Editor

JLLT edited by Thomas Tinnefeld
Journal of Linguistics and Language Teaching
Volume 7 (2016) Issue 1


Melissa Baralt / Roger Gilabert / Peter Robinson (eds.): Task Sequencing and Instructed Second Language Learning. London: Bloomsbury 2014 (ix + 237 pp.) (ISBN 978-1-6236-6276-2).

Task-based learning in Second Language Learning (SLL) can be said to have started in 1984 with Candlin’s radical proposal for new forms of syllabus design, which would shift the focus of language learning from pre-determined learning sequences towards language learning activities, such as tasks, which place meaning and real-world language use in the foreground, while engaging the learners’ cognitive faculties in the search for solutions to complete the tasks appropriately. Such tasks are claimed to be both motivating and challenging as well as serving to show the learners where they have gaps in their knowledge and need to do something about filling those gaps. In the intervening thirty years, research into task-based learning (TBL) has been prolific (Ellis 2003, Robinson 2011), but there is still a great deal of uncertainty among practising teachers as to how TBL can be implemented successfully. Hence the present volume on task sequencing, co-edited by one of the leading researchers into TBL, Peter Robinson, provides a welcome contribution to the current discussion.
Task Sequencing and Instructed Second Language Learning presents a number of empirical studies into sequencing, all of which have been devised to investigate Robinson’s Cognition Hypothesis in instructed SLL. It should be emphasised at this point that the volume, following Robinson’s hypothesis, is concerned with sequencing tasks in terms of cognitive complexity rather than linguistic complexity, meaning that the demands placed on learners’ conceptual, attention, memory and reasoning faculties should determine the sequence of tasks. Put simply, learners should start with cognitively simple tasks and gradually move on to more complex versions. The theory predicts that the simpler tasks will permit greater fluency while the more complex tasks will push the learners towards more complex output and accuracy, and ultimately lead to greater progress in L2 acquisition.
To operationalise the Cognitive Hypothesis within language learning, Robinson has elaborated what he calls the Triadic Componential Framework, which enables tasks to be graded systematically. This framework is intended to determine (1) task complexity, i.e. features which can be manipulated to make the task more cognitively demanding, (2) task difficulty, i.e. how the learners will cope with the task with the personal resources they have at their disposal at the time of working on the task and (3) task conditions, e.g. the interactive demands in the classroom context.1 In a pedagogical context, only cognitive complexity is sequenced, so tasks which involve the learners having to use, for example, knowledge of the world or call upon rational thought beyond the input features of the task are considered more complex than those which do not require more than having to exchange information. The language learning dimension is captured in a model with the acronym SSARC: Stabilize, Simplify, Automatize, Restructure, Complexify, which provides for consolidation and change in the interlanguage system. This model forms a kind of blueprint for drawing up the syllabus in terms of the current state of interlanguage and how it can be developed until it corresponds more closely to native-like language norms. The Triadic Componential Framework and the SSARC model form the point of departure for the studies assembled in the book.
The volume is divided into two sections:
  • Experimental Studies of Task Sequencing
  • Educationally Situated Studies of Task Sequencing in Natural Classroom Contexts.
These two sections are preceded by an Introduction (Ch. 1, pp. 1-34), which takes the readers through a potted history of Task Sequencing in SLL and is essential reading for a better understanding of the scope of the book and the terminology it adopts. Although the concept of task is not defined2, there appears to be agreement among the authors whose work is reviewed in the Introduction that cognitive processing is the outstanding feature that distinguishes TBL from other holistic approaches such as a communicative approach. The editors are critical of many studies undertaken so far as they fail to operationalise the concept of complexity in a reliable manner, e.g. Prabhu (p. 6), who used common sense judgement to sequence the tasks and devise the syllabus, or Long & Crookes (p. 7-8), who fail to “provide specific suggestions about how to implement sequencing based on the complexity level of a task” (p. 8). The need for explicit criteria for sequencing tasks was articulated by Skehan in the 1990s, but, despite the pedagogical merits of his scheme, the editors state that it lacks a “principled set of criteria” (p. 9) for sequencing and empirical testing. Furthermore, though Ellis’ task grading proposals include the consideration of classroom procedures adopted by teachers to teach the task (p. 12), they lack clear guidelines on how complexity can be determined.
