Blame Chomsky!

I was just wondering what to post when this link showed up in my inbox (thanks, Maureen!): Harry Ritchie of The Guardian saying “It’s time to challenge the notion that there is only one way to speak English”:

Did you see that great documentary on linguistics the other night? What about that terrific series on Radio 4 about the Indo-European language family tree? Or that news report on language extinction? It is strange that none of those programmes happened, or has ever happened: it’s not as if language is an arcane subject. …

There is, of course, Steven Pinker’s The Language Instinct – a bestseller that seems to have ticked the box for publishers and public alike as the book on linguistics. But The Language Instinct has a very specific agenda – to support Noam Chomsky’s theories about our language skills being innate; other areas of linguistics are glimpsed, if at all, fuzzily in the background.

I’m not blaming Pinker. He ultimately failed to justify his title, but he did reach a keen, large audience with a well-written book fizzing with ideas and examples. I’m blaming someone else, the person who, inexplicably, doesn’t exist – who should have written the book revealing how Pinker was so wrong and had a ding-dong with him on Newsnight; the ambitious, good-looking academic, who possibly had a spell in an indie band, with his or her own 13-part series about language on BBC2.

Of course, he’s plugging his own book, but it’s music to my ears, and when he goes on to blame my own favorite whipping boy, Chomsky… well, my heart swelled two sizes. There’s nothing new in the piece (and he gets a little carried away at times, talking about “the Proto-Indo-Europeans, that mysterious tribe whose homeland was recently located north of the Caspian Sea in about 3,300 BC”), but it’s an enjoyable read, and it’s my end-of-year present to the Varied Reader. A happy 2014 to you all!

Comments

  1. “It’s time to challenge the notion that there is only one way to speak English”: what a wag.

  2. Yeah, I know, but it’s a newspaper article; I’m just grateful it’s not another “words I hate” or “texting is killing English” thumbsucker.

  3. Steve Reilly says

    OK, I’m not really an expert on Chomsky, but isn’t this bit totally wrong? “This means, [Chomsky] concluded, that human languages have to be organised according to universal constraints and rules, “principles and parameters”. These constitute a “deep structure”, converted into the individual operations of a particular language by a series of “transformations”.”

    As I understand it, sentences have a deep structure that undergo transformations to get the surface structure. What Ritchie’s referring to by deep structure is actually “generative grammar.” Or am I off base?

  4. No, no, you’re quite right, and I wouldn’t suggest anyone run out and buy his book, which is sure to be full of similar errors. It’s just a journalistic rant, but its heart is in the right place!

  5. Well, give blame where blame is due. Chomsky shares the blame for many things mentioned in the article, but what does generative grammar have to do with annoying prescriptivism? What does “He turned grammar into an technical subject full of jargon and algebra studied on whiteboards by men with beards, leaving everyone prey to the pernicious drivel of the traditional grammar guardians, who belong to the 15% [richest Britons].” mean, anyway? Incidentally, some notable anti-Chomskyans are men with beards.

  6. Trond Engen says

    His point is that decades of vain search for Universal Grammar and Linguistic Wars took focus away from real work, both in documentation and communication. I think not. If academic linguists so badly wanted to be distracted, they would always have found something.

    But that linguist with a spell in an Indie band, isn’t that Geoff Pullum?

  7. I think not. If academic linguists so badly wanted to be distracted, they would always have found something.

    I disagree. Academic linguists were perfectly happy doing real work until Chomsky distracted them with delusions of omniscience. That’s like saying… no, belay that, I don’t want to Godwin my own thread.

    But that linguist with a spell in an Indie band, isn’t that Geoff Pullum?

    Yup, the Ram Jam Band!

  8. Trond Engen says

    I disagree.

    Yeah, so do I, actually. That was my faux-cynic side speaking. But wouldn’t quite a lot of those who went out hunting for unicorns have done so anyway? Different unicorns, and less organized in their efforts, but still ultimately unproductive. The huge gains for the profession would have been that with no internal organization they could have been ignored.

  9. And here I was under the impression, obviously false, that there was plenty of writing about language and linguistics out there, like, for example, Language Log, Omniglot, and, ummm…. pretty much the entire right-hand sidebar of Language Hat. Oh, and Language Hat. And I’ve always found plenty in book form, too. Obviously that was all an illusion, and I should rush out and buy that fulla’s* book.

    *Well, he did seem to be arguing that I should be allowed to speak non-standard English. But what, precisely, is standard English? I’m guessing by “standard English” he means RP, in which case I have to ask, since bloody when and who made that declaration? I spend an awful lot of time trying to convince my students that there is no standard English and they should learn whatever variety they like and should not be ashamed of their Chinese accents, they should just make sure they can communicate clearly. After all, English is no longer the sole property of its mother tongue speakers.

  10. And it was my impression that an awful lot of science is unicorn hunting and wild goose chasing and that an awful lot of discoveries happen accidentally in the process.

  11. And here I was under the impression, obviously false, that there was plenty of writing about language and linguistics out there

    Heh. Yeah, that’s why I snipped the bit about “Just as puzzling is the conspicuous lack of a properly informed book about language”; it’s a) silly and b) self-serving.

  12. Geoff Pullum aka Jeff Wright, organ (now a linguist)

    – from Wikipædia.

  13. Trond Engen says

    And it was my impression that an awful lot of science is unicorn hunting and wild goose chasing and that an awful lot of discoveries happen accidentally in the process.

    Fair point. But it seems to me to be an argument against the closed ranks during the Wars. If people chase unicorns individually, a chance discovery is much more likely than if everyone’s following the same map.

  14. Trond Engen says

    I almost forgot: Swedish TV has made a couple of popular documentaries about language. As a result, Sweden also has the TV personality that Ritchie is looking for.

  15. Chris Waugh: Standard English is a dialect, not an accent: it can be spoken in any accent, native or foreign. Your comment is written in Standard English (except for the spelling “fulla”), and so is mine. I’m even a native speaker of Standard English, but my accent is very very remote from RP.

    See Peter Trudgill’s excellent article, “Standard English: what it isn’t”.

  16. men with beards

    I wonder where the author got that idea. I don’t think there are more beards among male linguists than among other specialties. And linguistics is one area where women have done quite well, whether they followed Chomsky or not.

  17. John Cowan, let me put it another way: English does not have any official body that defines an official standard, unlike French or Mandarin as two examples. And it’s not difficult to find variations in what is considered acceptable grammar across the various English-speaking parts of the world, including in writing. Hence my objection to “standard English”.

  18. John Cowan says

    True that we don’t have an official standard dialect, but it is standardized de facto nonetheless. The worldwide variations in syntax are really really few, things like notional subject-verb agreement and the mandative subjunctive. The biggest difference worldwide is in vocabulary and of course orthography, though it almost never interferes with understanding. Standard English is quite a reasonable name for the dialect we are using.

  19. Just to add another point on the sloppiness of the actual linguistics: Ritchie gives the PIE word for ‘father’ as _pihter_. Aside from the lack of an asterisk, it has the look of a half-remembered form jotted down without actually checking on it (maybe getting the Sanskrit and actually reconstructed PIE forms blended?).

  20. Athel Cornish-Bowden says

    The thing that struck me is the statement that the person who was going to write the popular book showing that Chomsky and Pinker were wrong doesn’t exist. I thought he was called Geoffrey Sampson. OK, his book (Educating Eve) is nowhere near as a readable and convincing as Pinker’s, but it’s a bit hard to say that he doesn’t exist. I see that on his home page he tells us that his Erdös number is 5. There must be few academics in the world with Erdös numbers higher than that. Good grief — _my_ Erdös number is 3, but it’s not something I (usually) boast about. By the time you get above 2 it’s not much more than a game of chance.

  21. I’m 100% with Hat and most of you on this. It’s often said that Chomsky helped create so many academic posts, and he (whatever ‘Chomsky’ means) did; one might just regret they weren’t different posts, but, then, we needed a figure or a headline-grabbing type of linguistics to inspire that. I don’t regret it – the academic linguistics world is relatively small and unrepresentative, and minuscule when you get out in the everyday world of people and their fascination with languages. One of my best moves was to take early retirement from the academic world a few years early; utterly wonderful. I did try to ‘get on terms’ with ‘theoretical linguistics’, but it wasn’t me (my type of linguistics comes preceded by different adjectives). I remember being frozen out of chats with theoreticians; they were a clique and happy with themselves (or they hated each other) – good for them; it was great, one day, when I was chatting with one of the greatest recent Slavists, George Shevelov, and one of these theoreticians (not a Chomskyan) came up and introduced himself, expecting recognition, attention, even adulation (?) – Shevelov looked at him, looked back to me, and carried on chatting. Revenge is sweet. I find it refreshing to think that our type of linguistics somehow avoids, most often, awful cliques and personalities; and yet there are names to bear in mind: how about R.M.W. Dixon and his wonderful work (and that of his associates) and the late Larry Trask? I feel I got going in linguistics itself (not languages, I’d started those much earlier, as a teenager’s way of coping with not wishing to speak to anyone at all) with David Crystal’s ‘Linguistics’. But there are so many ‘names’. it was fun when I once went out ‘selling’ Languages at a school in the East End of London in the 1970s; Geoff Pullum was there too, selling Linguistics. I’ll try to ‘Speak My Mind’ here a bit more often – it’s a wonderful web site, thank you.

  22. Trond Engen says

    Guy Deutscher on TV! Plays the flute, not the base guitar, though.

  23. “And it was my impression that an awful lot of science is unicorn hunting and wild goose chasing and that an awful lot of discoveries happen accidentally in the process.” I used to tell my research students “We won’t stumble over anything unless we go walkies”. The “walkies” was intended to imply that we should act with the energy, high spirits, and inquisitiveness of puppies. Earnest, ambitious, ruthless, pot-hunting scientists would be aghast at the very idea. And look at what’s become of much of science at their hands.

  24. Pity it can’t be sung to the tune of Blame Canada.

  25. Athel Cornish-Bowden says

    “and yet there are names to bear in mind: how about R.M.W. Dixon and his wonderful work (and that of his associates) and the late Larry Trask?”

    sci.lang was never the same after Larry Trask died. I knew of his books before I knew of sci.lang, but it was only a few months before he died that I had any sort of interaction with him (at sci.lang). I found him helpful to non-specialists, and willing to talk with anyone without talking down to them. His contempt for Merritt Ruhlen knew no bounds, but he was able to justify it.

  26. John Cowan says

    Athel: 5 is in fact the median finite Erdős number, so I’m sure Sampson posts it out of geekiness, not boasting. The maximum known value is 15, but people with values greater than 8 are rare.

    There are 9779 people with ENs less than yours (as of October 20, 2010). WP says that Field Medalists have a mean EN of 3.21 with a standard deviation of 0.87, so 3 is not bad at all.

    Do you have a finite Erdős-Bacon number by any chance?

  27. Athel Cornish-Bowden says

    “Do you have a finite Erdős-Bacon number by any chance?”

    Almost certainly not! I’ve never acted on the stage (if we don’t count performances at school that are brest forgotten) and certainly never for the cinema.

    I’m impressed that you know how to type a Hungarian double acute! I suppose I could do it if I took the trouble to find the Unicode reference.

    I realize that “3 is not bad at all”, but it means very little. I have collaborated with a mathematician a few times, and in a moment of idleness I looked to see if he had a known Erdős number, and found to my surprise that it was 2, as he had written a paper with Winnie Li.

  28. “And look at what’s become of much of science at their hands.” Could you expand on that a bit, dearieme? When arguing with climate denialists, one of my points is that it’s ridiculous to claim there’s an Evil Worldwide Statist Climate Conspiracy when there’s nothing else like it in modern science. The only other potential examples I’ve come up with are string theory (which shows that it is a capital mistake to theorize before one has data, not a problem for climatology) and Chomskyism (which I think fails the ‘worldwide’ test–like creationism, it seems a particularly American disease, although it’s spread everywhere). What else? I’m not talking about externally imposed antiscience like Lysenkoism, nor, I think, are you.

