Larsen-Freeman Lost in Complexity: Bullshit Baffles Brains

role-of-complexity-in-business

I see that Diane Larsen-Freeman is doing a plenary at the 2016 IATEFL conference. She shares the “Invited Plenary Speaker” limelight with her admirer Scott Thornbury, who will wisely side-step theoretical stuff and entertain everybody with a history of ELT.  Thornbury often cites Larsen-Freeman’s incoherent ventures into science and postmodernism in his own forlorn attempts to promote emergentism, thus showing that he’s not only a dedicated follower of fashion but also a bad judge of scholarship.

Larsen-Freeman wants to re-direct ELT. In her plenary she will suggest that we ditch our outdated processing model of SLA and replace it with an ecological approach where, among other radical proposals, the construct of affordances takes the place of input.

Actually, I’m not sure that affordances is a construct here; maybe it’s just a metaphor, but anyway Larsen-Freeman is the last person to ask for clarification since she messes around with terms like metaphor, hypothesis and explanation with such gay abandon that it’s anybody’s guess what she means.

In an attempt to prepare everybody in the audience for what you’ll hear, I offer a summary of Gregg’s (2010) review of Larsen-Freeman and Cameron’s truly dire book on which the talk will be based. I’ve taken terrible liberties with Gregg’s text. You should read his article and I apologise to him for my clumsy summary.  Everything that follows is from the article.

Hix%20at%20the%20PixSCIENCE3-BW

Larsen-Freeman and Cameron (LFC) explain that complexity theory is the theory of complex systems, a complex system being a certain type of system

‘produced by a set of components that interact in particular ways to produce some overall state or form at a particular point in time’ (p. 26).

Complex systems are dynamical, non-linear, sensitive to initial conditions, open to input from outside the system, and adaptive.

Got it? Well good, because the book gives us nothing that could enable one to analyse SLA data or critique SLA research in terms of complexity theory. It’s hard to find any connection between complexity theory on the one hand, and the conclusions LFC draw from it on the other. AS Gregg puts it:

by and large, when they say ‘from a complexity perspective, P’, they could just as well have said more simply, and more accurately, ‘We think that P.’

Misunderstanding Science

LFC start by misunderstanding the nature of standard science and the relation of complexity theory to it. For instance, they say:

 Whereas positivist research is based on the assumption that there are universal laws and thus    sets predictability as a goal of the research process, from this complexity perspective, no two    situations can be similar enough to produce the same behavior; thus predictability becomes    impossible. (p. 16)

Some comments:

  1. Positivists are a (defunct) class of empiricists. The nativist linguists and cognitive scientists that LFC oppose are not empiricists, and yet they, too, believe in universal laws and value predictability.
  2. Complexity theory does not oppose mainstream science, as Gribbin (2004: 3) points out, ‘chaos and complexity obey simple laws – essentially, the same simple laws discovered by Isaac Newton.’
  3. Predictability in standard science is not always the goal – consider palaeontology or evolutionary biology, for instance – so non-predictability by itself cannot distinguish complexity theory from other scientific theories.
  4. Not only do LFC incorrectly characterize science as positivistic, they consider it ‘reductionist’, where ‘reductionism’ seems to mean ‘looking for an underlying cause or causes’. LFC reject ‘the common reductionist approach in science, which relies on a central principle that one can best understand an object of inquiry by taking it apart and examining its pieces’ (p. 231). How this differs from identifying the elements of a complex system and studying their interactions is not clear. In any case, on the standard meaning of the term it is LFC who could be considered reductionists, since they reject the position that language cannot be reduced to general cognitive or neurological phenomena.

