My latest Guardian column, "Why Philip Roth needs a secondary source," explains why it makes sense for Wikipedians to insist that Roth's claims about his novels be vetted by and published in the New Yorker before they can be included on Wikipedia:
Wikipedians not only have no way of deciding whether Philip Roth is an authority on Philip Roth, but even if they decided that he was, they have no way of knowing that the person claiming to be Philip Roth really is Philip Roth. And even if Wikipedians today decide that they believe that the PhilipRoth account belongs to the real Philip Roth, how will the Wikipdians 10 years from now know whether the editor who called himself PhilipRoth really was Philip Roth?
Wikipedia succeeds by "not doing the things that nobody ever thought of not doing". Specifically, Wikipedia does not verify the identity or credentials of any of its editors. This would be a transcendentally difficult task for a project that is open to any participant, because verifying the identity claims of random strangers sitting at distant keyboards is time-consuming and expensive. If each user has to be vetted and validated, it's not practical to admit anyone who wants to add a few words to a Wikipedia entry.