The European Union is linguistically diverse, with 23 official languages and 56% of its citizens who are able to converse in a language other than their mother tongue. However, European libraries do not share a subject heading or classification system or a bibliographic format. Thus, cross-language access to library collections is a complex problem involving not only parsing and translation, but also mapping subject headings, classifications and bibliographic formats, i.e. syntactic and semantic interoperability. The CACAO project (Cross-language Access to Catalogues and Online Libraries) will implement the following vision. A user enters a monolingual query, e.g. English cat, and retrieves relevant records containing, e.g., the German word for cat, Katze, French chat, or Italian gatto. CACAO will thereby refine and extend an implementation at the Library of the Free University of Bozen-Bolzano, Italy.