Lex minus quam voluit dixit

Logical models of legal reasoning often assume that norms give a complete description of their applicability conditions, an assumption which is too strong, due to the complexities and dynamics of the world. Norms cannot take into account all the possible conditions where they should or should not be applied, first of all because the legislator cannot consider all the possible contexts which are exceptional and he cannot foresee unexpected changes of the world [10]. Normative systems regulating real societies have two mechanisms to cope with this problem. First they distinguish regulative rules from constitutive rules. While the former, which are changed only by the legislative system, specify the ideal behavior, the latter ones provide an ontology of institutional concepts to which the conditions of regulative rules refer to. Second, the judicial system is empowered to change the constitutive norms, under some restrictions not to go beyond the purpose from which the regulative rules stem. This combination of rules and norm change allows the legislator to disentangle the specification of a behavior from the specific contexts of applicability. In this paper, we outline a logical framework which is able to capture the norm change power and at the same time the limitations of the judicial system in revising the set of constitutive rules defining the concepts on which the applicability of such a rule is based. Indeed, the distinction between regulative and constitutive rules (ontology vs norms) suggests that legal interpretation does not amount to revising norms, but to interpreting legal concepts, i.e., to revising constitutive rules [13]. While the distinction between constitutive and regulative rules has been already introduced in related computer science fields such as MAS, the interpretation process has been only addressed in the field of AI and law but only as far as it concerns case based reasoning for common law [2]. Also the relation between constitutive norms and contexts has been considered [9]. What is still lacking is a logical model of the interpretation mechanism which leads to dynamically revising constitutive norms to make the normative system flexible in multiagent systems. This issue breaks down into the following subquestions: How to model the meaning of an institutional concept? How to decide which constitutive norms to introduce to either shrink or extend the extension of the institutional concept? How to reason about the interaction between norms and goals? The methodology we use is to adopt an extension of Defeasible Logic [5], which allows us to model constitutive and regulative norms, the goals associated to norms and the norm change process.