Over the years, numerous scholars have attempted to gauge the degree to which national discourses about the European Union are connected. However, given problems in comparing discussions about a topic across different languages and media contexts, much of this work has focused on identifying whether the same issues are discussed at the same time, rather than the more important (but also much more difficult) question of whether they are discussed in the same way. In the past few years, new developments in machine learning and machine translation have, for the first time, made it possible to systematically compare the meanings of different texts across languages. I apply these latest innovations to an assessment of the European public sphere.

Using a corpus of more than a million newspaper articles about the EU published in leading newspapers across the five largest European Union member states — Germany, France, the United Kingdom, Italy, and Spain — I systematically assess both the key topics discussed and the framing of those topics when the media discuss the European Union. The results offer a better measure than has been available until now of the degree to which a European public sphere can be said to exist. In particular, they make clear that national discourses have become increasingly aligned over time, to the point that it now makes sense to talk about a European public sphere.