Misunderstanding language

LFC say

  • Language is no longer perceived as an idealized, objectified, atemporal, mechanistic ‘thing’. (p. 20)
  •  There is no need to conceive of language in a decontextualized terminal state of frozen animation. (p. 91)
  • A complex systems view of language rejects a view of language as something that is taken in a static commodity that one acquires and therefore possesses forever. (pp. 115–16)
  • Language is not a single homogeneous construct to be acquired. (p. 116)
  • Learning is not the taking in of linguistic forms by learners. (p. 135)

LFC offer not one single example of a researcher, nativist or otherwise, who perceives language as an idealized thing, or as something taken in, etc. This sort of preaching to the choir may satisfy the choir, or at least the less intellectually rigorous of the singers, but it is a disservice to the disinterested reader looking for information about complexity in language.

Here is one whole paragraph from the five paragraph section on ‘First language acquisition from a nativist position’ (p. 117; the numbers are added for reference):

Nativist accounts (1) rest on the assumption that the capacity to learn language is a unique    property of the human mind that is represented as (2) a separate module in the brain,  conceived of as (3) an organ within the brain that performs specific kinds of computation (4) (Chomsky, (1971). Nativists believe that this modular architecture allows the shape and form of I-language to be largely independent of (5) other aspects of cognitive processing or social functioning.  They also believe that the fact that (6) a UG is contained within the module (7) accounts for the evolution [sic; of language?] within the human species and explains how native language acquisition can take place so expediently, given what they feel is (8) a rather degenerative state [sic] of the input, filled with pauses and inchoate utterances and other dysfluencies, (9) referred to collectively as the ‘poverty-of-stimulus’. The impoverished input, combined with what is alleged to be an absence of negative evidence (i.e. evidence of what the system will not permit),  leads nativists to argue ‘that the complexity of core language cannot be learned inductively by general cognitive mechanisms and therefore learners must be (10) hard-wired with principles that are specific to language (Goldberg, 1995: 119), although quite naturally, the search for what these principles are is an ongoing one, which has gone through several stages so far.

Very briefly:

  1. Nativist accounts rest on assumptions that have nothing to do with the putative uniqueness of language.
  2. The module – again, not a specifically nativist assumption – is a module of the mind not the brain.
  3. Nor is it taken to be an organ of the brain.
  4.   Chomsky (1971) says nothing about the brain, or of the language faculty as an organ.
  5. The language faculty or UG is not an aspect of cognitive processing, let alone social functioning.
  6. UG is not contained in the putative module, it is the module.
  7. Nobody, but nobody, claims that UG accounts for the evolution of language, or of anything else, for that matter.
  8. Dysfluencies have never been of any importance in Poverty of the Stimulus arguments. And no nativist has ever said that input is filled with dysfluencies.
  9. And of course nobody refers to the dysfluencies themselves as the poverty of the stimulus.
  10. Nobody talks of ‘hard-wiring’ principles.

Misunderstanding acquisition research

This failure to engage with the facts shows up as well in LFC’s rare accounts of putative acquisition phenomena. They tell us, for instance, that a complexity perspective can explain the so-called ‘vocabulary burst’ in child language acquisition (pp. 129–30). They fail, however, to tell us what sort of burst this is; they do not provide any documentation of the burst; they do not even suggest what about the burst would be accounted for, or how, by a dynamic systems account. And they do not consider the claim by Bloom (2000) – who studied the data in detail – that there is no burst to account for (see also Ganger and Brent, 2004; McMurray, 2007).

Even more striking is LFC’s bizarre explanation for so-called U-shaped behavior in the first language acquisition of the English past tense (p. 129).

Initially, learners fail to mark past tense morphologically (i.e. they fail to use the verb + -ed construction) due to the frequency of irregular verbs in the language addressed to them. Later, the irregulars disappear in their production … As the number of verbs in the competition pool expands across the course of learning … the irregular forms reappear.

The only source LFC give for this idiosyncratic account is Ellis and Larsen-Freeman (2006), which simply says the same thing. LFC, in other words, make no attempt to examine the data, which clearly show that:

  1. Initially learners fail to mark the past on both regulars and irregulars.
  2. Irregular pasts, once they appear in production, do not ever disappear; and
  3. The U-shaped behavior LFC propose to account for does not exist (see, inter alia,  Marcus et al., 1992; Maratsos, 2000; Maslen et al., 2004).

