Argumentation Scheme-Based Argument Generation to Support Feedback in Educational Argument Modeling Systems

Abstract

This paper describes an educational argument modeling system, GAIL (Genetics Argumentation Inquiry Learning). Using GAIL’s graphical interface, learners can select from possible argument content elements (hypotheses, data, etc.) displayed on the screen with which to construct argument diagrams. Unlike previous systems, GAIL uses domain-independent argumentation schemes to generate expert arguments as a knowledge source. By comparing the learner’s argument diagram to a generated argument, GAIL can provide problem-specific feedback on both the structure and meaning of the learner’s argument, e.g., that the learner’s argument contains an irrelevant premise. To generate arguments, the argumentation schemes are instantiated from causal domain models specified by lesson authors. Thus, this approach to generating expert arguments has the potential to be used in other domains. In this paper we describe use of GAIL’s Authoring Tool to create the domain model and content elements to be provided for a specific lesson, how expert arguments are generated in GAIL, and how the feedback is produced. As GAIL is a work-in-progress, the paper also describes plans for the next design iteration.