The proper functioning of Intuition, Sinequa's platform for information retrieval, depends on the development and integration of a large number of linguistic resources. To improve their coherence and management, these resources containing very diverse knowledge have to be unified. As Intuition is available in most of the European languages, the multilingual factor makes this unification more complex. To overcome the problems caused by the structural differences among languages, a new linguistic architecture has been designed and modelled using UML. This meta-model is the starting point for the new database that will be the core component of a new workbench, centred on its user, the linguistic expert. This workbench will provide a centralised management of all the linguistic resources of Intuition.
