Offering 0pen and Honest Professional Advice
Ants In My Head - Complexity Theory and Language Acquisition
The following is a copy of a keynote address delivered (mostly in English) at the Spanish national conference of the EOI (Escuelas Oficiales de Idiomas – Official Language Schools) in Gran
Canaria in 2009.
Hormigas en la Cabeza
Lo que quisiera hacer es hablar sobre el ‘flip' o salto lingüístico que ocurre naturalmente cuando alguien ha embebido una cantidad enorme de palabras y frases de un nuevo idioma. El 'flip’ ocurre muy de repente – es un salto grande y se llama oficialmente un fase de transición.
Para mi el método de mas éxito de hacer este flip es Suggestopedia y una de sus carectarísticas mas fundamentales es la cantidad enorme de las palabras que se presenta a los estudiantes – un mil de palabras desconocidas, mas o menos – en el primer lección
Quisiera concentrarme en discutir sobre Suggestopedia y el proceso de adquirir un nuevo idioma desde el punto de vista de la Teoría de Complejidad (y teorías similares).
Elegí el título Hormigas en la Cabeza porque intento usar la formación y el desarrollo de una colonia de hormigas como una metáfora del proceso de adquirir un nuevo idioma.
Ants In My Head
“(A)nts do their learning at the colony level.. The “colony brain” is the sum of thousands and thousands of simple decisions executed by individual ants. The individual ants don’t have anything like a personality, but the colonies do.
Replace ants with neurons and pheromones with neurotransmitters, and you might just as well be talking about the human brain."1
”….la tremolor emocionant d'entendre una altra llengua, que va de la paraula al cervell i del cervell al cor, la manera com un idioma nou pot bellugar-se, cargolar-se i giravoltar ple de vida davant els teus ulls, el salt gairebé salvatge de la comprensió, l'instantani i joiós alliberament del significat, la manera com les paraules despullen els seus cossos impresos en un flaix de llum i calor.“2
I have been involved in teaching Welsh to Adults in North East Wales since 1985 and currently oversee a fairly large community programme catering for some 1,300 adults per annum. The majority of our students follow fairly traditional communicative language courses of 2 taught contact hours per week.
Since 1997, however, for reasons which I will shortly elaborate upon, I have personally channelled my energies into researching and experimenting with other teaching methods, methods commonly called Alternative or Experimental or Innovative. In 1997 I started using Suggestopedia, the pedagogic method developed by Dr. Georgy Lozanov of Bulgaria and currently this is the only teaching methodology which I personally, as a teacher, employ.
This being the case, it would seem that I am of the opinion that Suggestopedia is in some way superior or more effective than the other methods which I have had the opportunity to use. This is indeed true. And the purpose of this paper is to explain why I am of this opinion.
But let me start at the beginning, namely my first experience of learning another language as an adult.
Back in the summer of 1978 I enrolled by post on a 10 day residential Breton course for beginners (as I thought at the time). Breton is a sister language to Welsh, about as close linguistically as English is to German. I filled in the French registration form as well as I could (I had studied French up to the age of 16) and in due course turned up for the course.
When I got to the semi-derelict farmhouse where I would be spending the next 10 days to my horror I realised that I had booked myself onto the wrong course.
This course was for learners who had already achieved a high level of fluency – and I was, to all intents and purposes, a total beginner.
The course leader was actually in favour of not letting me attend the course at all but, thankfully, he was persuaded by the other tutors to allow me to stay. For the next ten days I was saturated with Breton input.
I lived, breathed, ate, slept and dreamt Breton.
The only thing I couldn’t do was speak Breton. Breton was hitting me from all directions and bits of Breton kept on popping uninvited into my head and scurrying through my consciousness like unruly ants. Yet, despite all the massive input I was receiving and all the entomological activity between my ears, when it came to trying to speak I could hardly manage more than a few very basic, stumbling utterances.
After the course ended I returned to Wales absolutely exhausted, (initially having great difficulty in reverting to English after we had docked in Plymouth). I stayed home for a week, not really doing much and then, the week following, I went to the National Eisteddfod in Cardiff (the National Eisteddfod is Wales’ foremost Welsh Language Cultural Festival held the first week in August every year).
