Read an Excerpt

CHAPTER 1

Criteria for Classifying and Sequencing Pedagogic Tasks

PETER ROBINSON

Introduction: Task Taxonomy, Classification and Programme Design

A fundamental pedagogic question underlies much of the research done into second language (L2) tasks. How do teachers design and deliver a sequence of tasks that sustains learner effort to use the L2, from beginning to end, and which simultaneously leads to L2 learning and development? The Bangalore Project, described by Prabhu (1987) was the first large-scale attempt to operationalise an answer to this question following a syllabus based entirely on tasks. The Bangalore Project's criteria for grading and sequencing tasks, as described by Prabhu, were intuitive, based on teacher decisions about whether and to what degree their 'reasoning demands' posed a 'challenge' to learners. But while intuitions can be shared, often they are not. Consequently, subsequent research has examined these, and other, claims made for the proposed effects of task demands on L2 learning and performance – the aim being to establish an empirical (non-intuitive) basis for decision-making about L2 task design and sequencing. Some characteristics, such as whether tasks are planned (Ellis, 2005); whether there is no fixed solution (open) versus only one (closed) (Long, 1989); or whether the task involves reference to events happening in a currently shared context (Here-and-Now) versus in the past, elsewhere (There-and-Then) (Robinson, 1995), have been quite extensively studied. Other characteristics have been less researched, and a number still remain at the proposal stage. While replications of what empirical findings there are, and some standardisation of measures used to arrive at them, continue to be needed, what has become clear is that there are multiple task effects on L2 learning and performance, as has also been found in the many task domains studied in the broader fields of behavioual, and psychological science.

Task taxonomy

What is still unclear, however, is how these different characteristics relate to each other categorically, and what effects they might have in combination during classroom performance – issues which are essential to theoretically and empirically informed L2 task and programme design. For example, giving learners time to plan, or designing a task where only one solution is possible, or a task that requires reference in the Here-and-Now, all seem to be characteristics that are different in 'kind'. And giving learners time to plan an open versus a closed task, or a Here-and-Now versus a There-and-Then task may produce very different results. Do these different 'kinds' of task characteristic, or combinations of them, affect not only performance, but also second language acquisition (SLA) processes in different ways? And how do we go from information like this – if we knew it – to using it in the design of pedagogic tasks for groups of L2 learners? What is needed to help promote answers to such questions is a classificatory system which introduces order into the current data base of findings, enabling categories of the learning and performance demands of tasks to be related to SLA processes, and also to be used in instructional programme design.

Taxonomic descriptions in many areas, such as personality traits (Matthews & Deary, 1998); educational objectives (Bloom, 1956); or cognitive abilities (Carroll, 1993), provide theoretically motivated, testable, rationales for how findings from laboratory studies can be generalized across each other, and to operational settings (e.g. clinical diagnoses of personality; formal settings for L2 instruction). Taxonomic descriptions can therefore guide research and application, but they must also develop to accommodate findings, and theoretical progress. In the early period of SLA research into tasks, many lists of characteristics thought to affect learning and performance were proposed. Understandably, at that time, these lacked to varying degrees, either theoretically motivated taxonomic structure, or empirical support. One notable exception was a taxonomy described by Pica et al. (1993) which classified task characteristics on the basis of interactional criteria, such as the goal of the task (open versus closed) and the direction of information flow (from one person to another, or mutually, both ways). This taxonomy was motivated by theoretical claims, and research into them, about the value of interaction for SLA. Since that taxonomy was proposed, SLA theory and research in information processing frameworks (e.g. Johnson, 1996; Robinson, 1996; Skehan, 1998) has resulted in a number of quite different proposals for relevant and operationalisable L2 task characteristics. These 'cognitive criteria' now need to be accommodated with 'interactional criteria' in an expanded taxonomic description of L2 tasks. With these issues in mind a taxonomy described later in this chapter identifies categories of L2 task characteristics that have been researched (as well as some, so far, only proposed), and classification criteria that can be drawn on to systematically assign them to one category or another.

Three approaches to task classification

One specific outcome of such a categorisation of pedagogic task characteristics is that it should offer guidance as to how to classify and sequence tasks in task-based approaches to syllabus design. Task classification is logically prior to task sequencing, but at least three different approaches to classifying tasks are apparent in the broader educational and psychological research literature on developing taxonomies of human learning and performance. A choice of one best methodology, however, is not a problem, since all are necessary, and complement each other conceptually, being most appropriate (though not exclusively so) at three different stages, and levels, of implementing L2 instruction, as is illustrated in Table 1.1.

