Tuesday, January 22, 2013

The Cult of Chomsky

In college, I certainly knew the name 'Noam Chomsky', though why I knew it, I wasn't sure. I worked security, patrolling the dark, deserted halls and subterranean passages which made the campus a labyrinth of shadows and half-heard sounds each time night enclosed it. My fellow denizens of the twilight realm included my nemeses: hobos, bike thieves, wall tagging skateboarders, drunken hedgebound copulators--and also my allies, my kine, those fellow security personnel who waited out the long shifts at their desks.

So I flitted from one to the next, patrol this building and visit one, then moving on to the next, spending a little time with each before returning them to their isolation. I still recall walking down a dim and lengthy hallway and hearing a clipped, minute voice echoing down it, reduced by boombox speakers to a buzzing, inhuman tone. When I arrived, my compatriot made to pause the recording, but I held up a hand and asked "what's this?"  

"Noam Chomsky," she replied, "the world's most famous intellectual".

Now there were several things which struck me as odd about this: firstly, that she was listening to a CD of him instead of reading his works, as was the normal collegiate method; secondly, the phrase 'famous intellectual' seemed to me to be a contradiction in terms. To be an intellectual means being highly regarded by your peers, to be judged by a small group of discerning, well-informed people to have accurate and defensible views. Thus, to some degree, it means being inaccessible, being an expert, a specialist capable of impressing those who already know a great deal.




Would you walk right past a Stradivarius?
To be famous, however, means to be widely accessible, to be easily understood and quickly encapsulated in a concept or persona. It means being judged by a wide group of ignorant people to represent something they aspire to, or something which touches them emotionally. But these people do not have the skill or experience to judge who is an expert, who is a master, as evidenced by world-renowned violinist Joshua Bell playing anonymously in a subway, thousands of people passing him by and almost none recognizing that they are in the presence of brilliance--but then, most people are not capable of telling the difference between a skilled street performer and a world-class musician.

So, how does one become a 'famous intellectual'? Well, Freud set up a fairly effective blueprint: first you must make a great deal of remarkable and unusual claims--ones which will give pause even to the layman--then you take the notoriety granted by the fervor you have caused and make it a podium from which to speak about politics and the human condition. And indeed, as I listened to the disc, Chomsky said nothing about linguistics--his special area of expertise--but rather the Bush administration, American colonialism, the Free Market, and all of those touchpoints which form the central contention in political soundbites. Yet all the soundbites he was presenting were terribly old: the talking points of last year, of two years ago. "Oh, is this an old lecture of his?" I asked. It wasn't.


Indeed, for all the times I have been pointed to one of Chomsky's articles or talks on politics since, I have never seen him express an idea that had not already been hotly debated over a period of months a year ago. I've since come to regard him as a sort of Reader's Digest Condensed Books version of a pundit: bringing to you a collection of ideas that everyone else is already tired of, in a pared-down, simplified version.

Well, he's noted for being a linguist, anyways, I thought to myself, hardly surprised that the political rantings of an old academic might be a bit out-of-touch--having seen a few of my professors get hopelessly off-topic. Not long afterwards, I found myself in a linguistics class--once more face-to-face with Chomsky's influence.

It was a terrible class, with a terrible professor, though she had a great deal of clout and seemed respected in her field. Perhaps she was just one of those who does their best work outside the classroom? Yet, again and again, the class would come to a standstill. She would be speaking at length on some point or another and you would hear everyone in the room getting antsy, murmuring, turning to their neighbors, confused, until finally, someone would raise their hand and ask a pointed question about some odd contradiction in what she had just said. She would always stand blankly for a moment, the gears working--and then, without fail, she would repeat whatever she had previously said and ask 'does that clear things up?'


Inevitably, it did not, so another student would ask the same question in a different way, trying to get to the heart of the matter, only to receive the same response. It soon became clear, inexplicably, that our linguistics teacher was fundamentally incapable of comprehending precise differences in how phrasing affected meaning. It was inconceivable to us that a linguist should be so incompetent at determining the meaning of words, but if we knew more about the state of Chomskyan linguistics, we might not have been so shocked.

I still recall a specific instance where we had just read about the Sapir-Whorf hypothesis, which states that the specific words we use affect the way that our brains process information. Our professor told us that this theory was completely and entirely wrong, and when a student asked why, she irately responded 'It just is!' The rejection of this idea goes back to Chomsky, and our teacher, like many hardline Chomskyites, felt no need to defend his conclusions.

Back in the fifties, Chomsky laid out a number of theories, but the most influential and transformative for the field of linguistics was 'Universal Grammar', the idea that the structure of language is pre-programmed into the brain before birth, and that all languages the world over follow the same rules which evolve from those biological presets. Now, I am not a linguist, so I cannot tell you whether his theory is true or not, but I can tell you that it is not the most simple theory and so, according to Occam's Razor, it cannot be considered the leading theory, independent of how popular it may be.

Occam's Razor states that if we have two competing theories and we cannot disprove either one, then we should consider the more straightforward theory to be the correct one. One illustration of this is conspiracy theories. My uncle is a paranoid schizophrenic, and believes that aliens move things around in his room at night. I have a competing theory: he sometimes misplaces things and doesn't realize it. His theory requires that highly advanced, complex creatures with scanty evidence for their existence have crossed the vast reaches of space in order to personally inconvenience a random unemployed dude. Since his theory is more complex than mine, it is more unlikely to be true, therefore we should assume that mine is correct unless we have some strong evidence.