Given the lack of a phenomenon to explain by appealing to dynamic systems theory, it is probably superfluous to point out that, in any case, the proffered explanation does not in fact depend in any way on dynamic systems.

This casual disregard for the actual phenomena of language is reflected in LFC’s reference list: something under 500 items, and not one single paper on language acquisition, first or second; not one single paper from the theoretical linguistics literature proposing some account of some specific phenomenon of language. This, of course, is because there is no discussion anywhere in the book itself about specific linguistic phenomena (unless you count the two examples just mentioned), even from their ‘complexity perspective’. Rather, what we get is ex cathedra, take-it-or-leave-it statements of doctrine:

  • Dynamic systems theory does away with the distinction between competence and performance. (p. 17)
  •  Complexity theory thus [sic] brings about a separation of explanation and prediction. (p. 72)
  • If language is conceived of as a complex system, then it is entirely possible for novel complexities to emerge. (p. 98)
  • A language-specific innate mental organ is not consistent with seeing language as a complex    adaptive system. (p. 156; pace, for example, Plaza-Pust, 2008; Hohenberger and Peltzer-Karpf,    2009: 482)

And so on. Sometimes LFC do not even bother to disguise their opinion as a claim. Here, for instance, in its entirety, is their discussion of the work of William O’Grady, arguably the one emergentist in the field of language acquisition to take seriously both the pheomena and the nativist account of the phenomena, and to offer a detailed, wide-ranging, and at least potentially competitive analysis that rejects the idea of a UG:

While we find O’Grady’s (2005) syntactic carpentry account intriguing, we do not feel that a computational account is necessary or desirable to account for the emergence of order (p. 159, footnote 1).

How, one might ask, do LFC propose to account for ‘the emergence of order’ without computation? Why would this be desirable? I have no idea and, if LFC do, they are keeping it under their hats.

Not only do LFC ignore the work that has been done by ‘positivist’, ‘reductionist’ acquisition researchers (i.e. researchers with a commitment to empirically sound science), they do not present any examples of research from their complexity perspective. As Ionin (2007: 28) says,

It does not seem very fair to criticize existing, developed theories for not incorporating all possible factors, when no alternative theory is presented that does incorporate these factors.

LFC do give an extended account of Larsen-Freeman (2006) – a study of five Chinese speakers writing the same story four times over a six-month period – but they do not show the relevance of dynamic systems theory or complexity theory to that research.

Gregg concludes

Complexity theory and dynamic systems theory are important, well-established, and productive parts of the physical sciences. Although it is still a matter of some controversy whether these theories can be applied to the biological and cognitive sciences, work such as that of Thelen and Smith is certainly promising, to say the least. So it is worth trying to find out to what extent, if any, complexity theory can be applied to the domain of language. LFC have failed utterly to address the question, let alone resolve it. The question is an empirical one, and the answer – if answer there be – will be found not by taking a perspective and expatiating on it, but by doing the science.

Conclusion

We must be critical. We must be on guard against bullshit. So often we’re lulled into acquiescence by big names and by academic credentials. What Gregg makes clear in his article is that the book he reviews is bullshit.

Scott Thornbury swallows this bullshit and attempts his own popular version of it (see https://criticalelt.wordpress.com/sla/emergentism-2/ for example). Like the LFC book, Thornbury’s attempts to argue the case for emergentism are hopelessly argued. Actually, they’re not argued at all: they’re just asserted with lots of illustrations.

So much of this new wave of relativist bullshit, most of it weirdly inspired by an unconsidered embrace of empiricism (why do so few of them appreciate the consequences of the epistemology they sign up for??) rides on the back of what is now considered to be Larsen-Freeman’s  “seminal work” of 1997, an article where she does exactly what Gregg criticises in his review of the 2007 work. In general, Larsen-Freeman

  • misrepresents scientific method; in particular she fails to distinguish between description and explanation and demonstrates an ignorance of the differences between a theory, a hypothesis, a theoretical construct, and a metaphor
  • misreprents positivism
  • misrepresents empricism
  • misrepresents reductionism
  • misreprents nativism
  • misreprsents cognitive SLA research
  • offers no convincing connection between general aspects of complexity theory and her views of second language learning and teaching.