I had been there a couple of days when one evening, outside a night club in Cardiff, who should I happen to meet but some of the students and teachers who had been on the Breton course with me a fortnight before (I knew they had been considering visiting the Eisteddfod sometime during the week ).
We greeted each other warmly and much to my (and their) surprise I suddenly flipped. When I say I flipped what I mean is that I suddenly, without warning, started to articulate myself quite fluidly in Breton.
The language was gushing out of me with practically no conscious effort on my part.
And I remember one of the teachers turning to one of her colleagues and saying in amazement "Ma Doue! Desket 'neus buan!” (Good God! He’s learnt quickly!).
If I were totally honest I’d have to admit that I, personally, didn’t feel that I had really done anything at all. I hadn’t looked at any Breton books or thought about the language very much for over a week. Yet in that time, out of conscious sight, all those disorganised, scurrying little linguistic ants in my head had surreptitiously and suddenly ordered themselves into an efficient, working collective. Without my knowing it a Breton ant colony had taken residence in my head.
The primary linguistic flip which I experienced that night, outside the night club in Cardiff, was not merely linguistic, however.
It was both somatic and affective as well. It was a felt flip. But it wasn’t a unique event because, since that time, I have personally experienced it (or subsequent secondary mini-flips) on more than one occasion. I have also spoken to numerous other successful language learners who have experienced their own flips and they can usually remember quite vividly where and when it happened. (The flip is not the final act of SLA, of course, but it is, I believe, the most essential part of the process. And it’s probably true to say – even if it is a tautology - that if learners don’t flip they’ll never flipping learn the target language successfully.)
As a teacher over the last two and a half decades I have seen some learners suddenly flip before my very eyes – but none as dramatically or as decisively as those students who have been on my Suggestopedic Welsh courses.
Why should that be the case? Why does the flip occur more regularly and with greater visibility on Suggestopedic courses than on traditional courses? The answer to that, I think, is twofold.
Firstly, it has to do with the basic assumptions underpinning traditional language teaching to adults (in the UK, at least) and secondly, it is linked to the word complexity. Let me elaborate upon these two points.
Traditional Language Teaching – Basic Assumptions
For me, three dominant beliefs underpin traditional second or foreign language teaching to adults in the UK and these three beliefs have a debilitating effect on syllabus design and the ultimate success, or otherwise, of language teaching programmes for adults.
The first is that the largely mono-lingual UK Anglophone population (i.e. those individuals whose mother tongue is English, irrespective of whether they are English, Welsh or Scottish) tends to believe that language learning is a very difficult undertaking.
Ordinary people may be able to learn very basic functional communicative skills in another language but complete mastery of another language is something that only 'clever’ people can achieve. This belief is culture specific and doesn’t exist, of course, in countries where people are naturally pluri-lingual.
The second belief is the assumption that languages are best introduced gradually in a linear fashion, one function or grammatical structure at a time, with the lexical content of lessons restricted to about 20 to 30 words or phrases.
The third basic belief is that students, from the word go, should be pushed to try and speak the target language (before they have had the chance to develop their ability to understand the language, what James Asher calls their comprehension literacy).
This unfortunate tendency is often strongly backed by government inspectors despite the fact that research by Asher, Postovsky, Krashen, VanPatten, Lightbown & Halter, Winitz and others clearly shows that delaying speech production is extremely beneficial for beginners.
Put all three of these assumptions together and what you get is a general approach which gives rise to linearly structured, gradualist, low-content syllabi which force early production at the expense of comprehension. The other thing which you get is an enormously high drop out rate as students vote with their feet.
I’m not saying, of course, that students on traditional, gradualist courses, never flip (a minority do). What I am saying, however, is that the majority of students on community language courses in the UK have usually given up on learning their target language before that flip happens.
Complexity Theory and Second Language Acquisition
In recent years there has been a small but growing shift to treat language itself as a complex, dynamic, adaptive, self-organizing system, appropriating ideas and concepts stemming from related fields such as Chaos / Complexity / Dynamic Systems Theory to describe language development and usage.
