On Feb 25, 2007, at 4:19 PM, Ray Saintonge wrote:
If the conclusion from being only unable to find secondary sources is that we can't have an article on a subject which many will agree to be otherwise valid that indicates that there is something seriously wrong with our criteria.
Exactly.
Another important thing to recognize about the way print encyclopedias use citations is that they not cite every statement. They give a summary. If they're using data or something they'll source the data - particularly if it's time sensitive. (i.e. "Top ten oil producing countries" or something, to use an example where I had the job of finding that data.)
But more often they summarize the situation and then put some "see also" type links at the bottom.
That's the gap that gets crossed when you move from verifiability to "cite sources." In verifiability, all I have to do is be able to point to places where you can go look and learn more, confirm this isn't a hoax, etc. Cite sources I have to tell you where you can find each and every piece of information.
Verifiability is acceptable for most of the information on Wikipedia. Claims that attract particular skepticism (and we can easily tell these by the skepticism they attract) should be sourced directly. But compare that to something like http://en.wikipedia.org/wiki/ Oriel_College (our featured article from yesterday), where the name of the college is sourced. Really. There's a citation two words in. This was not a useful citation to add. Whoever added it is very silly.
Compare also to [[Jacques Derrida]]. Here the problem is different. In the humanities, nobody writes a significant secondary source on Derrida that is not A) nearly as impenetrable as the original, and B) deeply involved in the same debates as the original in such a way as to be dreadful resources for providing a general overview. I don't mean this just in terms of NPOV issues either. I mean that the secondary sources are not meaningfully distinct from the primary sources in their quality or usefulness. There are sources intended for novices, but they're crappy and would lead to an article of questionable accuracy. And, really, there's something profoundly silly about an article sourced to books like "Derrida for Beginners" and "Derrida: A Very Short Introduction," or to introductory notes in a textbook. That's just not reputable in the humanities, and would make the article look laughable to any actual subject expert.
The way to write a good article on Derrida is to have a few people who have done some work using Derrida (there are thousands) to hash it out. The talk page should be used to smooth out debates. Good faith should be assumed - when one person says "Actually, you should really have another look at The Post Card where Derrida says X," the other person should. Classically thorny points and controversies should be sourced - especially the criticism section. But for the most part, it should be written by some people who know a decent amount about the subject going "OK, what needs to go into a general overview here."
And then at the end you should have a bibliography of books on the subject.
The problem is that somewhere along the line we went from a definition of "good article" that was based on looking at what other encyclopedias did to a definition of "good article" that was based on an internally decided principle.
Somewhere along the line, [[WP:NOR]] lost what was a vitally important line - that use of primary sources was acceptable so long as it did not produce "novel interpretations." That is to say that we never used to hold that the way to write an article was to consult entirely independent sources. And thank God, because that's no way to write an article, and not the way anybody else writes articles.
-Phil