Kevin Ashley on Reasoning with Hypotheticals

In a recent series of talks and papers, Professor Kevin Ashley and his colleagues have reported on new findings in their NSF-supported research on hypothetical reasoning. Their tutoring system LARGO (Legal Argument Graph Observer) teaches students about legal reasoning with hypotheticals using Supreme Court oral arguments as examples. In experiments at the Law School, they have demonstrated that features of argument diagrams law students constructed with LARGO are correlated with, and to some extent predict, students’ LSAT scores, years in law school, and performance on a posttest dealing with hypothetical reasoning skills. At the same time, evidence shows that students’ use of LARGO’s feedback is correlated with learning and that practice with LARGO is helpful to first year volunteer students with lower LSAT scores. The researchers, however, were not able to reproduce the latter result in a subsequent study with non-volunteer first year students.

Ashley reported the results in three invited talks in the last three months:

(1) “Some Thoughts on Using Computers to Teach Argumentation”, at the Special Track on Intelligent Tutoring Systems of The 21st International Conference of the Florida Artificial Intelligence Research Society (FLAIRS), Coconut Grove, FL, May 16, 2008,

(2) “Twists and Turns in Teaching Argumentation with Computers”, at the Twenty-Third AAAI Conference on Artificial Intelligence, Workshop on What Went Wrong and Why: Lessons from AI Research and Applications. Chicago. July 13, 2008, and

(3) “LARGO, an ITS for Teaching Argumentation with Hypotheticals”, Stage Setting Talk at Sub Tech 2008, William and Mary Law School, Williamsburg, VA July 24, 2008.

Three recent papers presenting the work are:

(1) “Reevaluating LARGO in the Classroom: Are Diagrams Better than Text for Teaching Argumentation Skills?” Pinkwart, N., Lynch, C., Ashley, K., and Aleven, V. (2008) in Proceedings of the 9th International Conference on Intelligent Tutoring Systems, Montreal, June,

(2) “Graph Grammars: an ITS Technology for Diagram Representations”, Pinkwart, N., Ashley, K., Aleven, V. and Lynch, C. (2008) in Proceedings of The 21st International FLAIRS Conference, Special Track on Intelligent Tutoring Systems, Coral Gables, Florida, USA, May, and

(3) “Argument Graph Classification with Genetic Programming and C4.5”, Lynch, C., Ashley, K., Pinkwart, N., and Aleven V. (2008) in Proceedings of the 1st International Conference on Educational Data Mining, Montreal, June.