    I got out of linguistics by 1976 because it seemed more and more to be symbol-pushing with no visible connection to how the mind could actually work. I don’t know if my judgement of semiparochiality was valid then or afterward.

  29. John Cowan says

    My self-created keyboard driver, Moby Latin U.S., interprets AltGr+” followed by o as ő, just one of the 900+ characters I can type with it.

  30. marie-lucie says

    Treesong: I got out of linguistics by 1976 because it seemed more and more to be symbol-pushing with no visible connection to how the mind could actually work.

    “No visible connection to how the mind could actually work” is exactly as I thought about how syntax was developing. I didn’t get out of linguistics, because by that time I had been concentrating on historical aspects, and then I got involved in a native language program (in Canada). Since then I have been combining these two interests, for which I have no use for Chomsky-derived linguistics.

  31. marie-lucie: “No visible connection to how the mind could actually work” … Chomsky-derived linguistics

    As you know, I have no stakes in, or understanding of Chomsky-derived linguistics. I just wonder what kind of “visible connection” you might be thinking of ?

    Is there a visible connection between psychology and the ways human beings work ? Between marijuana and the way pain works ? Between astrophysics and how the universe works ?

    Is it that Chomsky-derived linguists fail to do experiments to confirm predictions, experiments repeated in popular science programs on tv ? Is it that they cannot point to a monitor showing flashing lights in a region of the brain, and say “this is where deep structure X is located” ?

    Offhand, I see no reason why the production of language, which is not a simple matter, should not be describable in terms of complex mathematics. One problem many people may have with this is that it does not assume a building-block etiology of the type “first there were grunts, then there were gerunds”. You can point to grunts and gerunds, but not to a deep structure.

  32. The behavior of many Chomsky-derived linguists towards other kinds of linguist, as criticized in the post and comments here, is a visible connection to the way their minds work.

  33. Is there a visible connection between psychology and the ways human beings work ? Between marijuana and the way pain works ? Between astrophysics and how the universe works ?

    Well, yeah, at least the way ML is using the word “visible,” i.e., figuratively. “Tangible” would work there the same way. Maybe “apparent” would be best. If I were studying something claiming X to be connected to Y, and yet after some time no connection became apparent, I might just stop studying it. I hope it doesn’t need saying that a connection need not be obvious, simple, easily-pointable-to, etc. for me to discern it.

  34. That’s my point, Jim. “Visible”, even in a figurative sense, is perhaps not the best word, because it is too specific. But “discern” is no better, in my opinion, because it is too vague.

    To put it crudely: I think what basically chaps here is that one group of people runs around earning money doing stuff that another group doesn’t understand.

  35. Also: if the manners of Chomsky-derived linguists were not so ass-chapping as described here, perhaps their theories would not excite so much ire.

  36. To put it crudely: I think what basically chaps here is that one group of people runs around earning money doing stuff that another group doesn’t understand.

    Really? I don’t know enough about Chomsky to have an opinion myself, but I doubt Hat and ML, to name two relevant denizens, just don’t understand Chomsky; my bet is they understand it well enough but just consider it a pseudoscience. Now, we could spend all day trying to distinguish pseudosciences from “real” sciences; philosophers of science still haven’t come up with a thoroughgoing definition for the latter. But I know why I don’t go in for astrology, and it’s not ’cause it’s too tough.

  37. I put that badly: not “don’t understand” in the sense of not having enough smarts, but “just don’t see any point in it”. That’s why I asked if the opinion here is that there is not enough prediction and experiment in Chomsky-derived linguistics to qualify it as a science.

    Mathematics by itself is a “pure science” of which not prediction and experiment are required, but proofs of theorems.As I said, I see no reason why the production of language, which is not a simple matter, should not be describable in terms of complex mathematics.

    Imagine a world in which Chomsky-derived linguists did not dominate the field (do they still ?), but simply held “their fair share” of academic positions, along with historical linguists, philosophers, phoneticians, theologians, sociologists of whatever schools of thought. Wouldn’t that be nice ? There would still be bickering, but all bellies would be full and every minion could rejoice in his opinion publicly.

    It seems to me that people are blaming the rats for carrying the plague. But rats are just one part of the picture, along with fleas and lack of public hygiene. What academic, social and political conditions contributed to the rise of Chomsky-derived linguistics ? Has only the field of linguistics been ravaged by higher mathematics ?

    There must be connections, they’re just not visible to me. Surely there is more at work here than hateful exchanges between Democrats and Republicans.

    .

  38. “Could you expand on that a bit, dearieme? When arguing with climate denialists ..”: the adoption, on somewhat Fascist principles, of the term “denialist” is quite good evidence for what I mean.

  39. To put it crudely: I think what basically chaps here is that one group of people runs around earning money doing stuff that another group doesn’t understand.

    I’m glad you retracted the most insulting part, but you still seem to be mired in your belief that you can pronounce from your armchair upon every field under the sun without bothering your head about actually trying to understand it. Why not assume as a working hypothesis that those of us who have spent years studying this stuff actually know something about it?

  40. Hat: I didn’t retract an insulting part, since I did not tend to insult. As I have already written, I had unfortunately used the expression “doesn’t understand” to mean “just don’t see any point it”, without thinking that it could be understood in another way.

    What exactly do you imagine I am “pronouncing on” ? Certainly not on any topic of linguistics. Consider how it looks from my outsider end: time and again at this site people go on and on about the horrid face of Chomsky-derived linguistics. They talk about having given up a career in linguistics because of it. They complain about not being able to find jobs in other areas of linguistics. They accuse Chomsky-derived linguists of intrigue and manipulation.

    How to make sense of this ? How can such a thing happen in a scientific discipline ? Scientists should be able to reach agreement on the basic principles and methodologies of their field, should they not ? Is this a Lysenko scenario ? Is it comparable to the field of economics, which also was ravaged by higher mathematics ?

    Surely you see that these questions are not about linguistic topics, but about linguistics as a social phenomenon like other sciences ? I’m not pronouncing, but rather asking questions.

  41. But we’ve discussed Chomsky, the ravages he caused, and the response of other linguists many, many times on LH. This is not tabula rasa.

  42. No, it’s definitely not tabula rasa. That is exactly what I am talking about The same slate has been scribbled on so many times that it has become interesting in its own right.

    It seems that the rough-hewn sociology-of-science questions that I am asking are just not your thing, at least not when linguistics is the object. When I ask questions about certain phenomena in linguistic science, you think I am pronouncing on linguistic phenomena.

  43. “Is it comparable to the field of economics, which also was ravaged by higher mathematics ? ”

    I don’t know about economics but a lot of the enthusiasm for it in linguisitcs looks like simple notation envy. It was a way for a guy at MIT to look like he was doing serious work.

    And this goes to M-L’s point. How much of this “serious work” has had any real descriptive value or predictive power?

  44. J. W. Brewer says

    Stu, I’m not sure why the default sociology-of-higher-ed assumption should be that any given academic discipline is a “science” and a presumptively well-functioning one. How many decades did it take before the average U.S. university psychology department was completely free of Freudians, and they had to go hide out in Comp Lit, or Philosophy, or someplace similarly non-rigorous (maybe Applied Math, for all I know)? I was myself as an undergraduate taught philosophy of language by a self-identified Freudian (in the Phil dep’t not the Ling dep’t; I don’t think he got tenure and he subsequently wound up at the U. of Chicago).

    One of Paul Postal’s books (he having been one of the first of Chomsky’s inner circle to have a violent break with the Master, way way back when) purports to explain at some length why in his view the discipline is screwed up (or was as of the time he was writing) via the mechanism of many of the people with hiring/grantmaking power not employing criteria reasonably calculated to distinguish potentially productive scholarship from bunkum and bogosity. I can’t speak to the accuracy of his account.

    To go way way back to the block quote in the original post, it is not clear to me that Chomsky’s nativism is responsible for all of the other defects of theory and interpersonal relations often ascribed to him. At a minimum, it seems that one could take a generally nativist view of the human language faculty and not be doomed to follow through in the ways that Chomsky did. Nativism may be wrong as an empirical matter (and/or it may be the case that the current state of the field is such that arguing for or against nativism is not the most productive scholarly activity to engage in until better data is obtained and better analytic tools are devised), but it-leads-to-the-horrors-of-Chomskyism doesn’t seem to me the strongest argument in that direction.

  45. J.W.: I’m not sure why the default sociology-of-higher-ed assumption should be that any given academic discipline is a “science” and a presumptively well-functioning one.

    I for my part don’t assume that. I not sure that I know what a default sociology might be. What you write makes it sound like a Booster Club (Babbitt, Sinclair Lewis).

    The kind of sociology of science I have in mind is that of Bruno Latour (“science studies”), Luhmann, and historians of science such as Hans-Jörg Rheinberger. “Science”.here is a descriptive, not a normative term. Whether a given science at a given time is “well-functioning” is a matter of indifference to these people in their published work. They investigate particular sciences, and how scientists and ideas function there.

  46. marie-lucie says

    I agree with JWB’s last paragraph (I don’t mean that I disagree with the others). The idea of ‘nativism’ is easy to understand, whether or not a person has a background in linguistics, so non-linguists might think that the controversies about Chomsky’s work and influence revolve on that idea, but it has nothing to do with the nitty-gritty work of linguists on existing languages.

    About the early years of Chomsky’s career, starting when he was still a student, there was a very interesting (not recent) article by E.F.K. Koerner which I read several years ago on the internet, but which I was not able to find again (I started on Google but did not have the energy to follow every possible link). Perhaps someone knows or can find the one I am talking about. I think it was actually the text of an invited talk Koerner gave at a unviersity.

  47. Jim: And this goes to M-L’s point. How much of this “serious work” has had any real descriptive value or predictive power?

    Actually, that was my point, or rather question. marie-lucie had endorsed Treesong’s remark: “[linguistics by 1976] … seemed more and more to be symbol-pushing with no visible connection to how the mind could actually work”, but I was not sure what this meant. So I asked: “I just wonder what kind of ‘visible connection’ you might be thinking of ? … Is it that Chomsky-derived linguists fail to do experiments to confirm predictions, experiments repeated in popular science programs on tv ? “

  48. J. W. Brewer says

    A ways upthread, Athel C-B nominated Geoffrey Sampson as the Anti-Pinker, if not the Anti-Chomsky. I don’t really know Sampson’s work, but it appears that he has an ambitious new co-authored book out just a few months ago titled Grammar Without Grammaticality, the intro of which can be read here: http://www.grsampson.net/BGwgCh1.htm. It’s a fairly bracing claim that a quite wide variety of people have been Doing It Wrong, perhaps ever since Saussure, and that e.g. all those example sentences you’re in the habit of sticking asterisks next to on account of you don’t think them syntactically well-formed are evidence of a conceptual snare and delusion. So at a minimum I do not think think Chomskyism and Sampsonism jointly exhaust the range of possibilities, and perhaps quite a lot of us here would not fall into either camp.

  49. J. W. Brewer says

    By the way, without any disrespect intended to Prof. Pullum’s rock and roll bona fides, it is not really coherent to call any of the bands he played in in the ’60’s “indie,” both because neither the word (well, more precisely, the sense of the word being used here) nor its referent existed yet, and because if you were going to engage in the somewhat questionable and anachronistic task of picking out a few ’60’s acts as having been “indie” avant la lettre, Geno Washington et al. wouldn’t be particularly plausible candidates for such a list.

  50. > How much of this “serious work” has had any real descriptive value or predictive power?

