A Computational Framework for Formalizing Rules and Managing Changes in Normative Systems

Legal texts are typically written in a natural language. However, a legal text that is written in a formal language has the advantage of being subject to automation, at least partially. Such a translation is not easy, and the matter is even more complex because the law changes with time, so if we formalized a legal text that was originally written in natural language, there is a need to keep track of the change. This thesis proposes original developments on these subjects. In order to formalize a legal document, we provide a pipeline for the translation of a legal text from natural to formal language and we apply it to the case of natural resources contracts. In general, adjectives play an important role in a text and they allow to characterize it: for this reason we developed a logical system aimed at reasoning with gradable adjectives. Regarding norm change, we provide an ontology to represent change in a normative system, some basic mechanisms by which an agent may acquire new norms, and a study on the problem of revising a defeasible theory by only changing its facts. Another contribution of this thesis is a general framework for revision that includes the previous points as specific cases.