The main instigator of this shift, especially with regards to SLA, has been Diane Larsen-Freeman who in 1997 published a paper entitled 'Chaos / complexity science and second language acquisition’. In collaboration with others, she has since elaborated upon those early, tentative ideas in subsequent publications and by now the general idea of language as a complex, dynamic, adaptive, self-organizing system – nested within a hierarchy of other complex, dynamic, adaptive, self-organizing systems (the human brain, the human body, discourse partners, broader speech communities etc.) is rapidly gaining credence.
I outline below some of the main tenets of Complexity Theory.
According to Complexity Theory complex systems are very different from simple systems in that they are continually changing, self-organizing and adapting in response to information from the environment and from within the system itself.
All the elements in a complex system are interconnected and the development of a particular system is heavily influenced by the initial conditions at its inception. Due to the enormous amount of variables affecting complex systems it is nigh on impossible to precisely predict their short term behaviour (unlike simple, non-dynamic systems).
With simple systems (such as traffic signals) change is linear, regular and predictable. With complex systems, on the other hand, change is heavily influenced by a particular system’s history and it is non-linear and sigmoidal.
What this means is that complex systems can sometimes undergo gradual changes but at other times they experience sudden bursts of enormous change. These sudden bursts are called phase transitions (or phase shifts) and they are a core aspect of a complex system’s evolution.
A phase transition occurs when something changes suddenly from one discernable state into something markedly different. It occurs in the physical sciences, in the biological sciences and in the social sciences.
Oft cited examples are : when water boils or freezes; in superconductivity, when electrical resistance in mercury suddenly drops at close to absolute zero; when an avalanche suddenly occurs; when certain species rapidly displace others in an ecosystem; when economic markets rise or fall dramatically.
. The central point of my thesis (and a point underlined in the literature on Complexity Theory and SLA) is that the linguistic flip which I and countless others have experienced as part of our second or foreign language acquisition processes could be quite accurately described as a phase transition or phase shift.
It’s all well and good, however, to reflect on the fact that phase transitions happen in a wide variety of contexts, especially in complex, dynamic systems (including SLA). But the interesting question for me initially was why do they occur. Having spent some considerable time ploughing through some of the more accessible pieces of literature on the subject (for a lay person like myself, at least) it would appear that the reason for the sudden, dramatic leaps in self-organization regularly seen in complex, dynamic systems is the size (or bulk, or volume, or number of interconnected agents) of the network itself. This message turns up again and again. Indeed, some experts would dismiss my question as being rather puerile in the first place.
The answer to complexity turns out to be fairly obvious and not, in itself, especially interesting : If you have a lot of simple interactors, and let them interact, then the result can be rather complicated.
The core message would seem to be this. You simply cannot have complexity and you cannot have a phase transition unless you have enough agents tied into the system. Once you do have an increasing level of bulk in the system, however, you will eventually reach a critical mass, a threshold, a bifurcation point, after which the great leap to a different kind of behaviour occurs.
This idea has filtered through to popular consciousness in recent years and has entered common parlance as ‘The Tipping Point’. (The phrase was initially coined by Morton Grodzins in 1957 and was used by Thomas Schelling in his influential 1971 paper “Dynamic Models of Segregation”. It was further popularised in 2000 following the publication of Malcolm Gladwell’s book of the same name.)
Below is a description of a visually dramatic event (the violent wobbling of the then newly erected Millennium Bridge in London) which underlines this point quite succinctly (it’s an event which you can actually watch if you go to :
http://www.youtube.com/watch?v=gQK21572oSU
The bridge opened to the public on a sunny Saturday, June 10, 2000. As soon as police gave the word, hundreds of excited Londoners surged onto the deck from both ends. Within minutes it began to wobble, 690 tons of steel and aluminium swaying in a lateral S-shaped vibration like a snake slithering on the ground. Alarmed pedestrians clung to the handrails to steady themselves but the wobbling grew ever more violent, ultimately reaching deflections of 20 centimetres from side to side.
