Bendy Bedrock, Part 2


The post on Scientific American’s article on Chomsky has prompted some suggestions for further reading. Here’s a summary.


Kevin Gregg recommends Evans, N. & Levinson, S.C. (2009) The myth of language universals: language diversity and its importance for cognitive science. Behavioral & Brain Sciences 32:429-492.

The main article makes an argument that I don’t think stacks up (more importantly, neither does Gregg) but it’s well presented and it’s followed by “Open Peer Commentary”, where a very wide selection of scholars , including Baker, Bevett, Christiansen and Chater, Croft, Adele Golberg, Habour, Nevins, and Pinker & Jakendoff respond.  Very highly recommended.


Scott Thornbury confesses that his enthusiasm for Everett’s awful book has dwindled. (See here for the post on his A to Z blog where he cites Everett. A lively discussion followed.)   He now recommends Christiansen and Chater (2016) Creating Language: Integrating Evolution, Acquisition and Processing. MIT. A review on the  website   says:

Because children learn language quickly and easily, many theorists have believed this means there are specialized brain mechanisms specific to language acquisition. This has led them to ask how the brain has changed to accommodate language. Christiansen and Chater flip the question around, asking, “Why is language so well suited to being learned by the brain?” Taking a cultural evolution approach, they conclude language is easy for us to learn and use because language, like a living organism, has evolved in a symbiotic relationship with humans. Language has adapted to what our brains can do, rather than the other way around.

“We view language as piggy-backing on older pre-linguistic abilities,” Christiansen says. “Results from my lab indicate that there’s likely to be some biological differences in how people are able to process sequences of information and ‘chunk’ that information together into larger units. These differences interact with variation in linguistic experience and give rise to individual differences in language processing. The importance of experience is further underscored by the many studies showing that there’s a strong correlation between the number and variety of words that children hear and their language abilities. It can make a huge difference.”


Phil Chappel points us to a web site called The Conversation where a mum with a Ph.D. begs to differ with Chomsky’s UG theory, based partly on her experiences with her baby. She cites Vivian Evan’s book The Lnguage Myth (which is almost as bad as Everett’s), goes on to cite  “the growing body of research on infant and mother communication” and then claims that “babies need joyful, responsive human company”.  The article illustrates the danger of not appreciating  the strictly-defined domain of Chomsky’s theory, namely language competence. That babies need human company is a fine example of a motherhood statement, but it has nothing to do with the POS argument or UG theory.

More seriousy, Phil Chappell recommends Lee, N., Mikesell, L., Joaquin, A. D. L., Mates, A. W., & Schumann, J. H. (2009). The interactional instinct: The evolution and acquisition of language. Oxford University Press.  You can download a pdf file of Lee & Schmann’s presentation of the book at this web site. They say:

In this presentation, we outline a perspective on language acquisition based on evolutionary biology and neurobiology. We argue that language is a cultural artifact that emerges as a complex adaptive system from the verbal interaction among humans. We see the ubiquity of language acquisition among children generation after generation as the product of an interactional instinct that, as Tomasello indicates, is based on an innate drive communicate with and become like conspecifics.

The “cultural artifact” line is common to thise working on emergentist models. Note the reference to Tomasello who figures in so much of the lit. these days, including the book Scott mentions.


Ms. Westerlund invites “those with an open mind and more importantly, critical mind” (sic) to read an article by Tom Wolfe about Darwin and Chomsky (which is partly aimed at promoting his new book on Chomsky) in Harper’s magazine.

I don’t know whether Ms. Westerlund is implying that only those who possess an open critical mind will apreciate Wolfe’s argument, or that only they will recognise it for the blustering tosh that it is. I personallly think that just about anybody will quickly conclude that Wolfe has no idea what he’s talking about. What’s intersting is that Wolfe relies on the aforementioned Daniel Everett, he of the Pirahã study, to argue the case against Chomsky for him. Wolfe likes Everett becuase he’s a real macho man, a man who poses for the front jacket of his book up to his neck in water in a dangerous river while a Pirahã fisherman sits safely in his boat. Unlike the nasty, arrogant, desk-bound cissy Noam Chomsky, Daniel is a proper linguist who thinks fieldwork is essential, and “winds up in primitive places, emerging from the tall grass zipping his pants up”.