    As a Chomskyan (with a beard), I thought I should take a stab at this one, or rather at the implication that the answer is ‘none’. Look, there are a number of facts about language that were discovered and first studied by linguists working in the Chomskyan tradition (the Coordinate Structure Constraint; the distinction between control and raising; the Final-over-Final Constraint; the c-command condition on anaphora; the Path Containment Condition for English, and its opposite for Bulgarian and Japanese; the ban on tough-movement from subject positions; Holmberg’s Generalization; the double -ing filter in English; the existence of scrambling as an account of certain instances of free word order; etc, etc, etc.).

    The above will just look like a string of technical terms to a lot of folks, but I’d be happy to discuss any of those at length, with anybody who’s curious to hear more about them; they’re all generalizations about how language generally, or particular languages, are structured, with concrete consequences for particular examples (the Coordinate Structure Constraint, for example, is the observation that you can’t say, among other things, “What did you buy beer and?” (meaning “what is the thing that you bought, along with the beer?”))

    I’m sure we work on topics that don’t interest everybody, and our host is of course free to wish that we worked on different things, but it’s ridiculous to claim that we don’t really work on anything, or haven’t made any discoveries.

  51. Athel Cornish-Bowden says

    Just a small clarification. I’m not at all competent to judge how well Sampson is regarded in serious linguistics circles (not very, I suspect, but, as I say, I’m no expert). I mentioned him in the context of of the claim that no one had tried to write a _popular_ book to compete with The Language Instinct.

  52. I’m sure we work on topics that don’t interest everybody, and our host is of course free to wish that we worked on different things, but it’s ridiculous to claim that we don’t really work on anything, or haven’t made any discoveries.

    That’s fair enough, and I absolutely wouldn’t claim that you don’t really work on anything or haven’t made any discoveries. It’s true that my own interests don’t lie in those directions, but that of course is irrelevant; I’m not interested in entomology either, but I don’t go around slagging people who study insects. My problem with Chomskyans (as a group in my mind, present company of course excepted) is not that they study syntax or focus too much on theory, it’s that they behave (or rather, behaved when I was in grad school long ago) like members of a cult, disparaging everyone else who was doing linguistics and trying to take over every department they could. I think, and will always think, that Chomsky himself did terrible damage to a field I love and think is important, so I am incapable of being coolly objective about him and his theories. But that does not mean that I tar everyone with the same brush, and I appreciate your taking the time to comment and being so nice about it.

  53. marie-lucie says

    I agree with LH.

    From a theoretical viewpoint, perhaps the major problem is that the so-called Universal Grammar (UG) and its various rules, constraints and parameters originated from a description of English. Studies of other languages within this framework seem to concentrate on discovering whether or not rule (etc) X applies in languages A, B, C, etc, sometimes finding a way to reconcile them through some modification of the theory or formal representation, but not paying attention to other phenomena which occur in those other languages but do not fit the structure or components of UG. In particular, English is a very analytical language, so its syntax is highly developed, but some languages work synthetically, achieving through higly developed morphology what English and others express through syntax. This is true of Eskimoan, Salishan and Wakashan languages, among others (the latter two spoken in the Pacific Northwest, straddling the US-Canada border). There is some syntax in those languages, but it does not play the central role that it has in European languages, especially English. Instead, it uses morphology in ways that have no counterparts in English.

    If the concept of UG had been developed starting from one of those languages, or from an Asian or Australian (or whatever) language, it would probably have taken quite a different form.

  54. John Cowan says

    In short: Universal Grammar is a parochially defined concept.

  55. I agree with Hat and Marie-Lucie, but I think neither go far enough.

    Marie-Lucie is quite correct that Chomsky-type UG is hopelessly anglocentric, but even within Europe there are plenty of languages where morphology does a great deal of the work which syntax does in English: I very much doubt a UG based on Russian, Basque or Hungarian (or better still, all three!) would have much in common with UG as it exists today.

    Moreover, Chomsky himself first worked on the morphophonemics of Hebrew, a language which he knows quite well. Why has Hebrew played practically no role in establishing UG? Ancient Hebrew, especially, as a non-Indo-European language with a rich, well-studied and stylistically diverse written corpus, would have served as an excellent initial testing ground for any “Universal” theory of language.

    The answer is to my mind simple: Chomsky’s work isn’t science. It’s academic politics at its ugliest. On the basis of E.F.K. Koerner’s work (disclaimer: a former professor of mine) I believe that the case is unassailable: it owes its success to political machinations, and to nothing else. In this light I find it telling that Labovian sociolinguistics shares with UG a strong anglocentricity, and I find it even more telling that both “schools” came to dominate American linguistics starting in the sixties. Both marginalized all forms of linguistics which actually dealt with languages other than English. This at a time when the study and knowledge, in the United States, of Latin as well as of living languages, for all intents and purposes collapsed.

    I would thus answer my own question above as follows: if we accept that Chomsky’s goal was to dominate the field, and not advance knowledge, then the anglocentricity of UG simply shows that Chomsky was well aware that a “theory” which involved any language other than English would lack the popularity he sought. The arrogance which he and many of his disciples/followers/fanatics are definitely guilty of can be explained in the same way: if political power within academia, as opposed to scientific truth, is what you are after, then non-generativists are indeed a foe to be destroyed. Not colleagues with whom a dialogue is desirable in order to find the truth. It all fits.

    Indeed, more speculatively, I wonder whether it might be possible to link Chomsky’s political writings with his linguistics: his denunciations of the American government often become so strident that I suspect that to his mind the American government is a kind of Jungian shadow, whose lust for power and amorality he can denounce vitriolically and thereby convince himself (and others?) that those traits are (of course) wholly absent in him.

    @Norvin: I am willing to grant you the benefit of the doubt and assume that you are simply an honest man who fell in with the wrong crowd. But your comment above makes this difficult. This is because you are not answering the question asked.

    The question was whether your cult –err, I mean school; ‘SCUSEZ, LAPSUS! as we say in my own surface manifestation of UG– has produced anything with “Real descriptive value” and “(Real) predictive power”. You have not even attempted to give a satisfactory answer. Your answer that “there are a number of facts about language that were discovered and first studied by linguists working in the Chomskyan tradition” sidesteps the issue.

    The question is whether Chomskyans have described and accounted for these phenomena in a (scientifically) more satisfactory manner than other schools have accounted for them or could account for them. Unless you are indeed a most atypical Chomskyan who is familiar with other schools of linguistics, I very much doubt you could give a satisfactory answer.

    I hope you will surprise me.

  56. article by E.F.K. Koerner

    Perhaps Etienne recognizes what you have in mind off hand. There have been several notable ones on the general subject of “Chomskyan Revolution,” such as this here or the one in his Practicing Linguistic Historiography. But I sense that they are covering a later period than you recalled. There were also a couple on the timeline of the influence of Saussure, which would be earlier, but much more specific.

  57. marie-lucie says

    Etienne, thank you for your contribution. I have a disclaimer of my own: Koerner was my classmate when he was doing his doctorate on Saussure (not really a classmate, as he was past taking courses).

    Do you remember that article/talk of Koerner’s about Chomsky’s early years I was talking about? It showed how some of MIT’s profs deliberately set out to “push” Chomsky and give him privileges not afforded other students.

    About my choice of non-English languages above, I am well aware that I would not have needed to go that far afield for examples: even French is difficult to fit into the UG template without some serious distortions and omissions. I once had to teach French syntax using a UG-oriented textbook and found it very frustrating. Many linguistics profs deplore the fact that students have to take syntax courses using this framework, otherwise they will not be taken seriously if they want to go on to further linguistic studies. Such syntax courses tend to crowd out others, especially morphology which was long neglected in favour of syntax. Of course morphology is quite simple in English, compared to a large number of other languages, if not even the majority of them.

  58. John Cowan says

    Etienne: The study of languages other than English in the U.S. collapsed around 1916, if not 1776. It was artificially maintained on life support until the 1960s, but I don’t think you can blame the ending of mandatory (and useless) language requirements for high school and college students on Chomsky. As for his politics, Chomsky has pointed out that he would have had the same politics if he had become an algebraic topologist. Whether he would have engaged in the same intellectual imperialism in U.S. math departments, no one can say.

    Labovian sociolinguistics shares with UG a strong anglocentricity

    Whatever may be the case for his disciples, Labov himself has never, as far as I know, made universal claims for his work, or extended it beyond English. (MMcM perhaps knows otherwise.) Furthermore, he has a grounding in empiricism that the Chomskyites generally lack.

  59. John Cowan says

    The study of languages other than English in the U.S. collapsed around 1916

    Americanists are an honorable exception to this, of course.

  60. marie-lucie says

    MMcM,

    Thank you for trying to locate the text I was talking about. Looking at your first reference I recognize some things I have read before, but the article is much longer than what I remember, and the format is more polished, more suitable for a book. It is possible that what I read earlier and still remember parts of was the text of the talk Koerner refers to in a footnote, a presentation in Vienna in 1982, and that he used and enlarged this unpublished text for publication.

  61. marie-lucie says

    JC: the study of languages other than English collapsed

    This is about compulsory study of European languages in high schools, not academic research in these and other languages. The study of Native American languages was long classified as a branch of anthropology, until linguistics was established as a separate discipline. Chomsky’s emphasis on English changed the character and typical personality of linguists: many students were happy that they could study “language” without having to study “languages”, and rely on their own “native speaker intuition” instead of having to rely on the intuitions of actual speakers of other languages.

  62. John Cowan says

    This is about compulsory study of European languages in high schools, not academic research in these and other languages.

    Sure. But I realized that what I wrote could be misinterpreted.

  63. the adoption, on somewhat Fascist principles, of the term “denialist” is quite good evidence for what I mean.

    Oo, ‘somewhat Fascist’! I repeat: it’s ridiculous to claim there’s an Evil Worldwide Statist Climate Conspiracy when there’s nothing else like it in modern science. Your nonresponse is quite good evidence for what I say. Almost all the anti-AGW argumentation I see is godawfully weak; that and the religious or libertarian fundamentalism of so many of the denialists is why I class it with creation-ism rather than climat-ology.

    My original post was apparently too off-topic for anyone else to respond, but I think the comments since then do indicate that Chomskyism is a particularly American disease. And that people who claim ‘warmism’ is an ism are thus obliged to explain how this one single field of modern science somehow got almost universally corrupted everywhere.

  64. Treesong: … I think the comments since then do indicate that Chomskyism is a particularly American disease. And that people who claim ‘warmism’ is an ism are thus obliged to explain how this one single field of modern science somehow got almost universally corrupted everywhere.

    I don’t understand. What is the “one single field of modern science” that you see as almost universally corrupted ? What connection between Chomskyism and warmism requires certain people to produce explanations of what ?

    Is this science climatology ? Squabbles over global warming are not a sign that the whole field of climatology is corrupted. Is this science linguistics ? Why should squabblers over global warming have to explain why the whole field of linguistics is corrupted, if it is ?

    Possible further examples of American derailments: “Marxism” in political science departments, “Theory” in literature departments, “learning to count via set theory” and “learning to read by whole words not analyzed into syllables” in departments of education.

    My impression over the last 40 years is that German university departments have not been subject to such crazes. Fashions, yes, but not derailments. As one might expect, squabbles in Germany take place on a higher plane. The education system from top to bottom is a shambles, reorganized every 5-10 years. People can’t agree on anything.

  65. marie-lucie: Chomsky’s emphasis on English changed the character and typical personality of linguists: many students were happy that they could study “language” without having to study “languages”, and rely on their own “native speaker intuition” instead of having to rely on the intuitions of actual speakers of other languages.

    This is the most enlightening comment I have yet encountered on this subject.

  66. I just wonder now: what percentages of linguists in the various camps (Chomsky-riding, Chomsky-deriding) actually command more than one language ?