The violent wobbling turns out to have been the result of a phase transition brought about by the unconsciously synchronised movements of the large number of walkers on the bridge on the day in question. It was simply the case of the numbers on the bridge having passed a critical threshold. And it could, theoretically , happen again.
Once the crowd is large enough, there’s a chance that at some stage, enough people will step in sync by accident that a critical threshold will be crossed and the bridge will begin to wobble slightly. Once that happens, the feedback effect kicks in and reinforces the swaying…….there’s no trouble if there are fewer people than the threshold number. It’s not as if the bridge shakes a little for a small number of people and gradually builds up as the numbers increase. Either it doesn’t shake at all, or it wobbles violently and without warning, once the threshold is crossed. Like the straw that breaks the camel’s back, the onset of wobbling is a nonlinear phenomenon.
Another phrase from the study of complex systems which seems to have crept into common currency over the years is More Is Different.
It was originally coined by the Nobel Prize-winning Physicist P.W. Andersen as a title for a paper he published in 1972. In his book Emergence Steven Johnson connects this “old slogan of complexity theory” to the behaviour of ant colonies as follows:
More is different………..the statistical nature of ant interaction demands that there be a critical mass of ants for the colony to make intelligent assessment of its global state. Ten ants roaming across the desert floor will not be able to accurately judge the overall need for foragers or nest-builders, but two thousand will do the job admirably.
But going back to phase transitions and language flips, we actually should be a bit more precise if we want to label one phenomenon in terms of the other. The reason I say this is because there are actually two kinds of phase transitions, first-order and second-order. First-order phase transitions are either-or situations – either water is water or it’s ice (or steam). Second-order phase transitions are slightly different, they refer to situations whereby a system changes from a state depicted as islands of order within a sea of chaos/disorder to its mirror image, namely, island of chaos/disorder within a sea of orderliness.
It is possible that SLA flips are sometimes more like second-order phase transitions than first-order ones, whereby a learner having a few disconnected pockets of linguistic ability in a specific target language (e.g. me and my sparse Bulgarian and Greek) suddenly evolves into a speaker who has a rough command of a second or foreign language, even though there are still gaps in his or her knowledge (holes within the whole, so to speak). (This would probably be a fair description of my abilities in Breton, Spanish and French and, to a much lesser degree, Catalan.)
Whether specific linguistic flips are more like first or second order phase transitions could be a matter of debate but it would seem that one of the key factors determining whether phase transitions of any kind are first or second-order is the size of the network experiencing the phase transition, as explained by the complexity biologist Stuart Kaufman. He is talking here about the steepness of a Sigmoidal or S-curve which depicts, by means of a graph, the phase transition which occurs in a simple network model which he had designed. (Sigmoidal curves are curves which are relatively flat and horizontal at both their lower and upper ends but can display a sudden, narrow, almost vertical leap in the middle.)
All his model entailed was a varying number of buttons with threaded connections made between single buttons one at a time. What he found was that as the ratio of threads to buttons increased small connected clusters began to form. Then as the ratio passed the 0.5 mark all of a sudden most of the clusters became cross-connected into one giant structure.
(U)sing 400 buttons the sigmoidal curve rises steeply when the ratio of edges to nodes passes 0.5. The steepness of the curve at the critical 0.5 ratio depends on the number of nodes in the system. When the number of nodes is small, the steepest part of the curve is “shallow,” but as the number of nodes in the toy system increases – from, say, 400 to 100 million – the steep part of the sigmoidal curve becomes more vertical. Were there an infinite number of buttons, then as the ratio of threads to buttons passed 0.5 the size of the largest component would jump discontinuously from tiny to enormous.
The connection between volume/size and phase transition could be underlined by looking at studies from various fields. I have opted for this medical study of physical response to drug doses which can be found at
: http://www.curvefit.com/introduction89.htm
What you can see from this particular graph is that physical response to drug dosage is not a linear affair, it is in fact sigmoidal. You can give small doses of drugs to patients and no response is seen. Then suddenly, once the dosage reaches a certain threshold, response increases rapidly, only to reach a plateau once you pass the optimum level of dosage.
