Collaborative Dialogue during Tasks in Synchronous Computer-Mediated Communication

This dissertation, drawing on Swain's (2000; 2005) notion of collaborative dialogue, aimed to explore whether learners participating in SCMC based tasks engage in collaborative dialogue and, if so, to investigate the extent and characteristics of this behavior. In order to provide insights into the task type that may promote collaborative dialogue most effectively, the study compared two task types that are claimed to be contrasting (Swain & Lapkin, 2001) with regard to the degree of attention to form they promote: jigsaw and dictogloss. The following research questions were addressed: (1) To what extent do learners attempt to solve linguistic problems by way of collaborative dialogue during SCMC based tasks? (2) What are the characteristics of the collaborative dialogue with regard to focus, outcome, verbalization, collaboration, and source? (3) Does task type (dictogloss and jigsaw) have an effect on the amount of collaborative dialogue? (4) Does task type (dictogloss and jigsaw) have an effect on the characteristics of collaborative dialogue? The study involved 54 Turkish speaking English-as-a- Foreign-Language (EFL) learners. Learners were enrolled in four intact classes at two Turkish universities in Istanbul. Two jigsaw-dictogloss pairs were designed and matched for content. Following a practice task session, randomly formed dyads of students from each class performed one jigsaw and one dictogloss by using the MSN Messenger instant messaging and CoWord collaborative editing software. Language-Related Episodes (LREs) were identified in learner-learner interaction and categorized based on five levels: focus, outcome, verbalization, collaboration and source. It was found that LREs did occur during task-based SCMC. However, the frequency of LRES seemed to be low in comparison F2F studies. Results also showed that the dictogloss task elicited a higher number of LREs than the jigsaw task. At descriptive level, tasks were different with respect to the characteristics of LREs. While the dictogloss was characterized by orthographic and correctly solved LREs, the jigsaw was characterized by unresolved LREs. The results were discussed in light of the arguments about the effects of limited information-processing capacity on task performance (Skehan, 1998).