Balancing for Agent Decision Making through Argumentation

This paper builds bridges between reason-based normative reasoning and formal argumentation, by formalizing Tucker’s ethical theory and evaluating the strength of arguments via weighted reasons and tournament semantics. We characterize conditions under which various formulations of permissibility and detachment coincide, and we analyze independence properties that make these equivalences hold. We show that Tucker’s input–output “balancing” mechanism, developed for ethics, provides a domain-neutral model of practical reasoning that explains when multiple options are permissible and how to choose among them. As an outlook, we sketchhowthemechanismappliestodiscretionaryjudicialdecision making (e. g. , child custody), without claiming that Tucker’s ethical theory itself applies to law; instead, the argumentation-based abstraction transfers across normative domains.

Authors

• Liuwen Yu Luxembour Institute of Science and Technology
• Chenyang Cai Zhejiang University
• Réka Markovich University of Luxemburg
• Leendert van der Torre University of Luxemburg

Keywords

• Artificial Intelligence
• Knowledge representation and reasoning
• Balancing
• Normative reasoning
• Discretionary decision making