Second Language Task-Based Performance

Theory, Research, Assessment

Peter Skehan

• Publisher: Routledge
• Year: 2018
• ISBN:
  9781317245001
• Edition: 1
• Topic: Langues et linguistique
• Subtopic: Linguistique

Second Language Task-Based Performance is the first book to synthesize Peter Skehan's theoretical and empirical contributions all in one place. With three distinct themes explored in each section (theory, empirical studies, and assessment), Skehan's influential body of work is organized in such a way that it provides an updated reflection on the material and makes it relevant to today's researchers. Also in each section, an early publication is matched by at least one later publication, followed by a newly written commentary chapter, the combination of which provides the important function of offering a wider-ranging discussion. This book is an invaluable resource for researchers interested in second language task-based research or SLA more generally.

1
The Road to This Book

My career in applied linguistics didn’t start with an interest in tasks. At the very beginning, I had a post at the University of Birmingham, in a unit concerned with English for Academic Purposes. I had studied economics (as part of a dual honours degree), and was working with groups of Iranians who were studying economics but whose English was not of a sufficient level for the courses they were doing. So initially my focus was on the specialist language that they needed, drawing upon what I had studied of economics, linked to my previous experience as a language teacher. This was done in an EAP unit headed by Tim Johns and Tony Dudley-Evans, at a time when EAP had immense vitality as a sub-field within applied linguistics. So my earliest publications actually concerned how team teaching (with subject specialist and language specialist working in tandem) could best be organised.

But I was also working on a PhD on foreign language aptitude, at Birkbeck College, University of London. So in parallel to the focus of my day job, I was capitalising on the other part of my dual honours degree—psychology. This had, in fact, been followed by a Master’s in psychology at the University of Western Ontario (and more about this can be found in Skehan, 2013). I was particularly interested in the role of memory in foreign language aptitude. Another facet of my work at Birmingham was that my head of department, John Sinclair, wanted to introduce an MA in applied linguistics, and so, on the basis that I had studied psychology and knew more about statistics than other people in the department (and the bar was not particularly high), I was tasked with teaching language testing. So most of these early interests were somewhat thrust upon me.

Not that they were uninteresting, but my individual focus, from amongst them, was definitely foreign language aptitude. I continued to research this area for the next decade, after I obtained my doctorate. But it was undeniable that there was hardly anyone else, at that time, who was interested in aptitude (a situation which is now very changed). I needed, in other words, to cast around and find more sociable areas of applied linguistics within which to work! My interest was raised by the interchange between Rod Ellis (Ellis, 1987) and Graham Crookes (Crookes, 1989) on the impact of planning on second language performance. At that time, the move was taking place from discussing communicative language teaching to using the term ‘task-based learning’ and the exchange between them seemed central to the way not simply that tasks could be discussed, but also, how tasks could be researched. The exchange was extremely suggestive as to how different lines of research could be pursued which (a) considered potential differences between tasks, (b) explored an additional variable, planning, in its effects on tasks, and (c) raised issues as to how to conceptualise and measure performance on tasks.

The debate between Ellis and Crookes was the catalyst which stimulated a whole new research area, one which was central to second language acquisition research at the time, and one which had a community of active researchers. But there were some other major influences. First of all, I was, by this time, working at the Polytechnic of West London, which became Thames Valley University. We formed there the Centre for Applied Linguistic Research, and this Centre, through the UK Research Assessment Exercise, attracted funding. Some of that funding was used to employ Pauline Foster as my research assistant, and this in turn led to research grants from the Economic and Social Research Council. The funding, and having Pauline as a research colleague, enabled several research studies to be done. CALR also attracted doctoral students, one of whom, Uta Mehnert, also did distinguished work in the task-based field. This activity was maintained after a move to King’s College, London, where Parvaneh Tavakoli did a PhD with me. So my initial interest in tasks expanded and was supported, and I had the immense good fortune to work with excellent colleagues and doctoral students. My interest in the task-based area was firmly cemented.