(The dose-response curve above actually looks very similar to proficiency/vocabulary acquisition curves and phonetic change curves included in Larsen-Freeman and Cameron 2008).
Coming back to language, according to de Bot, Lowie and Verspoor (who have developed a Dynamic Systems Theory approach to Second Language Acquisition) volume of input does indeed greatly affect SLA.
What we can gather from almost all studies and theories (also those not compatible with DST) is that the amount (my emphasis) of meaningful input is of crucial importance in the acquisition process.
(Research done by Bates and Goodman in first language acquisition has shown, interestingly, that very young children first go through a vocabulary burst and that this is then followed, a few months later, by a grammar burst. They also found that the size of a child’s vocabulary at 20 months is a very strong determinant of the grammatical complexity of its language at 28 months.)
One of the most well known and influential proponents of the idea that input size (and complexity) drives SLA is, of course, Stephen Krashen. From the end of the 70s onwards, as part of a broad 'natural’ approach to SLA, he has argued for increasing the amount of comprehensible input given to students. In their discussion of his various hypotheses de Bot, Lowie and Verspoor state : “Compatible with DST is the heavy emphasis on input, the idea that the learner needs to be exposed to things he does not know yet in order to progress.”
According to Krashen optimal input should be comprehensible, interesting / relevant, not grammatically sequenced and provided in sufficient quantity.
Since the aim of my talk is to link Suggestopedia with Complexity Theory (and in particular with the concept of critical mass as a catalyst for linguistic phase transitions) it is interesting to note that in his summing up of the strengths and weaknesses of a range of teaching methodologies and approaches Krashen states that it would appear that : “Suggestopedia comes very close to completely matching the requirements for optimal input.”
Keeping all the above comments in mind let us now turn our attention to Suggestopedia.
Suggestopedia (Reserve Capacity Communication)
Suggestopedia has, historically, had a mixed press. Not because of its detractors but, rather, because of the somewhat fanciful descriptions put around by some of those claiming (without a solid enough grounding in the method) to be genuine aficionados.
Some of the descriptions have been rather mystical when actually the method is most eminently rational and reasonable (and those elements which may appear strange at first to Western Europeans have more to do with a mix of Soviet psychology and Bulgarian culture than New Age mysticism). Let me therefore give you a brief outline of the core concepts linked to method and a short description of how it is utilised.
1.1 Suggestopedia starts with the assumption that an individual’s normal, cognitive capacities are greater than is recognized in most if not all societies. Dr. Lozanov claims that people in general possess untapped reserve capacities of mind and that all of us could learn to absorb and creatively use far more information than is considered normal.
1.2 The problem, according to Dr. Lozanov, is that we are held back by the results of personal past conditionings and collectively held self-limiting beliefs which he calls Social-Suggestive Norms. These beliefs harden into psychological barriers which need to be overcome if people are to be enabled to achieve their true potential.
1.3 Rather than trying to tackle these beliefs head on through rational discourse Dr. Lozanov suggests 'harmonizing’ with them (in effect, side stepping them) and leading individuals through a transformative, self-liberating experience via education.
1.4 The method he developed for this purpose is Suggestopedia, which uses a combination of aesthetic, psychological and didactic means (you could add affective but I tend to include that in the term psychological).
1.5 As a method Suggestopedia targets both the conscious and non-conscious mind, puts great stress on affective as well as cognitive aspects and makes deliberate and creative use of teacher prestige to achieve its stated aims.
1.6 It is not a holistic method since it stresses both the global and the detailed aspects of the material presented to students. Dr. Lozanov refers to it as a “Global-partial, Partial-Global” method.
I list the most important procedural aspects of Suggestopedia below (but for a fuller account go to : http://dr-lozanov.com/en/en1.htm . For a copy of Dr. Lozanov’s original presentation to UNESCO go to :http://unesdoc.unesco.org/images/0003/000300/030087eb.pdf .)
