Explaining Discretionary Judicial Decisions Using Answer Set Programming

Abstract

We introduce a method for analyzing discretionary judicial decisions by testing the internal consistency of judges’ reasoning and uncovering their underlying assumptions. The approach represents both the facts of a case and the judge’s stated evaluation in Answer Set Programming (ASP) and uses a Python analysis to examine the resulting answer sets. Applied to a child custody case, the method reveals the weights implicitly assigned to different factors, highlights when the reasoning cannot support the outcome, and identifies alternative weightings that would make the decision consistent. The contribution is a formal tool for making judicial reasoning more transparent, analyzable, and open to critique.