Fascinating. Is this likely to be because of a smaller corpus (seems unlikely - there must be millions if not billions of pages in most European languages)

Or is it that search algorithms don't cross languages easily - or that the core developers and testers use English as a lingua Franca anyway?

Still a fascinating point - can you put some kind of example or test up?