Diane Larsen-Freeman is a Professor of Applied Linguistics in the Department of Language Teacher Education at the School for International Training in Brattleboro, Vermont. She received both her M.A. and Ph.D. degrees in Linguistics from the University of Michigan, and her B.A. in Psychology from the State University of New York. She is the author of many books and articles on second language acquisition research, English grammar, language teaching methodology, and teacher education. She has recently written an article titled "Chaos/Complexity Science and Second Language Acquisition," which will appear in the summer 1997 edition of Applied Linguistics. She was interviewed about this and related topics at the 1997 TESOL Conference by Craig Sower.
Sower: How did you get interested in chaos theory?
Larsen-Freeman: It was really happenstance. About 5 years ago a stranger in a bookstore at the Boston Museum of Science thrust into my hands the book Chaos by James Gleick and said it was a good read. I bought the book, but it sat on my bookshelf for a couple of years. Then one day I picked it up and found it very compelling reading indeed. Gleick tells a good story, he writes well--very accessibly. But more striking than the accessibility was the fact that I kept encountering themes that reminded me of issues in language acquisition that I had been wrestling with for some time. As I read more, I began to realize that the behavior of systems that are studied in chaos/complexity science provide, at the very least, metaphors for us in applied linguistics.
What is chaos/complexity science?
It is the study of dynamic, complex, nonlinear systems--systems that heretofore scientists knew about but didn't pay much attention to because they didn't have the tools with which to study them. With the advent of supercomputers and the tremendous potential for multivariate analysis they afford us, scientists have been able to do a lot more rapid number crunching, model these systems, and learn something about their behavior.
To a lot of people chaos/complexity science may seem somewhat remote from second language acquisition. What connections do you see between chaos/complexity theory and second language acquisition, and what are the implications both for teachers and learners?
Well those are two huge questions, but I'll try to be brief. My answer to your first question is the theme of a paper I delivered at the Second Language Research Forum in Montreal in 1994, a revised version of which will appear in the journal Applied Linguistics later this summer.
Again, very simply, the systems studied in chaos/complexity science are dynamic, complex, and nonlinear. We can certainly attest to the fact that second language acquisition is dynamic. The change in learners' interlanguage is often so rapid that attempts to describe grammars of interlanguages fail because they are far too categorical and static. Further, interlanguage is complex in two ways. Normally when we say something is complex, we mean it has a lot of parts. But there is another meaning to complexity in chaos/complexity theory: not only do complex systems have a number of parts, but the parts interact such that the behavior of the system as a whole emerges out of the interaction of its parts. I think complexity in both senses applies to second language acquisition. Interlanguage is a system that changes over time and one that involves many moving, interacting parts. A change in one part of the interlanguage system has an impact on other parts.
Finally, a linear system is one in which an effect is proportionate to its cause. In a nonlinear system, the opposite is true. You can have a big stimulus and get just a little response, or conversely, you can have a small stimulus and get a tremendous response. It seems to me language acquisition is nonlinear in the same way. On the one hand, we can see language study going on for a long time with little or no apparent return. On the other hand, we can sometimes see the penny drop after just one more lesson, or one more time that a rule is presented, or learners take in one more positive instance of a target form. "Eureka!"--the interlanguage system is convulsed into chaos. Because complex/chaotic systems are also self-organizing, eventually the disorder or chaos subsides, and systematicity is restored, a process in second language acquisition we call restructuring.
So just these three characteristics alone--dynamic, complex and nonlinear . . . and there are many others--suggest to me that there are a lot of parallels between the complex natural systems that physical scientists study and the study of second language acquisition. I think you're often helped in looking at your own field by examining what is going on in another field and seeing if there are insights you can gain.
As to the second part of your question, having to do with its implications for teachers and learners, I understand that this is the very issue that a group of my former SIT students and others are discussing here in Japan right now. I think they are intending to share some of their thoughts at the JALT Conference next fall. Anyway, it seems to me that if chaos/complexity theory enhances our understanding of the SLA process, then there must be implications for the language teacher/learner. For example, maybe recognizing that the learning/acquisition process is nonlinear will encourage teachers who might otherwise feel disheartened by the fact that their students seem to be making such slow progress. I feel certain that there are implications at this level for the teacher/learner, but since I have yet to work out all the details, I prefer to use my understanding of chaos/complexity theory at the moment to call attention to some of my pet peeves in second language acquisition.
Such as?
Well, one of my pet peeves is SLA research on affective variables which uses simple univariate analysis. Researchers study one phenomenon, see if it correlates with or predicts another through parametric statistics, and then decide if we can or cannot call it a causal variable. While I am sympathetic to the need to come up with a research focus that is manageable, I don't think we're going to find out about second language acquisition by focusing on one little piece at a time and trying to build up our understanding piece by piece. I think chaos theory tells us we can't isolate just one piece and try to understand the whole because the whole is not to be known in a piecemeal fashion. The whole reveals itself through the interaction of its parts.
