A Brief History of Task-based Language Teaching

Introduction

The first edition of Richards and Rogers Approaches and Methods in Language Teaching was published in 1986. It included 'Communicative Language Teaching (CLT)' but not 'Task-based Language Teaching (TBLT)' in the list of methods/approaches it considered. However, in the second edition, published in 2001, 'Task-based Language Teaching' was now listed under the general heading 'Current Communicative Approaches'. Between 1986 and 2001, TBLT emerged as a sufficiently well-defined approach to warrant separate treatment by Richards and Rogers.

In this chapter, I will first consider the major influences on the emergence of TBLT followed by an account of the proposals for a task-based approach that appeared in the 1980s. I then examine how TBLT subsequently developed, focusing on key issues relating to the design of a task-based syllabus and the methodology for implementing tasks. I will also briefly consider how TBLT has been adapted to computer-mediated environments and take a look at task-based assessment. The chapter concludes with an account of what evaluations of task-based programmes have shown about the effectiveness of TBLT.

Background to the Emergence of TBLT

The importance of including tasks in a language curriculum was affirmed in the CLT movement of the 1970s and 1980s. TBLT grew out of this movement with further inputs from early research in second language acquisition (SLA).

Communicative language teaching

CLT drew on theories of language that viewed language not just as a set of formal structures but as a means of communication. Hymes (1971) outlined a theory of communicative competence that accounted for both what is formally possible in a language (i.e. grammatical) and what is feasible and acceptable in terms of performance. Halliday's (1973) model of language took as its starting point the functions that language served rather than the formal properties of a language. He distinguished three broad functions (the ideational, interpersonal and textual), each of which was elaborated into a series of semantic networks which were then related to their linguistic exponents. Hymes and Halliday's theories fed directly into the recognition that 'there is more to the business of communicating than the ability to produce grammatically correct utterances' (Johnson, 1982) and to proposals for teaching language as communication.

Wilkins (1976), drawing on Halliday's theory of language, proposed replacing the traditional structural syllabus with a notional syllabus consisting of an inventory of language functions (e.g. expressing agreement or disagreement), semantico-grammatical categories (e.g. expressing time, quantity and space) and modal-meaning categories (e.g. expressing certainty and commitment). He argued that a notional syllabus provided a basis for an 'analytic' way of learning. That is, learners pass through a series of approximations to the target language, gradually accumulating the linguistic resources required to perform the various notions. In this respect it differed radically from the traditional 'synthetic' approach where linguistic items are taught and mastered incrementally and not synthesized until the final stages. A notional approach was seen as having high 'surrender value' as it helped learners to communicate from the start. It also afforded an ideal means for defining learners' communicative needs and therefore appealed to course designers concerned with specific purpose teaching (e.g. Swales, 1987). Wilkin's ideas informed the work of the Council of Europe's unit/credit system for teaching foreign languages at different levels of proficiency (e.g. the Threshold Level and Waystage Level), where each level was specified in terms of notions and the linguistic exponents for expressing them. This led ultimately into the Common European Framework (Council of Europe, 2011) in which different levels of proficiency are described in functional (i.e. 'can do') rather than linguistic terms.

First attempts at developing teaching materials based on a notional syllabus (e.g. Abbs & Freebairn, 1982) utilized the existing techniques and procedures of structural courses. That is, the linguistic forms for expressing each notion were presented in situations and then practised in mainly controlled exercises. Thus, although the organizational framework of a language course had changed, the methodology had not. It was still what White (1988) called 'Type A' – it was 'other directed'. There was, however, a growing recognition of the need for a communicative methodology reflecting White's 'Type B' approach, where the emphasis is on the process of communicating and on 'doing things with or for the learner'. Johnson (1982), for example, advocated what he called the deep-end strategy, where 'the student is placed in a situation where he may need to use language not yet taught' so as to activate 'the ability to search for circumlocutions when the appropriate language item is not known' (1982: 193).

Publications began to appear with ideas for communicative tasks (e.g. Klippel, 1984). Describe and Draw, for example, involved students working in pairs with Student A attempting to draw a picture or diagram described by Student B. These tasks were to be judged not in terms of whether learners used language correctly but in terms of whether the communicative outcome was achieved (i.e. whether Student A succeeded in drawing the picture/diagram accurately). Tasks had arrived as a major tool for language teachers.

