We see all kinds of articles and youtube vids about claimed polyglots who speak X languages "fluently", and where by my estimation the definition of "fluent" is often a debased one. So the next time someone writes an article about or makes a youtube video interviewing another person in X number of languages, I'd like to see them add a reading component. I.E. for each language for which that the person claims spoken fluency, take a random mid-length article from the front section and features section of 2 types of newspapers: highbrow (NYT/FAZ/etc) and mid-brow (daily smaller-city newspaper even a tabloid variety). And then ask the polyglot to give a literal (non-free) translation of those articles into his/her L1. For every unknown word just insert the L2 into the L1 translation and underline it, rather than try to ignore it. How would these polyglots do with something like this? If this could be done as a competition at some polyglot gathering, I'd like to place my wager on the Iversen horse.