Bug, Buggy model, Cognitive model, Constraint, Constraint-based model, Constraint-Based Modeling, Declarative knowledge, Learning by rote, Learning from error, Learning mechanism, Learning with understanding, Procedural knowledge, Skill acquisition, Tutoring
The ideas behind the constraint-based modeling (CBM) approach to the design of intelligent tutoring systems (ITSs) grew out of attempts in the 1980’s to clarify how declarative and procedural knowledge interact during skill acquisition. The learning theory that underpins CBM was based on two conceptual innovations. The first innovation was to represent declarative knowledge as constraints rather than chunks, propositions, or schemas. The second innovation was a cognitive mechanism that uses the information in constraint violations to revise and improve a partially mastered skill. This learning theory implied that an ITS could be built around a set of constraints that encode correct domain knowledge, without an explicit or generative model of buggy versions of a skill. Tutoring systems based on CBM have proven effective in multiple educational settings. CBM is limited in its focus on learning from errors. A broader learning theory, the Multiple Modes Theory, is outlined, and its implications for the design of more powerful ITSs are discussed.