Language learning in complex virtual worlds: Effects of modality and task complexity on oral performance between virtual world and face-to-face tasks

Virtual worlds have been identified as a potentially beneficial domain for language learning due to various cognitive and affective affordances such as immersive content, access to native speakers, and motivating properties. However, research on computer-mediated communication (CMC) has largely ignored the use of virtual worlds as a possible domain for communication. Additionally, the game-based language teaching (GBLT) sub-field of CALL has focused too narrowly on specific virtual world affordances, overlooking how communicating in such complex domains may affect learner output, particularly in comparison with face-to-face communication. Thus, the main aim of this study is to explore the potential differences in learner oral performance as they conduct tasks via two oral modalities: within a virtual world and face-to-face. Twenty participants (10 dyads) conducted six dialogic tasks, organised by modality into three task-pairs. Quantitative data was collected via transcribing audio recordings of all sessions. The data were analysed in terms of learners’ output complexity, accuracy and fluency using appropriate measures for each. Post-task questionnaires were employed to gauge perceptions of task difficulty, and therefore validate the researcher’s presumptions of task complexity. This data was also used to provide insight into findings from the quantitative data. Results suggest that virtual world tasks may hinder output fluency. However, complexity and accuracy were not significantly affected by mode. Instead, task complexity and type had a more considerable influence on these constructs. Lexical density was higher when conducting virtual world tasks, and, regardless of the increased cognitive demands posed by the virtual world, participants preferred to undertake tasks in this domain. Implications are provided regarding virtual world task design and the cognitive and affective affordances of virtual worlds for language learning, specifically for classroom contexts. Finally, the limitations of this study inform avenues for future research