Subsequently I moved to a post at the Chinese University of Hong Kong and I continued to be fortunate in the work that I did with tasks. This was in two ways. First, the Hong Kong system offered scholarships to doctoral students and while I was there I was able to supervise a very talented group of young researchers, all interested in aspects of task performance: Gavin Bui, Dai Binbin (Amy), Christina Li, Sheila Luo, Zhan Wang, and Edward Wen. It was a wonderful experience supervising such a talented bunch, but it was also really useful because we managed, as a group, to conduct fairly coordinated research, as indicated in the 2014 volume that I edited (Skehan, 2014a). But second, the Hong Kong Research Grants Council was a wonderful source of funding (and I shouldn’t forget the internal funds offered by Chinese University). This money enabled me to conduct far more research than I would otherwise have been able to do, to develop my interest in tasks. It also enabled me to work with the research assistants I was able to appoint, Francine Pang, Sabrina Shum and my former doctoral students, Zhan Wang and Christina Li.

These are all the background influences to my interest in tasks. They provide a context for the set of articles and chapters on task research which are reprinted here. It has been interesting and rewarding to work in an area which has grown considerably in breadth and vitality during that period. For my part, the attraction is the way the area unites a concern for theory, a connection with psycholinguistics more generally, and the prospect of pedagogic relevance. The theory (in my case) relates to performance rather than acquisition (although extensions to acquisition, not covered much here, are not difficult to achieve), and the enduring fascination as to how we can communicate as effectively as we do. Psycholinguistics is represented by the existing models of first language speaking, and how productive it is to apply them to the case of second language performance. Allied to this is the consequence, for second language performance, of working memory and attention limitations which provide the background to the way second language use can proceed. Finally the connections to pedagogy are not at all difficult to make. The days of an exclusive focus on grammar in language teaching are now long gone (I hope!) and the challenge is to find teaching techniques which produce effective learning which translates into effective communication. The need to use tasks then becomes central.

But allied to these different motivations is the need for a balance between theory and empirical research. One of the attractions of the task area is that there are, essentially, two pressures to conduct empirical research. If theory is to be useful, it has to make testable predictions, and this requires empirical research. But pedagogy, similarly, looks to research justification for any claims that are made which might impact upon classrooms. My original training (in psycholinguistics, at the University of Western Ontario) was very much directed in this way, and so the confluence of all these forces seemed ideal to me. A training in research methods and statistics, an interest in psycholinguistics more broadly, and then an eye for pedagogic relevance seems to lead inexorably to the need to do task research!

The set of interests also connects with the structure of this book. There are three main sections, and these are concerned with theory, empirical research, and assessment. I have to admit that I am, basically, ambivalent about theory, or more exactly, its immediate role in shaping research. Obviously, it is desirable, but I also think that theory needs a solid empirical base as its starting point. I doubt that task research has quite achieved this level. So I am drawn to the use of frameworks, which are below the level of theory or even hypothesising. Frameworks attempt to systematise what is done, so that there is greater prospect of cumulative progress. The 1996 article from Applied Linguistics, reprinted here in the theory section, epitomises this. It attempts to structure the way research might be done, and data collected, but it attempts to be relatively neutral with respect to theory. Yet theorising and the precision and prediction that accompanies it are desirable, and so the later chapters in the section on theory do try to go beyond such a more limited starting point. The purpose of the two remaining contributions to the section therefore is to explore just how ready we now are to theorise about second language task performance.

The second major section of the book is concerned with empirical research. Two articles on planning are included, followed by one which focusses on lexis. Then, there are two ‘reflections’ chapters, one on planning and one on empirical research itself, on the actual measures of second language task performance. I have included planning because it seems to me one of the success stories of task research: we now understand much more about consistencies in performance and also the processes which occur during planning. It is also an interesting example of harmony between quantitative and qualitative approaches. Then the reflections chapter on measures of second language performance also tries to draw attention to progress, but in this case in understanding the measures that can be used, and consistencies in their functioning, and options which are available.