  67. The syntax-centrism of Chomskyan accounts of grammar no doubt is closely related to the historical accident that the movement emerged in the US among native English speakers. But syntax plays a significant role even in languages like Russian (and Hebrew word order, notwithstanding all that morphology, is actually almost as fixed as English’s). Syntax is a problem of universal significance for linguistics that happens to be extra-important for English, not a parochial problem for students of English, and – as Norvin comments above – focusing so hard on syntax did let Chomskyans notice at least some things of very general relevance that hadn’t been noticed before, and might never have been noticed.

    The central problem, as Marie-Lucie notes, is rather the imperialism of the research program: the idea that any study of languages must be justified by its contribution to the study of Universal Grammar – that is, of language – or it’s just “butterfly collecting”, to use what used to be a favorite phrase of theirs. This is lethal to a lot of disciplines within linguistics, but most especially to historical linguistics, where only non-universal facts about a given language can be relevant to its classification.

  68. Well, fieldworkers from the descriptive-typologist camp typically know four languages: your mother tongue, English, the local lingua franca where you do your fieldwork and the target language. Language-learning is a quite accepted hobby in the cultural atmosphere too, I remember seeing a well-known scholar of Modern Tibetan speaking Persian, with everybody watching in approving admiration.

  69. What Lameen said.

  70. marie-lucie says

    Lameen: excellent comments. I did not mean to downplay syntax, but it is not all there is in linguistics.

    the idea that any study of languages must be justified by its contribution to the study of Universal Grammar – that is, of language – or it’s just “butterfly collecting”, to use what used to be a favorite phrase of theirs. This is lethal to a lot of disciplines within linguistics, but most especially to historical linguistics, where only non-universal facts about a given language can be relevant to its classification.

    I could not agree more. Butterfly collecting indeed! Doesn’t entomology count as a scientific endeavour? As a primarily historical linguist I can’t find anything in UG that has relevance to my work.

  71. Stu: What is the “one single field of modern science” that you see as almost universally corrupted ? What connection between Chomskyism and warmism requires certain people to produce explanations of what ?

    I rerepeat my point: it’s ridiculous to claim there’s an Evil Worldwide Statist Climate Conspiracy when there’s nothing else like it in modern science.

    I don’t see any field of modern science as universally corrupted, but plenty of denialists do; they claim anthropogenic global warming is a hoax by warmists or, less crazily, an exaggeration by alarmists. I took dearieme’s original comment to mean that he had other examples. They were not forthcoming. And the examples you give are both parochial, as you said, and not science, so I feel my argument remains strong that such denialists got some ‘splainin’ to do.

    The real point, of course, is that the facts are so solidly against them, but to get to that point you need to clear away the conspiracy-theory underbrush.

    And now I will shut up. Sorry, everyone.

  72. No need to apologize; it’s hard not getting infuriated with denialists.

  73. John Cowan: I entirely agree with you on Labov: he is a fine, open-minded scholar who has never attempted to promote his brand of sociolinguistics as the alpha and omega of linguistics, but all too many other prominent sociolinguists (I speak from experience here) are as dogmatically ignorant and intolerant as your typical generativist.

    However, I never blamed Chomsky for the decline in foreign language teaching in the U.S.A.: my claim was that, perceiving this decline, he deliberately made his brand of syntax anglocentric in order to maximize its appeal. In like fashion, I referred to the stridency of the tone in his political writings, not his political writings per se.

    Echoing some comments here on the imperialism of Chomskyan linguistics: to my mind one of the most deplorable consequences of the rise of generative grammar is the oblivion which pre-Chomskyan American linguistics has fallen into. This is a huge pity, as pre-Chomskyan American structural linguistics, with its heavy empirical grounding in the study of (especially) Native American languages, represents to my mind a high point in the history of linguistics. Sapir’s descriptive work, or Bloomfield’s 1946 sketch of Proto-Algonquian, for instance, would deserve to be far more widely read than they are today, by all linguists, whatever their nationality.

    Alas, because opposition to generative grammar outside the United States is often (unsurprisingly) rooted in various kinds of anti-Americanism, there is an unfortunate tendency for those non-American linguists who would find pre-Chomskyan American work highly stimulating not to engage with any kind of American linguistics.

    Stu: in my experience typologists and field workers tend (on average!) to be the most polyglot linguists, while historical linguists win the prize (again, on average) for the number of languages they have a reading knowledge of.

    Lameen: Ah yes, butterfly collecting. A good satirist could probably imagine what entomology would be like as a field today if over the course of its history it had come to be dominated by some Chomsky-like guru figure: I can see it already…generations of scholars furiously debating the properties of the “Universal insect” solely on the basis of the guru’s writings on his impressions of a crushed fly he cast a glance at decades ago.

    More seriously, though, Norvin seemed to defend generativism on the basis of the fact that generativists has been the first to notice certain phenomena…so in effect, butterfly collecting is okay, as long as the butterfly collecting is done by followers of generativism: butterfly collecting by non-generativists, on the other hand, is, of course, both trivial and uninteresting.

  74. marie-lucie says

    Etienne: another great comment.

    butterfly collecting is okay, as long as the butterfly collecting is done by followers of generativism: butterfly collecting by non-generativists, on the other hand, is, of course, both trivial and uninteresting.

    But when generativists are searching for examples validating their “discoveries” in non-English languages, they rarely do fieldwork but instead search those trivial, uninteresting grammars previously written by non-generativists!

  75. John Cowan says

    Etienne, I think that overstates the case. Norvin merely pointed out that Chomskyists have discovered interesting things, which nobody denies. Nothing is gained by exaggerating the case against Chomskyism, which is about intellectual imperialism, by adding to it an unsustainable claim that Chomskyist linguists aren’t doing science at all.

  76. John Cowan: actually, I indeed do not believe that generativists, as a rule, practice science.

    Here’s an example of a type of linguistics which is definitely a science: historical linguistics. In the nineteenth century the theory of the regularity of sound changes was expounded. This theory, of course, made specific predictions on language change: if sound change is regular, and not, as had been believed earlier, random, then predictions could be made as to what earlier stages of various languages were like.

    Thus, on the basis of the theory that sound change is regular Saussure claimed that Proto-Indo-European must have had phonemes of some kind in some very specific positions in some very specific words. When the Hittite language of Asia Minor was discovered it was realized (by Kurylowicz) that Hittite showed at least some of the phonemes whose existence Saussure had predicted. There are other examples, too (the Proto-Greek labio-velars whose existence had been predicted and which were discovered when Linear B was cracked, the discovery of the Gaulish word DUXTIR “daughter” on an inscription, confirming what had earlier been predicted by Celticists to be the reflex of the Proto-Indo-European word in Gaulish and Early Celtic…)

    Now, it would have been utterly impossible to make such accurate predictions about aspects of extinct languages before the nineteenth century, before the theory of the regularity of sound change was presented. Even the best pre-modern tradition of descriptive linguistics, Panini’s description of Classical Sanskrit, would have failed, because it had as its focus the description of a particular language, not cross-linguistic comparison. In this sense nineteenth-century historical linguistics represented a huge step forward because it made testable predictions, which later archeological and/or philological discoveries typically confirmed. It made very specific predictions that were falsifiable, in principle as well as in real life. In short, it was a genuine science.

    Pointing out the existence of interesting things does not a science make. Considering the amount of funding and institutional support generative grammar has enjoyed over the past several decades, it would have been utterly astonishing had they all failed to notice some kinds of phenomena which previous generations of linguists hadn’t. The same would have been true of any similarly blessed school of linguistics. Or, tellingly, of pseudo-linguistics.

    But inasmuch as, *despite all this institutional support*, generative grammarians have failed to show any way in which their school is in any measurable way better than others, have failed to show that generative grammar is a better heuristic tool than any of the many other grammatical traditions out there, have failed to make a single tangible, verifiable prediction about anything of a linguistic nature, I indeed think the default hypothesis ought to be that it is not a science.

  77. Etienne, you’re a man after my own heart!

  78. I agree with a lot of the criticisms here, but isn’t this beating a rather sickly horse? Up until, I don’t know, 20 years ago, the generativist monopoly was a significant problem. These days there are islands of Chomsky-style generativism, but is the influence of this school still as strong?
    However, I would say that other programs, historically born of the generative tradition but now independent of it, are causing similar though lesser harm. For example, it’s hard to avoid optimality theory, which like GG has at its heart a search for deep universals, and which likewise has generated a lot of work of dubious lasting value. I don’t know whether OT is going as strong as it used to, but it still seems hard to avoid.

  79. John Cowan says

    Pointing out the existence of interesting things does not a science make.

    Interesting particulars, no; interesting generalizations, yes. To be doing science, it suffices to aim for descriptive adequacy; that is, theories that account for what happens without requiring that they explain what doesn’t happen. Evolutionary biology, for example, cannot account for why certain hypothetical creatures have not evolved; they didn’t, and that’s all we know. (In other cases, we do know; macroscopic animals can’t move on wheels and axles, for example, because they would be required to be discontinuous.)

    It was Chomsky himself who said that linguistic theories must have explanatory adequacy to be interesting, and it seems to me that you have been seduced by this. “[I]f you turn your back on Mishnory and walk away from it, you are still on the Mishnory road.” (Le Guin, The Left Hand of Darkness). Of course explanatory adequacy is nice if you can get it, but you cannot in fact always get it if you try.

    I also must make a distinction between doing science (which I referred to) and being a science (which I didn’t). Literary criticism is not yet a science, nor is memetics, but it is possible to do science in these fields, as evidenced by Michael D. C. Drout’s work on tradition. It’s also possible to do pseudo-science or not-even-wrong science, which is part of why they are not yet sciences, and unfortunately most people do.

  80. It sounds like Etienne would like to hear some predictions that we’ve made, so here are a few.

    The Final-over-Final Constraint predicts, among many other things, that there will be no languages in which the standard word order is Subject-Verb-Object-Auxiliary.

    We predict that in every language, if the comparative form of an adjective is suppletive, then the superlative form is also suppletive (for example, English good has the suppletive comparative and superlative forms better and best, but there could not be an adjective good that had a suppletive comparative better but a regular superlative goodest).

    We predict that if there is a grammatical process that might in principle apply to any of several words in the sentence, the choice of which word to apply the process to is made on the basis of structural considerations, rather than by appealing to linear order.

    The approach to the Path Containment Condition developed in Kitahara (1997) and Richards (1997) makes a prediction about the treatment of examples in which wh-movement paths overlap; it predicts that in a language (like English) in which just one wh-phrase moves in a multiple wh question, for languages in which there is any difference in acceptability between crossing and nested paths, the preference will be for nested paths. Moreover, the prediction is that in languages that do not exhibit the English pattern for multiple wh-questions (that is, languages like Japanese, in which no wh-phrases move, and like Bulgarian, in which all of them do), if there is a preference, the preference will be the opposite of the English one.

    Current approaches to reconstruction predict that in all languages, if there is any material which can avoid reconstruction for Principle C in A-bar movement, that material will be contained in adjuncts rather than arguments of the head of the moved phrase. We expect not to find languages in which the opposite is true.

    We also predict that if any type of movement can avoid reconstruction for Principle C, it will be A-movement rather than A-bar movement; no language will exhibit the opposite tendency.

    Our theories of islands predict that in every language, translations of questions like ‘Why did you meet a man who resigned?’ will, if acceptable at all, only be able to be understood as questions about the reason for meeting, and not about the question for resigning.

    I could go on, at tedious length, but, well, I probably already have. Again, happy to discuss any of these in more detail.

    Oh, and Etienne wants evidence that generative approaches are in a better position to handle these than non-generative ones, which in a sense is easy; as far as I know, non-generative approaches have never attempted to deal with any of them, apart from the last one. The last one has sometimes been approached in terms of ‘conditions on working memory’; I think such approaches have been very convincingly critiqued, for example, here: http://ling.umd.edu/~colin/research/papers/sprousewagersphillips2012.pdf

  81. John: macroscopic animals can’t move on wheels and axles, for example, because they would be required to be discontinuous.

