Abstract/Summary

A paramount discussion on the cognitive approaches to Task-Based Language Teaching (TBLT) is the issue of predicting the systematic effects of cognitive task complexity (TC) on second language (L2) performance and the factors that interact with the effects of TC (Baralt, 2013; Kormos & Trebits, 2012; Révész, Michel, & Gilabert, 2016; Robinson, 2007; Skehan & Foster, 1999; Tavakoli, 2014). Two competing models of TC, i.e. the Cognition Hypothesis (Robinson, 2001) and the Limited Attentional Capacity (Skehan, 1998) have informed this research agenda. However, there is still a need to more carefully define and systematically operationalise intentional reasoning (IR) as a TC variable within task-based research. More importantly, more research is needed to investigate the interaction between the effects of TC and L2 learners’ individual differences. This thesis draws on the findings of two inter-related studies. Study One aimed to investigate whether increasing TC through IR demands would be associated with an increase in syntactic complexity, lexical complexity, and accuracy, and a decrease in fluency of L2 learners’ oral performance. This study further investigated whether the +IR task would be perceived as more difficult. IR was operationalised on two levels i.e., task instructions and task content. A mixedmethods within-participants study design was conducted with 20 Jordanian secondary school students who performed two video-based oral narrative tasks with varying degrees of IR and completed a retrospective questionnaire on their perceptions of task difficulty (TD). The design was counter-balanced to avoid any practice or order effects. Following the analysis of the participants’ oral performance which was operationalised through a number of CALF measures, the findings of Study One revealed a systematic positive impact of IR on syntactic complexity and accuracy, and a negative impact on lexical complexity. However, fluency was not significantly affected by the IR demands. The participants perceived the +IR task as more difficult than the -IR task. They further attributed the difficulty to the IR demands which were required by task instructions and to the unfamiliarity and unpredictability of the content of the +IR video clip. These mixed results acknowledged the need to consider a possible interaction between the learners’ individual differences and the effects of IR demands on L2 speech production. Study Two was then designed to examine: 1) the effects of manipulating TC by IR in oral narratives on learners’ L2 speech performance and perceptions of TD; 2) whether learners’ individual differences in language proficiency (LP) and working memory (WM) mediate the effects of IR; and 3) to what extent LP and WM can predict performance on tasks of different degrees of TC. Employing a mixed-methods approach, the study had a 2 x 2 within-between-participants factorial design. The participants were 48 learners of English at a secondary school in Jordan. They performed the same two video-based oral narratives of Study One and completed a retrospective questionnaire to rate their perceptions of TD. A counter-balanced design was used to control for any impact of order or practice. Oxford Placement Test (Alan, 2004) and a set of elicited imitation tasks (Wu & Ortega, 2013) were used to measure the participants’ LP, and backward-digit span tasks in L1 and L2 (Kormos & Trebits, 2011; Wright, 2010) were used to test their WM. The participants’ oral performance was analysed in terms of a number of CALF measures. The quantitative and qualitative data obtained from the questionnaire were carefully analysed. The results confirmed that IR demands resulted in significantly producing more syntactic complexity, accuracy, speed fluency, and filled pausing, whereas lexical complexity decreased in the +IR task. However, no effects were evident on silent pausing or repair fluency. The participants perceived the +IR task as more difficult than the -IR task. The same themes which were mentioned in Study One emerged as the main factors that contributed to the perceptions of TD, i.e. task-induced and task-inherent cognitive demands as triggered by task instructions and content. Even though main effects were detected for LP and WM on some aspects of L2 performance, no interaction effects were significantly observed between TC and LP or WM. The findings further designated LP as a reliable predictor of speech performance with respect to lexical complexity, accuracy, speed fluency, and pausing fluency. However, WM did not statistically explain variations in any aspect of L2 oral performance but correlated significantly with accuracy and lexical complexity. These results imply that considering cognitive task complexity in isolation may provide a too simplistic picture of what is happening during task performance. Furthermore, the study highlights the importance of considering the interrelation between the cognitive demands of a task and its linguistic requirements to explain intentionality when making decisions on what analytic measures of CALF to employ. The findings have also substantial implications for L2 pedagogy and research.