| The construct of cognitive complexity has played an increasingly important role in studies on task design, which aim to explore how increases in the cognitive complexity of tasks differentially mediate interaction and learning outcomes (Kim, 2009; Gilabert, Baron, & Llanes, 2009; Kim & Tracy-Ventura, forthcoming; Nuevo, 2006; Revesz, 2009, forthcoming; Revesz, Sachs & Mackey, forthcoming; Robinson, 2001, 2007). The Cognition Hypothesis (Robinson 2001a, 2003, 2005a, 2007b, 2010; Robinson & Gilabert, 2007) predicts that more cognitively complex tasks will result in greater incorporation of forms made salient during interaction, and that cognitive individual differences will affect learners' performance as the tasks increase in complexity. In addition, researchers have posited that modality may play an important and differential role for SLA; however, studies on computer-based interaction have thus far been tangential to task-based research. The research on the effects of increases in cognitive complexity on learning is so far inconclusive, with no study to date comparing its effects in different modes.;The current study sought to fill this gap by operationalizing the Cognition Hypothesis, looking at the effects of increases in task complexity and modality on L2 development alongside the provision of recasts. Learners engaged in two-way interactive tasks for which they had to come up with the intentional reasons of peoples' actions (+complex) or not (-complex). In addition, learners carried out the task with the researcher in either the face-to-face (FTF) or computer-mediated communication (CMC) mode. 70 intermediate-level learners of Spanish were randomly assigned to one of the following groups: FTF+C, FTF-C, CMC-C, and CMC+C. The targeted linguistic item was the Spanish past subjunctive. Uptake was explored as a mediating variable for learning, and working memory capacity (WMC, measured via the OSPAN, CSPAN, and RSPAN; cf. Conway, Kane, Bunting, Hambrick, Wilhelm, & Engle, 2005) was explored as a moderating variable. Independent measures of task complexity were also collected, including time judgments of the tasks and anxiety and perceived difficulty questionnaires.;Results indicated that engaging in more cognitively complex tasks yielded higher development, but differentially so according to mode. The +complex task resulted in the highest gains for the FTF mode, but hardly any development for the CMC mode. The -complex task in the CMC mode led to the highest amount of development. Contrary to expectations, uptake and WMC did not predict learning, and in fact were negatively and significantly related to development in the FTF+C group.;In order to explicate these findings, a deeper probe into the concurrent processes demonstrated by the participants during interaction was carried out. The follow-up analysis revealed that while some participants noticed the form, others demonstrated exemplars of hypothesis testing and rule formation, features associated with awareness at the level of understanding (cf. Leow, 1997; Rosa & Leow, 2004; Rosa & O'Neill, 1999). In fact, awareness and production of the form during the treatment appeared to be the clinching factors that explained the superior performances of FTF+C and CMC-C.;To conclude, it was found that (1) increases of cognitive complexity in the FTF mode appear to promote deeper processing and subsequent higher level of awareness, which was found to significantly predict L2 development in this study, (2) L2 development in CMC can be extended to the FTF mode, (3) modality and task complexity interact in unique ways for SLA, (4) neither uptake nor WMC was found to predict L2 development, and (5) a more fine-grained operationalization of what constitutes uptake after feedback may be needed in future research employing this concept. |
