| Some recent SLA studies have used verbal protocols to gather data on learners' cognitive processes (e.g., online think alouds in Bowles, 2003, 2004; Leow, 1997a, 1998a, 1998b, 2000, 2001a, 2001b; Rosa & Leow, 2004a, 2004b; Rosa & O'Neill, 1999). However, the potential effects of verbalization on learners' performance have only been investigated in two studies, Bowles and Leow (in press) and Leow and Morgan-Short (2004), leaving this methodological issue open. Additionally, an ever-present topic of debate among second language acquisition researchers and language teachers is how and when to correct students' errors in the L2. Recently, SLA studies (e.g., Nagata & Swisher, 1995; Rosa & Leow, 2004b; Sanz & Morgan-Short, 2004) have produced inconsistent results regarding the effects of explicit versus implicit written feedback on L2 development in computerized tasks. This study addressed both the reactive effects of verbal protocols on L2 development and effects of type of feedback in a CALL problem-solving task.; Participants were first-semester students of Spanish randomly assigned to interact with a series of task-essential (Loschky & Bley-Vroman, 1993) CALL mazes in one of the following experimental conditions: [+metalinguistic + explicit feedback; +metalinguistic - explicit feedback; -metalinguistic + explicit feedback; -metalinguistic - explicit feedback]. The targeted structure is the dative experiencer construction involving the Spanish verb gustar, in which sentences display non-canonical word order.; In this study, results showed that there was a significant interaction between verbalization and feedback on the production of old exemplars, although there was no effect for verbalization on production of new exemplars, regardless of the type of feedback provided. Type of feedback did have a differential effect on participants' written production of both old and new exemplars of the targeted structure, with those participants who received explicit feedback performing significantly better on the immediate post-test than those who received implicit feedback. By the time of the delayed post-test, however, the two feedback groups performed statistically similarly. Finally, verbalization was reactive for latency in this study, due to significant differences between the metalinguistic group on one hand and the control and non-metalinguistic groups on the other. |
