In the fourth run of the SLB course on Task-Based Language Teaching, which starts in late October this year, we’re very lucky to count on the collaboration of Peter Skehan, whose work earned him, along with Mike Long, the IATBLT’s Distinguished Achievement Award in 2017. His more recent work includes
- Skehan (2018) Second language task-based performance: theory, research, assessment. Routledge.
- Ellis, R., Skehan, P., Li, S., Shintani, N., & Lambert, C. (2019). Task-Based Language Teaching: Theory and Practice (Cambridge Applied Linguistics). CUP.
There’s also this collection of papers in his honour:
- Wen & Ahmadian (2019) Researching L2 task performance and pedagogy: in honour of Peter Skehan. Benjamins.
CAF & CALF
Over the last thirty years, Skehan has conducted research (often with Pamela Foster) which focuses on how various factors affect the complexity, accuracy and fluency (CAF) of students’ L2 production (mainly speaking) when performing tasks. We should note first that structural complexity and lexical complexity are now treated separately by Skehan (hence CALF), and second that, although there are generally recognized ways of operationalizing these constructs (e.g., structural complexity is usually measured in terms of subordination, in the form of the total number of clauses divided by the number of AS (Analysis of speech) units (Foster et al 2000) and accuracy is measured through the percentage of error-free clauses) all of them are open to a wide range of interpretation.
The Trade-off Hypothesis
Skehan claims that, due to the limitations of attentional resources and working memory, when students are performing tasks there is competition between complexity, accuracy, and fluency for attentional resources, leading to a “trade-off” between them, such that one is prioritized at the expense of the others. Since meaning normally takes priority when students perform communicative, meaning-based tasks, fluency is usually given priority, and thus complexity and accuracy suffer. Furthermore, even when there is attention available for form, there is a still further tension between attention directed to complexity and attention directed to accuracy. To this, we need to add a consideration of interlanguage development.
Learners can be seen to be developing their interlanguages in terms of the increased complexity, accuracy and fluency of their production, and a second perspective on how the three areas of CAF inter-relate focuses on acquisitional sequence. Skehan and Foster (2007) offer Table 1, which captures how new language is developed and how, subsequently, greater control is achieved over this new language:
Table 1: A developmental sequences for complexity, accuracy, and fluency
|Complexity ↓||This represents new cutting edge and possibly risky language, and foreshadows growth in the interlanguage system|
|Accuracy ↓||This represents a striving for control and error avoidance, possibly by the avoidance of cutting-edge language, and by avoiding fluency to enable more time to be used to achieve higher accuracy|
|Fluency||This represents a focus on meaning, automatisation, lexicalization and a push for real-time processing|
Skehan sees the central challenge for task-based instruction as discovering how task characteristics and task conditions can be manipulated to produce performance which maximises complexity, accuracy, and fluency even though these three areas may enter into competition with one another. Through a series of studies, Skehan established that different task characteristics and different task conditions exert systematic influences on performance.
Various studies (most of them by Skehan and Foster) used three different types of task:
- Personal: a personal information exchange task where one student tells another student how to get to their home to turn off a gas oven which had been left on.
- Narrative: subjects had to construct a story based on a sequence of pictures.
- Decision making task: subjects had to role play a judge and decide on appropriate punishments for wrong doers.
Changes were made to these tasks in follow up studies in order to further explore the implications, but let’s stick to basics.
In terms of task characteristics, “The three tasks essentially opposed familiar with unfamiliar propositions, and clear structure for the information required with progressively less predictable structure and information” (Skehan, 1998, 108). It was found that
- Accuracy was significantly higher on the personal and decision-making task than on the narrative task
- The personal task produced less complexity and higher accuracy than the other two.
- A clearly structured task promotes accuracy and fluency.
- Tasks based on familiar or concrete information lead to greater accuracy.
- Fluency, and dialogic tasks lead to greater accuracy and complexity, while monologic tasks generally produce the reverse results.
Interesting in themselves, these conclusions are also important in supporting Skehan’s argument that accuracy and complexity can and should be separated when examining the effects of task characteristics. Pace Robinson’s Cognitive Hypothesis (Robinson, 2001a, 2001b, 2005), which Long’s version of TBLT relies on to some extent, Skehan argues that when accuracy and complexity do, in fact go together, they do so for reasons that can be better explained by his LAC approach than by Robinson’s Cognitive Complexity framework.