The chapter continues with an introduction to Robinson’s work on task sequencing (outlined above) and summarises the individual contributions in the book.
Chapter Two, “Task sequencing in the L2 Development of Spatial Expressions” (pp.37-70), describes an empirical study on the L2 development of spatial expressions in L2 English by learners at B1 level during an intensive summer school. Levkina & Gilabert start their contribution with a survey of TBL and the selection of cognitive complexity as the rationale for sequencing tasks. The survey goes over the ground already covered in the general introduction to the volume but is well worth reading as it adds more detail in parts and gives a slightly different take on the previous research. It also reprints a table drawn up by Robinson showing the main features of the Triadic Componential Framework (p. 43), allowing the reader to take in all of the design characteristics of the Framework at a glance (or two).
The research questions here concern the impact of “Task Sequencing in the L2 Development of Spatial Expressions”, and in particular whether the order of presentation of three task formats (simple – +complex – ++complex) affects the learning and long-term retention of new target items related to expressions of spatial relations. In the study, 48 learners took a pre-test to establish their familiarity with the spatial expressions in question and were then randomly divided into three treatment groups with three task sequences: (1) simple – +/++complex, (2) ++/+complex – simple, (3) randomised (+complex – simple – ++complex). There was no control group. After completing the tasks, the learners took a post-test and a delayed post-test two weeks later. The tasks centred around the arrangement of furniture in a new flat, with a written (email) description of a room already furnished, drawing the location of the items on a floor plan, a multiple choice test of the spatial expressions and, at the final level of cognitive complexity (++), the furnishing of a new kitchen. The cognitive complexity is summarised in a table on pp. 48-49, and the task items themselves are described in a series of appendices (pp. 57-63).
The results turned out to be somewhat ambiguous, though the participants clearly did learn something. The hypothesis that the SSARC model with the complexity sequence simple – +/++complex (Group 1) would produce better results could not be confirmed in the first test condition: Group 2, with the ++/+complex – simple order, scored higher. All groups showed significant vocabulary gains in the delayed post-test, but this time, Group 1 members were able to improve their scores, which would speak for the SSARC model. The authors conclude that sequencing the tasks from simple to complex has the potential to aid retention over time, but more research is needed. Nevertheless, the study is a useful starting point and the task material used for study is a good example of how cognitive complexity can be increased while keeping the language demands relatively simple.
Chapter Three discusses “The Role of Task Sequencing in Monologic Oral Production” (pp. 71-93). In her introduction, Malicka raises the topic of skill learning and the question of how the criteria for deciding on what is simple or complex can be determined. Her discussion centres around two competing approaches, namely Robinson’s Cognition Hypothesis and Skehan’s Trade-off Hypothesis. By now, the basic premises of Robinson’s ideas should be clear, but the juxta­position with Skehan’s work permits further insights. Task complexity comes about through the interplay of attentional, memory, reasoning and other information-processing demands. Whereas Robinson differentiates between cognitive, interactive and learner factors, Skehan distinguishes between code complexity, cognitive complexity and communi­cative stress. In any case, though both call for the sequencing of tasks in a principled manner, they disagree on how learners distribute their attention in a complex task. For Robinson, attentional resources are unlimited so linguistic accuracy should not suffer when cognitive complexity is increased, while for Skehan, they are limited and so accuracy and complexity may be in competition for resources, leading to a trade-off in language production (see pp. 72-74 for a more detailed explanation). This section of the chapter continues with a review of studies pertinent to both hypotheses, finally calling for further research. XXX
Malicka’s own research project attempts to search for more evidence for or against these two hypotheses as well as to ascertain out how performance is affected by sequencing conditions. Her participants (N = 50) were learners of L2 English with L1 Spanish / Catalan. Half of the group were given a series of tasks with a sequencing order of simple, complex and +complex3. The other half were given the tasks in five different randomised sequences. As the learners were in the tourism field, they were asked to play the role of a hotel receptionist and leave a message in English on answering machines belonging to clients whose room reservations had to be changed. Complexity was increased by giving the “receptionists” a client profile (in Spanish) with the requirement to choose whether the client should be given a different room or moved to a different hotel.
