Abduction and legal reasoning

In this paper we present LAILA+, an extension of the LAILA coordination language for abductive logic agents, i.e. reasoning agents that collaborate towards the solution of a given problem exploiting a set of distributed, possibly partial, knowledge of the application domain. The extension consists of i) the possibility for agents to communicate with each other hypotheses while devising a coordinated solution, and ii) a relaxed consistency mechanism based on a given agent hierarchy: stronger agent coherence may overcome weaker agent inconsistency. We argue that the framework well adapts to legal reasoning, with agents that try to prove/disprove evidences from different, possibly partial and hierarchical, viewpoints, as often happens for instance in a trial.