CONSTRUCTIVIST USES OF EXPERT-SYSTEMS TO SUPPORT LEARNING

Production rule expert systems are used most frequently to represent t he knowledge of an expert in order to provide advice to novices or mod el expertise in intelligent systems. We argue that these uses of exper t systems, while useful and productive of learning, supplant much of t he cognitive processing in the user or learner and impose a model of t hinking that requires learners to replicate that model in their own th inking. Because meaningful learning depends upon the depth and quality of processing by the learner, we contend that production rule expert systems best support learning, not as models of thinking to emulate, b ut rather as formalisms for representing the knowledge of learners or as tools for modeling the content that is being studied in a causal ma nner. When learners create expert systems to represent their own think ing or to represent the causal logic in a content domain, they are pro cessing content at a deeper level and constructing knowledge that is m ore meaningful to them. In this paper, we first contrast objectivist a nd constructivist epistemologies of learning. Then we analyze two obje ctivist applications of expert systems - advisors and expert/student m odels in intelligent tutors. Later, we describe and analyze three incr easingly constructivist applications of expert systems, including stru cturing personal feedback, representing personal knowledge through sim ulating cognitive processing, and engaging critical thinking as cognit ive study tools.