This paper presents two new corpus-based studies of feedback in the domain of teaching Spanish as a foreign language, concentrating on the type and frequency of different feedback moves, as well as what happens in the moves that follow the feedback. In particular, as well as looking at positive feedback, it concentrates on two general kinds of negative feedback strategies: (1) Giving-Answer Strategies (GAS), where the teacher directly gives the desired target form or indicates the location of the error, and (2) Prompting-Answer Strategies (PAS), where the teacher pushes the student less directly to notice and repair their own error. Investigating the GAS/PAS distinction sheds light on the relative importance for Intelligent Computer-Assisted Language Learning (ICALL) systems of knowledge construction from interaction, which many believe is crucial for effective learning from ITS. The main finding here is that, although GAS occur more frequently than PAS in both corpora, it is the PAS that are more effective, in terms of eliciting explicit repairs by the students. The first study takes place in a classroom context, whereas the second, smaller, study looks at tutorial interactions. This makes it possible to investigate the extent to which the mode of interaction influences the frequency and effectiveness of feedback moves, as well as to look at how concepts such as "wait time" are relevant to explain moves that are ineffective. The paper concludes by using these results to make recommendations about how to choose appropriate feedback moves in ICALL systems.
|Number of pages||34|
|Journal||International Journal of Artificial Intelligence in Education|
|Publication status||Published - Nov 2007|