Grandma knew how to have fun
Likewise, the notion of Universal Grammar is more complex than the idea that we might acquire language through the process of 'what fires together, wires together', the theory that whenever our senses experience data, those data are connected to other things that are happening at the same time, which gradually builds up a mental process. For example, if we see a bird and hear its call and see our mother pointing at it and saying 'bird', then the visual of the bird, the sound it makes, the sound our mother makes, and her gesture all tie together in our minds. Every time this happens again, those ties become stronger. Meanwhile, if we happen to smell a pickle while seeing a bird, that will also tie those experiences together, but if it never gets reenforced again, then that tie will waste away with disuse. This theory explains why, whenever Mary smells Chanel No 5, she thinks of her grandma.

Now, I'm not saying Chomsky is wrong, but just assuming that he is correct is destructive to the field of linguistics. Along with Universal Grammar, Chomsky developed the idea of the phrase structure rules, which lay out a system for how all language works and how all phrases operate in language. However, this structure is all based upon related European languages, and was never thoroughly-tested with fieldwork or observations of how other, non-Western languages work. Yet, for half a century, this has been the method by which linguists study all languages, which is a problem.

To see why, let's look at the progression of phonetics in linguistics through history. In English, we have twenty-six letters that we use to represent the sounds of language. There are, of course, other alphabets--like Cyrillic, Arabic, or Japanese Kanji--which represent sounds in different ways. During the colonial period, Western scholars were studying and translating many foreign languages, and there was much debate over how to represent them.


Many different systems were developed for approximating foreign sounds using the roman alphabet, but these were only approximations, which is how 'Beijing' and 'Peking' came to refer to the same Chinese place name. Eventually, language scholars realized that the human mouth is capable of producing far more distinct sounds than appear in English, or French, or any other single language or language group, and so they developed the phonetic alphabet as a way to try to represent all the sounds that a mouth might make, and as many of the subtle variations between those sounds as was feasible.

But let's imagine for a moment that instead, phonetics had taken a Chomskyan route: a fellow comes forward and declares that, due to the preset structure of the mouth, it can only produce certain sounds, and that these sounds are common to all languages. He then presents a method of representing all languages using the roman alphabet. Students take this method and go out into the field, writing down foreign words phonetically and sending them back for scholars to study. Since all these words are written in roman characters, the scholars then declare that the theory is right: clearly, all foreign languages are made up of the same sounds as English.

Of course, we can quickly see the error here: the roman alphabet is not capable of representing non-Western sounds. It cannot distinguish the tonal differences of Chinese, or the Indian 'd', which is made by touching the tongue to the center of the roof of the mouth, or the clicks of some South African languages. Therefore, any representation of foreign languages in this alphabet will be approximations, unable to demonstrate the unique sounds present in them, or the key differences between these languages.


This is the same problem with Chomsky's phrase structure: it makes many assumptions about how phrases work, how they are related, and about how meaning is produced in language, and these assumptions are based upon the scholarly study of Western languages. Any fieldwork we get back which presents a foreign language in terms of sentence trees will only be able to depict the ways in which that language is similar to Western tongues, and will have no way to represent fundamental differences.

This is especially problematic in a field like linguistics because the most unique and unusual languages are dying out every day--the few languages which have not yet been altered by Western influence are constantly under threat from missionaries, relocation, loss of habitat, and biased researchers. Imagine for a moment that all of our knowledge of Chinese was recorded in the Wade-Giles romanization method and then Chinese died out, so that the only knowledge we had of the language was what we had written about it. Clearly, we would not be able to determine from this record the ways in which pronunciation and tone fundamentally differed, and once a language is gone, there is no going back.

That is why it is vital to ensure that our methods of data collection contain minimal bias. If our methods would tend to confirm one hypothesis, or to miss important data, then they need to be restructured. We cannot look at Chinese through roman letters, nor can we expect all languages to conform to the structure of English phraseology. Indeed, there are field researchers like Dan Everett who are saying that this bias is already firmly entrenched across the whole of modern linguistic study, and that a cult of personality has sprung up around Chomsky which is not merely resistant to change, but is invested in a system which is incapable of measuring the relevant data. Indeed, if you read that article, you will find it contains many direct quotes from Chomsky and his followers about their unwillingness to look at any data that do not confirm their preconceptions--and that the majority of heads of linguistic departments are Chomskyites.


And this is not merely a case of old ideas which were good when they were developed, but which are now starting to show their age,  because if we look back at Chomsky's methodology, we will discover that it has been deeply flawed and unscientific from the start. Peter Norvig, director of research at Google, has put together a devastating analysis of the fundamental flaws in Chomsky's work, showing how from in his earliest writings to his most recent statements, Chomsky demonstrates a remarkable ignorance of the scientific process.

In his post, Norvig demonstrates how Chomsky places theory above data, ignoring the fact that a theory which is unable to account for the data is worthless. Chomsky was never a field researcher, he never had to deal with actual languages as they were used. Instead, he spent his time in academia, developing his theories based on second-hand accounts and assumptions. He made many grand statements and then developed a system which can only confirm them--but that's what 'famous intellectuals' tend to do.


Freud did the same thing: he laid out a system of radical theories which were not based upon a wide collection of data, but upon Freud's own selective experience. He then developed a system based upon his theories which was insular and unable to deal with information that did not match up. Scientifically, both he and Chomsky jumped the gun: they put the theories before the data. In Freud's case, the result was that nearly every idea he had about the brain has turned out to be completely wrong, and now it looks as if Chomsky is headed for the same end.

