The Neuroscience of Task-Based Language Teaching
By: Curtis Kelly
Among the Best of 2022
Readers chose this article as one of the best from all of 2022! Check out the other nominees with the button below.
The more I learn about task-based teaching, the more I realize that it is especially suited to the brain. But how? I could think of many reasons why Task-Based Language Teaching (TBLT) fits discoveries in neuroscience, but none that stood out as being specific to TBLT…until, that is…. a month ago. That’s when I found a particular theory that explains the power of TBLT perfectly: The Killer Theory.
So why is a connection to neuroscience so important? Neuroscience is highly relevant to our field, language learning, but that is not always obvious. Indeed, language and learning are both basic functions of the human brain, so a teacher not knowing how the brain works is like an electrician not knowing how electricity works. That electrician might still be able to do basic wiring, but would risk doing harm if going off blueprint. The same holds for teachers (which is obvious in the poor wiring of many textbooks). Brain studies might not revolutionize the way we teach–experience has helped us develop many brain-friendly ways of teaching–but it still helps if we learn the reasons why some methods work better than others without having to electrocute a few learners in the process. And new findings that we should learn about are coming out almost every week. So let us learn.
As for TBLT, it is particularly impoverished in discussions involving neuroscience. I have only found a few articles on TBLT that even mention the brain, and fewer that include recent neuroscience,[1] still a foreign language in our field. Twice, when I was invited to write chapters for books on TBLT, I proposed writing about the neuroscience behind it, and this topic was rejected both times. That was sad, because so much written on TBLT involves speculating about why it is so effective, why students enjoy it, and how it surpasses traditional ways of teaching, but these discussions seem to be observations on the wiring, not the underlying electricity.
[1] You might be wondering what I mean, since numerous articles talk about uptake, short- and long-term memory, acquisition, etc, but these are all “pre-schism” concepts: established long before the nineties when we literally became able to see how the brain works. They are not what we consider neuroscience.
The Output Hypothesis and Neuroscience
A good example is Merrill Swain’s output hypothesis, a widely debated theory in the eighties and nineties. The star child of the day was Krashen’s input hypothesis which implied output-oriented TBLT was going down the wrong road.[2] Swain and others argued that output was still important because it helped learners become aware that their language knowledge is flawed, thereby creating language needs. Those needs then fuel the acquisition of the related forms (Swain & Lapkin, 1995). It was a lovely theory that made sense to anyone who has ever tried to learn a language. So, I do not mean to detract from it, but like most theories in language teaching, it was built on observation and conjecture about the brain, rather than an understanding of the actual processes identified by neuroscience. And so, the output hypothesis was vulnerable to refutation, with Krashen the main critic (2001, Chapter 3).
[2] Long (2010) argues that the claim TBLT has an emphasis on output (and thereby interaction) is a fallacy, but, from my experience, at least in Japan and other Asian countries where English is an L2, the assertion holds.
Neuroscience might not be able to settle the controversy, yet once we understand how almost all cognition, including language, is predictive, with a focus on learning from prediction error, we get a richer picture of the main tenet of the output hypothesis, that errors lead to learning. Predictive processing tells us that errors–where our predictions do not correlate with incoming information–are the only things we pay attention to. Prediction error leads to dopamine release and rewiring, i.e., learning. Understanding this process helps ground the output hypothesis (electrically?), because it reveals the actual physiology behind how language needs, focus on form, and negotiation of meaning lead to increased acquisition. (See our Think Tank on Predictive Processing for more.)
