Section Break: Reply to Mperel's and Sarvagnya's important questions
editSarvagnya asks "which policy or guideline?" says that "The fact that this is used [by reliable sources] is a direct evidence of its reliability." Mperel asks "is a source made reliable because other reliable sources regularly cite it"?
As Relata refero points out, this idea has a long history here at Wikipedia. The phrase "Find out what other people say about your sources" and the suggestion "to cross-check with an independent source" appear in Beland's original version of the page and Radiant!'s first guideline version
But the idea that "cit[ation] by other reliable sources ... is indeed the best possible sign (and our original standard!) that it is reliable" is still there, hiding in plain sight, implicit in our basic guideline, our lex generalis:
Articles should rely on reliable, third-party published sources with a reputation for fact-checking and accuracy.
The rest of the WP:RS guideline and parts of WP:V is lex specialis explicating this sentence, but as WP:RS wisely states, not exhaustively, as some editors seem to argue. The dictionary meaning of "reliable" - able to be relied on (by whom?) is clearly relevant. (And the source at hand, BOI seems prima facie reliable and unexceptional - Would we be having this discussion if it contradicted known reliable sources?) That accepted source A in fact does rely on source B is evidence of nonzero weight that B is reliable in the ordinary, and I and many others hold, according to Wikipedia rules. Look at "with a reputation for fact-checking and accuracy". How do we evaluate reputation? Our practice is not to only use sources for which we could find another source explicitly saying this is a great source with great fact-checking and accuracy - and how do we evaluate "the another source" then?
No, the action of citing and using a source speaks louder (and more frequently) and directly about "reputation" than the rarer, but implied, words "this source is good (accurate, fact-checking), we've checked it out." And citations are easy to mechanically google and find, objective and practical.
As Sarvagnya says "a source doesn't become reliable by 'proclamation' or from editors' assurances or 'judgement' that they are reliable." But he is wrong when the he states that "this exercise" of determining a source's reputation "starts by telling us precisely who runs a site, a newspaper, a journal etc." That is simply not what the key word "reputation" means. Reputation doesn't mean "biographical" "credentials" information about a person or institution, which is a useful but indirect means of establishing reputation. Reputation means what other people or institutions think about him, her or it. Were Michael Scheuer's books unreliable sources when he was still anonymous?
Indeed, reputation of a journal, person or text is often based on mecahnical computation of citation indices based on the raw fact of citation alone, and not only here at Wikipedia, where it is often used as a criterion of notability. We should not be reinventing the wheel when we think about source reliability. Looking at "who cites it" is second nature and best practice in source evaluation in many fields, in academia, journalism, law (where there's even a word for it - Shepardizing), to many editors and in the real world in general.
This idea is entirely in line with how other core policies and practices work at wikipedia like NOR - we should endeavor as far as is possible to not substitute our own judgment for that of published sources. Another example was recently at AfD - two of the most acute and respected editors (who argue here pro -BOI) argued to delete, or only weakly to keep Sara Roy - she's only a Research Associate at Harvard (biographical, credentials, objective info) - why have an article on her? But it turned out that she is citably well known as the world's #1 expert in her field. "Reputation" data does and should trump "objective" data.
However, it might be a good idea to make this more explicit and less reliant on talmudic/jesuitical disputation.
So my proposal is a line in WP:RS like
"That an accepted RS cites a source is evidence for its reliability for use in the same manner as the accepted RS." The "in the same manner" part is there to satisfy Nishkid64's and Sarvagnya's "RS-hood by association" concerns.
An maybe we should also have a sentence somewhere to the effect that we are not trying to reinvent the wheel, but just to adapt the best and standard academic, journalistic, legal, real world practices in source evaluation, which certainly includes this idea, to the wonderful world of Wikipedia.John Z (talk) 09:12, 14 May 2008 (UTC)