In all of the complexity levels, the number of elements was kept the same, but reasoning demands were manipulated in terms of the number of different speech acts needed to complete the task, e.g. describing, apologising, justifying. The simple task condition only required the transfer of information; the complex task required an apology, explanation of the options and recommendation for one of them. The +complex task required a justification for the change in addition to requirements of the complex task condition. The planning time was limited so that the learners could familiarise themselves with the task but not plan the answer. After completing the task, the participants filled out an affective variables questionnaire. The student monologues were scored on fluency, lexical complexity, structural complexity and accuracy. The results make for interesting reading: in three of the measures, fluency, lexical complexity and accuracy, there was no significant difference between the two groups’ performance, but in structural complexity the complex task triggered more complexity in the simple – complex group, whereas the simple task triggered more structural complexity in the randomised group.
The second question concerned interaction between task complexity and sequencing from simple to complex. Here there the sequencing order had no influence on production. Malicka concludes that “overall performance was heavily affected by cognitive complexity” and that fluency deteriorates on the +complex task (p. 88). Since the patterns of performance varied across the four measures, she suspects that other phenomena known in SLA research, such as Kellerman’s U-shaped behaviour, were at work but also that attention switches may have taken place. Importantly, the differences between the complex and +complex tasks were enough to trigger performance changes (p. 89). The results, however, did not provide support for the hypothesis that the simple – complex sequence is the ideal order. Despite the somewhat inconclusive results, the study was well-planned and implemented, opening up new channels of research. Like the previous study, it shows how the concept of cognitive complexity can be operationalised in a meaningful way, not just for research purposes but also for classroom pedagogy.
Chapter Four deals with a new challenge to practitioners, namely the role of task sequencing and online pedagogy. Like all of the studies reported on so far, Melissa Baralt’s contribution “Task Complexity and Task Sequencing in Traditional Versus Online Language Classes”, starts with a presentation of Robinson’s Cognition Hypothesis and the SSARC model of task sequencing. Her introduction emphasises two variables, which have not been mentioned in this review so far but are part of Robinson’s cognitive complexity construct. These are: resource-dispersing variables which affect the learner’s ability to direct attention to non-linguistic features of the task and access the L2 system, for example, planning time, which would reduce the pressure and make the task “simpler”. The second variable encompasses resource-directing features, governing how attention is directed to language and the need for particular language forms: in a task involving intentional reasoning features such as causality or conditionality, which require specific language forms, make the task more complex. With these two variables, the teacher can adjust the complexity of the tasks in preparation for real-world communication.
One of the main characteristics of task-based learning is interaction, so one may well ask how this can be achieved in online learning. Baralt has opted for online chat as a medium to guarantee interaction in online tasks in her research. In particular, she adopts synchronous computer-mediated communication (SCMC) which not only permits interaction but also allows focus on form and feedback. The fact that there is a delay in response times may even be an additional advantage as the learners have more time to plan. In a previous study, Baralt looked into the efficacy of feedback in face-to-face and online situations, using a simple – complex task design. In the complex task, learning was better in the face-to-face mode, whereas in the simple task, the SCMC mode worked better. Studies by other researchers, summarised in a table (p. 102), indicate that complex tasks lead to more clarification requests, language-related episodes (LREs) and self-correction, all of which lead to language development.
In the present study, Baralt investigated the impact of different sequences of cognitive complexity on learning opportunities and L2 development. The tasks were intended to force the learners to use the Spanish past subjunctive, a form which had been covered in regular textbook-based class sessions. The learners were required to read part of a story in their L1 (English) and retell it to their partners in Spanish with the help of comic strip pictures. Complexity, in terms of +/- intentional reasoning, was increased in some of the tasks, which were completed in different complexity sequences (p. 108).
XXX The results showed that whenever the learners in the classroom group had to deal with complex tasks, they produced more LREs, and this affected the learning outcome positively. The learners in the online group, however, produced no LREs at all regardless of the task complexity sequences, meaning that the results of Baralt’s previous study could not be confirmed. Baralt concludes that more complex tasks are more effective than simple ones for language development and reports that the learners in the classroom group had enjoyed the tasks but missed getting feedback since they did not construe the learner-learner dialogues as “feedback”. As far as the online group is concerned, it may be the case that learners simply need a different type of task which matches their online experience outside the classroom. This would make comparability across groups more difficult to achieve, but would be worth trying out. This study is well planned and implemented and stands out not just for its scientific merits but also because for once, we are not just reading about the acquisition of L2 English!