But the real disappointment about these 'famous thinkers' is the effect that they had on their field. By taking a leap and proposing some strange theories, they got a lot of people interested, they brought vitality to their areas of study--but then, all of that extra energy went into trying to prove ideas that were fundamentally flawed. They jump-started their field, set it ahead ten years, but then ensured that it would stagnate in a morass of weak ideas and bad methodology for the next half century, maintained stalwartly by the next generation of theorists who don't want to think that perhaps they have based their entire careers on error. Indeed, many of Freud's ideas are still hanging around, more than a century later, despite the fact that they've been discredited--though they aren't hanging around in psychotherapy circles, but as a series of popular misconceptions.


And that seems to be the same route Chomsky is taking: shedding any claim on scientific relevance and becoming a popular figure. Indeed, of the up-and-coming young linguists I know, none of them take Chomsky seriously, and of the people I know who take Chomsky seriously, none of them have the least comprehension of linguistics as a science. In the end, it seems most likely that Chomsky's theory will end up becoming a literary lens, just like Freud's, and no professional linguist will give him a second thought.

I only hope that, by the time linguists develop a less biased method of data collection and linguistic analysis, it won't be too late, because while Freud's missteps meant that he mistreated some mentally unwell people, Chomsky's might mean that we lose out on hundreds of languages that are quickly dying, and once they are gone, the field of linguistics will be that much more bare. The more diverse and accurately-recorded data a field has to draw upon, the stronger that field and its theories will be, and yet there are few fields more ephemeral, more easily damaged than linguistics. All you have to do is wait and do nothing--it's shrinking every day.

31 comments:

  1. Replies
    1. I've angered an incoherent idiot--I must be doing something right.

      Delete
  2. Very interesting, informative and thought-provoking. I must say I had a similar "situation" with Chomsky. He was the "famous one". Somehow it has never crossed my mind to question it. I feel ashamed, I must admit, because being at bit interested in linguistics I should have. Thank you for giving me the opportunity to realise it and to change my approach :)
    Great piece of writing!!!

    ReplyDelete
    Replies
    1. Thanks, I'm glad you enjoyed it. It can be odd how sometimes, we just take someone's authority for granted because we know their name and reputation, but of course, the work has to be able to stand up on its own, big name or no.

      Delete
  3. Overall I agree with your critique of Chomsky, particularly the cult of the public intellectual that he uses to lend weight to his otherwise uninspired political ideas. However, I do think that you may be setting up a something of a straw man when addressing Universal Grammar. You seem to be arguing against the idea of Universal Vocabulary which as far as I know is not postulated by any mainstream linguists. Unless I am misunderstanding you, you’re example is of a child learning the word for a bird.

    ReplyDelete
    Replies
    1. Chomsky and his followers do postulate the existence of Universal Grammar, and though the specific instance I used was with the word 'bird', I was trying to suggest the larger theory that there is no necessary substructure for language in the brain, but that language can be learned and developed purely by example and experience--that we don't need a built-in structure for 'nouns' in order to associate the sound our mother makes with the bird we see, any more than we have to possess a grammar in order to associate the smell of a flower with that flower--the fact of their constant and repeated proximity is enough for our brains to make the connection.

      Delete
  4. Nothing to say about the "cult of personality". But "Chomskyan linguistics" includes not one but three different research programs (the phrase structure rules were abandoned in the 70s in favor of the theory of X-bar). The innatist argument is very strong (I would recommend "The Language Instinct", by Steven Pinker, a far more understandable and kind author than Chomsy, and a critic of him). The Sapir-Whorf hypothesis was indeed very biased, as demonstrated by authors like Kay and Malotki, but there are still very interesting defenses of linguistic relativism, like the monumental work of George Lakoff, "Women, Fire and Dangerous Things".

    ReplyDelete
    Replies
    1. Thanks for the response and the suggestions. I supposed I had classed Pinker along with authors of pop pseudo-neuroscience like Lehrer and Gladwell. I know Sapir-Whorf was not developed using the best methods, and had many flaws, but there have been leaps and bounds in the family of non-innate theories since then, such as the statistical and probabilistic learning models used in the study of AI and search algorithms. These seem to produce better results with less labor, as Norvig's article points out, which to my mind makes that theory a strong contender, if only by means of 'elegance' and Occam's razor--those early indicators which often lead us to developing an effective approach.

      Delete
  5. I can not help but find a resonant tone with the emotional content of the above sentiments. I can not, however, concur with your dire conclusions. Not all, or even most, of Freud's ideas were without merit. Indeed, he made excellent observations and was reported as having meaningful, beneficial relationships with many of his patients. He was certainly a notorious arm-chair theorist, but he was paying attention through his cocaine-induced haze. I have no defense of Chomsky, but I believe that there are great thinkers in linguistics minding the store. Further, I find it compelling and comforting that modern technology allows for systematic, extensive documentation of most dying languages. The tone of your essay is well-received, but the message may be more dire than the evidence would warrant.

    http://faculty.fortlewis.edu/burke_b/personality/readings/freuddefense.pdf

    ReplyDelete
    Replies
    1. Curious--of all the things I thought this post might dredge up, an attempt to defend Freud as relevant was not something I expected. Even in the few places I encountered his work in academic study--mostly literary theory--his approach did not seem to be taken seriously as a coherent method, but was used more like a random pattern generator, something like tarot cards, which force the analyst to look at a work in a fresh way.

