This paper discusses the NLPmodule of Condorcet, an information retrieval project of the University of Twente, the Netherlands. The Condorcet system indexes scientific documents by mapping title and abstract of the documents to concepts and relations, defined in modern versions of classical indexing thesauri, i.e. ontologies. After a brief sketch of Condorcet’s approach to document indexing and the design and development criteria used in the research, the linguistic principles are discussed that form the conceptual basis of the NLP presented in this paper. As engineering issues are of equal interest in this application-oriented research, formalization and implementation of the principles are discussed as well.