Chapter Five, the last in this section of the book, deals with “The Effects of Guided Planning, Task Complexity and Task Sequencing on L2 Oral Production”. In this contribution, we return to the competing hypotheses of Robinson and Skehan. Colin Thompson reviews a study by Mochizuki & Ortega on the production of relative clauses by Japanese learners who were given guided planning to complete the tasks. He concludes that the results of this study can be used in support of both of these hypotheses, but detects a limitation in the study in that the effects of planning were not monitored adequately. In his replication study, Thompson focused on the notions of task complexity and task difficulty, which, according to Robinson, are distinct constructs (p. 128). He hypothesised that with guided planning the learners would be more likely to produce the particular form (relative clause), that learners would find the more complex tasks more difficult and that the task completion would be less successful.
The learners, who were divided into a guided planning and an unguided planning group, were asked to recount a story to a researcher with the help of a series of pictures. The story was always a narrative, but the storylines were varied so that there was no element of repetition in subsequent trials. Increased complexity was achieved by increasing intentional reasoning, which resulted in having to produce more relative clauses. Following the tasks, the learners took an immediate post-test, followed by a delayed post-test, and completed a questionnaire. The whole study spanned a seven-week period. The results showed that the learners produced more relative clauses in both post-tests than they had in the pre-test, although they did not produce all of the expected relative clauses, partially confirming the hypothesis. Regarding the benefits of guided planning, the results could only be partially confirmed. Although the learners who had been given planning help showed greater gains in the immediate post-test, there was no substantial difference between the groups’ performance in the delayed post-test. The perception of task difficulty did not vary between the two groups: both found the simple task the least difficult while the most complex task was judged to be more stressful.
Thompson concludes that the elicitation task was unsuccessful in the pre-test, but this is not so surprising since of the seven example sentences given on p. 145, only two require an obligatory relative clause. However, after completing the tasks, the learners produced more relative clauses in subsequent narratives. Thompson states that with guided planning and increased task complexity, both groups showed significant gains, pointing out that his learners could draw on their declarative knowledge and did so once they were primed towards producing relative clauses. He also claims that sequencing tasks in terms of increasing intentional reasoning along the lines of the Cognitive Hypothesis encourages learners to turn their efforts towards producing more complex output. Sequencing the tasks and offering guided planning may have facilitated proceduralisation of knowledge enabling the learners to produce the intended form in natural speech. Hence, even with high proficiency learners, it may be useful to provide guided planning in more complex tasks.
The results of the questionnaire of the learners’ perceptions of task difficulty showed that more complex tasks are considered more difficult, but that there was no advantage for the group which had received guided planning help. The study once more underlines the importance of increasing task complexity as a means of promoting language knowledge, especially with learners who are capable of stretching themselves to meet the challenge.
Chapter Six opens the section on task sequencing in natural classroom contexts. Kim & Payant start their contribution, “A Pedagogical Proposal for Task Sequencing: An Exploration of Task Repetition and Task Complexity on Learning Opportunities”, by summarising Robinson’s Cognition Hypothesis and SSARC model, reminding the reader of the main types of syllabus design: synthetic and analytic. They recommend developing tasks that will generate large numbers of LREs, which will focus attention on form and so ultimately benefit the learners’ language use outside the EFL classroom. They also review the secondary literature on task repetition but find that the issue of the interaction between task repetition and task complexity has been largely ignored. Their contribution attempts to fill this gap and discover how task complexity and task repetition interact, creating opportunities for LREs, which should lead to further learning.
Although the study was set in a normal classroom context, it did not actually reflect teaching conditions in a primarily task-based classroom, as the learners worked on the task in only three teaching sessions of a syllabus which was normally four-skills-based with communicative elements (p. 159). Since the task-based teaching was presumably quite a novelty, the learners may have been more motivated to work under these conditions than in the regular context.