    I would be interested to learn where you picked up that idea. Something very like it is an argument used repeatedly by Peter Janich (very short English WiPe entry here) in his work on “protophysics” and later “methodic culturalism”, especially in the 2005 Kultur und Methode which I read last year.

    I met Janich briefly in my last year at UT Austin in 1969/70 when he was there. The background was that I had picked up a copy of Logische Propädeutikby Lorenzen/Kamlah at Garner and Smith’s bookstore, and was reading it with my minimal German at the time. Janich had studied under and worked with Lorenzen. That’s another reason for acquiring several languages: you meet people and learn first-hand about ideas that you might never have encountered otherwise.

  82. Norvin—FWIW, I once used ‘goodest’, meaning a person of high character, instead of ‘best’, which would have signified to me a sense of skill or overall quality. To me, ‘better’ carries the comparative meaning for both senses. In other words, Two separate meanings for good-better-best and good-better-goodest. The person I was talking to, a native English speaker, understood me exactly as I had meant.

    I am available as a subject for a paper on the matter, in an appropriately prestigious journal, for a reasonable fee.

  83. John Cowan says

    We predict that in every language, if the comparative form of an adjective is suppletive, then the superlative form is also suppletive.

    I’m working my way through Bobaljik 2007; if this is discredited, please let me know. The question looks very interesting, and involves comparative evidence as well as a diachronic vs. synchronic question. I’ll discuss it further when I finish the paper.

    One thing that has come to mind so far, though, is that without a decent theory of the origin of suppletion, the claim that it is UG and not historical accident that compels the superlative to be a morphological extension of the comparative doesn’t convince me. Anglophones acquire better/best, worse/worst because that’s what they hear. But I’ll try to withhold judgment until I’ve finished reading.

    [I]f there is a grammatical process that might in principle apply to any of several words in the sentence, the choice of which word to apply the process to is made on the basis of structural considerations.

    Any theory that involves hierarchical decomposition at all will predict this.

    Our theories of islands predict that in every language, translations of questions like ‘Why did you meet a man who resigned?’ will, if acceptable at all, only be able to be understood as questions about the reason for meeting, and not about the question for resigning.

    I think you mean “and not about the identity of the one who resigned”; if so, the same remark applies.

  84. : if the comparative form of an adjective is suppletive, then the superlative form is also suppletive.

    That fails for Algerian Arabic: mliħ “good”, xiṛ (mənn-u) “better (than him)”, əl-mliħ (fihǔm) / *əl-xiṛ (fihǔm) “the best (of them)”. (I would say it fails for Standard Arabic as well, but in Standard Arabic it’s not entirely obvious which adjective xayr “better” is a suppletive form of.)

    : Our theories of islands predict…

    I’ve never seen a theory of islands that could adequately account even for English-internal variation – half the supposedly ungrammatical examples in my textbook were completely acceptable to me. I’d be glad to hear that the situation has progressed since, but it hasn’t been that long. (An old Chomsky book – much older than my textbook – once prompted me to gather some experimental data on this, which illustrates just how unrepresentative introspective judgements are on the issue: Subjacency judgments.)

    And this is one of my biggest problems with the Chomskyan tradition in practice: not that it doesn’t make predictions – it makes them at the drop of a hat – but that it tends to declare victory for its predictions well before they have been adequately tested.

  85. I should emphasise, to be fair, that Norvin did not actually claim victory for the predictions he listed. My complaint in that respect applies more to the theory of subjacency, which is routinely claimed as an achievement of generative grammar and taught as established fact while glossing over serious empirical problems.

    Ironically, however, judging by Bobaljik 2007, Bobaljik’s comparative-superlative claim hardly counts as a prediction anyway. He found (with a good deal of help from the typological literature) that for almost all languages, “if the comparative form of an adjective is suppletive, then the superlative form is also suppletive”, and proceeded to propose a semantic analysis – going against the majority of prior literature – from which that would follow. Quite apart from the generalisation not actually being universal (even he cites the case of Karelian), it’s entirely post facto.

  86. : The approach to the Path Containment Condition developed in Kitahara (1997) and Richards (1997) makes a prediction about the treatment of examples in which wh-movement paths overlap…

    A little background: Many languages move wh-words (specifically) to the start of the sentence. Some, like English, only let you do this with one wh-word at a time: “Who gave whom the book?”. Others, like Bulgarian, let you do this with several at a time: “Who whom gave the book?”. Still others, like Chinese or Japanese, don’t do it at all. If I understand correctly after my brief look at Roberts 1997, the prediction Norvin is reporting here is that all languages like English, where only one WH-word can be moved at a time, should judge:
    a. √What do you wonder who saw? [with the reading: For what thing is it the case that you wonder who saw it?]
    as grammatical, and:
    b. *Who do you wonder what saw? [with the reading: For whom is it is the case that you wonder what he saw?]
    as ungrammatical.
    Whereas all languages that allow multiple WH-movement, like Bulgarian, should do the opposite, and judge their equivalents of a. as ungrammatical and b. as grammatical.

    There should be a few people on this thread familiar with languages that allow multiple WH-movement – care to weigh in on this prediction for your languages? If it comes to that, how do other English-speakers rate a. and b.?

  87. Lameen, thanks for spelling out that prediction (and for the Algerian counterexample to Bobaljik’s theory, which I’ll bear in mind as I talk and think about that prediction in future). One note of caution; the prediction about overlapping wh-paths only holds for speakers who have any preference between your (a) and (b) at all. There are speakers for whom both are irredeemably bad, and speakers for whom they’re both fine. But if there’s a preference, then we predict that the preference will run a certain way.

  88. I agree with a lot of the criticisms here, but isn’t this beating a rather sickly horse?

    Sure, and a good thing too! I’m delighted the horse has fallen out of the running, but it still gives me pleasure to beat it on a regular basis. I am a bad person.

    And this is one of my biggest problems with the Chomskyan tradition in practice: not that it doesn’t make predictions – it makes them at the drop of a hat – but that it tends to declare victory for its predictions well before they have been adequately tested.

    Exactly! And then when someone goes to the trouble of digging up counterexamples, they shrug, go “Oh well,” and proceed to the next exciting prediction, without worrying their heads about why this happens on such a regular basis. It’s almost as if they were, I don’t know, astrologers or palm-readers.

  89. Er, no offense, Norvin! I have nothing against individual members of the sect, just Chomsky himself and the failings of his theory (and the way it was rammed down the collective throat of the field). You seem like a fine fellow.

  90. [peeve]

    on such a regular basis

    What’s the matter with “so regularly”?

    [/peeve]

  91. What’s the matter with “so regularly”?

    Nothing. There is also nothing wrong with “on such a regular basis.” Every language has various ways to say things, differing in style, connotation, rhythm, or what have you. There is not One Right Way to say things.

  92. No argument on the One Right Way. Just annoyed at the overuse.

    “So regularly” gets 762K hits on Google. “On a regular basis” gets more than 1 billion.

  93. Well, that shows you it’s far more popular than your preferred alternative. You might consider whether that’s “overuse” for some incomprehensible reason or just a feature of contemporary English. To me, as (clearly) to a lot of people, it’s an attractive phrase with a good rhythm. My advice to you is to let go of your irritation, because the phrase isn’t going away.

  94. Maybe a fairer comparison would be “regularly” versus “on a regular basis”, or “so regularly” versus “on such a regular basis”.

  95. Also, what does counting Google hits tell us? I get more ghits for “so regularly” than for “regularly”.

    In Google N-gram “regularly” beats “on a regular basis” by a mile, and “so regularly” similarly trounces “on such a regular basis”.

  96. marie-lucie says

    Sorry if I am not as familiar with the literature as some of you are, but why is the comparative-superlative case interesting from the point of view of syntax rather than morphology? It seems to me that in languages that mark “degree” morphologically, like the Germanic ones, if one stem (called suppletive) only occurs in the “degreed” context, while a semantically equivalent stem does not occur in that context, the “degreed” stem is normally inflected for however many degrees can be marked (two in English, German, etc, but other languages might have fewer or more than two degrees).

    Again without referring to published literature, I think that “suppletion” arises from the existence of near-synonyms which in the course of time become associated with specific contexts. For instance, in English, Latin, and a number of languages, some stems meaning ‘good’ have become used exclusively with suffixes marking degrees of (relative) goodness (so better/best), while good itself has been restricted to the generic, unmarked meaning. Adding the degree suffixes to the restricted stem (as a child or adult learner might do, or an adult speaker for a joke) would result in perfectly understandable forms (although the opposite will not necessarily be true: *bet ‘good’ would not now be understood). In the Algerian case, it is likely that in older stages of the language there were not two but three or four near-synonymous stems, which were used in mostly different contexts (either semantic or grammatical or both) and ended up being used exclusively in those contexts.

    In languages like French or Spanish which do not inflect for degree but use extra words (Fr plus, Sp mas ‘more’), suppletive forms exist but that is because they continue suppletive Latin forms which were themselves inflected (for just one degree!) eg Fr bon/bonne, meilleur/-e, Sp bueno/-a, mejor ‘good, better/best’ from Latin bon-us/-a/-um, melior.

  97. Athel Cornish-Bowden says

    ‘Maybe a fairer comparison would be “regularly” versus “on a regular basis”, or “so regularly” versus “on such a regular basis”.’

    That’s what I thought. It gives preference in the same direction, but to a much smaller degree — about 3 million for “on such a regular basis”.

  98. Athel Cornish-Bowden says

    Ø: Sorry, I see you’d already followed up your comment before I did.

  99. marie-lucie: The Algerian case is partly parallel to the Romance one. Classical Arabic had a highly productive morphological comparative, along with a couple of suppletive comparatives of which the commonest was xayr “better”, seemingly derived from the noun xayr “goodness, benefit”. However, it’s not obvious which adjective this is the comparative of, because there are several near-synonymous common words for “good” (ħasan, ṭayyib…), and all of them also allow regular comparatives (‘aħsan, ‘aṭyab…) Algerian Arabic has lost almost all morphological comparatives in favour of an analytic construction using ʕla “than”, but retained suppletive xayr > xiṛ, along with a handful of common formerly regular comparatives like kθəṛ “more” < 'akθar. Unlike Classical Arabic, it does not allow the productive comparative formation to be used with the adjective mliħ "good" (*mliħ ʕli-h "better than him" is unacceptable, though comprehensible).

    Hat: And then when someone goes to the trouble of digging up counterexamples, they shrug, go “Oh well,” and proceed to the next exciting prediction, without worrying their heads about why this happens on such a regular basis.

    Yes, pretty much. You’d think after all these years, and with the growth of typology, there would be a moratorium on proclaiming barely tested (or even already falsified) speculations as established universals, but it keeps on happening. Chomsky himself has been known to do it, as Postal points out with regard to the A-over-A principle (previously discussed here at Language Hat). And in many cases, the framework offers so many potential ways of reanalysing any given counterexample that it’s hard to tell whether a given prediction is even falsifiable in principle. Honest question (I haven’t followed the history closely enough to know the answer): is there any theory within the Chomskyan tradition that has been retired simply because its predictions turned out to be wrong?

    The thing is, though, “Chomskyanism” isn’t a theory so much as a research program plus a paradigm. The paradigm is that the human mind has an innate, more or less universally shared faculty of language, with language-specific adaptations, which all linguistic activity imperfectly reflects. The research program is that linguistics should be a branch of psychology dedicated to understanding this faculty of language better. If specific theories proposed within this program (or even most such theories) turn out to be wrong, that doesn’t change anyone’s mind about the program’s validity. Even conclusively proving that the paradigm is wrong – say, that language learning relies entirely on general human learning capacities – wouldn’t actually doom the research program, any more than chemistry became useless once we realised it could all be theoretically reduced to physics. And research programs can’t be refuted; they can only be rejected. I don’t have any objection to people who want to do linguistics for the sake of understanding psychology, as long as they’re prepared to let me keep doing linguistics for the sake of understanding history, or for that matter for the sake of understanding language(s).