      To me, Freud is just a continuation of an older theory of psychological understanding--the Greek and Roman model--and I think Freud's appropriation of this older model was usually flawed and short-sighted. I think one of the clearest example is the Oedipal theory.

      Freud encountered symptoms of this condition often in his work, but he failed to detect the fact that it was only prominent because of the particularly unusual cultural setup in which he lived. For a middle class European of his time, it was usual to spend the early years away from one's own parents, being raised by nurses and servants, meaning that during the vital imprint period before about age six, these Europeans would not process their parents as close blood relations.

      This is why, when they interacted more with their parents later in life, there would be conflicts and attractions there that would not exist under normal child-rearing conditions for human beings. Yet, if you look at the original myth, you'll see that Oedipus himself spent his imprint years away from his parents--meaning the Greeks had a better grasp on the condition than Freud did, and that Freud missed a significant portion of the original theory.

      I tend to see works like Ovid's Metamorphoses as an early form of the DSM: a list of conditions, symptoms, and causes to aid people in thinking about how the human mind works. Indeed, looking through the history of philosophy and literature, this is precisely how great thinkers used these myths, referencing the Narcissus type, the opposition of the Apollonian vs. Dionysian models, and other such comparisons. There's a rather interesting analysis of the original Narcissus myth and its parallels with modern theories of the mind here.

      To me, Freud is just a small step in that much larger and older tradition, and I did not come to the conclusion that he really understood the material he was trying to interpret.

      I do hope you're right about the state of the linguistic community, but the articles I linked from Norvig and Everett paint a much more troubling picture. Sure, it's good that we can now better capture language with technology, but that doesn't help the languages that have already been lost or adulterated over the past half century.

      Thanks for the article, I'll have to look it over and give it my careful consideration.

      Delete
  6. From where did you draw the motivation to write this article? Also, are you a linguist? More is known about quantum physics and subatomic particles than the origin and computation of a metaphor. You should really tread carefully when you are critiquing this kind of subject matter. Professor Chomsky is extremely humble, and has never once asked for a following. His work in linguistics has done remarkably more work for mankind than any blog you will ever write. You should really check your framework of study, and ask yourself from where you draw your motivation to criticize this man.

    ReplyDelete
    Replies
    1. "From where did you draw the motivation to write this article?"

      Well, I'd heard of Chomsky be reputation, and knew a number of people who liked to name-drop him, so I decided to do a bit of research on my own and see if I could learn something from the work he's done.

      Instead, I found the work and methodology to be shoddy, and the discourse surrounding his work likewise flawed. So, I wrote down my findings as an article. If there are any particular points you'd like to dispute, you're welcome to do so, but I'm afraid merely stating that something is complex and that little is known about it does not constitute a refutation.

      Indeed, the lack of knowledge we have about linguistic concepts and how they come about is central to the argument I put forth in my article: if this is something we know less about than quantum physics, then isn't it presumptuous for Chomsky to put forth a total theory that denies other options, despite the fact that we cannot yet write them off based upon the data?

      "His work in linguistics has done remarkably more work for mankind than any blog you will ever write."

      This blog will do more than any slipshod comment you happen to put together. I hope what you say about Chomsky is true, despite your inability to support it--because otherwise, his work is leading us down the wrong road and setting the field of linguistics back.

      Delete
  7. You reference Norvig's "devastating" analysis but do not include Chomsky's response to Norvig. This seems an oversight. Furthermore, a statement like "Indeed, of the up-and-coming young linguists I know" calls for a Wikipedia-like [Who?] and [citation needed].

    ReplyDelete
    Replies
    1. "You reference Norvig's "devastating" analysis but do not include Chomsky's response to Norvig."

      I'm not familiar with Chomsky's response, and googling 'Chomsky's response to Norvig' just produces a lot of third-party articles talking about Norvig's critiques.

      "a statement like "Indeed, of the up-and-coming young linguists I know" calls for a Wikipedia-like [Who?] and [citation needed]."

      So you'd like to know the names of my friends who are studying linguistics? I'm not sure how that would help.

      Delete
  8. How is postulating some presumably complicated structure/mechanism in the brain by which humans acquire language by detecting patterns in raw data ("what fires together...") any simpler than postulating the structure of "universal grammar"? Your "Occam's razor" is pure rhetoric.

    ReplyDelete
    Replies
    1. Well, the fact that it's not a 'presumably complicated structure', I suppose. It's merely an extension of the observations that when neurons fire at the same time, the connections between those neurons are automatically strengthened. It's not a purely linguistic premise, it's a theory about how all structures in the brain develop over time.

      The reason it is simpler is because it does not presuppose that the underlying structure of language comes pre-loaded into the brain, as Chomsky posits. Let me make an analogy using another piece of data storage hardware: the DVD. If we possess a film DVD, that means that its physical structure is such that it contains the data for the film in a usable form. So, how do these data come to be on the DVD? The Chomskyan theory would say that when the DVD is made in the factory (as the brain is made in the womb), the data for the film are hard-coded into the disc. The 'fires together' theory would say that the disc is made blank, and that the information is added to it later.

