@rabLearn: Towards Intelligent Answer Analysis and Feedback in ICALL

This paper describes the development of an intelligent computer-assisted language learning (ICALL: Intelligent Computer Aided Language Learning) system for learning Arabic, especially, the advanced method of feedback and the approach for analysing learner’s answer. The corrective feedback is undoubtedly the attribute generally associated with the ICALL and the interactive language learning environment, it forms the essential component to evaluate the quality of a teachware. As it is the case for other concepts, several traditions converge to lead to the concept of feedback in ICALL: we talk about feedback in the area of the technology, the area of information and communication and other fields. After a brief presentation of the general denomination of this word, we show what the theories of learning languages say on the feedback and the correction of errors produced by learners. We will study aspects of feedback specific to the field of ICALL by describing our language learning environment called @rabLearn which is based on the use of NLP (Natural Language Processing) tools such as our Morphological analyser [7, 8] and we seek to know in what ways analysis of learner’s answers by the computer improve the quality of the feedback.