  100. John Cowan says

    I am a bad person.

    Perhaps you should add “I am a bad person who loves to beat Chomskyites on a regular basis” to the right margin of the blog.

    I wonder if it’s been looked at, how often novel suppletives replace existing suppletives, like ME-ModE went for OE eohde, both suppletive of go, or much/many as the suppletive positive of more, most, displacing mickle (OE mycel) ‘big’.

  101. I don’t have any objection to people who want to do linguistics for the sake of understanding psychology, as long as they’re prepared to let me keep doing linguistics for the sake of understanding history, or for that matter for the sake of understanding language(s).

    It would help if they’d call their field “psychology of language” and let us reserve “linguistics” for what it should properly (and yes, I’m being prescriptive here) mean.

  102. Norvin-

    Thank you very much for replying.

    “The Final-over-Final Constraint predicts, among many other things, that there will be no languages in which the standard word order is Subject-Verb-Object-Auxiliary.”

    Grab your passport and toothbrush and get thee to rural Newfoundland right away! In some older speakers’ English you have such pairs as:

    1-That boy’s doin’ too much work he is
    2-That boy’s doin’ too much work he was

    The final “he is/he was” are tags, historically, of course, but their frequency is such that the above two sentences definitely struck me as the unmarked structure, doubly so since the final “he is”/”he was” is the only tense-marking (present versus past) element of the sentence. It certainly looks like Subject – Verb – Object -Auxiliary order to me. There may be other varieties of English which exhibit similar structures (Conservative Hiberno-English varieties might be worth examining in this light).

    Please understand that I am not giving this counter-example out of some kind of angry nihilism. As a historical linguist I would dearly love it if theoretical work allowed us (for example) to exclude certainly structures when having to select between competing possibilities when reconstructing earlier stages of historically attested languages or reconstructing proto-languages. I would be nice if the generative seminar I was forced to take in grad school turned out to be worth more than nothing to my research. But the fact that different branches of generative grammar make different predictions and that the rise/decline of different such branches appears unrelated to any close engagement with linguistic data is, to put it diplomatically, not encouraging.

    And speaking of data, I would urge you to be very careful. I know from personal experience that much of the non-English data used by at least some generativists was obtained from graduate students who, on account of their relative powerlessness, knowledge of what their professor was hunting for, as well as their immersion in an anglophone environment, cannot be said to have been ideal informants.

    The fact that you yourself are willing to discuss these matters does put you light-years ahead of the pack, I must say. I sense you would have made an excellent historical linguist, dialectologist or typologist. You’ll forgive me if the phrase “”¡Dios, qué buen vassallo si oviesse buen señor!” comes to mind.

    Y: I truly would love it if generativism were a sickly horse, but alas, it is not. Most non-linguists, if they have heard of linguistics at all, think of Chomsky as the prototypical linguist, and indeed many a “linguist” seems to believe that pre-Chomskyan linguistics is a contradiction in terms. So it is worth beating this horse, because it is most vigorous and healthy, sadly.

    John Cowan: Oh, I agree we can examine something scientifically without making predictions. But inasmuch as the “Chomskyan revolution” is supposed to have ushered in a Brave New Age of Linguistics so far ahead of pre-Chomskyan linguistics that the latter may be consigned to historical oblivion, it is only fair to compare generativism to the hype: as I tried to show above, its “explanatory adequacy” is measurably inferior to that of the Classical Historical Linguistics that was in place when Chomsky was born.

  103. Etienne,

    I was referring not to the status of Chomsky among non-linguists, but to the status of generative grammar within North American linguistics departments. Thirty years ago, you absolutely had to at least pay lip service to generativism to get your Ph.D. in all but a few universities. These days, you can be utterly anti-generativist and look forward to a faculty job (as much as anyone, anyhow). Am I wrong?

  104. > Honest question (I haven’t followed the history closely enough to know the answer): is there any theory within the Chomskyan tradition that has been retired simply because its predictions turned out to be wrong?

    Sure. One of the first moves in Minimalism, for example, was Chomsky’s proposal to get rid of the idea of D-structure and S-structure, which he had previously invented as a way of implementing a hypothesis that all movement operations should take place after the entire tree was constructed (so you had a series of operations that created a tree without movement, which was the D-structure, and then you performed movement operations that changed the positions of the phrases you’d created, generating the S-structure). Chomsky was convinced by work by David Lebeaux and by Danny Fox which demonstrated that it was useful to allow movement and non-movement operations to be interleaved (in particular, they showed that it was useful to be able to construct, for example, a DP, then move it, and then attach a relative clause to it–so-called “Late Merge”). So, that did it for S-structure and D-structure.

    Lots more examples where that came from, but that’s one.

  105. John Cowan says

    Most non-linguists, if they have heard of linguistics at all, think of Chomsky as the prototypical linguist

    Most non-physicists thought of Einstein as the prototypical physicist, but he could never accept quantum mechanics.

  106. Aside from these particular cases, using the existence of particular universals as a test of universal grammar is unreliable. Every linguistic conference brings counterexamples to claimed universals from new studies, and there’s no reason to believe that the current body of documented languages covers more than a small part of linguistic possibility. A better test of innate universals would be to ask the question, “can a child (or a talented L2 learner) learn to produce and understand this particular feature?” I can easily learn to use “good-better-goodest” (which may exist in some natural languages), or to stress every third syllable in a sentence (probably does not), or maybe figure out how to use an SVOS sentence structure, repeating the subject phrase at the end (I doubt that exists either.)

  107. What follows will be a quick series of posts, meant to address some of the points that have come up, and surely missing others, since this conversation has gone on for a while!

  108. Etienne: I will stop by Newfoundland sometime and have a look, thanks (the goal will be to find out whether those things at the end of the sentence are actually auxiliaries, since, as you seem to acknowledge, it’s something of a stretch). Lots of questions arise: how do they say ‘I’m doing too much work’? And what happens in sentences with multiple auxiliaries? And do they do this trick in embedded clauses?

  109. Marie-Lucie:
    why is the comparative-superlative case interesting from the point of view of syntax rather than morphology?

    There’s an open question about what kind of connection there is between syntax and morphology; one proposal out there (that of Distributed Morphology) tries to get syntax to do as much morphological work as possible, and on that standpoint you expect morphological phenomena to shed light on syntax, and vice versa. Doesn’t mean they’re right, of course, but that’s the hypothesis they’re pursuing.

  110. Lameen, if you’re not already familiar with it, you might be interested in this paper by Sprouse, Almeida, and Schütze:

    https://files.nyu.edu/ad151/public/papers/sprouse.schutze.almeida.2013.pdf

    What the authors did was to replicate the English judgments from a decade of papers from the journal Linguistic Inquiry, using a variety of experimental methods from other branches of cognitive science, getting judgments from large numbers of speakers. Their results agreed with the judgments in the original papers, 95% of the time. Another paper by Sprouse and Almeida does the same for Adger’s syntax textbook (a leading textbook in Minimalist syntax):

    https://files.nyu.edu/ad151/public/papers/sprouse.almeida.2012.pdf

    This time the replication rate was 98% (and yes, there are some Subjacency judgments in there). I think our data are pretty reliable, actually.

  111. Finally, just a note about anglocentrism. That’s a charge that gets levied against Chomskyans on a couple of grounds, I think. One is that, especially early in its history, Chomskyan syntax tended to concentrate on data from English. As has already been suggested, I think that was largely because the practitioners, at the beginning, were mostly native speakers of English. Etienne suggested that it was sinister that Chomsky hadn’t made a lot of use of Biblical Hebrew in constructing his theory; I think a simpler explanation is that, even today, we don’t tend to make much use of corpora at all, in any language.

    This is because the data we study are vanishingly rare in corpora. I have a long-standing interest in multiple-wh questions: these are questions like “Who bought what?”, which ask about more than one thing at the same time. Such questions don’t often appear in conversation, or in corpora; there are none of them in the Bible, for example, which is a fairly large text (and the only corpus we’re going to have for a large number of languages, at the rate we’re going). Lameen has an earlier post where he invites us to consider the difference between ‘What do you wonder who saw?’ and ‘Who do you wonder what saw?’ I’m fairly confident that I won’t find either of those sentences in a corpus, unless it’s a corpus of example sentences from syntax textbooks. But I also think that any native English speaker who gets a contrast between these is going to prefer the first. That’s the kind of fact we’re choosing to build our theory on, and the only way we can get that kind of fact is by asking people; we’re not going to find it in a corpus, or in a descriptive grammar.

    At any rate, if Chomskyans once wrote mainly about English, we don’t do that any more. Over the past ten years, for example, MIT graduate students have written dissertations (http://web.mit.edu/linguistics/graduate/dissertation/list.html) about the syntax of Chol, Kiowa, Korean, Mandarin Chinese, Mebengokre, Slovenian, Tagalog, Tlingit, Tseltal, Uyghur, Zulu…okay, and English. Did you really expect us to stop writing about English? There’s plenty left to discover. But ‘English syntax’ dissertations are now the exception, rather than the rule.

    The other version of ‘anglocentrism’ that you hear about is that our theories about other languages are really intended to make them look like English; that’s an idea that marie-lucie mentioned at one point. I don’t think that’s true either, though of course it’s much harder to prove one way or another. A fact that seems significant to me is that there are many aspects of the theory currently accepted among people of my ilk for which the best evidence is from languages other than English.

    For example, successive-cyclic wh-movement (the idea that wh-movement proceeds in a series of small hops) was first proposed as part of an account of island phenomena, but the best evidence for it, I think, comes from languages like Irish, Chamorro, Tagalog, and Dinka.

    We standardly posit a distinction between VP and vP, (a phrase headed by the verb, and a phrase just above it which is responsible, among other things, for introducing the agent of a transitive verb and for assigning case to objects) and the best evidence for this that I know of is from German (from the work of Susi Wurmbrand).

    Seth Cable developed a very influential theory of pied-piping (the observation that movement can move objects of variable size, subject to certain restrictions: that, for example, “[Which table] did he hide under?” and “[Under which table] did he hide?” are both more or less fine, modulo objections from English teachers, and both clearly preferable to “[Hide under which table] did he?”) which attributes the choice of how much material to move to the conditions on the placement of a morpheme which attached to the moved phrase. The morpheme is invisible in English, but visible in Tlingit, and it is on the Tlingit facts that he bases his argument.

    We now assume, contra much earlier work, that the null subject of embedded infinitives (the (inaudible) subject of the bracketed clause of a sentence like “John tried [to lift the piano]”) has the same case morphology on it that any other nominal would, and that’s not something you can learn from English—you learn it from languages like Icelandic.

    We really are in the business of learning from all of the world’s languages.

  112. John Cowan says

    Chomsky was convinced by work by David Lebeaux and by Danny Fox which demonstrated that it was useful to allow movement and non-movement operations to be interleaved.

    Well, that’s as may be, but saying theory B is more useful than theory A is not to falsify theory A; that happens only if evidence which cannot be accounted for by theory A appears. Now utility is often the reason for preferring one theory to another, to be sure. The Keplerian heliocentric solar system is just plain easier to reason about than the Tychonic geocentric solar system, but the Tychonic system has not been and cannot be falsified, because the two systems are interconvertible. (The Ptolemaic and pure Copernican systems can be and have been falsified, on the other hand: no amount of superimposed circles amount to an ellipse.) But Lameen’s request was for a theory within the generative paradigm that had been actually refuted by evidence not merely discarded as inelegant.

  113. John–

    Yes, I was speaking inexactly, sorry. The Lebeaux/Fox facts are baldly incompatible with the D-structure/S-structure approach, which is why that approach was discarded.

  114. John Cowan says

    Okay, excellent. “How absolute the knave is! We must speak by the card, or equivocation will undo us.” —Hamlet.

