Toward modeling and teaching legal case-based adaptation with expert examples

Abstract

Studying examples of expert case-based adaptation could advance computational modeling but only if the examples can be succinctly represented and reliably interpreted. Supreme Court justices pose hypothetical cases, often adapting precedents, to evaluate if a proposed rule for deciding a problem needs to be adapted. This paper describes a diagrammatic representation of adaptive reasoning with hypothetical cases based on a process model. Since the diagrams are interpretations of argument texts, there is no one "correct" diagram, and reliability could be a challenge. An experiment assessed the reliability of expert grading of diagrams prepared by students reconstructing examples of hypothetical reasoning. Preliminary results indicate significant areas of agreement, including with respect to the ways tests are modified in response to hypotheticals, but slight agreement as to the role and import of hypotheticals. These results suggest that the diagrammatic representation will support studying and modeling the examples of case-based adaptation, but that the diagramming support needs to make certain features more explicit. © 2009 Springer Berlin Heidelberg.