In terms of task conditions, Skehan and his collaborators looked at the effects of pre-planning and the conclusions are:
- pre-task planning consistently produces greater complexity and fluency
- pre-task planning sometimes produces more accurate language
- pre-task planning is more effectively done when led by the teacher, and least effectively done in a group of learners
These conclusions about task conditions lead us to a consideration of Skehan’s view of how tasks should be implemented.
Skehan’s view of how tasks should be implemented are summarized in Table 2 (Skehan, 1996).
The general purpose of the pre-emptive, or pre-task activities is to encourage restructuring of the students’ interlanguages – to incorporate new elements or re-arrange existing elements. These activities aim to make salient language which will be relevant to task performance. They can also aim to ease the processing load that learners face when actually doing a task, releasing more attention for the actual language that is used, so that more complex language can be attempted and greater accuracy can be achieved.
The main factor affecting performance during the task is the choice of the task itself, Tasks should not be so difficult that excessive mental processing is required simply to communicate any sort of meaning. Nor should tasks be so easy that learners are bored, and do not engage seriously with the task requirements, with the result that no gain is made in terms of stretching interlanguage or developing greater automaticity. Skehan considers that the mayor area for adjustment while tasks are being completed is in the area of stress (or communicative pressure) put on students. This can be manipulated by considering these factors
I’m afraid I can’t go into these here, but I hope they at least to some extent, speak for themselves.
Then there are the post-task activities. Skehan argues that learners’ knowledge of what is to come later can influence how they approach attention-management during an actual task. While a task is being done, the teacher needs to withdraw and allow natural language acquisitional processes to operate. But the danger is that communication goals will be so predominant that lexicalized communication strategies will become so important that the capacity to change and restructure, to take syntactic risks, and to try to be more accurate, will not come into focus as serious goals. Post- task activities can change the way in which learners direct their attention during the task by reminding learners that fluency is not the only goal.
In ‘Post 1’, Skehan suggests three general post-task activities: public performance, analysis, and tests. In public performance, learners are asked to repeat their performance, publicly, in front of an audience – the rest of the class, the teacher, or a video camera. Skehan, following Willis (2009), suggests that the knowledge while the task is being done that a task may have to be re-done publicly will cause learners to allocate attention to the goals of restructuring and accuracy. Analysis can involve the teacher or the students reviewing the performance, and tests can be whatever the teacher thinks is appropriate. In Post 2, the teacher decides what follow up work is required.
The important thing here is to emphasise Skehan’s use of something close to the Willis (Willis, 2009) framework, much at variance with Long’s (2015) approach to TBLT.
Let me finish this brief summary with a quote from Skehan (2007)
“First, it is assumed that the CALF [Complexity, Accuracy, Lexis, Fluency] categories can represent an acquisitional sequence, and so tasks which promote greater complexity are pushing for new language, while tasks which promote accuracy or fluency are supporting control of an existing interlanguage level. In this view, first there is destabilization, and then there is a concern for control (eliminating inaccuracy first, and then achieving fluency second). But second, and more fundamentally, LAC regards the task itself as having the important function of making some aspect (or aspects) of language salient. The teacher records what language has been made salient in this way, Then it is assumed that important acquisitional work takes place at a post-task stage.“
This rushed attempt to summarise Skehan’s LAC paves the way for our forthcoming discussion with Peter, scheduled for next week, when Neil McMillan and I will talk to Peter about his work and how it relates to Mike Long’s version of TBLT.
The first thing I’d like to draw attention to is the way that Skehan has gone about his research. Rather than attempt to test some wide-ranging hypothesis, Skehan has worked bottom-up. In his studies, he selects a few tasks and then tweaks them, concentrating on this variable or that. He finds that variable X affects outcome Y and then does further studies to check and slowly extend his investigations. Only when he introduces Levelt’s model of speech production (Levelt, 1989) – which I haven’t even mentioned!!! – does Skehan attempt an explanatory theory to support his work. Of course, Levelt’s theory is an important addition, but somehow it isn’t central, to me at least, when it comes to appreciating his huge contribution to SLA research.
Second, narrowing things down (a lot!) what marks Skehan out in the world of ELT is his dogged pursuit of a reasonable (sic) way to go about ELT. He’s done so much to promote an alternative to coursebook-driven ELT, to raise the bar when it comes to how we can apply SLA research findings to teaching practice. Whatever differences there are between Long’s approach to TBLT and Skehan’s, they both agree: ELT practice must change!