In 1999 I was in a bar in San Cugat with my good friend Connie O’Grady and George Yule. George was on a roll. “Bullshit baffles brains, George”, intervened Connie, “but don’t suppose for a second that we don’t see through it. We don’t  like dressed-up crap, and we recognise good stuff on the rare occasions that we get it”.

 

Gregg, K. R. (2010) Shallow draughts: Larsen-Freeman and Cameron on complexity. Second Language Research, 26(4) 549–56.

Larsen-Freeman, D. ( 1997 ). Chaos / complexity science and second language acquisition. Applied Linguistics, 18( 2 ), 141-165.

Larsen-Freeman, D. and Cameron, L. (2008) Complex systems and applied linguistics. Oxford:   Oxford University Press-

Advertisements

16 thoughts on “Larsen-Freeman Lost in Complexity: Bullshit Baffles Brains

  1. I ‘liked’ this post but not because I think DL-R and Thornbury are promoting “bullshit”. I haven’t read the book, but have read articles and posts and heard talks from them on the topic. Personally, I found the presentation of these ideas highly compelling and sense-making as a teacher-practitioner looking for compelling input and sense-making for my own reflection and inquiry (for example, at this keynote in Portland in 2014: http://www.tesol.org/attend-and-learn/international-convention/convention2014/featured-speakers/diane-larsen-freeman-keynote-video). My question is, I suppose: is the popularization of an idea not good enough or even irresponsible “bullshit” unless the idea is presented in its full scale? I’d be interested to read how DL-R addresses her audience in the preface of that book. Does she frame her presentation in some way that might condition our expectation of less or indeed no “messing around with terms”, or of the inclusion of tools that would “enable one to analyse SLA data or critique SLA research”? I guess all I’m saying is one man’s bullshit is a another man’s fertilizer. Or is that all too relative?

    Like

  2. Hi Matthew,

    Sometimes I despair!

    When I say that the LFC book is bullshit, I mean that it’s nonsense, not well-argued, not substantiated.

    I also mean that it uses terms with no regard for their consistency and that it tries to persuade the reader with lots of spurious academic crap. It’s the worst kind of academic text: pretentious and empty.

    Nothing you say even dents the arguments against it presented by Gregg: nothing whatsoever!

    You ask “is the popularization of an idea not good enough or even irresponsible “bullshit” unless the idea is presented in its full scale?” LFC had a full book to “popularise their idea” and in my opinion it’s rubbish. If you disagree well fine: say so. Give your reasons.

    Then you ask “Does she frame her presentation in some way that might condition our expectation of less or indeed no “messing around with terms”, or of the inclusion of tools that would “enable one to analyse SLA data or critique SLA research”?

    Well, let’s wait and see. All I ask is that when she blathers on about ecology and affordances you don’t melt in admiration but rather call her bluff. Because a bluff is what it is. What does it all mean?

    What, in the end, is all this stupid posturing about? Your experience of me experiencing you? The complex dance of ideas in an uncertain universe? The leaf on the ground of Larsen-Freemen’s well cut lawn somehow explaining the Napoleonic wars? The misrepresentation of nearly everything to explain precisely nothing?

    It’s bullshit, Matthew.

    Liked by 2 people

  3. Thank you for the reply, Geoff. And let me add I’m glad your blog is back in its new guise.

    Well aware that I’m swimming in waters far deeper than I can really swim in, I’ll respond…gulp…

    You said: “All I ask is that when she blathers on about ecology and affordances you don’t melt in admiration but rather call her bluff.”

