Rivière or fleuve? Modelling multilinguality in the hydrographical domain

The need for interoperability among geospatial resources in different natural languages evidences the difficulties to cope with domain representations highly dependent of the culture in which they have been conceived. In this paper we characterize the problem of representing cultural discrepancies in ontologies. We argue that such differences can be accounted for at the ontology terminological layer by means of external elaborated models of linguistic information associated to ontologies. With the aim of showing how external models can cater for cultural discrepancies, we compare two versions of an ontology of the hydrographical domain: hydrOntology. The first version makes use of the labeling system supported by RDF(S) and OWL to include multilingual linguistic information in the ontology. The second version relies on the Linguistic Information Repository model (LIR) to associate structured multilingual information to ontology concepts. In this paper we propose an extension to the LIR to better capture linguistic and cultural specificities within and across languages.