Towards Better Tutorial CALL: A Matter of Intelligent Control

Abstract

Although drill and practice exercises continue to account for a substantial proportion of CALL production, most available foreign language programs suffer from expedient programming and poor pedagogy. At the heart of the problem is the failure of language teaching specialists to involve themselves in CALL research and development. In the absence of AI-derived natural language processors, programs based on anticipated student responses remain the only viable means of creating sophisticated CALL lessons. Special purpose authoring languages offer the most accessible approach to such programming, but are seriously hampered by their lack of flexible feedback capabilities. In particular, much greater author control is needed over pattern matching devices to accommodate the processing of non-differential variations in predicted answers. Provided that basic command structures incorporate variable parameters with pre-set, but alterable, default values, increased author management of response handling need not place unreasonable programming denmands upon the courseware designer. In any event, variable lesson parameters are required to enable students to more intelligently exploit the computer-aided learning environment. To demonstrate the feasibility of implementing flexible response handling control, the essential design features of an Australian developed authoring language, EMU, are examined in some detail.