But, according to Wolfe, Everett is not just a proper macho man, he’s a brilliant academic. Everett’s fieldwork, Wolfe assures us, revealed that the Pirahã language lacked recursion, thus refuting Chomsky’s claims for the universality of his UG. Of course, Everett’s work did no such thing. Lots of linguists have since replied to Everett’s claim, pointing out that the Pirahã language indeed had recursion (e.g., “I want the same hammock you just showed me”) and that “recursion” is not that central to Chomsky’s theory anyway.


Robert Taylor offers “some interesting research about sounds for common ideas being the same across languages (roughly 2/3rds)” The link takes you to an article about the study, not the study itself, which is published in the journal Proceedings of the National Academy of Sciences, and is written by Morten H. Christiansen – Scott’s man turns up yet again!  Christiansen and colleagues argue that basic concepts, like body parts and familial relationships, and the sounds that people around the world turn to in describing them have a robust statistical relationship.


Finally, let me recommend an article from one of favorite web sites, the Stanford Encyclopedia of philosophy. The article, Innateness and Language gives an interesting discussion of some of the arguments for and against Chomsky’s theory,

28 thoughts on “Bendy Bedrock, Part 2

    1. For some reason your posting that I wanted to reply to doesn’t have a ‘reply’ thingy to click on, but this should do. You cite the Reali & Chater paper in Cognitive Science in reply to my question whether Bates has offered a concrete example of a linguistic phenomenon that she could account for. And I’d say that the R & C paper is such an example–although not by Bates!–although it’s a simulation with a connectionist model, not an investigation into actual human subjects. This is the sort of thing one wants to give some content to the emergentist position(s).


      1. Hi Kevin,

        In this post (Bendy 2) Mura cites the article Linguistic explanation and domain specialization: a case study in bound variable anaphora by David Adger and Peter Svenonius.. In the previous post (Shifting sands…) he cites the R & C paper in reply to your challenge to me to say what Bates had ever proposed …


  1. My enthusiasm for Everett may have dwindled, as you say (but only because others state his case so much better), but I still stand by the quote in my blog, i.e. ‘No one has proven that the poverty of the stimulus argument, or Plato’s Problem, is wrong. But nor has anyone shown that it is correct either. The task is daunting if anyone ever takes it up. One would have to show that language cannot be learned from available data. No one has done this. But until someone does, talk of a universal grammar or language instinct is no more than speculation.’


    1. You mention Plato. Plato liked to use examples from geometry. As a matter of interest, do you accept Plato’s argument, made most clearly in The Meno, that experience alone is insufficient to explain our knowledge of geometry; our knowledge, for example, that the internal angles of every and any triangle necessarily amount to 180 degrees (knowledge that certainly cannot be accounted for by appeal to experience alone since obviously no one can have experienced any and every possible triangle)? Or do you consider this unproven?


    2. A couple of quick comments:
      1) Elfnotes provides a lovely example of just how real linguists do real research on real phenomena of language, thus advancing the investigation of what the language faculty consists of, and by doing so illustrating the poverty of the stimulus.
      2) Scott’s quote — I can’t tell if he’s quoting himself or Everett, but no matter– talks about ‘proving’ the POS argument. But we’re not talking about logic or math, we’re talking about an empirical science, which deals in evidence not proofs. What scientists, linguists included, do is infer to the best explanation. The linguistic literature is filled with examples of inference to the best explanation, where the best explanation is that the human language facility includes unlearned–innate, if you will– language-specific principles or rules or capacities; knowledge, in short. Those who would deny the existence of such unlearned knowledge have an obligation to show that there is a better explanation. By and large, they have failed to meet this obligation. It’s not clear to me that e.g. Tomasello or Bates have ever really tried to.
      3) Christiansen and Chater, though, are for real, and need to be taken seriously, as does, say, O’Grady. I just got C&C’s new book, and I hope I can find the time soon to read it.


      1. Kevin, the words are Everett’s, not mine. There is another explanation, which may not be ‘better’ but requires fewer assumptions about the nature of cognition and of learning, hence is consistent with the law of parsimony. Digo yo.


