Fatal attractors

sauleda 01Me, in a pastelería (cake shop)

¿Qué es eso?


Ese pastel, allí.

Tarta de manzana.

¿Y aquella aquí… ¿Qué lleva?


Si, quiero una de este.

Which translates, more or less, as:

What is that?


This cake, over there.

Apple tart.

And that one here, what’s in it?


Yes, I want of one of this.

The translation is meant to convey the fact that the demonstratives (eso, ese, aquella, etc) are all over the place – literally – and only accidentally coincide with where I’m pointing, and with the gender or number of the thing pointed at. And, more tellingly, were I to have the same conversation again tomorrow, I might well use a whole different combination, plucked off the shelf, as it were, in a similarly random fashion.

This is as good an instance as any of how my Spanish, or pockets of it, functions in a state of ‘free variation’.

At some point, I must have been aware of the range of choices available for identifying objects near to me, near to you, and distant from both of us, and how these demonstrative adjectives and pronouns are also sensitive to number and gender, such that any decision to use one requires making a selection from 15 different options (see table below).

Understandably, I resisted learning the rules of a system that seemed impossibly complicated (three deictic points, three genders, as well as singular and plural). Instead I simply pulled demonstratives out of a hat, applying them indiscriminately, and sometimes throwing in a Catalan one for good measure. Of course, the physical context of the situations in which these items are typically used meant that pointing and eye-gaze made up for whatever incoherence resulted from my capricious grammar. Hence, there was little or no feedback as a result of miscommunication. I got the cakes I wanted.

pastelesAnd so the system (or lack of system) became entrenched. The same thing, more or less happened with past tense forms, with dependent prepositions, with clitic pronouns, with por and para, and with ser and estar. There are even sets of like-sounding words that I also deploy in free variation, the verbs planear, plantear, plantar, and planificar being a case in point. These wildly chaotic sub-systems seem to co-exist alongside other systems that are relatively stable.

In complexity theory, the fluctuations between relative stability and instability within dynamic systems are well attested. As Lewin (1993. 20-21) puts it:

Most complex systems exhibit what mathematicians call attractors, states to which the system eventually settles, depending on the properties of the system. Imagine floating in a rough and dangerous sea, one swirling around rocks and inlets. Whirlpools become established, depending on the topography of the seabed and the flow of water. Eventually, you will be drawn into one of these vortexes. There you stay until some major perturbation, or change in the flow of water, pushes you out, only to be sucked into another. This, crudely, is how one might view a dynamical system with multiple attractors.

Language learning is similarly ‘chaotic’: as Larsen-Freeman (2006: 592) notes, ‘There are no discrete stages in which learners’ performance is invariant’. And she adds (p. 593) ‘Learners do not progress through stages of development in a consistent manner. There is a great deal of variation at one time in learners’ performances and clear instability over time’.

The transition from one attractor state to another is called a phase shift, and one definition of fossilization might be ‘the absence of phase shifts’. Thus, Ellis (1999: 472) argues that ‘fossilization arises when learners fail to resolve the inherent variation in their interlanguage’. Witness my demonstratives.

But proponents of complexity theory would argue that any such stasis is an illusion: there is no permanent ‘end state’ in evolving, dynamic systems.  Hence ‘if there is no end state to language, it may be unhelpful to think in terms of fossilization as an end state to second language learning’ (Larsen-Freeman & Cameron, 2008: 10).

In a study of five Chinese learners of English, Larsen-Freeman tracked the ‘messiness’ of individual learning trajectories – a messiness not dissimilar to my random use of demonstratives. Larsen-Freeman conjectured that this very messiness might be indicative of the ‘instability that precedes a phase shift’ (2006: 611). And she adds, suggestively, ‘It is here where a pedagogical intervention might be optimal’ (ibid.).sauleda

What kind of pedagogical intervention?

I am drawn back to one of my favourite language learning accounts. Christopher Isherwood (1977:76), the writer, describes how he overcame a gap in his linguistic competence:

…Humphrey said suddenly, “You speak German so well – tell me, why don’t you ever use the subjunctive mood?” Christopher had to admit that he didn’t know how to. In the days when he had studied German, he had left the subjunctive to be dealt with later, since it wasn’t absolutely essential and he was in a hurry. By this time he could hop through the language without its aid, like an agile man with only one leg. But now Christopher set himself to master the subjunctive. Very soon, he had done so. Proud of this accomplishment, he began showing off whenever he talked: “had it not been for him, I should never have asked myself what I would do if they were to — etc., etc.” Humphrey was much amused.

