 1:34 pm

Ah, ok. Are you by chance familiar at all with recent lexicalist work on Multiword Expressions? Might be what you're hinting at the possibility of toward the end there. E.g., http://mwe.stanford.edu

I haven't kept up myself over the last year, but I find it reassuring that some theory group is working toward solutions on overlooked problems like this. But I guess that's what happens when the group's goals include broad-coverage grammars (unlike a certain linguist named by the thread, ahem).

13tomcatMurr
Jan 2, 2010, 7:44 am

Accccccxxxxhem!

I also have not kept myself up in the field, as I am too busy with my Russian literature studies (which, franchement, are what I live for). Most of my observations and knowledge reflects the current state of play at around 2000, which is when I was researching this stuff. Occasionally whispers of what's going on reach me in my mountain eyrie through the fog of Pushkiniana.

if you're interested, I can send you a paper I presented on multiword items at AILA in Singapore in the early noughties.

Anyway, thank you for drawing my attention to this. It is a fascinating area of research, one that overlaps CL, grammatology, literary use of language, and language pedagogy, main interests for me.