  115. Norvin: Could you give a citation to where Chomsky discusses Lebeaux/Fox and the abandonment of d-structure?

  116. At any rate, if Chomskyans once wrote mainly about English, we don’t do that any more. Over the past ten years, for example, MIT graduate students have written dissertations (http://web.mit.edu/linguistics/graduate/dissertation/list.html) about the syntax of Chol, Kiowa, Korean, Mandarin Chinese, Mebengokre, Slovenian, Tagalog, Tlingit, Tseltal, Uyghur, Zulu…okay, and English. Did you really expect us to stop writing about English? There’s plenty left to discover. But ‘English syntax’ dissertations are now the exception, rather than the rule.

    I am very glad to hear that, and in general I would like to warmly thank you for venturing into this den of anti-Chomskyism and being so informative and reasonable. Despite my prejudices, I am always glad when a member of the other team shows up and chats; I don’t like existing in an echo chamber!

  117. I don’t like existing in an echo chamber!

    I’m glad you feel that way! I always feel sort of guilty about trying to interfere with the echoes.

  118. No, no! If I ever get to where I only want to hear from the like-minded, just put me out on an ice floe.

  119. Lameen, the elimination of D-structure and S-structure is one of the big themes of Chomsky’s paper “A Minimalist Program for Linguistic Theory,” generally regarded as the starting point for Minimalism, which appeared in two books:

    Chomsky, Noam. 1995. The Minimalist Program. MIT Press, Cambridge, Mass.

    Hale, Kenneth, and Samuel Jay Keyser (eds.) 1993. The view from building 20: essays in linguistics in honor of Sylvain Bromberger. MIT Press, Cambridge, Mass.

    In the 1995 version, which is the one I have handy, the discussion of Lebeaux starts on p. 204. Looking at it again, I see that he doesn’t cite Fox there, which makes sense–should’ve remembered that Fox’ contribution was later. He does also cite Freidin, van Riemsdijk, and Williams for related observations.

  120. J. W. Brewer says

    Part of the difficulty of this dialogue is the long time scale involved. Norvin was probably in nursery school when hat was having his unpleasant encounter with earlier versions of Chomskyism in grad school in the ’70’s, and had probably made it to 9th grade when I had my own encounter with a no doubt somewhat different version in a Syntax class in 1985, in which we were taught that, at least according to the dominant school of thought, “deep structure” (the transition to the more euphemistic “D-structure” may have been underway but was not complete) was an actual thing – as real as phlogiston had once been thought to be. The author of one of the texts we used (Andrew Radford) apparently subsequently did a text on Minimalism, so I assume he shifted with the times; the author of the other (C.L. Baker) is said by the internet to have died in ’97, so I don’t know if he made the transition. So Norvin is valiantly trying to do his this-is-not-your-father’s-Oldsmobile pitch when (to shift makes of cars mid-metaphor) it seems that everyone else wants to rehash old stories about exploding Pintos and crappy Edsels and old man Ford’s distribution of anti-Semitic propaganda. Although on the other hand it sounds like Lameen and perhaps Etienne have independent knowledge of more recent models from the same manufacturer and believe they have grounds for considering them defective in their own right.

    Two other points: 1) there is a distinction to be made between the organized Chomsky cult/movement with its controversial academic-political behavior and the Chomskyan or “generative” tradition or research program more broadly, with some notable individuals (e.g. the late James McCawley, as discussed in an interesting piece someone linked to a few dozen comments back) having broken with, or been expelled by, the former, without necessarily having left the latter. 2) Y’s assertion that this is all ancient history and the cult/movement has lost its hegemony is heartening if true. It strikes me that a good way to check that (or more generally the current impact of the alleged “imperialism” of the movement) would be to look at the lengthy program for the big annual LSA confab just concluded. (I mean, of course, for someone who is not me but has both more time and more knowledge to look at that program . . .) What proportion of the panels were on the sort of issues and subfields the Chomskyans (allegedly) thought were beneath their dignity and should be left to amateur lepidopterists? (Although I suppose it’s possible that the Empire realized that those subjects weren’t going away because people were interested in them and thus launched a belated program of colonization/subjugation?) What proportion of the papers for panels on Chomskyan-approved Serious Subfields seem (from the titles, abstracts, etc.) to be coming from other-than-Chomskyan perspectives?

    2b) If in fact there has been a decline in such hegemony over the last few decades, it would be fascinating or at least amusing to try to work out a timeline of that decline and graph it against a timeline of the shift (increase?) in Chomsky’s general celebrity in non-specialist circles as a polemicist on non-linguistic topics to see if there is, in fact, an inverse correlation.

  121. John Cowan says

    I will also add that what I have said about Chomskyism is not intended as a reflection on Chomsky himself, who was nothing but kind to me when I met him. Freud supposedly asked God to save him from the Freudians.

  122. John Cowan says

    I’m glad to hear of these acceptability tests of sample sentences, though I expect that most of the participants were WEIRDos, which somewhat undercuts the results.

  123. WEIRDo sampling: now that’s a concept I’ve needed for a long time, to describe a technique used to obtain the “experimental data” reported in several pop-psychology books and articles I have read.

    The last such was Daniel Kahneman’s Thinking, Fast and Slow. He writes about certain non-“rational actor” behavior displayed by people who assess probabilities and make bets. Much of the “data” consists of responses by groups of university students and professors to questions like “if you are offered a choice between a 40% chance of winning $40, and a 70% chance of losing $70 – which choice would you make ?”

    As the book progresses, the questions get ever more complicated and tricky. I stopped trying to think them through, and instead started to think about what kind of people have the time, interest and even conceptual background to deal with such questions. The answer, as I would now put it, is: WEIRDos.

    Don’t get me wrong: the book contains lots of interesting ideas. The problem is that Kahneman manages to give the impression that his experimental results describe Human Behavior In General.

    Around the same time as I read that book I was reading Simmel’s Philosophie des Geldes. There, on three or four pages, Simmel briefly sets out the same ideas that Kahneman inflates into a book (in which there is no index entry for “Simmel” and, as far as I can remember, no reference to Simmel).

  124. marie-lucie says

    Norvin, I don’t doubt that many linguists on the Chomskyan tradition are writing on the syntax of languages from A to Z rather than just about English. My point was/is that in describing such languages within the Chomskyan-derived framework they will tend to impose that framework (whatever it is at the time) on those languages, leaving aside features which do not fit or are irrelevant at the particular stage of the theory. Since the framework has changed so much in the last 50 years or so since it became widespread, the view of the structure and character of those languages has also changed, so that previous works are now obsolete. It is one thing when theoretical articles or textbooks become obsolete, but work on languages which are not otherwise well-known often do not contain much in the way of usable data for a person trying to learn them rather then test a theoretical point.

  125. An excellent point. With those boring old butterfly-collecting language descriptions, I could read a short account of a language and get a decent idea of how it worked. With Chomskyan stuff, all you tend to get is material chosen to illustrate some bit of theory, which is of interest only to the few people who care about that bit of theory.

  126. Theoretical questions – from the Chomskyan tradition or otherwise – can play a useful role in prompting grammarians to describe phenomena that traditional grammars rarely covered. But, even when working on previously unstudied languages, most theorists prefer to focus their energy on deep coverage of some small part of the grammar (which maximises the chance of advancing theory), rather than a shallower description of the whole grammar (which maximises the work’s general interest). Good general descriptive grammars are certainly still being written, but their production is motivated more by typology than by the Chomskyan research program. There was a slew of language description theses written as transformational grammars in the early 70s or so (you practically have to be a historian of science to read them now), but after that the theorists’ abandonment of transformational grammar taught most language documenters to steer clear of rapidly changing frameworks.

    Norwin: Thanks for the reference – I had skimmed those works, but I didn’t pay enough attention to the role of this particular argument. As for the sentence acceptability judgments, I’m very glad to see that someone’s testing this, but I’not at all surprised that they get 95-98% on all example sentences put together. In the textbook I used (Carnie), as in the Chomsky book I discussed, the English problem cases were practically restricted to (and very frequent among) illustrations of subjacency, which account for very few of the examples in either of the papers you linked.

  127. Etienne: the Newfoundland tag (or tag-like) structure you mention (‘That boy’s doin’ too much work he is’) is familiar to me from my childhood in Waterford, which I think was the source of most Irish migration to Newfoundland in the first half of the 19th century. But I don’t think I’ve heard anything like your second example (?’That boy’s doin’ too much work he was’). (And a Waterfordian would probably say ‘lad’ rather than ‘boy’ in that context, I think).

    More widespread in Ireland is a variant with ‘so’ – ‘That lad’s doing too much work, so he is’.

  128. This seems to be the right place to mention that Pullum has an Erdos-Bacon-Sabbath number of 10.

  129. J.W. Brewer says

    I would make it 11, on account of disagreeing with their Sabbath-number calculation methodology.

  130. January First-of-May says

    Somewhat more on-topic, Chomsky’s EBS number appears to be 11, but it’s a much shakier Sabbath path (someone should find a proper non-sampling connection, presumably through MIT Gangnam Style).

    I see that on his home page he tells us that his Erdös number is 5. There must be few academics in the world with Erdös numbers higher than that.

    My mom’s Erdős number is 5, and she has only one publication (that I could find) – an article from the late 1990s coauthored with a bunch of people in her biology lab (including my dad, who also appears to have Erdős number 5 – I suspect it’s actually 4, but I wasn’t able to find a path).

    But I agree, as far as modern scientists are concerned, beyond 2 (and definitely beyond 3) it has more to do with luck than with just about anything else (though, yes, 5 is fairly typical – especially outside mathematics – and higher numbers aren’t that rare, though often just because improvements are hard to research; there’s probably a mathematician out there who has an Erdős number of 6 or 7 through mathematics only and 4 or 5 through an article they coauthored with a biologist).

    Bacon numbers are easier – if you have a finite one that’s higher than 3, either you didn’t research it properly, or you’re a real obscurity.
    Sabbath numbers are tricky because there are so many obscure connections, some of which might well be unverifiable (but even then, if you have more than 7 steps – or more than 3-4 steps from a major musician – you’re probably doing it wrong).

  131. J.W. Brewer says

    There are a certain number of hard-rock musicians who have also worked as or at least been trained as research scientists, so I was not surprised to see some of them on the list of individuals with interestingly-low EBS numbers, but I was struck that Jackie Fox a/k/a Jacqueline Fuchs (probably best known non-musically for having been a classmate of Barack Obama at Harvard Law School) got her Erdos number started via her post-rock-career undergraduate work in linguistics, namely a 1988 piece titled “A Restructuring Analysis of English “To”-Contraction and Romance Clitic Placement.”

  132. J.W. Brewer says

    Although Ms. Fuchs’ chain brings up a difficulty w/ what counts as “collaboration.” Her piece was supposed to be her undergraduate honors thesis, and thus supposedly her original work w/o co-authors, so treating her faculty adviser as a collaborator seems a little shaky. Similarly the next link in the chain from her to Erdos is her adviser’s doctoral dissertation (again supposed to be his own original work), with the chair of the dissertation commitee (Noam Chomsky, as it happens) treated as a collaborator. Does that work both ways? If I were perchance to have a lower Bacon or Sabbath number than the faculty adviser for my own (unmemorable and unimpressive!) undergraduate senior essay, could he trace his own path to a lower EBS number through me?

  133. Yeah, those are B.S. There is no wiggle room in the calculation of Erdős numbers. They need to involve published scientific papers listing both individuals as authors. My own Erdős number is 4 (which is unremarkable in theoretical physics), but if hazier publications (e.g. course materials) were allowed, it would drop to 2.

  134. January First-of-May says

    There is actually surprisingly much wiggle room in the calculation of Erdős numbers. For starters, do we include non-mathematical papers? (Pretty sure they’re often excluded, though IIRC that’s not what the Tom Odda article says.) Can we exclude hyper-large collaborations with hundreds or even thousands of official coauthors? (They’re pretty common in high-end physics and biology.) What about coauthored books? (Usually not allowed.) And so on.