Task Types and Neuroscience
Another long-standing argument in TBLT is about the degree to which classroom tasks should reflect what students are likely to experience in real encounters. This concern seemed to evolve out of an early recognition of the classroom tasks as “pedagogic” and the outside experiences as “real-world” or “target”) tasks (Nunan, 2010). Nunan also separated pedagogic tasks into two types: “rehearsal tasks” that mimicked probable real-world situations (such as filling in hotel registration forms) and “activation” tasks designed for general language practice, but which are less likely to happen in the real world (such as comparing pictures to find a difference). Some experts, like Long, seem to favor rehearsal tasks, with a touch of disdain for “English for nebulous purposes (ENP)” (2016, para 49), while others, such as Bygate et al. (2015), seem accepting of both. A few of the TBLT textbook authors I have interacted with have stronger opinions, arguing for as much “real-worldliness” as possible.
So, are rehearsal tasks superior in preparing learners for the outside world? Are activation tasks too far from reality to be of value? How much real worldliness should we strive for? At the heart of this argument is the thorny issue of learning transferability (Perkins & Salomon, 1992). To what degree can the basic processes learned in one unlikely–but probably more fun–type of language practice (such as planning a mission to Mars) be used in potential future real-world situations (such as planning a sales campaign)? Various studies have found cases of both good and poor transfer of learning (Haskell, 2000), making it hard to compare the value of rehearsal vs. activation tasks.
While, again, neuroscience might not settle the dispute, knowing how the brain does transfer of learning gives us a good start. In particular, knowing that all learning occurs through transfer helps. A brain is particularly good at two things: committing all its resources to a single task, and assembling neural teams to do so (Anderson, 2010). In the last two decades, we have abandoned the view that each brain area has a single function, now understanding that they are called up for various kinds of service. For example, when comparing number sizes, the brain calls on a particular neural area that normally keeps track of our fingers, the Intraparietal parietal sulcus in the motor cortex. Since this bit of brain is good at keeping track of things, it is recruited for other tasks that require its organizational skills, such as comparing things.
The more times we recruit particular neural teams to accomplish a task, the stronger their connections become, and that is what learning is all about. Basically, then, the brain is solving new problems by using parts that have previously had other functions. This, in itself, is transfer. So, it makes sense that the skills gained through tasks that are not real worldly can be transferred to new problems. Transfer of learning is a brain basic. What matters is the distance involved: Knowing how to order a ham sandwich transfers easily to ordering a salad, but not as easily to ordering new copy machines for a company.
Again, neuroscience gives us a new perspective on an old problem. And there is more.
Emotion and the Social Brain in TBLT
One of the advantages to using tasks to teach language is that they are inherently fun. People enjoy solving problems and, since most of the TBLT activities involve interaction, the socializing makes it fun, too. Is this just a special bonus we get from TBLT, one that keeps students engaged, but otherwise has little to do with language learning?
We now know that the enjoyment is at the heart of learning. From the perspective of neuroscience, fun is a signal. It is the brain’s way of telling us that this experience is good for us and we should remember it and do it again. That is why dopamine, the neurotransmitter that gives us reward (fun), also causes motivation (wanting to do it again), and increased synaptic connections (learning). Our brains are built to remember things that make it feel good. And even though it might not be the language itself that makes the activity fun—it’s the look of glee on your partner’s face—the brain is still primed to learn everything connected to that experience, including the language. (See Think Tank on Emotion.)
In fact, according to Immordino-Yang, we can only learn things that have emotional valence, things that are meaningful to us:
[It is] literally neurologically impossible to build memories, engage complex thoughts or make meaningful decisions without emotion. (2016, para 1)
So, the fun in TBLT is more than just fun. It is the precursor for deeper learning.
And, as we are coming to understand, the socializing in TBLT plays an important role too. Socializing uses the recently discovered mentalizing network, a part of the social brain. This network allows us, and drives us, to understand others. In a fascinating TED Talk, neuroscientist Matthew Lieberman calls the social brain our superpower, one that has great potential for learning as can be seen in pairwork and tutoring. But we have a “kryptonite” too: education. Since most of education has students studying things on their own, just using the analytical brain, we lose the learning benefits made available by the social brain.
Therefore, the socializing often present in TBLT too, is not just a frivolous add-on. It also instigates deeper learning.
