Access to legal knowledge is particularly challenging to information retrieval systems. Not only is legal knowledge usually expressed in linguistically complex forms, but it is also structurally sophisticated (e.g pieces of legislation applicable to a case, version in force of a legal document, other related sources). Modeling the collection of documents in such complex domains requires taking into account the semantic content of the documents as well as the intertextual structure of the collection since documents are usually related to each other by various types of links.
In this talk we will present two approaches for modeling and querying a collection of interlinked legal documents. We will briefly introduce a first approach based on Formal and Relational Concept Analysis. We will then present a second approach based on semantic web techniques (RDF, OWL and SPARQL) and show how it allows overcoming the limitations of the formal approach. Different types of relational queries are discussed.