    I’ll do neither, thanks! I won’t melt in admiration because I’ll be busy a) using what is presented to reflect on my lived experience as a teacher and b) looking into this whole ‘complexity’ business further for myself. I may allow myself some low-level admiration, but I’m not particularly interested in worshipping the messenger. I think I’d agree that there’s some degree of ‘hero worship’ in the field;

    I won’t call her bluff because…well, I can’t. Gregg and you have done the extensive scholarship; seemingly a good deal more than those whose work you are critiquing, or in a better way. I certainly have not! All I can do in that area is read what a) gets to me b) can be accessed and then read and understood to some reasonable degree by someone who isn’t paid or supported to do all that much of either. Pretty sure I did read Gregg’s paper maybe a year back, likely because you’ve mentioned it before. Obviously I’ve read the post above.

    Having done so, I’m better informed but remain uninterested in the calling of the bluff myself – not because I’m an uncritical lemming but because of the above, and…well…I’d say: *it’s not to my benefit to attempt it*. What Diane’s 2014 keynote did was make me all the more interested to find out more about the topic (again, in my own fashion…I’ll never become an expert as I can’t read O’Grady, McMurray, etc. etc. and stay employed not to mention married). This inquiry IS to my benefit. I think…

    If ^that’s all way below the discourse level here, and if my expressed concern about ‘what X does for ME’ rather than ‘what X does for science’ serves no purpose here – or worse, provides but an embarrassing distraction that can only shot down as further evidence of the hoodwinking of the masses, maybe don’t approve it. I certainly don’t have the ability to offer good critique of your critique and the above isn’t meant as any attempt at one.

    That said, I would like to add one more bit here:

    You said: “You ask “is the popularization of an idea not good enough or even irresponsible “bullshit” unless the idea is presented in its full scale?” LFC had a full book to “popularise their idea” and in my opinion it’s rubbish. If you disagree well fine: say so. Give your reasons.”

    That question wasn’t me implying any disagreement with your assessment of LFC’s book. It was me wondering what your answer to that question would be beyond this book alone. I’m not actually sure a project of ‘popularisation’ is even at play here, but that’s where my mind went as I consider the critique and imagine a defense for myself. Anyway, I’m thinking there’s nothing of substance to the question or maybe it’s actually stupid: asking if there might be, in theory, good reasons for ‘bad scholarship’.

    Like

  4. The line “I think I’d agree that there’s some degree of ‘hero worship’ in the field;…” was the beginning of something I was going to add there before hitting ‘post comment’ – ooops. Well, it was simple: agreeing with you about the dangers of being uncritical around the ideas of those with power and influence and charisma after clarifying my own intention not to fall sheepishly asleep at that wheel myself.

    Like

  5. Steven Pinker remarked re Chomsky that ‘Anyone who goes to cocktail parties knows that one of Chomsky’s main contributions to intellectual life is the concept of “deep structure,” together with “transformations” that map it onto “surface structure.” When Chomsky introduced the tems in the behaviourist climate of the early 1960s, the reaction was sensational. Deep structure came to refer to anything that was hidden, profound, universal, or meaningful and before long there was talk of the deep structure of visual perception, stories, myths, poems, paintings, and so on.” It is the fate of certain of academe’s coinages, probably because they are deemed ‘up to the minute’ and therefore impressively ‘difficult’, to become immensely fashionable. They are then bandied around without regard to what they were originally used to mean, though always with an implicit appeal to the authority conferred upon them by their abstruse, technical origins. Such, I fear, has been the fate of the terms ‘chaos’ and ‘complexity.’ ‘Deconstruction,’ of course, is probably the example par excellence.

    Liked by 1 person

    • Hi Patrick,

      One good thing about returning to blogging is that I get to shoot the breeze with you again.

      Your comments put me in mind of Shaw’s remark to the man in the audience who shouted out

      “Utter Rubbish!”

      when he came on stage to take the general tumultuous applause at the end of the first night of Major Barbara.