    3. Hi Scott,

      This is a real disappointment. . I’m honestly surprised that you continue to say that until somebody proves that the POS is “correct”, talk of a universal grammar is no more than speculation. It’s impossible to prove that ANY general hypothesis or theory is correct, so to that extent, all of science is mere speculation, as Hume pointed out some time ago. The way scientific theory construction works is through conjecture and refutation and through inference to the best explanation.You need to get a grasp of these methods. I recommend Popper’s “Logic of Scientific Discovery” and Peter Limpton’s “Inference to the best explanation”,

      As for your remark that the only reason that your enthusiasm for Everett has dwindled is because others state his case so much better, this is also surprising. You think Everett’s case is OK, but Christiansen and Chater’s is better, is that it?. The problem is that they’re not 2 versions of the same argument! It seems from your writings over the past few years that you’re “sold” on some version of emergentism, and you’re shopping around for the best version you can find. You get all excited by emergent grammars (“grammar for free”you used to say before you cooled on that one), you endorse all the incoherent nonsense that Larsen-Freeman and Cameron propose, you get all excited about Everett before coming across Christiansen and Chater. Evaluating texts by Everett, Larsen Freeman & Cameron, Christiansen and Chater, MacWhinney, O’Grady, Lee, Mikesell, et. al. , V. Evans, and others as if they were all good emergentists, and all you have to do is find the one that states the case best, is a strange way to go about critical reading.


      1. ” You think Everett’s case is OK, but Christiansen and Chater’s is better, is that it?. The problem is that they’re not 2 versions of the same argument!”. No, they’re not – and I didn’t actually say that. One refutes the idea of UG on the grounds that at least one language exists that does not have recursion (and, admittedly, this has been challenged by his critics), The other offers a ‘non-magical’ explanation of recursion (and other linguistic features that PoS proponents argue could not possibly be acquired). I find these arguments attractive because they explain language learning in terms of general cognitive faculties without the need to invent a language-specific one (the location of which has eluded even quite sophisticated fMRI and PET scanning, btw). (In the same way, I buy into a Darwinian theory of evolution (even though I am unqualified to asses the evidence) because it doesn’t require me to conjure up a Divine Architect.) Moreover, usage-based theories handle ‘rogue’ phenemona (of the type ‘Her an actress?!’) that Chomskyists consign to the ‘not applicable’ category, which seems to me to be cheating (in the same way that they argue that speakers of Piraha ARE hard-wired for recursion, but they have ‘opted out’ of using it). Finally, I like the way the usage guys use empirical data (e.g. animal behaviour, in the case of Tomasello, or corpus data in the case of Bybee) rather the made up sentences of the type John donated Tom a photo of himself.
        But we’ll never agree on this, so I will just wait until you’ve read Christiansen and Chater and look forward to your incisive, albeit possibly enraged, demolition job! 🙂


      2. Hi again, Scott,

        Please don’t say we’ll never agree on this; I think we might manage to agree about a lot of it, actually. Anyway, for the time being, let’s continue to examine the arguments and the evidence.

        1. I notice that you don’t reply to comments about your claim that those who propose explanatory theories must prove them if they’re to be more than mere speculation. We can lay that one to rest then, I hope.
        2. Everett does not refute the idea of UG on the grounds that at least one language exists that does not have recursion.
        3. UG’s explanation is not magical; it’s an example of an inference to the best explanation.
        4. Attempts to explain language learning in terms of general cognitive faculties without the need to invent a language-specific one have so far failed.
        5. A Darwinian theory of evolution is consistent with UG theory.
        6. Usage-based theories handle what Chomsky’s theory sees as performance data because they work in different domains, often including aspects of sociolinguistics. Consigning ‘rogue’ phenomena (of the type ‘Her an actress?!’) to the ‘not applicable’ category, is not cheating, it’s simply saying that the data is not relevant to Chomsky’s theory of linguistic competence. While I think your examples are bad ones, nevertheless, as I’ve said, I think some moving of the goal posts, and “that doesn’t count” maneuvers have been used by Chomsky.
        7. You say you “like the way the usage guys use empirical data (e.g. animal behaviour, in the case of Tomasello, or corpus data in the case of Bybee) rather the made up sentences of the type John donated Tom a photo of himself”. But what do they use it for? What follows from their use of this empirical data?