In much the same way, I’m approaching the ‘holes’ in my own competence. How? A good old-fashioned students’ grammar, and a workbook of exercises. Hopefully, I will be able to find opportunities to activate, in real contexts (like the pastelería), the explicit knowledge gained from this mechanical practice, and trigger some kind of phase shift.

I’ll let you know how it goes.

Formas del pronombre demostrativo (from Sánchez Pérez & Sarmiento González 2005)

Formas del pronombre demostrativo (from Sánchez Pérez & Sarmiento González 2005). Click to enlarge.



Ellis, R. (1999) ‘Item versus system learning: explaining free variation’, Applied Linguistics, 20/4, 460-80.

Isherwood, C. (1977) Christopher and His Kind: 1929-1939, London: Eyre Methuen.

Larsen-Freeman, D. (2006) ‘The emergence of complexity, fluency, and accuracy in the oral and written production of five Chinese learners of English,’ Applied Linguistics, 27/4, 590-619.

Larsen-Freeman, D. & Cameron, L. (2008) Complex Systems and Applied Linguistics, Oxford: Oxford University Press.

Lewin, R. (1993) Complexity: Life on the edge of chaos, London: Phoenix Books.

Sánchez Pérez, A. y Sarmiento González, R. (2005) Gramática práctica del español actual, Madrid: SGEL.

About Scott Thornbury

I write books about ELT methodology and teach on the MA TESOL program at the New School in New York. I live in Barcelona. View all posts by Scott Thornbury

20 responses to “Fatal attractors

  • Amadeu Marin

    “Language learning is similarly ‘chaotic’: as Larsen-Freeman (2006: 592) notes, ‘There are no discrete stages in which learners’ performance is invariant’. And she adds (p. 593) ‘Learners do not progress through stages of development in a consistent manner. There is a great deal of variation at one time in learners’ performances and clear instability over time’.”

    My 30-year experience teaching EFL more than confirms that empirically. My ‘B2 students’ make similar mistakes (oh, sorry, I mean… show similar attractor-induced interlanguage –better? LOL) to my low B1’s. Then, why do we keep paying attention to that mumbo-jumbo of pre-intermediate, intermediate, upper-intermediate, etc.? Why do we insist on trying to classify students by discrete levels when our experience tells us that so-called ‘mixed ability’ classes is the norm?

    Wait it moment… is it perhaps because ELT publishers need to keep selling 6 or 7-level courses? I wonder…

    • Amadeu Marin

      Sorry to show up again, but I just discovered a typo in my last paragraph… The beginning should read ‘Wait a moment…’.

      Don’t want other contributors to think I’m trapped in a twirling vortex… ;-). By the way, I love that Lewin metaphor. How clearly the concept is conveyed. I’m more and more into metaphors for explaining seemingly complex ideas.

    • Scott Thornbury

      Thank you, Amadeu. Yes, I think the fact that all the research into learning and interlanguage suggests that learning (of grammar, at least) is NOT linear and incremental, and the fact that coursebooks, syllabi and exams assume that it is, is the single biggest headache in our profession.

      I’m glad you liked the vortex metaphor – I stumbled upon it as I was flipping through Lewin’s book in search of a definition of ‘phase shifts’.