      Now, because DVDs can contain many kinds of information, not only films (and films themselves are all different), it would make more sense to start with a blank disc and then let users customize it later. Likewise, the brains of human beings can contain many different types of data--they are very adaptable to various situations. Hence, it would be simpler for the brain to include fewer hard-wired processes at creation, since the outcome is variable.

      An example of such a variable outcome would be a blind person's brain showing activity in the visual-spatial region in conjunction with proprioception, which does not occur in sighted people. That area of the brain has been written with a different dataset because the neurons were affected by different stimuli, just as one can write different data to different parts of a DVD by modulating the laser stimuli.

      CONT.

      Delete
    2. Of course, there are simpler brain structures which are hard-wired--the 'sphexish' behaviors of insects and lesser life-forms--but language is not a simple, basic structure, it is a recent development, not centered in the 'lizard brain', but the prefrontal cortex, an area which is widely plastic and adaptable.

      The 'fires together' theory is also paralleled by recent experiments in 'self programming' computers, where a system is set up with only the most basic fractal, recursive process, and then is set to run with various limitations and parameters placed on it (a 'development environment' which mirrors the way brains learn). Through random trial and error, keeping only those parts which work within the given parameters, the system eventually develops a program capable of performing within those parameters.

      What is most fascinating about this process is that it is not necessary for the computer to 'know' how to code, or how its own systems work--it does not require a whole in-built, pre-existing structure to guide it from the first to last step. Originally, with human programmers, each programmer could only program what they absolutely understood would work--they were not able to write programs outside of their theoretical structure. However, by letting the operating environment solely dictate the outcome, it is possible for the machine to organically develop a working process that a human programmer cannot understand, and could not have developed himself, based on a current knowledge of theory.

      The surprising aspects of how these self-learning computers develop often stem from aspects of electromagnetic theory which human knowledge has not yet been able to make sense of. In studying high-level electrical engineering, there are many known processes which have been discovered by trial and error, but which science has no practicable theory to explain, and so they are learned by rote.

      This is because an evolutionary process can develop a working solution without requiring that every detail of the solution be perfectly understood, as touched on in the Norvig article linked above in the section about working from datasets vs. working from theory. For example, it's entirely possible to calculate the speed of acceleration of gravity and use this information accurately in predictions without knowing what gravity is, how it propagates, or what causes it.

      These are some of the reasons that 'what fires together' is a more elegant, less complex way to account for the observed data.

      Delete
  9. You have to treat Chomsky as a linguist and Chomsky as a "public intellectual" as two distinct entities for the purpose of a critique. So let me address them separately:

    1. Linguistics: I do think that the (fairly recent) success of statistical machine learning models has brought up new data which in addition to work by Hudson-Kam and Newport does seem to point towards the fact the Universal grammar unnecessary. However, historically one of the reasons that motivated the idea of UG was the problem of animal language acquisition. There were several cases of chimpanzees and oragutans taught sign language and amassing vocabularies of hundreds and sometimes even over a thousand signs. However, they never seemed to form language like structures in their signing. Whereas, human children very quickly mentally conceptualize the difference between verbs and nouns during language acquisition. This was one of the problems that UG was hypothesized to explain. There is still work being done on language in animals which does pose the question, why do humans find it so easy to learn language? Also, your claim that UG was never tested on non-European or non-PIE languages isn't true. Grammars of languages from all language families have been formulated along the lines of a generative grammar.

    "I only hope that, by the time linguists develop a less biased method of data collection and linguistic analysis, it won't be too late ... Chomsky's [missteps] might mean that we lose out on hundreds of languages that are quickly dying out, and once they are gone, the field of linguistics will be that much more bare." Umm ... this is kinda, a non sequitur, considering both theorizing about language acquisition, field linguistics and documenting less well-known languages can both be done, and are done simultaneously. Chomsky has never spoken against field work. This would be like saying that a physicist who worked on particle physics has held back work on astrophysics.

    2. Statements on politics and US foreign policy: Here, I am more or less in complete agreement with Chomsky, and he has done significant work to expose the motivation/objectives behind US foreign policy. I'm not sure if some political theorist or economist has made the points he is making before, but at least Chomsky is the only one who seems to address a lay audience with these findings and concerns. He can be somewhat dense and repetitive, but do you actually disagreed with him on any of the points he raises?

    ReplyDelete
    Replies
    1. "There is still work being done on language in animals which does pose the question, why do humans find it so easy to learn language?"

      Yes, and I'm not arguing against the presence of a language center in the human brain which is structured to take in data differently from the animal brain, I'm rather saying that the ability to learn language, however remarkable, does not imply a universal phrase structure.

      "Also, your claim that UG was never tested on non-European or non-PIE languages isn't true. Grammars of languages from all language families have been formulated along the lines of a generative grammar."

      Looking at the structure of your statement, it seems to support my assertion. You say that grammars of languages from all families have been formulated along the lines of a generative grammar, which I think is precisely the problem: generative grammar was not formed based upon those languages, rather they were made to fit the generative theory.

      This is also the problem with his influence on field work: his language structure rules and sentence tree formulas are used in field work to record all languages, they are used in the analysis of those languages, in the articles and books written about them, and so the entire dialogue of linguistic research and interaction is built upon these assumptions. It's more like saying that the work of an astrophysicist on Earth building a machine for analyzing minerals would affect the the results of a mineralogical study by an astronaut using said machine on the moon.

      " I'm not sure if some political theorist or economist has made the points he is making before, but at least Chomsky is the only one who seems to address a lay audience with these findings and concerns."