      “I agree with you sir”, said our Bernard, “but who are we against so many!”

      There is, as you say, something inevitable about the fate of terms that start out well and are then taken up by those who know so little of what they speak.

      Like

  6. While I’m at it, why do people, and the ELT world is particularly bad for this, insist on saying ‘paradigm shift’ when they mean simply ‘big change’ or ‘cognitive dissonance’ when they just mean ‘saying one thing and doing another’?

    Liked by 1 person

  7. Thanks for putting this blog back up. I think any word with -ist on the end used as an adjective is prone to highfalutin misuse, and reductionist is a great example. To some people any reference to more basic components of a complex thing is reductionism, and any connection drawn between the physical and the human is the same.

    Like

  8. Good to see you back, Geoff. Two quick comments, if I may:
    1) I think there’s a useful distinction to be made–and I think I’m not alone in making it–between bullshit and horseshit. ‘Bullshit’ for me implies intention to deceive, and I wouldn’t accuse Larsen-Freeman of having any such intention.
    2) Matthew evidently was impressed by L-F and others’ presentations and papers, which is fine; and if those presentations led to, say, improvements in his teaching, even finer. But being impressed and being stimulated are not the same as being convinced, or shouldn’t be. I first saw Krashen talk at a TESOL conference decades ago, and I was impressed, nay, wowed. I immediately ordered his books (there were just 2 back then), and eagerly started to read them. And that is when the proverbial scales started falling from my eyes.

    Liked by 1 person

    • Hi Kevin,

      If “bullshit” implies intention to deceive then it’s not the word I was looking for. I’ve used it to refer to others’ work and I’ve always meant it as something like “pretentious nonsense”. I’m glad to have the opportunity to clarify that I’m not suggesting for a second that L-F – or anybody else I’ve mentioned on a blog – is trying to deceive.

      Like

      • Doesn’t pretentiousness necessarily involve an intention to deceive, or at least a wickedly irresponsible indifference to whether what one says (and seeks, presumably, by saying, to promote) is true or not? Saying what you honestly hold to be true is never pretentious, surely, though it may be mistaken.

        Like

  9. Thanks for this post. My field of study/research is language assessment and quantitative methods. I have read most of LF papers and her jointly written book. I am always happy to see that some scholars attempt to launch a new interdisciplinary topic and expand it. I began to read LF with elacrity but got disappointed. For one, they have converted a wholesome field of science, ie, dynamical systems theory, into metaphors or what the call dynamic systems etc. Terms like chaos, attractor, SDIC, basin of attractors etc are innately mathematical and do not Concern eddies, starlings, or galloping horse etc that are used by LF et al. While I recognize that metaphors are useful , the problem is what I have already pointed to: reducing a field of science to metaphors and then adapting those metaphors in applied linguistics. The reason, from what I can gather, is the lack of a profound understanding Of mathematical and quantitative data analysis, methods. I cringe when I see the claim that input – output models are useless, whereas all mathematical functions including those that can reach chaos are all input-output based. In fact, lowlie et al tried to fit a dynamical math function into a data set for language learning patterns and in most parts it failed to emulate the underlying structure of the data, meaning that this and such models, just like other regression models can be far away from the kind of perfection that LF has assumed; see
    Lowie, W., Caspi, T., van Geert, P., & Steenbeek, H. (2011). Modeling development and
    change. In M.H. Verspoor, K. de Bot, W. Lowie (Eds.), A dynamic approach to second language development: Methods and techniques (pp. 22-122). Amsterdam: John
    Benjamins Publishing Company.

    LF’s suggestions for language assessment (LF and Cameron, 2008) are very naive and show that the authors are unaware of the current status of language assessment. I respect these colleagues but Overall, I think the dynamical systems or what they call dynamic systems literature is hardly anything more than pseudo science.

    Liked by 1 person

Leave a Reply

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

WordPress.com Logo

You are commenting using your WordPress.com 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 )

Google+ photo

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

Connecting to %s