But what about those differences? Well, they’re not as great as they might appear. Skehan’s view that tasks should be the syllabus, and that they should be designed in such a way that they “destabilize” the learner’s interlanguage before helping to eliminate inaccuracy and achieve fluency is completely in accord with Long’s view. The pre-task -> task -> post task sequence is a major difference, and in the end, I side with Long, because I think his sequence of pedagogic tasks, based on Target Tasks, is better: each pedagogic task in the sequence includes attention to all three phases of Skehan’s methodolical phases. However, when it comes to the design of the pedagogic tasks, I think Robinson’s framework is unwieldy and based on an over-ambitious hypothesis. Skehan’s Task Sequencing Features, which pay attention to Code Complexity, Cognitive complexity, processing and familiarity, and factors that affect Communicative Stress, seem a beter guide.
In the first three runs of our course, Neil and I have learned that Long’s TBLT has to be adapted to fit the very different contexts of ELT which confront course designers and teachers worlwide. We continue to develop our course, and we anxiously (sic) await Peter Skehan’s contribution to it.
Finally, here’s a bit from Ellis, et al (2019)
“One key distinction that Prabhu (1987) makes is between syllabuses which are intended to function as operational constructs in the curriculum and those that are intended to function as illuminative constructs. When a syllabus is designed to function as an operational construct in the L2 curriculum, the aim is to provide course content as a resource for teachers to construct their lesson plans and reach curricular goals with different learners in the varying classroom contexts in which they teach. This type of syllabus has low internal structure and leaves methodological and implementational issues to be determined by teachers based on their experience of what works at the local level and the problems that different learners face in the classroom. In short, the syllabus specifies only what will be taught, not how it will be taught. The content of syllabus is fixed, but how the teacher uses this content is flexible. By contrast, when a syllabus is designed to function as an illuminative construct in the L2 curriculum, the aim is accountability for both what will be taught and for what will be learned as a result. The focus at the planning stage is on ensuring accurate prediction so steps are taken to bring what is taught and what is learned into careful alignment. In achieving this end, the line between syllabus and methodology naturally blurs, and the syllabus takes on a much broader role within the curriculum.”
Ellis’ position, like Prabhu’s, is that an illuminative syllabus would be undesirable, even if it were possible, as “it limits what teachers and learners bring to the learning process in terms of the intuitive decisions and adjustments that they make in optimizing learners’ mastery of syllabus content”. Now there’s the rub. It sounds reasonable, but I couldn’t disagree more with this position, and I’ll be interested to hear what Peter has to say when Neil and I talk to him.
Peter’s chat with us will form part of the TBLT course run by SLB which starts on October 22, 2021. For more info, see here: TBLT: From Theory to Practice.
Ellis, R., Skehan, P., Li, S., Shintani, N., & Lambert, C. (2019). Task-Based Language Teaching: Theory and Practice (Cambridge Applied Linguistics). CUP.
Levelt W.J. (1989). Speaking: From intention to articulation. MIT Press.
Long, M. (2015). SLA and TBLT. Wiley.
Skehan (2018) Second language task-based performance: theory, research, assessment. Routledge.Ellis, R., Skehan, P., Li, S., Shintani, N., & Lambert, C. (2019). Task-Based Language Teaching: Theory and Practice (Cambridge Applied Linguistics). CUP.
Skehan, P. & Foster, P. (1997) Task Type and Task Processing Conditions as Influences on Foreign Language Performance Language Teaching Research, 1, 3, 85-211.
Skehan, P. & Foster, P. (2007). Complexity, accuracy, fluency and lexis in task-based performance: a meta-analysis of the Ealing research. In: Van Daele, P.; Housen, S.; Kuiken, F.; Pierrard, M. and Vedder, I., (eds.) Complexity, Fluency and Accuracy and Fluency in Second Language Use, Learning and Teaching. University of Brussels Press, Brussels, pp. 207-226
Robinson, P. (2001a). Task complexity, cognitive resources, and syllabus design: A triadic framework for examining task influences on SLA. In P. Robinson (Ed.), Cognition and second language instruction (pp. 287-318). Cambridge University Press.
Robinson, P. (2001b). Task complexity, task difficulty, and task production: Exploring interactions in a componential framework. Applied Linguistics, 22, 27-57.
Robinson, P. (2005). Cognitive complexity and task sequencing: A review of studies in a componential framework for second language task design. International Review of Applied Linguistics in Language Teaching, 43(1), 1-33.
Wen & Ahmadian (2019) Researching L2 task performance and pedagogy: in honour of Peter Skehan. Benjamins.
Willis, J. 2009. The TBL framework: the task cycle. In K.Van den Branden, M. Bygate, J. Norris (eds) Task-based Language Teaching – a reader. Benjamin.