At this time, it was common to distinguish two types of language work depending on whether the focus was on 'accuracy' or 'fluency', with both seen as important (Brumfit, 1984). Providing opportunities for students to use their linguistic resources freely by performing communicative tasks in small group work catered to 'fluency', which Brumfit defined as 'the maximally effective operation of the language system acquired by the student so far' (1984: 57). However, Brumfit also stressed the importance of accuracy work involving more traditional types of instruction. The question that then arose was how to combine fluency and accuracy work in a language curriculum. Johnson (1982) suggested the answer lay in a 'communicative procedure' consisting of three stages. In Stage 1 the students perform a communicative task using whatever resources they have available (i.e. the deep-end strategy). In Stage 2, the teacher presents those linguistic items which the students' performance of the task showed they had not yet mastered. In Stage 3, these items are drilled if necessary. Brumfit went a step further by proposing an integrated language curriculum consisting of separate accuracy and fluency components, with accuracy dominant initially and fluency gradually taking over. Brumfit's ideas about an integrated curriculum are considered further in Chapter 10.

CLT was an 'approach' rather than a well-defined 'method'. Howatt (1984) distinguished a weak and strong version. In the weak version teaching content was defined in terms of the linguistic realizations of notions and functions but the methodology remained essentially the same as in the traditional structural approach – a Type A approach. In the strong version, the content of a language programme was specified in terms of communicative tasks and the methodological focus was on fluency – a Type B approach. There were also proposals for combining the two approaches as in Johnson's 'communicative procedure' and Brumfit's modular curriculum. As we will see, this attempt to encourage fluency while not neglecting accuracy figures too in TBLT.

CLT has had a major impact throughout the world. Nunan (2003), for example, reported that the educational policies of seven countries in the Asian region mandated the use of CLT. Its influence in Europe has been even greater with the European Common Language framework providing the basis for the teaching and assessment of languages throughout the region. Just about every course book emanating from major publishers today lays claim to being 'communicative'. As its influence has spread, however, CLT has become increasingly less well defined. Littlewood (2014) noted its vagueness, commenting 'the most common understanding has been that it means teachers including communicative activities in their repertoire' (2014: 350). Today even approaches based on a structural syllabus and a presentation–practice–production (PPP) methodology lay claim to being 'communicative' on the grounds they include a communicative task in the final production stage. Clearly, such an interpretation represents an even weaker version of CLT than Howatt's 'weak version', creating the space for TBLT to take over the reins of the 'strong version'.

Second language acquisition research

SLA research also contributed to the emergence of TBLT. The 1960s and 1970s saw a number of cross-sectional and longitudinal studies of learners acquiring an L2 naturalistically (e.g. Cancino et al., 1978; Dulay & Burt, 1973). The cross-sectional studies examined the accuracy with which a set of English grammatical morphemes (e.g. Ving, plural-s, Ved, 3rd person-s) were produced by different groups of learners. They resulted in the claim that there was an acquisition order that was common to all learners irrespective of their first Languages (L1) or their age. Furthermore, a very similar order was found in classroom learners suggesting that instruction did not have a major impact on the developmental route learners followed. The longitudinal studies focused on individual learners' acquisition of specific grammatical structures (e.g. negatives, interrogatives and relative clauses). They showed that learners did not progress directly from zero knowledge of these structures to target-like use but instead passed through a series of stages involving 'transitional constructions' en route to the target form. Progress was gradual and often very slow and at any one stage of development considerable variability was evident in the use of those constructions that had been acquired up to that point. Furthermore, it was clear that learners did not set about achieving target-like use of one structure before embarking on another. Rather they worked on several structures concurrently. This research led to the claim that learners might have their own 'built-in syllabus' for learning an L2 (Corder, 1967).