      On the contrary, every political work of his that I've heard or read just sounds like a rehash of Daily Show episodes and Onion articles from three before, but without the humor.

      "He can be somewhat dense and repetitive, but do you actually disagreed with him on any of the points he raises?"

      Hmm, I've never thought of him as dense, if anything he seems to gloss over points too facilely. I don't necessarily disagree with his conclusions, no, but just because I agree generally with a guy who walks up to me on the street in 2014 and says "You know, I'm starting to think that Bush fellow wasn't the greatest president", that doesn't mean I find the statement to be timely, insightful, or a productive addition to the discussion.

      Delete
  10. "I do not know of any case of Chomsky's work actually impeding work by field linguistics."

    Come now, I link to and discuss one such case at length in this post.

    "This is a problem with observational disciplines, where it's much harder to make predictions or conduct experiments, but just come up with simple frameworks which explain a lot of data."

    It is a problem, and one made worse by simply accepting a methodology which was neither developed nor tested in the field. If we know going in that this is a field where predictions and experiments are difficult, then it behooves us to take caution and care in how we structure our work. If predictions are difficult, then basing our analysis on a series of narrow and specific predictions is not likely to turn out very well for the field.

    "I would strongly urge you to give it a try before you make up your mind on Chomsky."

    Thanks for the suggestion, I'll keep it in mind.

    "Don't let a terrible professor turn you away from a thinker who might have more insight than you give him credit for."

    Actually, I didn't know the professor was towing the Chomsky party line until years after graduation, when I read the articles linked in my post and realized that I'd actually experienced the effect firsthand.

    ReplyDelete
  11. Come now, I link to and discuss one such case at length in this post.

    I think you are being disingenuous with your discussion there though. Firstly, your analogy: "But let's imagine for a moment that instead, phonetics had taken a Chomskyan route: a fellow comes forward and declares that, due to the preset structure of the mouth, it can only produce certain sounds, and that these sounds are common to all languages. He then presents a method of representing all languages using the roman alphabet. Students take this method and go out into the field, writing down foreign words phonetically and sending them back for scholars to study," is a bit tenuous.

    Since you have linked to Everett talking about Pirahã, it needs to be said that his data on them is somewhat controversial, being reliant on data collected by just one individual. Hopeully, we will gain a lot more information on Pirahã in the near future to settle this question.

    Secondly, I think his claims on Chomskyian ideas afecting data collection in field linguistics are a bit overstated. Mainly because for more than a couple of decades, most field linguistics analyses include emic and etic information on grammar except for those languages. These emic descriptions do away with any external biases because they include whatever grammatical/phonemic categories and rules the speakers themselves utilize to make sense of their language. To say that field linguistics has been led astray by Chomskyian ideas is to discredit all the work on field work and frameworks that the thousands of anthropologists and field linguists have done.

    Even if Pirahã is and exception, the fact that thousands of languages from all the major language families _do_ show recursion makes it a language universal, and surely needs to be explained. I think the cult of personality around Chomsky (or any cult of personality, for that matter) is pretty useless and counterproductive in academia, but its existence doesn't necessarily discredit all of his work either.

    ReplyDelete
  12. "I think you are being disingenuous with your discussion there though. Firstly, your analogy ... is a bit tenuous."

    You think I'm being disingenuous? That I know better but am being deliberately misleading to some end? That is a rather strong and unflattering accusation, and not one you support by your arguments. Also, the fact that you find the analogy 'a bit tenuous' is irrelevant. Unless you can actually delve into it and point out where and how it fails, your passing claim that it is tenuous means nothing.

    "Since you have linked to Everett talking about Pirahã, it needs to be said that his data on them is somewhat controversial"

    Certainly true, which is why it is only one of the sources I call on for my larger argument. However, saying that a particular case does not convince you is very different from saying that you don't know of any such, especially after having just read my post. Claiming ignorance of one of my basic points did not make me feel as if you were actually following the argument.

    "These emic descriptions do away with any external biases"

    Really? They do away with any external biases?

    " To say that field linguistics has been led astray by Chomskyian ideas is to discredit all the work on field work and frameworks that the thousands of anthropologists and field linguists have done."

    No, it doesn't discredit all of those ideas. It is entirely possible for researchers and theorists to do great work even while working with inferior tools and false hypotheses. Eratosthenes was able to perform a great many highly accurate and useful astronomical measurements despite lacking both the use of a telescope and the heliocentric theory.

    "Even if Pirahã is and exception, the fact that thousands of languages from all the major language families _do_ show recursion makes it a language universal"

    Actually no, it does not make it universal if there is an exception to it. Also, if--as I posit--the Chomskyan method of data collection and analysis contains an inherent confirmation bias in favor of recursion, that would bring into question some percentage of those thousands of results.

    ReplyDelete
  13. Actually no, it does not make it universal if there is an exception to it. Also, if--as I posit--the Chomskyan method of data collection and analysis contains an inherent confirmation bias in favor of recursion, that would bring into question some percentage of those thousands of results.
    In a field such as linguistics if you find one exception and several thousand languages which exhibit those universals, I would still think that a framework which posits those properties as language universals has some merit and contributes to our knowledge of the nature of language. Of course, if exceptions arise, they need to be studied as well, but that doesn't invalidate the utility of the earlier theory. Newton's laws of gravitation, and Galilean kinematics are still very useful theories even though they are only approximate limiting cases of General relativity and the special theory of relativity.

