An epistemological approach to intelligent tutoring systems

Abstract We advocate a new approach to building Intelligent Tutoring Systems (ITS) based upon the concept of familiar schemes. We describe the pedagogical shortcomings of approaches that do not sufficiently take into account the structures that learners bring to bear in new learning situations. By determining the nature of the schemes of the individual learner, and then attempting to provide the types of problems, analogies, and explanations that will best fit with that individual, an ITS can best assist a learner in developing the heuristics and ‘administrative skills’ to solve non-trivial problems. We describe the components of a system built according to these goals, and discuss its advantages and shortcomings.