Task complexity, task condition & variation in L2 oral discourse

Abstract

Task-based approaches to language teaching are attracting increasing attentionnfrom teachers, syllabus designers, and SLA researchers. Task is viewed as anparticularly viable unit of syllabus design as it is seen to provide a more systematicnand more psychologically valid basis for syllabus design and implementation thanntraditional approaches to instruction. Of central importance to the use of task-basednsyllabi is the grading and sequencing of tasks, and these processes depend on a betternunderstanding of which factors contribute to task complexity. There is a recognisednneed for more empirical research into how various task factors interact with, and servento determine, second language learning outcomes. A number of studies addressingnthe issue of task complexity have appeared recently (Brindley, 1985, 1987; Long,n1985, 1989; Nunan, 1989, 1993; Robinson, 1995), and taxonomies of relevant tasknfeatures have been proposed (Pica, Kanagy and Falodun, 1993; Robinson, Rahimpournand Urwin, 1996). The present study builds on this body of research by investigatingntask variation as a function of task complexity, which refers to the linguistic andncognitive complexity of the task and is represented in terms of temporal referencen(Here-and-Now vs There-and-Then vs Here-and-Now/There-and-Then) and taskncondition, which reflects the task goal and participants (Open vs Closed). The factorsnare investigated in an experimental study that seeks to address the research question:How do the different hypothetical parameters of task complexity, alone and inncombination, affect the production of oral discourse of L2 learners (as measured bynfluency, accuracy and structural complexity of output)?nBy establishing systematic sources of task-induced variation, the present studynaims to contribute to the growing body of research findings upon which syllabusndesigners and researchers can draw to make principled decisions about grading andnsequencing tasks for both pedagogic and research purposes.nTo address the research question, oral narrative discourse by learners of Englishnas a second language was collected and analysed. Spoken protocols were elicited fromn50 international students, aged 18-40 years, from different language backgrounds. Thendata were collected individually from participants in either the Open or Closed taskncondition. Participants in the Open condition were presented with three narrative tasksnat three levels of task complexity: 1. Here-and-Now ; 2. There-and-Then ; and 3. Hereand-Now/There-and-Then.nParticipants in the Closed condition undertook only levels 2nand 3, due to the nature of the task. The data were analysed in two ways. First, a withininnsubjects design was carried out for the Open and Closed tasks respectively, with taskncomplexity as the within-subjects variable. A mixed design was then carried out, withnTask condition as the between-subjects factor and task complexity as the withinsubjectsnfactor, using only the There-and-Then and Here-and-Now/There-and-Thennresults in the analysis.nThe elicitation task consisted of narratives based on cartoon strips representingnthe two or three levels of task complexity. The recorded narratives were transcribed andnanalysed for accuracy (as measured by Error-free T-units and Target-like Use ofnArticles), linguistic complexity (S-nodes per T-unit, percentage of Lexical Words), andnfluency (Words per Pause). The results of overall MANOVA and univariate ANOVAsnrevealed that there were significant differences among the tasks in terms of accuracynand fluency for task complexity. There was no significant difference among the tasks innterms of linguistic complexity. A significant difference was also found between the Closed and Open conditions for the fluency measure, but not for accuracy andnlinguistic complexity. There was also a trend towards greater accuracy in the Closedncondition.nThe results of study have implications for task grading and sequencing in task-basednsyllabus design.n