    No, it doesn't discredit all of those ideas. It is entirely possible for researchers and theorists to do great work even while working with inferior tools and false hypotheses.
    Sure, but what you are doing is equivalent to blaming Newton for formulating a theory of gravitation based on some of the astronomical observations he had at that time for not predicting the correct rate of precession of the perihelion of Mercury. The experimental observations for estimating this date back to the end of the 17th century, and the solid belief in Newtonian gravity didn't prevent the collection of raw data which later showed a deviation from Newtonian predictions.


    "These emic descriptions do away with any external biases"

    Really? They do away with any external biases?

    That statement is something of a tautology, since emic descriptions are just collection of how the community itself views it's own grammar and rules. So, yes it does away with external biases, though of course it introduces internal biases of those communities themselves or of how they wish to be seen.


    You think I'm being disingenuous? That I know better but am being deliberately misleading to some end? That is a rather strong and unflattering accusation, and not one you support by your arguments. Also, the fact that you find the analogy 'a bit tenuous' is irrelevant. Unless you can actually delve into it and point out where and how it fails, your passing claim that it is tenuous means nothing.

    Ok, the reason I accuse you of being disingenuous and your analogy being problematic is this: Your whole analogy focusses on romanization schemes and inaccurate representation of phonemes not present in English. This has not a problem since IPA has been in use since before Chomsky was born. There's also the difference between using the IPA alphabet and other romanization schemes which are purpose built for a particular language (the Wade-Giles system for instance is completely suitable, though slightly cumbersome to use. A person familiar with the rules of romanization would read pinyin Kǒng Fūzǐ exactly the same as Wade-Giles: K'ung3 Fu1-tzu3).
    So let's assume you are also aware of that and are using the romanization analogy to stand in for other information which is lost due to pre-existing theoretical frameworks.

    continued ... (exceeded character limit) ...

    ReplyDelete
  14. ... continued ...

    Now here is where my argument runs aground :) I began composing this comment as a sort of stream-of-consciousness reply, and realize that I don't have much to back up what I was about to say (write) here. So, let me add a caveat that some or all of what follows here may be wrong (I am not a professional linguist, though I am a scientist with a fair degree of interest in languages and linguistics). I think that such biases from theory have not influenced field work that "strongly". The minimal introduction I have had to field linguistics and language documentation programs seems to indicate their methods and toolsets are more closely influenced by anthropological and sociological work as opposed to neurolinguistics and language acquisition.
    * * *

    Deciding to dig a little more on this idea I came upon an interesting page in a book by Everrett. I feel that Everett's accusation (or thesis to uses a less adversarial word) is not that Chomsky's influence actually affected the results of field work but rather re-oriented the focus of research departments more towards analytical study of language structure this reducing funding for field work. This is more like the analogy of a sudden flurry of research work on particle physics just making physics departments more likely to focus on that and reduce the effort directed towards astrophysics. In which case it's difficult to exactly lay blame anywhere as academia is as prone to fads and fashions as any other human endeavour.

    But, yeah, thanks for making me think a bit harder about some of these questions!

    ReplyDelete
  15. Chomsky's most noted influence is in the field of computer science. I suggest you look into that.

    ReplyDelete
    Replies
    1. What, meaning the Chomsky–Schützenberger hierarchy? My understanding is that the movements in modern computer language have moved rather far beyond those original descriptions, and that it isn't really considered pertinent anymore.

      Delete
  16. You quite heavily misportray and simplify the Sapir-Worf hypothesis. I am also unaware if any form of it is taken seriously by linguists of today. Certainly that language shapes thought in SOME cases can be considered, I guess.(And you seem to care what linguists these days have to say, insomuch as it pertains to your shared dislike of Chomsky, at least)

    To not be completely hand-waving at this point, I'll add that it is not the words that shape our thinking (this would imply that the connotational meaning of words or some such mechanism affects our thinking, is the claim of this theory, whereas it is not), but the structure of language. Of the values between words. (See Saussure for the meaning of "value" here) In general they speak of linguistic categories and the structure of language. A typical Sapir-Worf claim would be that a language that has no past tense results in different modes of thought about the past in the culture that uses that language. Or that Inuits, due to having more words describing different shades of snow, perceive the types of snow better, cognitively, thanks to language. And a whole host of other highly disputed claims. Mainly disputed, as far as I remember from my studies, in that there arises a question if words shape our categories, or mental categories give rise to how we use words.

    ReplyDelete
    Replies
    1. Romendacil said: "You quite heavily misportray and simplify the Sapir-Worf hypothesis."

      My only mention of the Sapir-Worf hypothesis is that one of my professors was unable to explain it to the class--I don't see how that equals a 'heavy misportrayal'.

      Unless of course you are conflating the 'fires together' theory with the Sapir-Worf Hypothesis, as some do--rather the same way recent discoveries like epigenetics and horizontal gene transfer have been compared to earlier discredited theories of Lamarckian evolution. These recent linguistic ideas, coming from discoveries in the fields of AI and neurological medicine, do have some commonalities with Sapir-Worf, but are part of a new theory based on modern experiments and data--they are no more based on that theory than epigenetics are based on Lamarck.

      These fields, based on studying the actual structures of brains, and of thought, are coming at the question from a different direction, and it's hardly surprising that there should be a fair amount of friction between them and academic linguistic theory as new data comes in and must be added to the existing structure.

