<p>This study investigates the influence of two task implementation features, the level of task structure and the use of language support, on learner language production during task-based text synchronous computer-mediated communication (text-SCMC) interactions. The study draws on two theoretical sets of claims concerning the process of second language acquisition (SLA). The first, broadly described as cognitive accounts of language learning, the Cognition Hypothesis (Robinson, 2001b, 2003b, 2005) and the Trade-off Hypothesis (Skehan, 1998, 2009), has generated a large body of research on the role of implementation features (a means of varying task complexity) in influencing learner language production. The second, the Interactionist Approach (Gass & Mackey, 2006) has also claimed the facilitative role of interaction in promoting second language production. Most of the studies in both these areas were conducted in face-to-face settings (e.g. Gilabert, 2007b; Michel, Kuiken & Vedder, 2007; Robinson, 2007b; Tavakoli & Foster, 2008; Tavakoli & Skehan, 2005). Because SCMC is growing more pervasive in academic and professional communication, it is timely for empirical research into the effect of task complexity on interaction and language production to be conducted in this setting (Lee, 2008; Smith, 2008). It is this gap that the current study aims to address. The participants were 96 engineering learners at a technical university in Malaysia in an English for Professional Communication course. Using a 2x2 experimental design, the learners were placed in one of four experimental groups defined by high or low task structure (+TS or -TS), and with or without language support (+LS or -LS). Each group was subdivided into teams of four. In each team, the students engaged in a 45-minute chat session performing a simulation of a decision-making task on an engineering problem. The chat exchanges were captured and then analyzed to determine the role of these task implementation features on the occurrence of focus on form sequences and on the accuracy, complexity, and quantity of language produced during the tasks. Results showed that the two task implementation features (+/-TS and +/-LS) influenced the occurrence of language-related episodes (LREs), accuracy, complexity and quantity of output. The findings on the effects of task structure (TS) revealed that the learners engaged in more LREs and their output was more accurate when task performance was highly structured (+TS). However, task structure did not have a significant effect on the structural and lexical complexity of the output nor on the amount of language produced and equality of participation. The findings on the effects of language support (LS) demonstrated that the participants engaged in more LREs and their language use was more accurate when performing the task with language support (+LS) than without it (-LS). In contrast, they produced more complex language when performing the task without language support (-LS). Without language support (-LS), the learners were also found to produce fewer turns but with more words per turn. The finding for equality of participation was non-significant which suggests that participation was not affected by language support. To summarize, the current study lends qualified empirical support to the Interactionist Approach (Gass & Mackey, 2006) and the trade-off effects proposed by Skehan (1998, 2009) in that cognitively simple tasks promoted more accurate, but less complex production than cognitively complex tasks as they apply to taskbased interactions in a text-SCMC context. Additionally, the finding demonstrates partial support for the Cognition Hypothesis (Robinson, 2001b; 2003b, 2005) in that increasing task complexity along the resource-dispersing dimension decreased the accuracy of language production. The visual salience of language in a text-SCMC setting may be an important explanatory factor in accounting for this finding. The study, therefore, provides evidence that the nature of text-SCMC may be facilitative to L2 learning, particularly as a medium for learning of form during communicative practice.</p>
Predicting head-marking variability in Yucatec Maya relative clause production
