The topic of this article is multilingualism and language choice in EU institutions. An analysis of whether the choice of working monolingualism penetrates progressively into the entire community and affects other languages and language use and if so, how this occurs. It comprises different sub-topics and perspectives such as the attitudes of EU officials, external communication of EU institutions with civil society through website presentations, with journalists through press conferences and with the German Parliament by the submission of papers. Also, internal institutional aspects were considered by analysing language choice for contributions in plenary and committee meetings in the EU Parliament and the Commission as well as the commissioners’ language skills. Special attention was paid to the unofficial institutional procedural languages, namely English, French and German, which have emerged in this function. English has gained outstanding prominence in all of the domains analyzed, though this does not seem to be welcome according to the officials’ professed attitudes and EU language policy principles. This research shows, however, that the prominence of English is the political reality, despite the seemingly adverse political will and in spite of the reduction in the number of procedural languages, the extension of interpretation services and the promotion of officials’ foreign language skills. Parliamentarians fear the negative impacts of such a development on the EU’s democracy and legitimacy. In addition, growing monolinguality causes language conflicts with, and within, national parliaments as shown by this research in the case of the German Bundestag. Neither the members of the German Parliament, however, or the EU commissioners seem to consider this a burning issue of EU politics. Finally, this research reveals the potentially positive effects of working plurilingualism in EU institutions as well as the stark discrepancy between the EU’s language policy and its actual language politics. The article is based largely on research planning and guidance by Ulrich Ammon and also research carried out by Jan Kruse and, formerly, Verena Wimmers and Michael Schloßmacher.