An intelligent tutoring system for helping children aged 7 to 15 to acquire a second language

✍ Scribed by Charles G. Button

• Publisher: Springer Netherlands
• Year: 1989
• Tongue: English
• Weight: 947 KB
• Volume: 18
• Category: Article
• ISSN: 0020-4277
• DOI: 10.1007/bf00121218

✦ Synopsis

Traditional approaches to teaching a foreign language in Artificial Intelligence (AI) systems restrict the computer to correcting grammar or vocabulary; this is similar to structural teaching strategies (STS) in which a second language (I.22) is leamed in relation to a first language (L1). By contrast, this paper suggests that the role of a computer tutor in language learning need not necessarily be confined to that of a foreign language expert. It investigates ways in which to represent both the learner's L2 skills and the foreign language expert. These lead to a novel computational method for teaching a foreign language, suitable for children aged 7 to 15.

The paper shows that Schank's conceptual dependency (CD) theory is one way of representing knowledge for a specific domain such as eating habits. It then goes on to show that the computer can monitor a dialogue about eating habits by pattem matching the Learner's and Corrector's conceptual dependency form against its own predefined script, and can then formulate further questions about the same topic, using a set of production rnies. This enables it to act as an intelligent tutoring system for second language acquisition. Its role is supportive but similar to a communicative teaching strategy (CTS) in that L2 is learned independently of L1. The paper then shows how this communicative teaching strategy would fit in with a leaming environment consisting of electronic mail penpals who act as both correctors and learners for their respective L2.