      "Inuits, due to having more words describing different shades of snow, perceive the types of snow better, cognitively, thanks to language"

      This example, while commonly brought up, is based upon a falsehood. The structure of Inuit language is such that is produces compound words, like in German, so that to say 'fluffy snow' or 'icy snow' or 'blowing snow' would all use single words, not because they have wholly unique words for each, but because they affix their adjectives--thus, it would be equally correct to say they have more words than we do to describe the sky, or faces, or a shaft of wood.

      However, there is another example of color difference in language seeming to shape differences in thought, namely the color blue. In William Gladstone's studies of the history of languages, this color is absent from the oldest texts--even in descriptions of the sky. In every language he studied, the development of words for color seemed to be the same: first black and white, then red, then green and yellow, and always blue last.

      If these 'basic colors' were encoded into our minds, then we would expect to see them emerge together, that the idea of naming things by color would be swiftly followed by a delineation of those basic colors--but that's not what we see.

      Of course, there are minor distinctions in what words are used for what range of color in every language--some have more colors than English, others fewer, and research like that of Jules Davidoff with the Himba tribe seems to confirm that this changes the way they look at the word. They do not have a word for blue, and struggle to recognize 'which square is different' when distinguishing between green and blue, underperforming compared to people from cultures that do distinguish blue.

      Now, we can point out that, since they don't have a word for blue, that means parents aren't pointing out the difference to their children--nor is that difference being reinforced in song or story, so they are never taught the skill of distinguishing that color. This is the sort of pervasive effect a difference in language can have on thought patterns.

      Certainly, these observations are controversial (just as Chomsky's were at the time of their introduction), but it also makes sense that they would be controversial, as they present data which is difficult to account for, under the current system and theory. As Norvig and Everett argue in the articles linked above, new discoveries are struggling to make headway in an entrenched system. The fact that they are disputed does not mean that they are any more likely to be right, or wrong, in the end, but these ideas have yet to be dealt with satisfactorily.

      Delete
    2. "a question if words shape our categories, or mental categories give rise to how we use words"

      Well, we know that language shapes thought in some ways. Political consultants like Frank Luntz have based entire careers on the fact that people will vote for or against something based purely on what you call it. Likewise, when you prime subjects with certain words, it will change their responses, their moods, and their opinions.

      Of course, we can always ask 'did the brain structure create the language, or the language create the brain structure'--and like any other nature vs. nurture question, the answer lies somewhere in between, and separating the two is not really possible.

      As I say above, occam's razor would suggest that a whole pre-existing, built-in system is more complex, and hence less likely--especially as humans are less sphexish than any other animal, and must learn even basic skills like basic motor control and recognizing individuals. It's entirely possible that our brains are 'learning machines', like the most successful modern AIs, which develop their own rules based on data rather than subjecting data to a pre-existing structure of complex rules.

      Delete
  17. On the subject of his cultism...

    Conspiracy theorists love him, the paranoid far left idolise him, anti-semites use his writings to justify their bigotry, every tyrant or dictator can count on him for support. As a linguist he knows how we process words - I think that is his strength - He rarely caveats his pronouncements or hedges statements with 'I can't be 100% sure but..." or "to the best of my knowledge...". He speaks like he has access to a vast repository of knowledge the rest of us do not. The truth is he's holed himself up in academia all his life. Speak to people on the front line of the real world and they rarely agree with him.

    For example, Chomsky holds the USA to an entirely different set of standards than the countries of 'brown people'. He marks down most of their aggression and violence as acts of defence. I'd love to know what methodology he uses to reach his conclusion - though I suspect the conclusions are what comes first. I also feel he projects his inherent distrust of authority on to world events - he's paranoid and doesn't appear to trust any authority - in the main the rest of us are not like this - that is something he won't entertain or accept.

    This is the thing, there are things he literally cannot *possibly* know that he says he does and pronounces them with a god-like confidence that is attractive to people who are intellectually lazy, who want a shortcut to 'the truth' and don't mind being led to it... Enter Mr. Chomsky. The more I've written the more I'm convinced that his popularity has all the hallmarks of a cult.

    Enjoyed your thoughts on this pompous little man.

    k

    ReplyDelete
    Replies
    1. Kevstar said: "He rarely caveats his pronouncements or hedges statements with 'I can't be 100% sure but..." or "to the best of my knowledge...". He speaks like he has access to a vast repository of knowledge the rest of us do not."

      Well, I certainly don't think it's necessary to start every statement with these sorts of caveats. As humans, of course we are only presenting our own thoughts, and can be wrong. I think it's fine to leave out 'I could be wrongs' in favor of presenting your arguments simply and clearly, and that it doesn't necessarily imply that the speaker has an undue notion of their own superiority of infallibility. Indeed, such weaselly phrases are often used to obfuscate the matter and avoid straightforward discourse.

      "he's paranoid and doesn't appear to trust any authority"

      Well, I can respect an iconoclast, because authorities are not worthy of our trust. There is no authority, no matter how beneficent the idea that began it, that has not become a tool for those on top to consolidate power and wealth for themselves. Every church, democracy, business, and charity becomes a source of entrenched corruption, given enough time--and the common man is a fool who imagines they have his best interests at heart.

      Indeed, my chief problem with Chomsky is that he has done the same thing, that despite all his railing against authority, he has set himself up as an unimpeachable authority at the top of a whole system which is resistant to change and which holds down those who do not support that system. He has himself used this method to consolidate power and discourage anything that threatens his position.

      Delete