Phil Sandifer wrote:
I actually think (and have argued both at conferences and on Wikipedia) that NPOV is more radical than that. NPOV is not just a goal, but an epistemology - it serves to replace the goal of metaphysical or ontological truth with a socially defined truth - presenting a mainstream overview. (In this regard it's much more allied with contemporary critical theory than I think most people realize) The nice thing about a socially defined truth is that it can be more readily checked, especially with a social editing system, since then the method of writing and the standard of evaluation are closely related. Hence our dependence on consensus and discussion.
An interesting question then is, to the extent that this is successfully captured in the articles we consider good ones, which is the "mainstream" that we're giving an overview of? In many cases we bias to specific types of sources---for example, supporting scientific sources over theological ones---in ways that a survey of the public at large might not. On the other hand, we don't tilt as strongly towards academic views as some encyclopedias do. I'd suspect that what we're summarizing is in effect the consensus view of the sorts of people who happen to edit Wikipedia in a particular area, which is itself a somewhat biased sample. This becomes particularly clear in certain sub-areas, and in looking at how articles differ across language versions.
-Mark