The final section focusses on testing and assessment, and includes two articles, one from 2001 and the other from 2009. They are both concerned with the same broad question: does task research have anything to say that is relevant to the testing of speaking? Naturally I argue that it does, although, as the section indicates, there can be some legitimate doubts about this. The ‘Reflections’ chapter tries to justify the claim of relevance, and basically argues that task-based research may have a lot to contribute in characterising the notion of ‘ability for use’, as a mediator between underlying competences and actual performance.

I have published a range of pieces over the years, and so one might think that the existence of these books and articles would be enough! But I have felt, and this was an idea originally encouraged by Rod Ellis, that bringing together and reprinting a selection of this material might have its value. One justification for this might be to juxtapose more highly cited publications with material that is less known, to try to bring out connections between these various writings (and this principle does apply to a limited extent). But in thinking of how such a collection might be put together, it occurred to me that it might be more revealing if, within the different sections just described (theory, empirical work, assessment), I juxtaposed something early and something late.

Fundamentally, I am hopeful of progress in the field, and one way of bringing this out might be to compare something from the beginning of my interest in task research and then something more recent (where ‘beginning’, of course, doesn’t go back that far!). This device can then become a structure to bring out more clearly where progress may have been made. Consequently the first section, on theory, includes three articles, from 1996, 2009, and 2015. These have a pleasing twenty-year span. The second section contains articles from 1996, 2009, and 2014, which might imply the same nicely separated choices. In fact, this is a little more complicated, in that the 1996 and 2014 inclusions do have the focus of planning research, but the 2009 chapter is actually about the different area of empirical research with tasks. They do, though, provide the starting point for discussion of a range of empirical results. An additional striking feature here is that all three chapters in this section have a collaborative nature. The two chapters on planning are jointly authored, and this reflects the way they were, indeed, joint ventures. In each case I was a research grant holder and my co-author was the actual data collector as well as vital participant in the research. The other chapter, on lexis, is single authored but really draws fundamentally on the joint research I have done with Pauline Foster, and even relies, in part, on the data she collected for her PhD since this was based on native-speaker baseline data. Really the conclusion is simple: research is a joint activity and the jointness of the chapters in this section simply follows from that.

The third section includes articles from 2001 and 2009, the smallest time interval in any section. There is also a chance element here. I haven’t done any direct research on language testing (despite my earlier interest in this area), and so I have only written on it when I have been invited to contribute something. In each case the chapter comes from an edited book, and I was responding to a need in that book to produce something on testing. So what I attempted to do in each case was take what I knew from the task research area and then try to apply it to assessment.

In a way, the remaining aspect of the book, which I have mentioned, but not covered explicitly yet, is the most important. Each section is followed by a reflections chapter (or in the case of the Empirical Work section, two reflections chapters!). These reflections chapters, or epilogues, are where I try to develop how, in my view, changes have occurred, and what the nature of these changes is. The epilogues (sometimes quite long), try to go beyond the individual chapters and articles, and most ambitiously, make claims about progress. Immediately, of course, I will contradict myself! Progress, in my view, is least evident in the first section, on Theory, not least because claims about theory can only be limited and provisional. They don’t amount to much more than saying that our understanding of tasks themselves is now deeper, and that we have some clearer hints about the areas where research effort might be focussed. With a spoiler alert, I would claim that thinking about the conditions under which tasks are done has been more rewarding, and that linking second language task research to a model of first language speaking, that of Levelt, has paid dividends, to a much greater extent than has research into task characteristics themselves. The latter seems to me to be beset with lack of consistency and lack of clear findings.

Regarding assessment, in other words jumping from the first to the third section, the epilogue claims that task research has huge potential relevance to testing, in several areas. Understanding task characteristic and task condition influences is very important, as is a greater appreciation of how measurement of second language task performance (through complexity, accuracy, fluency, and lexis) might influence ratings of spoken performance. But the greatest potential, in my view, is in using what we have learned about tasks and task conditions to gain inroads into defining what the construct of ability for use might mean. My assumption is that there are underlying competences, but that, overlaid on these, and mediating them as far as actual performance is concerned, is the capacity to mobilise, to access underlying resources, and to solve communicational problems and breakdowns. As the Testing and Assessment epilogue argues, task research is suggestive as to how such an ability for use can be defined.