In behaviour descriptive approaches to task classification, categories of tasks are based on observation (both participant and non-participant) and descriptions (which may be elicited by structured, or unstructured interviews from job performers, supervisors, etc.) of what people actually do while performing a task. Examples of these approaches are behavioural typology analysis (Williams, 1977); classification of common denominators of task performances required across industrial or other workplace jobs (McCormick, 1979); and descriptions of jobs on the basis of worker functions or behaviours (Fine, 1974). Behaviour descriptive approaches to identifying target tasks, their subtasks, and the steps needed to perform them (together with a sampling of target domain discourse), are important at the Needs Analysis stage and level of Table 1.1, as input to L2 course design (see e.g. recent papers in Long, 2005), and also to performance-referenced testing of abilities to accomplish either steps in, sub-tasks of, or global target tasks (for placement or achievement purposes) to some criterion measure of success (e.g. Norris et al., 1998; Robinson & Ross, 1996).

Information-theoretic approaches adopt a different level of description, classifying tasks in terms of the information processing stages, and the cognitive processes involved in mediating input to the task performer and the output (spoken, written, and/or other behavioural responses) required for successful task completion. These cognitive-task analytic approaches have been the focus of attempts at task classification, and proposals for task sequencing both outside (e.g. Hollnagel, 2003; Seamster et al., 1997) and within the L2 literature (e.g. Bygate et al., 2001; Robinson, 1996, 2001a, 2005a; Robinson et al., 1995; Skehan, 1998, 2001). While sharing the same information-theoretic approach to describing task characteristics, not all of these proposals share the assumption that classification of pedagogic tasks be based on the target tasks identified in the needs analysis, as will be described below.

A third approach to task classification, the ability requirements approach, classifies tasks in terms of the human cognitive abilities required to perform them effectively (Carroll, 1993; Snow et al., 1984). Clearly, L2 learners differ in their strengths in abilities drawn on during information processing (such as working memory capacity), and these differences, as well as differences in the information processing demands of pedagogic tasks themselves, will affect the outcomes of pedagogic task performance for individuals. Research into interactions between L2 learner information processing abilities, and the information processing demands of tasks has begun to be systematically approached in recent years, and this will be important to identifying aptitudes for different kinds of task performance and L2 processing, and so to matching learners to, or supporting performance on, various pedagogic task types and practice sequences, as will also be described below.

Programme design

Articulating L2 task classification and sequencing decisions, using these analytic methodologies and across the stages and levels of programme design illustrated in Table 1.1, therefore presents a number of implementational challenges. Based on the description of target tasks, pedagogic tasks need to be designed, classified using information-theoretic criteria, then sequenced so as to increase L2 processing demands to target task levels. These decisions themselves need to be formatively evaluated, and revised, based on evidence of whether and to what extent they lead to success on target task assessments. Aptitude profiles, once developed, need to be used to adjust sequencing decisions for individual learners, perhaps by increasing amounts of task practice for (some) learners low in the abilities (some) task demands draw on. The following sections further address issues introduced here: (1) articulating classification and sequencing decisions with needs analysis and assessment of learner abilities; (2) other constraints on pedagogically relevant classificatory systems, and a taxonomic description of L2 pedagogic task characteristics which aims to meet them; and (3) some implications of this taxonomy, and the classification criteria it includes, for task sequencing and L2 syllabus design.

Formal Settings for L2 Instruction: Articulating Needs, Task and Learner Analyses

An instructional-design theory offers explicit guidance on how to effectively articulate learner needs, instructional tasks, and learner abilities, and on how to implement and evaluate programmes of instruction, across a variety of formal settings (Reigeluth, 1999). Theories differ over how this articulation is best achieved. Similarly, various proposals for L2 instruction afford an important role to tasks but differ in their rationales for how task-based instruction can promote SLA processes, and also with respect to which of the stages in Table 1.1 are necessary to, and so should be articulated during programme delivery and evaluation. These theoretical and implementational differences are summarised below, since they have each led to a variety of proposals for L2 task classification and sequencing, some of which are complementary, and some of which are not.