The learners, Korean learners of L2 English, completed tasks under four experimental conditions involving repeating the content and repeating the procedure, each in simple / complex formats and with a new partner for each different session. The data revealed two types of LREs: lexical and grammatical, which were sub-categorised as resolved by learner-learner dialogue, by learner-teacher dialogue or incorrectly resolved and unresolved. The results are too extensive to be dealt with in detail here but show, for example, that while the procedural repetitions yielded more lexical LREs, the number of lexical LREs decreased in all groups after each task session, indicating that some learning had taken place and the learners had fewer problems. With grammatical LREs, the story is not so clear cut: with procedural repetition, there was an increase in grammatical LREs in Task 2 but, overall, procedural repetition tended to require more teacher-learner resolution than the task repetition mode where the learners were able to solve the problems themselves. In all conditions, the number of successful resolutions outweighed the number of incorrectly resolved or unresolved solutions.
Repetition of activities is a standard procedure in EFL classrooms whatever syllabus is chosen, as it allows the learners to consolidate their knowledge and attend to form. Kim & Payant claim that task repetition does not merely lead to duplication of production but may lead to increased automatization, promoting proceduralisation. It is still an open question, however, as to how many repetitions are optimal and whether content or procedural repetitions are preferable. The study focused on the number of LREs generated and while the authors point out that the number of LREs is not the same as language learning, there was no other measure of learning in the design of the study. No tailor-made post-tests, which would have indicated whether learning had really taken place, were administered.
This is an interesting study that reveals more gaps in our thinking about TBL and shows that other dimensions such as language proficiency and the ecological validity of the tasks would also need to be taken into account in the research design. The Cognition Hypothesis’s focus on complexity as the main sequencing principle could not be confirmed here: task repetition would have to be included as a factor in developing task sequencing proposals but the optimal type of repetition and the timing of such interventions are still topics for further research.
Chapter 7, “Teachers’ Application of the Cognition Hypothesis when Lesson Planning: A Case Study”, by Baralt, Harmath-de Lemos & Werfelli raises the question of how teachers perceive TBL and implement it in the classroom. The chapter reports on the experience of two teachers’ attempts to devise TBL units during an in-service training course. They had been introduced the Cognition Hypothesis and the notion of sequencing tasks in TBL according to complexity on Day One of the course: Day Two was reserved for practical work on preparing lesson plans for TBL sessions in their own classrooms.
The chapter summarises the Cognition Hypothesis and the features of the Triadic Componential Framework, as well as previous research on teachers’ operationalisation of these theoretical concepts. The research in this field is limited: Van den Branden reported that teachers were unhappy about using tasks they considered too difficult for their students and preferred traditional PPP approaches, while Gurzynski-Weiss found that teachers’ beliefs, perceptions of their students’ abilities and institutional factors were among the reasons for not implementing TBL (see pp. 183-5 for more details).
The two teachers in this study taught Spanish and Italian, respectively, at a US university. In the workshop, they were introduced to the idea of sequencing tasks from simple to complex, shown examples and then, working collectively with other participants and the researcher as well as on their own, prepared their own plans for up-coming class sessions. These TBL planning sessions were organised as a mixture of interactive group phases and individual think-aloud recordings, all of which were recorded and later evaluated by the researcher. The teachers had volunteered for the workshop and were positively disposed towards TBL. Their current teaching was described as communicative but both felt restricted by a grammar-based syllabus which was imposed on them by their institution. Both teachers drew on materials which they had already devised to teach a particular grammatical structure (present perfect) and adapted them to create a task-based unit consisting of simple and more complex tasks with pre-task, task and post-task activities. The results, however, revealed that their understanding of the Cognitive Hypothesis and the Triadic Componential Framework did not entirely correspond to the principles underlying the approach. Their interpretation was to extend these principles to their current teaching practice, which would involve teaching the grammatical structure explicitly before letting their students loose on the tasks. In other words, they doubted that the tasks could be a vehicle for learning without prior teaching and were simply using them as additional practice material.