Drawing on these studies, Krashen (1981, 1985) proposed a theory consisting of five hypotheses. The central hypothesis was the Acquisition-Learning Hypothesis. This proposed a distinction between 'acquisition', defined as the 'subconscious process identical in all important ways to the process children utilize in acquiring their first language' and 'learning', defined as 'the conscious process that results in "knowing about" language' (Krashen, 1985: 1). Krashen argued that true proficiency in an L2 depended on 'acquisition' and that the 'natural order' reported in studies of L2 learners was a manifestation of 'acquisition'. 'Learning' had some value for monitoring output derived from the acquired language system but only in a very limited way. Krashen attached most importance to the fourth of his hypotheses – the Input Hypothesis – which stated that 'humans acquire language in only one way – by understanding messages or by receiving "comprehensible input"' (1985: 2). The hypothesis had two corollaries. The first was that 'speaking is the result of acquisition, and not its cause'. In other words, Krashen saw acquisition as entirely input-driven and claimed that 'speech cannot be taught'. The second corollary was that 'if input is understood, and there is enough of it, the necessary grammar is automatically provided' and thus there was no need to teach the grammar of a language.

At the same time that Krashen was developing his theory of L2 acquisition, Terrell was experimenting with a new way of teaching Spanish as a foreign language in the United States that did not involve the presentation and practice of discrete grammatical structures. Krashen's hypotheses provided the theoretical rationale for this approach. In 1983 Krashen and Terrell published a book called The Natural Approach. The book outlined both the theoretical basis of the approach and a curriculum for implementing it. The stated goal was to develop the ability to communicate in the target language. Objectives were specified in terms of the topics the teacher and students would communicate about (e.g. clothing; sports and games; work activities). In accordance with the Input Hypothesis, the instructional activities for the early stages were comprehension-based and even later 'the great majority of class time is devoted to activities that provide input for acquisition' (1983: 58). Emphasis was placed on ensuring a constant flow of comprehensible input through teacher talk. Early production activities were an extension of the input-based activities, requiring simple one-word utterances from the students. As learners progressed, a greater variety of 'acquisition activities' figured, including dialogues, interviews, preference ranking activities, personal charts and tables, and 'tasks' where the teacher or students chose a specific activity such as 'washing a car' and then described the components of the activity. The emphasis throughout was on activities that cater to 'acquisition' but there were also suggestions for teaching monitor use. These included both direct grammar explanation and traditional practice exercises (e.g. blank-filling, sentence combination; substitution) aimed at 'learning'. However, Krashen and Terrell recommended that learning activities should be delayed until students were spontaneously producing short sentences and should focus on simpler grammatical rules such as inflections and simple word order changes. They commented 'learning activities will be used judiciously and in some cases not at all' (1983: 148).

The Natural Approach is based on the principle that L2 development does not require intentional learning on the part of learners. It emphasizes activities that focus learners' primary attention on meaning and thereby cater to incidental acquisition. TBLT is based on the same principle. However, in some other important respects The Natural Approach differs from TBLT. It was premised on the assumption that acquisition is an entirely subconscious process and it gave little recognition to the important role that production plays in language learning. However, the research on order and sequence in L2 acquisition along with Krashen's claim that 'acquisition' could not be taught fed into the early proposals for TBLT.

First Proposals for TBLT

'Tasks' figured in both early CLT and The Natural Approach but in neither were they conceived of as the units around which a complete language course could be built. In CLT they served either as the means for identifying which specific properties of language needed to be taught as in Johnson's (1982) 'communicative procedure' or as an adjunct to a traditional structural approach to cater for the development of 'fluency' as in Brumfit's (1984) integrated curriculum. In The Natural Approach, course content was specified in terms of topics and situations with tasks serving as one type of activity for generating comprehensible input.

It was not until the mid to late 1980s that the first proposals for a task-based approach appeared. These early proposals (Breen, 1989; Candlin, 1987; Long, 1985) were largely programmatic in nature. They focused on the rationale for a task-based syllabus and outlined how to design and evaluate a task-based curriculum. Prabhu (1987) provided the first complete account of a task-based course, whereas Nunan (1989) gave practical advice about how to design tasks along with examples of actual tasks.

Rationale for TBLT

From the start, therefore, there were multiple inputs into the rationale for TBLT. Long's (1985) advocacy of TBLT was premised on the need for any approach to teaching to be 'psycholinguistically based'. Drawing on research in SLA, he argued that 'there is no reason to assume that presenting the target language as a series of discrete linguistic or sociolinguistic teaching points is the best, or even a way to get learners to synthesize the parts into a coherent whole' (1985: 79). He saw an approach based on tasks as providing an 'integrated solution to both syllabus and methodological issues' (1985: 89).