Needs analysis

Depending on the domain, and so content to be learned, instructional-design theories differ in their recommendations for how units of instruction be classified and sequenced, but many agree that an important first step is the analysis of learner needs and that subsequent classification and sequencing of instructional units be based on it. For example, Gagne et al. (2005) describe the first two steps in the 'Analysis' component of their ADDIE (Analyse, Design, Develop, Implement, Evaluate) model of instructional design as follows: '1. First determine the needs for which instruction is the solution. 2. After identifying the needs to be served by the course and some of the situational factors, then conduct an instructional analysis ... an important outcome [of which] is task classification' (Gagne et al., 2005: 24–25). Consistent with this, Long (1998) has argued that methodological choice of task as a unit of analysis in the identification of learner needs and in syllabus design allows both of these stages of programme development to be articulated in theoretically coherent, and practically complementary ways, identifying six operational steps involved in the design and delivery of task-based language teaching: (1) conduct a needs analysis; (2) from the needs analysis identify target tasks; (3) from the target tasks derive pedagogic task types; (4) grade and sequence the pedagogic tasks; (5) deliver instruction using appropriate methodology; (6) evaluate instruction using criterion-referenced measures of task performance.

Needs analysis and pedagogic task classification

Articulating task-based needs analysis and task design has two consequences for task classification criteria. They should be consistent with (that is, map on to, in conceptually coherent, and operationally feasible ways) the behaviours involving language use identified by the needs analysis. And using the criteria to grade and sequence pedagogic tasks should facilitate both language performance (access to and effective deployment of existing L2 knowledge under increasingly demanding processing conditions) as well as language development (since in many cases learners will not only have to execute what they know of the L2 more effectively to be successful on target tasks, but also acquire new L2 knowledge).

While accepting that needs analysis is necessary, however, other approaches to L2 instructional design adopt different units of analysis, classifying the language (not tasks) needed to communicate in a 'purposive domain', and then using this (sociolinguistic, rhetorical functional) linguistic classification as a basis for the syllabus (e.g. Crombie, 1985; Munby, 1978). Consistent, in methodological principle, with language-based needs analyses some recent approaches to task-based instruction advocate using tasks as a means of delivering a linguistically (at least in part) specified and sequenced syllabus (Johnson, 1996; Nunan, 2004; Willis, 1990). Others have argued that a needs analysis, of any kind, is not necessary, opting to base syllabus design, and task classification decisions on intuitively determined estimates of the difficulty of pedagogic tasks (Prabhu, 1987), or on decisions about how sequences of tasks can lead to the balanced development of linguistic accuracy, fluency and complexity, defined independently of specific target task demands (Skehan, 1998).

Task classification and ability assessment

Proposals for the use of tasks in instructed L2 settings differ not only with respect to the nature, or necessity, of needs analysis, but also in the importance they attach to, and so the provision they make for, assessing learner abilities and individualising instruction. While task-based approaches have often been claimed to be learner-centred (e.g. Candlin, 1987; Nunan, 1993) very few have described in any detail how pedagogic tasks could be articulated with ability-requirement analyses, with the aim of matching learners to, and supporting performance during, instructional options in task practice, and sequencing. Elsewhere a distinction has been made (Robinson, 2001a, 2001b) between task complexity (dependent on differences in the intrinsic processing demands of tasks, such as simple addition versus calculus, or exchanging business cards, versus conducting business meetings in the L2) and task difficulty (dependent on the abilities learners bring to tasks, such as aptitude for math, or for L2 learning and use). Task complexity, and levels of cognitive abilities affecting perceptions of task difficulty clearly interact in differentiating success on tasks for L2 learners, and also the linguistic outcomes (accuracy, fluency and complexity) of L2 task performance. However, the relative lack of progress in researching the important issues of how they do interact, and of which abilities are most widely predictive of differential L2 success, and variation in linguistic performance, no doubt hinges on the absence of currently agreed criteria for classifying, and distinguishing between, the most pedagogically relevant information processing demands of tasks. This itself partly reflects the absence of an agreed set of superordinate task categories and a classificatory system that relates them to SLA processes and L2 performance demands in a principled way. This is the issue to which we turn next.

(Continues…)

---

Excerpted from "Investigating Tasks in Formal Language Learning"
by .
Copyright © 2007 María del Pilar García Mayo and the authors of individual chapters.
Excerpted by permission of Multilingual Matters.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.

---