EuroWordNet builds a multilingual database with wordnets for several European languages.Each language specific wordnet is structured along the same lines as WordNet (Miller et al. 1990): i.e. synonyms are grouped in synsets, which in their turn are related by means of basic semantic relations, such as hyponymy, meronymy, semantic roles.Each wordnet uniquely describes the lexicalization pattern of a language.Multilinguality is achieved by storing the languagespecific wordnets in a central lexical database in which equivalent word meanings across the languages are linked to a so-called Inter-Lingual-Index (ILI).In this paper, we will address the way multilingual relations are expressed in EuroWordNet and how they are (semi-)automatically extracted from the monolingual wordnets and bilingual dictionaries.Different kinds of equivalence mappings are described to deal with fuzzy mappings and gaps.The fact that these equivalence relations are established at a more global synset level and that it is possible to browse to closely related synsets, makes it possible to get a comprehensive conceptual match of concepts across languages, even when lexicalizations differ.Other evidence, such as co-occurrence probabilities and morpho-syntactic constraints, can then be used to find a translation or a correct combination of words in a target language that is appropriate in a context.