The goal of the project MULTIPLA is to develop and examine methods for learning multiple ontologies and integrate them. The purpose is that applications can exploit the different ontologies to switch between domains, foci and natural languages in a seamless way. The project will be devoted to developing new paradigms for ontology learning with a particular focus on combining different natural languages and domains for knowledge retrieval applications. This will require the development of new techniques for learning ontologies for different natural languages or domains, methods for aligning the different ontologies as well as knowledge retrieval approaches capable of using these aligned ontologies. The envisioned scenario is one in which users of a knowledge retrieval system can pose queries in some source language which can be either answered on the basis of information available in the same language, or answered on the basis of information available in some other language.