A Complex-System Approach: Legal Knowledge, Ontology, Information and Networks

Abstract

In this paper we first briefly summarize the process used for building ontology from a legal corpus given in natural language. Current ontology-building supposes a particular structure and a finite number of relation types. The corresponding architecture is mainly driven by tree-like structures that capture a part of the full complexity that is effectively at work in any legal system. We propose to endow a legal ontology with further functionalities related to its mapping in a given corpus. We define posterior probability functions related to the frequency of occurrence of any term or concept, and information functions that measure the mutual information shared by terms in the corpus, whatever might be the a priori links represented between them in the ontology. We then show how these probabilistic tools can be also associated with a scale-dependent view on the network structure of a legal corpus (from the larger scale of the network of all codes or laws of a legal system, to the much finer scale of articles). New perspectives mixing semantic web and some properties of complex systems are described.