    There’s even more wiggle room once we get to Erdős-Bacon numbers, and especially Erdős-Bacon-Sabbath numbers, since we expect the set of people who qualify at all (with verifiable paths) to be fairly limited, and it is thus still interesting to explore those who qualify if we relax the rules a little.

    Ignore Thomas Edison though; he barely qualifies on the Bacon number, doesn’t really qualify on the Sabbath number, and his Erdős number is a complete cop-out, but the list creators really wanted to include him in some way.
    (Though I think I did have some slightly more promising leads for Herbert G. Wells…)

  135. Any scientific paper is allowed. The authors are exactly those that are listed, regardless of whether somebody thinks the ATLAS author list is too long. Textbooks are not allowed, but reviewed scholarly monographs are. These are pretty well known rules in the mathematical sciences, at least in America.

  136. Large collaborations has to be discounted. We can built a graph with edges between co-authors of the papers (multiple edges between the same nodes are allowed). Each edge will have resistance N(N-1)/2 Ohm (N is number of co-authors) and define TRUE ERDOS NUMBER as equivalent resistance between a person and Erdos. It will make some numbers very small (for someone with multiple collaborations with Erdos himself or for someone very well connected).

  137. J.W. Brewer says

    One can use the EBS examples to determine that Bacon’s own Sabbath Number is no more than 4 (via several paths, some of which sidestep my methodological quibble I alluded to above), and therefore anyone with a known Bacon Number has a Sabbath Number that can be no greater than their Bacon Number +4. I guess the question is whether there’s a shorter path from Bacon to Sabbath, perhaps via an intermediate link who lacks a known finite Erdos number? (Obviously this may not work if you insist that Bacon-Numbering and Sabbath-Numbering require different and mutually exclusive modes of collaboration . . . .)

  138. There’s really no point in calculating Bacon-Sabbath numbers if the Bacon and Sabbath distances are computed using the same graph.

  139. January First-of-May says

    There’s really no point in calculating Bacon-Sabbath numbers if the Bacon and Sabbath distances are computed using the same graph.

    This. (Though I think it’s more that the Bacon graph is viewed as a subset of the Sabbath graph – but that isn’t quite true either.)

    The definition of Sabbath numbers involves the connections having something to do with music, and IIRC (non-musical) films don’t necessarily count, not even as soundtracks (musical films might, though I don’t think it ever came down to that).

    I wonder, incidentally what exactly the methodological quibble was (the only one I could think of is that perhaps short-time Black Sabbath members from the nineties shouldn’t have counted for zero).

  140. But the edges in the Bacon graph are movie collaborations, and in the Sabbath graph it’s music. There may be collaborative events that are represented in both graphs, but the S = B + 4 equation is not valid in general. EDIT: Second on the ball here, forgot to refresh before typing.

    Now a fun piece of trivia would be the existence of an event represented by edges in all three graphs.

  141. I don’t see how. A published paper can hardly be either a movie or an album.

  142. In the 21st century, I’ll bet it could be both.

  143. J.W. Brewer says

    Yeah, the methodology quibble is that guys like Cozy Powell who did a spell of duty as a hired hand in a well-past-golden-age version of a band marketed as Sabbath shouldn’t count – they should personally have Sabbath Numbers of one (because they did play with at a minimum Tony and in many cases Tony-plus-Geezer), not zero. So if your closest linkage is to Cozy (or Bev Bevan or Glenn Hughes or some random journeyman who, unlike those guys, lacks independent name recognition) your Sabbath Number should be number of links to Cozy plus one. Whether *anyone* other than the original four should count is arguable, but beyond those four there are only, I’d say, one to three plausible candidates out of the much longer list on wikipedia of the various guys who passed through various late-in-the-day line-ups.

  144. David Marjanović says

    In the 21st century, I’ll bet it could be both.

    Not really, but there is such a thing as a supplementary movie. Several papers I’ve read, and even one I’ve reviewed, have one or several.

    (Usually they show renderings of computed-tomography data spinning around.)

  145. You’re denying that it is both, but I’m saying it could be both. Use your imagination!

  146. Bob Dylan accepted his Nobel Prize yesterday in a closed meeting of the Swedish Academy, and several Academy members tagged along to the concert. Would that qualify as a shared event between E and S graphs at least?

    (Dylan may well have acquired an Erdős number, but other performers at the concert probably shouldn’t. Question is whether the Academy members get Sabbath numbers. It seems Dylan didn’t acknowledge their presence during the concert — if he’d called them on stage there might be a good case).

  147. @Lars:. Certainly not.

  148. John Cowan says

    A thought: What if we unified the graphs such that a link by movie, by paper, or by live co-participation counted as a link, thus determining a single EBS number as opposed to the current E+B+S approach? Numbers would probably drop quite a bit.

  149. David Marjanović says

    On January 4, 2014…

    Ah yes, butterfly collecting. A good satirist could probably imagine what entomology would be like as a field today if over the course of its history it had come to be dominated by some Chomsky-like guru figure: I can see it already…generations of scholars furiously debating the properties of the “Universal insect”

    Richard Owen’s archetypes. A very short-lived concept, related to the fact that Owen was the last creationist in comparative anatomy (…not that modern fundies would recognize him as a creationist).

  150. David Marjanović says

    Embedded in this comment there’s a delightful rant about Chomsky and especially some of his followers. I have no idea how some of the insults passed peer review, but still it was published in a journal in 1996.

  151. David Eddyshaw says

    “It goes without saying that Chomsky is a brilliant syntactician. If his syntax had a more straightforward connection with linguistic data, it would be a monumental achievement.”

  152. David Eddyshaw says

    Esa Itkonen’s criticisms are interesting as coming predominantly from the side of the philosophy of linguistics, which is his particular thing: he focuses on the questionable assumptions at the back of the whole edifice, e.g.

    Within the generative paradigm it was for a long time held to be a “truism that language is an infinite set of sentences” (Chomsky 1969: 57). Far from being a truism, however, the decision to view language in this way had far-reaching consequences concerning the overall conception of linguistics.

    He also prays Wittgenstein in aid of his criticism (pretty lucidly, I thought), which pleases me …

    The rant reminds me a bit of MAK Halliday’s response (in An Introduction to Functional Grammar) to a conference speaker’s comment “I take it for granted that the goal of linguistics is to characterize the difference between the human brain and that of an animal.”

    Halliday remarks: “That this should be one of a hundred goals one might readily accept; but that this – or anything else – should be ‘the’ goal of linguistics is hard to take seriously.” (He goes on to list a couple of dozen of important goals without breaking into a metaphorical* sweat.)

    * This is Halliday.

  153. a conference speaker’s comment “I take it for granted that the goal of linguistics is to characterize the difference between the human brain and that of an animal.”

    I would have neighed loudly and walked out.

  154. But then I took great pleasure in disconcerting the young and well-meaning guy who was trying to inculcate us with Chomskyanism in grad school by denying the acceptability or unacceptability, as the case may be, of his example sentences. Hey, I’m a native speaker too, buddy.

  155. I found one section particularly apt.

    “Let us consider the X-bar syntax. It is based on the assumption that all lexical categories share the same structure, namely specifier, adjunct, head, and complement. The head is obligatory, and whether or not it has a complement, is stipulated in the lexicon. The specifier and the adjunct are optional ….”

    “……”

    “The foregoing account was simplified insofar as theoretical analogy had already been at work in the definition of S: because the subordinate clauses are generally preceded by a complement (in English that, for instance), and because it is neat to assume that the subordinate clauses and the main clauses share the same structure, it seems ‘natural’ to make the generalization that the main clauses too are preceded by a complement…. Never mind that it is invisible (except in questions, in some languages). Notice how difficult it is to deny that something invisible exists.”

    “Such breath-taking reinterpretations and generalizations can only be performed by people who are obsessed with theoretical analogy. In view of this, it is rather ironical that of all people, the representatives of the generative paradigm should deny the existence of analogy (cf. Subsection 3.2). More importantly, there is no obvious way (and perhaps, no way) that the results of their theoretical analogy, although formalized with extreme precision, could be amenable to falsification.”

    I find it peculiar that generativists think it “makes sense” to posit (under X-bar) that all lexical categories share the same structure, “namely specifier, adjunct, head, and complement”. Well, yes, it makes sense, it’s “neat”, but is it real? Itkonen’s criticism strikes home. It’s just a theoretical analogy.

    His observation that it is difficult to deny that something invisible exists also rings true. Generativists are good at building whole theoretical structures on the basis of certain kinds of syntactic behaviour (‘islands’ are a favourite), often resulting in the positing of invisible elements, e.g., ‘trace’. But I find it hard to take seriously formalisms which seem to claim that the human mind raises a subject to a higher position in the tree while leaving a trace at the original position in order to preserve its theta role. It’s a wonderful and elaborate theory, but this depiction of what supposedly happens when speakers frame sentences simply doesn’t make sense to me.

    Perhaps the fault is mine since I find it both extremely difficult and most incongenial to envisage syntactic structures in terms of mathematical formalisms. All of the ‘advances’ in Chomsky’s theories have involved ‘refining’ the mathematical formalisms in order to make them ‘computationally simpler’. I simply can’t follow that.

  156. One of Chomsky’s great achievements is supposedly that he freed linguistics from the straitjacket of logical positivism and behaviourism, which rejected mentalism, psychologism, and any role for meaning in the analysis of language, and allowed linguists to pursue the holy grail of ‘universal grammar’. There has been a big price to pay for that ‘liberation’, the most conspicuous example being Chomsky’s foray into biology with his speculation that a mutation in the human gene was responsible for human linguistic ability. Really, I’m pretty sure people in centuries to come will wonder how such crackpot theories were ever taken seriously by their ‘scientific ancestors’.

  157. I’m ahead of the curve — I’ve been wondering that for years!

  158. John Cowan says

    Chomsky quoted by Itkonen:

    John wondered what stories about us we had heard
    *John wondered what stories about us we had told

    Itkonen says that the *, if it means ‘ungrammatical’, has no empirical justification for being there, as Chomsky himself also says. But to me it is the absence of * in the first sentence that is wrong: these sentences are unacceptable with us, but fine with ourselves.

    Itkonen for himself:

    But this is a contradiction because it is not reasonable to deny that the rules of chess determine how one ought to play.

    I think that’s an equivocation on ought, which has no deontic force here: it’s simply that if you are not playing by the rules, you are not playing chess, but something else, perhaps basketball. Certainly among those who are generally agreed to be playing chess (a social determination), their behavior could be used to infer the rules.

    Wittgenstein argues that knowledge is primarily intersubjective, regardless of whether it is knowledge about observable events or knowledge about rules (or norms) accessible to intuition. Why? – because in the domain of genuinely subjective knowledge, expressed by my private language, I cannot know whether or not I am mistaken.

    Those of us who actually deal in private languages know that this is not true. If I decide (in accordance with tradition) to express the accusative case in my private language P by a prefix, then if I am reflecting how to say “John went to the store” in P and I leave the prefix off the word meaning ‘store’, then I have made a mistake, and I can certainly know it. If I write down the case paradigms of P, then the knowability of this mistake is even more obvious, even if I show my writings to nobody. Of course I could decide that the fact that I have left off the accusative prefix means that my previous understanding of P was wrong, and then I can change P accordingly so that the mistake is no longer a mistake. But this is a distinctly different kind of move.

  159. Stu Clayton says

    Those of us who actually deal in private languages know that this is not true.

    Privileged access and intra-subjectivity for the win ! Of course I am merely free-associating here. You may not even understand what I’m saying, but I know that it’s true.

    I agree with you 100%, though you may never know it. What more can we wish for in this vale of crocodile tears ?

Speak Your Mind

*