Suggestopedic courses are structured according to cycles, each cycle linked to a very large act (supplied with translation) of a specially produced play which the students study. Each act is worked through as follows:
Lesson 1
Introduction – Students adopt new fictional identities
Preparation – The bare outlines of the story are presented by the teacher using props.
Concert Sessions – Reading the act twice to the background of music
Lesson 2 – 5/6
Elaborations – activities (games, songs, dialogues) based on the language presented in the act. Attention given fleetingly to grammatical aspects which have been presented peripherally on wall charts.
Production – Students act out parts of the act.
New act introduced.
Some commentators have over the years sought to pick out the essential and non-essential aspects of the method, isolating this or that particular procedure. In this respect they are very much like the blind men and the elephant because, by doing so, they fail to see the interconnectedness of all the different aspects. They also, crucially, fail to notice the size of the beast.
Suggestopedia and Volume of Input
From a purely objective standpoint, of course, the most obvious aspect of Suggestopedia, the aspect which truly distinguishes it from other methods, is the sheer enormity of the input introduced in a Suggestopedic class.
If that input were not ridiculously large and complex compared with traditional methods then, quite frankly, it wouldn’t be Suggestopedia.
To give you an idea of what I mean by ‘ridiculously large and complex’ compare lexical input size and complexity on traditional, 'normal’, community education language courses in the UK with Suggestopedia.
On normal courses input would be limited to usually no more than some 30 new words in the first lesson and one or two grammatical structures. This level of input would tend to be repeated every lesson for a year at least. On Suggestopedic courses, however, you would be introduced to some 800 (or more) new words in the first lesson and practically all of the salient grammar of the target language!
The voluminous nature of Suggestopedia was specifically underlined for me by Dr. Lozanov himself in two comments he made at the end of a training event which I attended in Bulgaria two years ago. The first comment which stuck in my memory was this :
S
uggestopedia is not just a 'happy’ method. If you do not introduce a very large amount of material on your courses you will be strengthening the (self-limiting) Social-Suggestive Norms. Always teach on the level of the Reserve Capacities of Mind.
And the second :
If your students are having difficulty in absorbing the material, give them more.
From a traditional perspective that would sound like a very ill advised strategy indeed. But, as we have learnt from Complexity Theory – More is Different.
(In many ways, Suggestopedic cycles parallel the creative process as outlined by the French mathematician Henri Poincaré – Saturation, Incubation, Illumination, Verification. Students on
Suggestopedic courses are saturated with input (both conscious and non-conscious) in the first lesson of each cycle; they then have an incubation period between the first and second lesson; following this they start to bring forth the language from their subconscious in the second lesson of the cycle; finally, at the close of each cycle, they are given the opportunity to verify their own acquisition in the production stage.)
To summarize, therefore : Over the years Suggestopedia, as a fully integrated, multi dimensional methodology, has enabled my colleagues and I :
a) to help learners get huge volumes of linguistic ants into their system, to absorb their target language (in my case, Welsh) 'joyously and spontaneously’, as Dr. Lozanov would say, without creating any kind of distress or strain.
b) to nourish and support adult learners until they, of their own accord (and because of the critical mass of language absorbed), suddenly flip.
Each student reaches his or her own personal Tipping Point (or Flipping Point) in his or her own time, of course, but when it does happen, as predicted by Complexity Theory or Dynamic Systems Theory, production skills suddenly erupt (as they did with me outside the night club in Cardiff).
Based on my own experiences (and some well documented research) compared with students on traditional language courses far more students on Suggestopedic courses succeed in achieving their flips and they do so :
a) in a much, much shorter period of time than would usually be expected - and crucially
b) before they have been tempted to give up on their attempts at language learning.
And that is the reason why I, as a language teacher and as the Director of a fairly big langauge school, am of the opinion that Suggestopedia is indeed superior or more effective than other methods which I have hitherto used.
1 Johnson (2001 : 115)
2 Kostova (2006 : 187-88)
3 Krashen (1987) ; Lightbown & Spada (1999 :128-35) ; Winitz (Ed.) (1981)
4 &