So you think reductionist efforts are inherently flawed?
Yes I do. I think chaos/complexity theory argues against reductionism. Since the process is complex, it is foolish to think that we can learn about it a piece at a time--that we can add them up somehow and at the end we'll have the puzzle figured out. I don't think it's going to work that way. Of course, we can't study the whole of SLA anymore than we can any other complex phenomenon. Thus, it seems to me that we need to think about a new unit of analysis--perhaps, as in Vygotsky, studying a microcosm in order to understand a macrocosm.
Equally simplistic, and therefore another pet peeve of mine, is an untheoretically motivated pretest/posttest experiment in which subjects are pretested, undergo a treatment, and then are post-tested to determine what the learning outcomes are. It is then assumed that the outcomes will tell us something about the effectiveness of the treatment. A whole stream of research in the second language acquisition literature is based on that kind of study.
Chaos theory suggests that studies of this sort are flawed. As I mentioned earlier, the whole system can be thrown into great convulsions, or chaos, by the simplest of stimuli. For example, students can encounter "-ing, -ing, -ing" over and over again; the teacher can go on about it, the students can read about it and do exercises supplying it, and still the students don't use it. They say "I am study," for instance. Then one day they might hear it one more time and the penny drops and they're able to use the "-ing" accurately, meaningfully, and appropriately. To attribute the learning to the proximate cause is faulty. Similarly, to say that students didn't learn from previous treatments is also spurious.
So I have a lot of concern about research that suggests that we can measure the effectiveness of one-time treatments based on measuring one-time learning outcomes. How to define learning is a serious problem in SLA research--as it is in language teaching. As I said before, I haven't yet gotten very far in thinking what this all means for pedagogy, but it seems teachers too would have difficulty in assessing when something has been learned. What I can say at this point to SLA researchers is, "Wait a minute, we have some problems here in the way in which we conceive and carry out our research."
You seem to be arguing for deductive rather than inductive processes; that it's not possible to prove a thesis, but it is possible to falsify a thesis. Is that fair?
Yes, I can see why you would say that. Each encounter with a phenomenon would simply be one more bit of evidence for it, but all conclusions would have to be provisional, subject to a disconfirming example.
In your article in Applied Linguistics, you wrote of "the camel's back effect," that you add that one last straw and . . .
In talking about students I'd rather not say "break the camel's back . . . "
Your dropping penny was a nicer metaphor . . .
Yes, you don't know in chaotic systems, as you don't know in the development of interlanguage, for example, which penny it will take to cause that great upset, that great restructuring of the system. Nevertheless, it is at that point that creative growth is possible. In this regard, I like the chapter title Waldrop uses when he discusses what happens when the system convulses: "Life at the Edge of Chaos."
Given that the systems studied in chaos/complexity science are dependent upon, or sensitive to, initial conditions, similar perhaps to individual differences among learners, what does this imply for centrally planned programs?
Again, although I am not entirely comfortable with applying chaos/complexity to language pedagogy at this point, I could speculate that central planning would play a role in establishing initial conditions, but that much of what is important after that would operate locally. Central planning can't tell us when the penny is going to drop for you, compared to me or anyone else. I think that's one of the reasons I like biological metaphors. Organic processes suggest to me that teachers cannot use a cookie-cutter approach; we can't mechanize teaching. I used to think efficiency experts were really wonderful, and I still think we need to become as efficient as we can in the classroom, but I don't think efficiency necessarily means centrally-planned mass-production in education. That's the disconnect for me.
So is this where the art of teaching comes in?
Yes, it would be ironic if an archetypical science forces us to conclude that teaching is an art, wouldn't it? I am not sure that I would want to conclude that. At last year's TESOL Convention in Chicago, I participated in the closing plenary event entitled "The Great Debate," organized by David Nunan. Elana Shohamy, Henry Widdowson, Dick Tucker and I were to debate the proposition that teaching is an art, not a science. Of course it is a false dichotomy, but an interesting way to get the issues on the table. Anyway, I was asked to argue that teaching is a science. The way I argued for the science of language teaching last year was to say that teaching, like science, is best served by individuals who possess an attitude of inquiry, who are curious about their students and about the whole teaching/learning enterprise. Curiosity, knowing that we don't have all the answers, is humbling and also motivating. It encourages us to continue to ask questions about what it is we do. However, now that I don't have to worry about debating Henry, I would admit that an attitude of inquiry is probably indispenable to art as well. In either case, what I think serves our learners best is when a teacher is really present, reading students, trying to figure out the next thing to do to unleash that learning--to throw the system into initial chaos out of which will emerge or self-organize a system that is more in alignment with the target language. So does it take an artist to perceive that or a scientist who is carefully monitoring his or her students? I don't know.
Well, that leaves us with good questions. Thank you very much.
References
Gleick, J. 1987. Chaos: Making a new science. New York: