Matt Brown wrote some thoughtful comments, some of which I deeply disagree with:
MO, those would only be legitimate sources to cite if the subject itself is obscure and known only to specialists. If it's a well-known subject, it would make more sense to use mainstream sources on the subject.
Yes and no. One reason for citing sources is so that readers who want to know more (or check facts) can go to the sources. In this sense, I do agree with Matt that it is important to provide sources that anyone who is on-line and has access to Amazon.com, or access to a good library, can find.
But I disagree with Matt's distinction between "obscure subject" and other subjects. A subject that is not obscure, for example the Holocaust or the Bible or the U.S. Civil War, obviously has loads of popular and easily acceptable sources we can cite. But there is always ongoing academic research, and much important and relevant information will come from relatively obscure sources. This is precisely the material we want to include in a high-quality encyclopedia, even if the cited sources are hard to find.
If the obscure source is indeed important, it will at least have been cited by someone else. If, for example, you find an obscure source on the Holocaust that is not cited in any mainstream work on the Holocaust, it would be original research to begin to build an argument based on it.
Matt, this is just 100% wrong. You simply do not understand our NOR policy. It would violate our policy to "build an argument" on any source, "mainstream" as well as "obscure." But adding material, including published data, published explanations, published interpretations, is NOT "original research" if it comes from sources that are, however obscure, reputable.
(If you thought mainstream Holocaust historians were ignoring some obscure but credible and important source, that would be an issue to take up with them; we're just here to report the consensus in the field, not to create it.)
Again, 100% wrong. We are not here to report the consensus. As a matter of fact, one of the most important functions of our NPOV policy is to ensure that diverse views (and if they are diverse, they obviously do not represent a consensus) . We report on different views, provide the proper sources, and any context about the sources that can help readers evaluate the views being represented. "Consensus" hotonly has nothing to do with it, it is antithetical to what we stand for.
Steve
Steven L. Rubenstein Associate Professor Department of Sociology and Anthropology Bentley Annex Ohio University Athens, Ohio 45701