  • Svetlana

    Dear Scott,
    Thank you very much for such a vivid example of complex systems at work, very frank confessions, e. g. “I resisted learning the rules of a system that seemed impossibly complicated”, even for you, an educated linguist (!). Then what can be expected from our poor students, bombarded with numerous tables, without a clue even of what those might be about? (e.g. the English verb tense forms) The end point is resistance, very strong resistance rooted in the basic instinct of survival :-). Yet at this point of development of your Spanish this table doesn’t look so frightening (did I get you right?). Why doesn’t it? Probably because this explicit knowledge gets now supported by the implicit data from the language accumulated through comprehensible input / output (i.e. interaction with the environment, which is only possible when it is “meaningful”)? In a number of comments to this blog Krashen’s Natural Approach is bitterly criticized, these comments sound like those you can hear from “deceived investors” – “I have invested so much time in reading / listening to gain “comprehensible input”, yet no gains (profit) in language production”. However, this input is a necessary condition to trigger “mirror neurons” responsible for imitation but which get activated only when the goal is clear; this leads to gathering the data for implicit subconscious processing. Once there is sufficient information collected the linguistic rules stop looking frightening and, all of a sudden, start making sense. Since the “sensitive period” for spontaneous language acquisition is over for most adults (the axons between neurons do not grow as quickly as they do before the age of five), it can be compensated by a conscious effort to set up a goal and develop a plan of how to achieve it (your plan of attack with the traditional grammar book) — a unique (?) human ability embedded in the function of the fully ripened and developed frontal lobes (A. R. Luria “Neuropsychology”).
    Sorry for a long comment, but I am desperate to find out – will your plan of attack work and will the naughty Spanish demonstrative pronouns start behaving properly abiding by the objective norms? Please let me know!

    • Scott Thornbury

      Hi Svetlana… yes, the fact that I no longer found the table of demonstratives daunting may owe to the fact that I (like Isherwood and his subjunctives in German) had recognized a gap in my competence and had embraced the need to fill it. This would seem to be an important role for the teacher. At the same time, I was ready to notice the gap – which may not always be the case with our students. As Hamlet said: The readiness is all!

      As for my plan of attack (as you delightfully put it): having studied the demonstrative table, and done a few exercises, I am starting to notice them a lot more, both in written text and spoken talk. I realise that I had never given them ANY attention in 30 years. If noticing is a prerequisite for acquisition (as Schmidt argues) than I am a little further down the track, it would seem. But I now need to do what Isherwood did, and USE them, even to the extent of over-using them. I need to order more cakes!

  • Language Garden (@DavidWarr)

    Very interesting as usual, Scott. Here is a brief account of my similar experience learning Spanish, amongst other things, “por” and “para”. http://languagegarden.wordpress.com/2012/06/17/reap-what-you-sow/
    Hope you like it.

  • hanaticha

    First of all, I love your visuals, Scott. This little bakery seems to be an ideal place for practising demonstratives 🙂 You give a classic example of seemingly trivial language items, which are terribly difficult to learn. I’m not a psychologist but it may be because adult learners skip certain phases when learning L2. Kids usually start speaking using only a few small words and they get plenty of opportunities to practise them, connecting them with the physical world. Adult learners, on the other hand, want to use sophisticated language right from the beginning since they need to express their sophisticated ideas. We are too busy to deal with such ‘unimportant’ words, and we can use gestures to convey the message anyway, as you point out. I’d say it’s not a question of conscious learning, though (of grabbing a grammar book). I agree that we may be trapped in a neural networks whirlpool for a certain period of time and only a lot of meaningful practice will save us in the end.

    • Scott Thornbury

      Thanks, Hana! It is certainly a puzzle why it is that kids acquire the seemingly ‘unimportant’ (and barely audible) grammar words and adult learners don’t. Kids can be no less fixated on meaning than adults, and yet, for kids, the little words, which contribute so little to the propositional content of an utterance, never ‘fall through the cracks’. Nor is it the case that they get much correction or explicit instruction. (And don’t come running to me with some spurious and untestable claim about universal grammar!)

      • paulwalsh

        Hi Scott,

        Interesting blog. Regarding child learners, don’t you think that learning songs has a big part to play in building automaticity and pattern-recognition in children? And learning and memorising songs might also help you with your Spanish too.

      • paulwalsh

        I’m also learning Serbian (and German, which is going slowly…) and find that songs really help! I describe my attempts at Decentralised Learning here http://goo.gl/3WjHcx

      • Scott Thornbury

        Hi Paul – yes, the usefulness of repeated routines like songs, nursery rhymes etc, can’t be over-estimated, and I think you’re right in suggesting that these help ‘entrench’ the patterns and rhythms of the language. I’m not sure, though, that these explain the capacity of the child learner to acquire the ‘small print’ details of the language – the inflexions and function words – since it’s just as easy to sing a song, as it is to utter a phrase, by paying attention only to the words and syllables that carry the ‘meaning load’. E.g. Three blind mice/ Three blind mice/ see how da run/ see how da run/ da all ran af-da farmer wife/ she cut da tail da carvin knife/ Three blind mice… etc. In fact kids do seem to go through this de-grammared phase initially. Why, then, does the non-salient grammar get mapped on?