        I’ve ordered Christiansen and Chater’s new book. Judging from their previous work and from reviews of this one, I doubt very much that, once I’ve read it, I’ll manage any incisive demolition job. And, in the event that I find it persuasive, I’ll say so – cheerfully. :-)

        P.S. Here’s the start of an article by Jeffrey Watumull, Marc D. Hauser, Ian G. Roberts and Norbert Hornstein called “On Recursion”. It’s from the Frontiers in Psychology journal, Jan. 2014.

        It is a truism that conceptual understanding of a hypothesis is required for its empirical investigation. However, the concept of recursion as articulated in the context of linguistic analysis has been perennially confused. Nowhere has this been more evident than in attempts to critique and extend Hauser et al’s. (2002) articulation. These authors put forward the hypothesis that what is uniquely human and unique to the faculty of language—the faculty of language in the narrow sense (FLN)—is a recursive system that generates and maps syntactic objects to conceptual-intentional and sensory-motor systems. This thesis was based on the standard mathematical definition of recursion as understood by Gödel and Turing, and yet has commonly been interpreted in other ways, most notably and incorrectly as a thesis about the capacity for syntactic embedding. As we explain, the recursiveness of a function is defined independent of such output, whether infinite or finite, embedded or unembedded—existent or non-existent. And to the extent that embedding is a sufficient, though not necessary, diagnostic of recursion, it has not been established that the apparent restriction on embedding in some languages is of any theoretical import. Misunderstanding of these facts has generated research that is often irrelevant to the FLN thesis as well as to other theories of language competence that focus on its generative power of expression. This essay is an attempt to bring conceptual clarity to such discussions as well as to future empirical investigations by explaining three criterial properties of recursion: computability (i.e., rules in intension rather than lists in extension); definition by induction (i.e., rules strongly generative of structure); and mathematical induction (i.e., rules for the principled—and potentially unbounded—expansion of strongly generated structure). By these necessary and sufficient criteria, the grammars of all natural languages are recursive.


      3. Human cognitive faculties, it is surely to be granted, are both marvellous and little understood. One thing we do know, though, is that (amazingly, and in sharp contrast to the cognitive faculties of any other creature anywhere that we know of) those faculties faciltate the capacity to use language. Chomsky, I think, does not countenance the idea of a divine designer, who might have purposely gifted us with a mental module specifically designed for language use. Rather, his invocation of a language acquisition device, I think, should be understood as a proposal for a research programme. Chomsky’s claim, it seems to me, is only the claim that the fact that our cognitive capacities are so well suited to language use is sufficiently remarkable as to warrant specific study. Is there seriously anything objectionable about this claim? The waters are muddied, of course, by the fact that we have no remotely satisfactory theory of mind, or of meaning.


  2. From the paper that elfnotes cites:

    principles to be
    evaluated for domain specificity should be principles that
    actually do explanatory work in capturing linguistic phenomena.
    That is, we need to understand the nature of the linguistic
    phenomena first, and use that understanding to ask more general
    questions of cognitive science. Any alternative approach that
    ignores or dismisses a vast range of empirically impeccable
    work, and attempts to show that some proposed principle of
    communication or learning explains something general about
    language is insufficient. Any such alternative needs to have, or
    at very least be in principle capable of extending to, the kind of
    empirical coverage and explanatory depth of current generative
    linguistic theory.


  3. I’ve just read the lively discussion at Scott’s blog. It’s…intense.

    Anyway, with regard to PoS, I know there was talk of some studies that indicated toddlers have already begun to develop grammaticality before they’re able to express it, to the extent that they can distinguish syntactical patterns. Aside from that, the only case of children not developing functional, complete language systems is in cases where the child lacked the capacity for communication, per se, or was a victim of severe abuse.

    Going by the Falsification principle, how would opponents of the PoS reconcile these issues against Chomsky? One could contend, as far as syntactical awareness goes, that the infant picks it up from external sources, but it seems odd, to me, that they could figure out this complex system so quickly.

    (And, if preferred, I can post references for any studies refer to)


  4. In a couple of postings above, Scott says, inter alia,

    (1) “Attempts to explain language learning in terms of general cognitive faculties without the need to invent a language-specific one have so far failed.” Says you.
    (2) There is another explanation, which may not be ‘better’ but requires fewer assumptions about the nature of cognition and of learning, hence is consistent with the law of parsimony. Digo yo.