This qualitative study reveals the gap between theory and practice in teacher training and indicates the inability of teachers to adjust their teaching methods in accordance with new insights into language learning pedagogy rapidly. A two-day training course is clearly not sufficient to change the habits of many years, even if the participants, like the teachers described here, find TBL a plausible and potentially effective means of learning language. It also indicates that TBL may have to be introduced as an institution-wide project with large teams of teachers cooperating to devise TBL units and share on-the-job insights. In the long run, textbook publishers with their greater resources will probably make the running here and flood the market with ready-made solutions.
The final contribution by Craig Lambert & Peter Robinson, “Learning to Perform Narrative Tasks: A Semester-long Classroom Study of L2 Task Sequencing Effects”, takes a macro-learning perspective as opposed to the micro-learning examples, e.g. spatial expressions, present perfect or relative clauses which have been the topics in the volume so far. The pedagogical goal was to teach students to summarise texts and give an opinion on what they had read. The research goal was to lend support to the Cognitive Hypothesis and its ramifications. To this end, the chapter starts with a summary of the Cognition Hypothesis and the SSARC model of task sequencing.
Two student groups taking part in regular basic courses were arbitrarily divided into an experimental group and a control group. They both received the same instruction in the first half of the term to enable them to analyse short stories (e.g. plot, characters) and organise their opinions on what they had read. In order to be able to compare the groups, the researcher also collected extensive data on language proficiency, motivation and educational background of all participants (pp. 220-221). In the second half of the term, the control group continued analysing stories and writing summaries, while the experimental group underwent a six-session treatment phase of narrating a story from pictures, then with the addition of further pictures attributing mental states to the actors (complex version) and then summarising an authentic story (real-world task). Before the task-based phase began, all students summarised a short story as a pre-test and at the end of the term, a post-test was administered. These summaries were analysed for syntactic complexity, explicit reasoning and grammatical accuracy. It is not clear how much feedback was given to the two groups during the experiment, if any at all. Here more information about the classroom routines would help readers to evaluate the role of the task-based intervention. As it is, there was little difference between the two groups, but so many other variables played a role in the experimental design that one cannot attribute this finding exclusively to the SSARC model. It may also be the case that narrating from pictures draws on different skills than summarising.
The study is an attempt to confirm the Cognitive Hypothesis under real-life conditions, but it showed that it is almost impossible to isolate one single factor in the language learning process for empirical testing. The range of dimensions investigated, such as working memory capacity, are an indication of what might be involved in learning gain, but it is well-nigh impossible to capture all of these in a single study.
On the whole, the volume gives much food for thought. While the notion of working from simpler versions of a task through more complex versions until target-like language is produced will intuitively appeal to most teachers (Ch. 7), even without the SSARC model, the finding that the reverse sequence of complex to simple (Ch. 3) turned out to lead to qualitative changes in output is more than intriguing. All of the contributions cover the Cognitive Hypothesis and SSARC model in more or less detail, but fail to give much detail about the learners concerned – are they beginners, advanced learners, schoolchildren, young adults? What is their previous learning experience? What are their needs? How much TBL have they already experienced? What is the learning culture in the country of the experiment? This kind of information would help the readers gain a more general perspective and enable them to judge the outcomes of the studies better.
The question of task sequencing is certain to be a subject of further research and is certainly one that is worth answering. The present volume makes an interesting starting point for anyone looking for a challenging research topic. It is to be hoped that more research is forthcoming, particularly with languages other than English: German with its morphological complexity and its increasing appeal as a second language would be an excellent candidate.

References
Ellis, Rod (2003). Task-based Language Learning and Teaching. Oxford: OUP.
Robinson, Peter (2011). Task-Based Language Learning: A Review of Issues. Language Learning, Vol. 61, Sup. 1, pp. 1-36.

Author:
Prof. Dr. Veronica Smith, MA
Alpen-Adria-Universität
Institut für Anglistik und Amerikanistik
Universitätsstraße 65-67
9020 Klagenfurt
Austria

1 The Triadic Componential Framework is described in more detail in the contribution by Levkina & Gilabert in Chapter 2, p. 43. The authors list factors which should be taken into account, showing how complexity can be increased in a step-by-step manner.
2 The reader is recommended to turn to Ellis (2003: 4-5), if help is needed in this respect.

3 The terminology has not been unified in the different contributions. The authors’ use of the symbols + and ++ has been maintained in this review, despite the potential ambiguity.