What nonsense. Wikimedia can and should do original research. It's only in the main namespace of Wikipedia that we have that restriction, and that's because we see an encyclopedia as a secondary source.
It's only since the addition of ISO 639-3 that it's even feasible to use ISO 639 as a canonical list of languages, and that's only because SIL was recognised as the most competent body to do such a thing. The Library of Congress was doing a poor job of it, and I would absolutely stand by our decision to add language editions to Wikipedia that they didn't recognise.
How would you suggest deciding what is and isn't a language?