ABSTRACT
Intelligent tutoring systems, by their name, are supposed to bring intelligence in some way to the task of computer-based instruction. There are two key places for intelligence in an ITS. One is in the knowledge the system has of its subject domain. The second is in the principles by which it tutors and in the methods by which it applies these principles. Clearly, human tutors are effective only when they possess both kinds of intelligence; lack of either component leads to instructional ineffectiveness. Humans cannot tutor effectively in a domain in which they are not expert, and there are also inarticulate experts who make terrible instructors.
