Toward a Theory of Curriculum for Use in Designing Intelligent Instructional Systems

Abstract

Implicit in the approaches being taken by current efforts to create intelligent computer-based instruction is the notion that curriculum is almost an epiphenomenon of knowledge-driven instruction. Early computer-based instruction had little control structure other than an absolutely rigid curriculum and was insensitive to the subtleties of different students’ partial knowledge. As a result there was a reaction in the direction of representing the students’ knowledge as a subset of the target or goal knowledge to be taught and simply deciding de novo after each piece of instruction what piece of missing knowledge to teach the student. I am convinced that goal knowledge is as important to intelligent machine activity as it is to human activity and that it also must be well understood and explicitly represented in an instructional system if that system is to be successful in fostering learning.¹ This chapter presents an architecture for representing curriculum or goal knowledge in intelligent tutors and is thus a first step toward a theory of curriculum that can inform the design of such systems. To illustrate one way in which such a theory can sharpen our ideas about learning and instruction, the later part of the chapter focuses on the concept of prerequisite that is the basis for existing computer-assisted instruction and shows how that concept has been inadequate in the past.