      • wezjohnson

        Hi Scott,

        Thanks for the stimulating and edifying blog post, as always.

        You say: “It is certainly a puzzle why it is that kids acquire the seemingly ‘unimportant’ (and barely audible) grammar words and adult learners don’t.” As you no doubt know already, according to Steven Pinker in his book The Language Instinct, there is no such puzzle. For Pinker, children have an innate need to impose grammatical order on language. This is why, according to Pinker, children are capable of transforming a pidgin into a creole in just one generation, and why deaf children are able to take crude signs and impose a standard grammar onto them. However, you suggest that you remain unconvinced by this thesis when you say: “don’t come running to me with some spurious and untestable claim about universal grammar!” I would say that Pinker’s research is not so spurious or untestable as you claim, so I am very curious to know why you confidently reject this theory. Pinker’s research is pretty persuasive from my point of view, but I would love to be challenged on this. Can you recommend any books that, in your opinion, successfully counter Pinker’s arguments?

        Thanks for your time. Have a good one. Wes

      • Scott Thornbury

        Thanks, Wes… sorry it’s taken so long to reply (I’m on the road). I don’t buy into the ‘innatist’ argument because I’ve been persuaded that general (i.e. non-language specific) learning capacities (including sociocultural learning) do account for the capacity of children to acquire a uniform grammar, and that there is therefore no need to post a mysterious and so far undiscovered ‘language acquisition device’ or any innate hard-wiring. The best counter-argument is Michael Tomasello’s Constructing a Language: A Usage-Based Theory of Language Acquisition, Harvard University Press, 2003. I haven’t got time or space to summarize the arguments here, but as far as I know Pinker has produced slender evidence with which to refute them.

  • Mark Makino

    Since by your admission this non-grammar usually works for you, what is your motivation for “fixing” this? Is akin to finally putting together that jigsaw puzzle you got as a gift years ago? Or not wanting “I’m a foreigner” to be part of the message you convey when asking about pastries? Something totally different?

    • Scott Thornbury

      Good question, Mark. Given I’ve managed up to now without the ‘little words’, and given that I’ll never NOT sound like a foreigner, why bother? That, at least, has been the principle I’ve been operating on up to now. I think, in the end, it’s a matter of pride rather than precision, and a certain sense of obligation, if not to one’s interlocutors, but to oneself – your jigsaw puzzle analogy is very apt..

  • Martin Sketchley

    Scott, I thought I would share a school’s experience of teaching and their learners’ experience of acquiring Spanish during this academic term. They incorporate magazines during lessons to supplement their current curriculum and perhaps it will provide you with an additional perspective to language learning: http://oneyearinspanish.wordpress.com/

  • Nick Bedford

    Hi Scott
    Thanks to your last post I’ve been thinking all week about a number of advanced students “mistakes” such as “Here you are my composition” – as opposed to “Here you are”/Here you have my composition”, which have long puzzled me. I was loathe to call something a mistake if there were two ways of righting it and I always felt that the students who still said things like ” I don’t remember the exactly place” – instead of “I don’t remember the place exactly “/ “I don’t remember the exact place”, had simply collapsed two forms into one. After reading your post I much prefer the idea of “free variation” – with its implication of dynamism – over the idea of “collapse” – with its implicit idea of stasis. This is because students are not always making the mistake – it’s just that on occasions they fail to make the complete swing from one choice to another e.g. He speaks fluent German :::::: He speaks German fluently. On occasions they’re apart from both poles. Thank you a lot!! 😉

    • Scott Thornbury

      Thanks, Nick…those are great examples of the kinds of errors I suspect I make in Spanish on a daily basis, where alternate forms vie with one another in a kind of random way. Some scholars dispute the fact the choice of one over the other is ‘free’, arguing that.specific contexts determine the choice of form, but Rod Ellis has argued that the choice is a bit like the rather random way we might pull a cardigan out of the drawer. (He’s the one who wears cardigans, not me!)

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 )

Facebook photo

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

Connecting to %s

%d bloggers like this: