ABSTRACT
Tasks can vary in terms of difficulty and implementation conditions, and these in turn may affect the nature of the language produced by the learner. There have been a large number of studies which have investigated the effects of these task properties and implementation conditions on the language the learner produces. To date, results have been mixed and much more detailed research is needed before we can conclude about the potential effects of task type on L2 performance. Of particular interest in this study is the need to investigate the effect of high-stakes environments( i. e. testing situations) on task performance, to determine which linguistic phenomena contribute to perceptions of levels of performance. In the research reported here, we compare the performance of two VCE students of French on the conversation section( section 1) of the oral examination. The nature of the language produced by both learners was analysed in terms of accuracy, complexity and fluency. Overall, findings suggest that fluency was favored by both learners although different strategies were used to achieve it. Pedagogical implications of these findings are discussed.
Key Words
speaking, oral assessment, accuracy, complexity, fluency, exemplar based language, rule based language system, trade-off hypothesis, cognition hypothesis, assessment task.
INTRODUCTION
The Victorian curriculum for Languages studies is informed by a communicative approach to learning, from the early primary years through to the final years of secondary study. Year 12 students completing their final year of language study sit the Victorian Certificate of Education( VCE) oral and written examinations, which also aims to incorporate communicative principles( Victorian Curriculum and Assessment Authority( VCAA), 2010a: 41). Thus classroom activities and the final exam utilize a range of tasks designed to resemble real-world activities which are meant to promote purposeful language use.
In VCE French, the oral examination contributes 12.5 % to the final study score and consists of a 15-minute interaction with two assessors. It is an externally assessed examination and is divided into two sections of seven to eight minutes each. The first section of the examination is labeled a‘ conversation’ and‘ consists of a general conversation about the student’ s personal world, for example, school and home life, family and friends, interests and aspirations’( VCE Study Design, 2010b: 31). The second section( entitled‘ discussion’) focuses on a theme selected and prepared in advance by the student( detailed-study), and aims to‘ enable the student to develop knowledge and understanding of, for example, historical issues, aspects of contemporary society or the literary or artistic heritage of the community’( VCE Study Design, 2010b: 24). The first section is usually considered less cognitively demanding as it focuses on topics familiar to students whilst the second section tends to be more analytical.
Tasks can vary in terms of difficulty and implementation conditions and these in turn may affect the nature of the language produced by the learner. For example, tasks may vary in terms of familiarity, level of abstractness, whether completed individually or in pairs, and whether they are preceded by planning time. There have been a large number of studies which have investigated the effects of these task properties and implementation conditions on the language the learner produces. The principle dimensions used to describe second language performance by researchers have been syntactic complexity, grammatical accuracy and language production fluency. To date research results have been mixed and the explanations offered refer to two contrasting models of language processing and production: Skehan’ s( 1998) Trade-Off hypothesis and Robinson’ s( 2001, 2005) Cognition hypothesis.
Skehan’ s( 1998) Trade-Off hypothesis contends that attentional resources and working memory are limited. The implications of this hypothesis are that when producing language, fluency, accuracy and complexity compete for attentional resources and better performance can be expected on less complex tasks. This hypothesis, for example, was used to explain Skehan’ s findings( 2001) where less demanding tasks( more familiar and concrete) led to greater accuracy and fluency but not complexity. In a more recent version of this hypothesis, Skehan( 2009) contends that fluency can be accompanied by either accuracy or complexity but not both. That is, learners may focus( consciously or subconsciously) on fluency and accuracy or fluency and complexity. Thus it seems that the trade off is mainly between accuracy and complexity.
In contrast, Robinson’ s( 2001, 2005) Cognition hypothesis contends that learners can access multiple attentional pools that do not compete with one another. Furthermore, drawing on Functional linguistics, Robinson claims that complex tasks will lead to more complex and accurate performance because the learner is pushed to use more complex and accurate language to meet the demands of such tasks. Thus according to Robinson, linguistic accuracy and complexity are in league with one another rather than in competition. Indeed, a large scale study by Kuiken and Vedder( 2007a) found positive correlations between challenging tasks and linguistic accuracy and complexity.
Thus, the potential impact of tasks and implementation conditions on learners’ performance is not clear. Another possibly confounding factor is the learners’ goals. In his more recent work, Skehan( 2009) suggests that mixed research findings may be due to the fact that adult learners can prioritize attention to particular areas and that task performance may be the result of participants’ goals and interpretation of task requirements. In other words, it is the learners who prioritize certain areas – complexity or accuracy – and what they choose to prioritize may be different for different learners and in different task situations. For example, Iwashita, Elder and McNamara( 2001) reported increased accuracy on a cognitively demanding task( abstract task) and attributed these findings to the fact that the task was completed as a test. The researchers suggest that learners may pay extra attention to accuracy because they feel that this aspect is paramount in a testing situation, to the detriment of fluency and complexity.
In addition to taking into consideration learners’ goals, researchers now also recognise the need to use a greater range of measures to assess the impact of task properties and conditions on second language performance. For example, some researchers have looked more closely at the relationship between specific task conditions and specific linguistic features( e. g. use of tense-aspect marking in Robinson et al., 2009) or type of errors( Kuiken & Vedder, 2007b). In particular, the focus has shifted beyond linguistic complexity to lexical choice and sophistication( Skehan, 2009). Skehan( 1998) suggests that learners have access to exemplar-based and rule-based
Volume 47 Number 2 5