But if the specific goal of progress is to be claimed, the most convincing reflections chapters, in my view, are those in the second, empirical research section. The two relevant chapters are concerned with first planning and then second measures of second language task performance respectively. In each case, I would argue that considerable progress has been made. In the area of planning, we now have a range of findings which allow a powerful set of generalisations to be claimed as to the effects of planning. We have also seen progress in the research methodologies that are used to investigate planning, with a blend of quantitative and qualitative techniques. We are reaching a deeper understanding of what happens in planning and how we can use this knowledge to have an impact on pedagogy. Regarding the measurement of second language task performance, a great deal has been learned over the last twenty years. There are options as to the measures which can be used. If the major approach has been to explore complexity, accuracy, lexis, and fluency, in each case we are clearer about what each of these constructs represents and how its different facets can be measured. There may have been a great deal of ground-clearing in this work, but it has paid dividends. Future researchers may therefore be in a more secure position in the measures that they use.

Part I
Theory

Introduction

The two articles and one book chapter in this section are nicely spaced (1996, 2009, 2016), and reflect my views on the theoretical underpinnings of task performance. The first, from Applied Linguistics in 1996, was an attempt to be broad in perspective and programmatic in nature. Task research, at this point, was gathering momentum, but seemed to me prone to individual studies which might not cohere. Instead they reflected the preoccupations and contexts of many individual researchers, and as a result, risked the danger of fragmentation. I felt there was a need to take a wider view, to provide a framework for research, and to make links to underlying psycholinguistic theory and also to pedagogy. I was attracted to the term ‘framework’ since it is relatively neutral, theoretically, and only provides a general and inclusive structure within which research studies can be located and then linked to one another. I did not feel that the research results available at that time justified any stronger term than this.

There was also the point that I wanted to find a perspective which provided an alternative viewpoint to the Interaction Hypothesis. I had nothing against this hypothesis at all (although I had, and still have doubts about how much interaction can contribute ‘on the fly’ to developing underlying interlanguage systems). But my interest was performance, primarily, and it seemed to me that Graham Crookes’s (1986) work, especially, had shown how performance itself could be rewarding as a research target. So this was what was uppermost in my mind when writing the 1996 article. I wanted to explore how research into second language task-based performance could be given a more secure underpinning.

From the range of material reprinted in this volume, and also the three reprinted articles in this section, the 1996 article has the greatest focus on pedagogy. This, perhaps, is the moment to add a few words on that subject, since it will not be developed elsewhere. The emphasis, in the first section, as just indicated, is on performance. But there are potential pedagogic links which can be clarified. First of these is that a greater understanding of performance can lead to more effective pedagogic work to achieve balance between the different areas of complexity (structural and lexical), accuracy, and fluency. One would like learners to make progress in all three of these areas, but it may be the case that some learners prioritise one or two at the expense of other areas. It may also be that certain tasks push learners in particular directions, e.g. fluency-oriented tasks. So one pedagogic advantage of researching tasks is for teachers and pedagogues to be better equipped, through knowledge of tasks and task conditions, to address any imbalances in the different dimensions of performance.

There is also the claim that there is an implicit pedagogic sequence built into the performance areas. New language and extension of an existing system connects with greater structural complexity. Such new language may initially be used haltingly and be error-prone. But as time passes, this new language may be automatised to some degree, and though still halting, be less error-prone. Then, as the new language becomes more automatised, accuracy and speed of production may be possible, and a more fluent performance results. In this view, the central value of tasks would be to foster this process of automatisation, and to ensure that new language is converted into usable and non-attention-demanding language. Knowledge about task characteristics, as well as the influence of task conditions, would be at the heart of how this is achieved.

But this still leaves unaccounted for the issue where new language might come from. One source might be task and condition influences which promote complexity. Planning would be the clearest exam...