Abstract:

This paper discusses the use of student models in computer-assisted language instruction (CAI) software. It examines three standard techniques for modeling student knowledge, and shows why they do not work well for CALI. It then argues that modeling which focuses on skills in specific contexts is particularly suitable for language learning because the ability to use skills appropriately is a crucial aspect of knowing a language. The way in which skills-in-context modeling can be applied is discussed in relation to DIAGNOSIS, a program developed recently by the author and his colleague Philip Thomson.

Author: Alan Bailin