Going back to what I wrote at the start, I understood that TBLT had some deep connection to how the brain learns, but I could not put my finger on any killer theory in neuroscience that truly fits TBLT. In the sections above, I touched on predictive error processing as our brain’s basic way of understanding the world, neural reuse as the basis of learning, how emotion steers learning, and the learning potential of the social brain. However, these theories fit almost every kind of language study. So, what did I discover a month ago that is so particular to TBLT?
The Killer Theory: Cognitive Control
The hallmark of task-based learning is the use of tasks. As other contributors to this issue have informed us (also see Ellis, 2019), tasks are problems to be solved, usually filling some kind of knowledge gap between learners. The key points are that the learner’s primary goal is completing a task, and language is just a tool for doing so. Language, then, is not the object of direct study as in other approaches, it is the conduit.
In other words, TBLT bridges the gap between knowing… and doing.
Knowing to doing, a gap we tend to gloss over in most areas of education. Even now, most teaching is oriented towards banking information in learners’ heads, so that someday they’ll have it handy in case they need it. We live in this fantasy world where we assume that just knowing equates to proficiency at using, or doing. And yet, we have all heard about those English students who have read all of Dickens, but sputter in the simplest interaction.
TBLT is different in that it assumes that the knowing is already there, and wholly shifts the emphasis to the doing (correctly recognizing that doing enhances the knowing as well). It is L2 internship.
From the perspective of neuroscience, closing the knowing to doing gap is actually training a hugely important part of the brain, located in the pre-frontal cortex and elsewhere, whose function is cognitive control. Cognitive control is the system that we have for orchestrating our goals into actions; it is the way we get things done. As neuroscientist David Badre tells us:
…knowledge and action are distinct things to some degree. So, knowing is not enough. You have to be able to bridge from what you want to do to how you behave. And that gap is not trivial. It’s not easy. You actually need a class of functions in the brain to bridge it. And that’s what cognitive control is all about. (Brain Science Podcast, 2021)
This is the killer theory I discovered a month ago. I listen to this podcast in which David Badre talks about cognitive control while I was taking a walk in Kyoto. His interview literally stopped me in my tracks. The brain function he
described was so obvious, and yet, so elusive. The concept has been around for a while (Miller, 2000) and we were able to identify many of its components years ago, executive functions, but it was hard to get the big picture of how these worked together.[3] Badre explains why. The context of cognitive control is so open-ended that it is hard for neuroscientists to study—how does one research things like the planning you might do on trip to find a good place to have lunch? Nonetheless, it is becoming quickly recognized as an important area of research in neuroscience, and hopefully, with time, education will recognize its importance as well.
After all, TBLT and medical training[4] have. And now we have this whole new field of study in neuroscience that validates these non-traditional, hands-on approaches. In a way, neuroscience tells us that TBLT is not just an alternative, it is mandatory, especially in EFL environments (help with acronyms here). It is the keystone in the arch that connects linguistic knowledge to communicative competence.
Hearing Badre a month ago finally answered a crucial question I had been wrestling with for most of my life: why task-based approaches are so important in education. With my students, that particular kind of wiring seemed to work well, but I could not conceive how the underlying currents flowed. Now I can. And I hope my new understanding of the electricity energizes yours.
[3] Wikipedia does not even have a separate page for “cognitive control.”
[4] In the sixties, the gap between knowing and doing in medical schools was recognized as actually being deadly. The doctors they produced tended to kill and maim their patients in the process of gaining hands-on experience. Howard Barrows solved this problem by moving “hands-on” directly into the medical curriculum, using a problem-based approach with actors and simulations. Medical students were given fictious patients to care for: listening to their ailments, giving them tests, and determining the best treatments (1996). This problem-based approach is now being used, to some degree, in most medical schools.
Curtis Kelly (EdD.), a founder of the JALT Mind, Brain, and Education SIG, is obsessed with gaps these days.
This is a spam to me, not to be left out