    On (1). Well, name some successes. Eflnotes has cited a couple of attempts, like Reali & Chater on structure dependence. What they show is that an artificial neural network can evidently produce the desired equivalents of “Is Max __ the man who is dancing?” rather than “Is Max is the man who ___ dancing?” They claim that their results “indicate that it is possible to distinguish between grammatical and ungrammatical AUX questions based on the indirect statistical information in a noisy child-directed corpus containing no explicit examples of such constructions.” I’m struck by their use of the terms ‘grammatical’ and ‘ungrammatical’, as the program has no knowledge of AUX, or Main Clause, for starters; what the network distinguished was between high and low probability.

    On (2) Bates says something similar, and is similarly wrong about parsimony. You don’t simply count ‘assumptions about the nature of cognition and of learning’ and award the prize to the theory that makes fewer assumptions. Questions of parsimony only come into play when you have fairly precisely comparable theories, and of course the question of what assumptions are being made is not one of number but of empirical adequacy. The explanation that Scott seems to be thinking of is one that assumes associationist learning; associationist learning has an abysmal track record, as has been pointed out since Kant. So the assumption that language learning only requires association–one assumption–is hardly to be preferred to the assumption that association plays no part in language learning–also one assumption. We can start talking about parsimony when two competing theories of language and language learning have comparable coverage of the phenomena; and that time is yet to come.


  5. Dear Geoff,
    You write,
    “5. A Darwinian theory of evolution is consistent with UG theory.”
    I am not quite sure I understand what this means, … But I can see how Darwin’s thoughts on the origin of species and Chomsky’s UG model to account for language capacity have some things in common. They start with the facts, we are, we speak, and work towards an acceptable way of giving reasons why. (I heard of Chomsky first as lit. student. Then I encountered him again when studying Krashen while trying to teach English “the right way”. At that time the idea of a language acquisition device seemed to be a flimsy argument. The LAD, really? I doubt. But then, I confused the conclusion for the grounds that would warrant such inference. We do not “observe” the LAD. Rather, the LAD turns out to be the last dike that holds it all together. All other explanations seem to leak. It is the best explanation we’ve got.) Curiously enough, to me both, LAD and Darwinism, have come under pressure with new facts about the fact—us–speaking. Molecular biology and computer powered language analysis throw up new information about the artifacts that make one “shudder”, at least, I feel intrigued. I have tried to harmonize molecular biology (as a point of interest) with SLA. If the LAD explanation has any connection to reality it must be found in DNA. It is the only way information about what a species is and does is passed on.



    1. Hi Thom,

      I’m not sure how to respond to this. My obvious response is “All I said was…”. And all I said was that Darwin’s theory doesn’t say anything that can be taken as an argument against UG theory.


  6. I think there are serious doubts about whether any naturalistic theory, such as Darwin’s of natural selection, can ever accomodate intentionality. Since intentionality is, it seems to me, obviously intrinsic to language we must doubt, surely, whether any naturalistic theory can ever accomodate language.


    1. Patrick, thanks for posting this; how did you download it? What Fodor is talking about (I wish as articulately as he writes) is to be found in detail in Fodor & Piattelli-Palmarini, “What Darwin Got Wrong” (Farrar, Strauss, & Giroux 2010), well worth reading. But I don’t think we need be as pessimistic as you. The use of language is of course steeped in intentionality, but that doesn’t mean the structure of language is.

      Liked by 1 person

      1. Dear Kevin, but what is at stake is the why of language, not the what, right? I have the Fodor book. I agree with his criticism, but am not sure I can see what better he has to offer. I have to go back to re-read.


      2. Hi Thom.

        You are right that Fodor isn’t offering anything ‘better’. He has long been sceptical, as am I, about whether the theory of natural selection can accomodate anything that we would recognise as a fair description of mental content (and I think we can be forgiven for being sceptical about any theory of language that does not accomodate a fair description of mental content). Lately, controversially, he has generalised his concerns about teleosemantics to develop a critique of adaptationism altogether. It is legitimate, surely, to raise an objection to a prevailing theory even if you can’t suggest a superior alternative. There’s a good response to Fodor here.

        I probably shouldn’t have even mentioned it, but Geoff mentioned Darwinian theory and I have a certain twitch.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s