Following, & participating in, the various threads about Original Research, I've noticed that a couple of points have not been touched upon. And I think we should consider them while we are all trying to develop a consensus on this matter.
1. Original research leads to the problem that some editors think that they "own" an article. One of the exciting things about writing for Wikipedia is being able to say "I put that text there!" And while we are aware of the problems surrounding people who do not want their purple prose to be deleted, the problems can only increase if we allow people to put their own research in Wikipedia & say "Wikipedia has something I discovered."
The Wiki way (as I understand it) is that one person adds their input, and another may change or add to it; the result is that the content of Wikipedia inevitably changes over time, gradually (hopefully) moving between viewpoints in search of a balance. Because original research is such a personal thing, it can only disrupt this desirable changing.
2. Original research is inevitably POV; allowing original research thus makes it easier for POV-pushers. I hope I'm not the only one who has noticed that if a person performs research & discovers a fact, they believe it to be true; they will not be receptive to permitting other POVs intrude on their research. As a tangible example, if I were to set forth my original research stating that cats are better than dogs[*], I would obviously do so only if I am convinced that it was true, & that anyone who disagreed with me were either ignorant or in denial of the obvious facts.
And if we were to look at the motives of many POV-pushers, I believe that they are acting in good faith, but are under the same misconceptions that are linked to people who insist on inserting some forms of original research.
3. To write some articles with a desirable quality, it is almost impossible not to resort to what can be called, in good faith, original research. And this point makes the matter a hard one.
To write about any literary work beyond the bare facts of its creation & its characters & plot easily leads to original research. After all, the Wikipedian is asked to explain perhaps the most important things about the work -- what is it about, why is it important, what does it mean -- based all too often on a personal interpretation of the primary source. I wouldn't be surprised that this is the reason why many articles about literature, music, cinema & the plastic arts are in need of more material. While in many cases there exists a usable amount of secondary literature about a work -- a prime example being the plays & poetry of William Shakespeare -- for a lot of authors who should have articles in Wikipedia, there is little or nothing. Looking at my own bookshelf, I'd be hard-pressed to think of critical literature on Craig Lesley, Kazuo Ishiguro, James Branch Cabel, Toni Morrison, L. Sprague de Camp, Harry Harrison, Hunter S. Thompson or Chinua Achebe.[**] And even some topics that should not be hard to find abundant secondary literature for -- say a history of the influence Homer's _Iliad_ had on Western Literature -- would be difficult to complete without doing research & drawing conclusions (e.g., just how far was Elias Loennrot influenced by the _Iliad_ in his compilation of _The Kalevala_?).
Another instance is the creation of lists on Wikipedia; in some categories, the compilation & sorting would involve some kind of POV, whether conscious or accidental. The example I had in mind was the project of compiling a list of Mayors of Addis Ababa, the largest city in Ethiopia. I hope we can all agree that this is a list worth having in Wikipedia; but can we be assured that the people listed weren't included or excluded on an NPOV basis? Or that if there is a conflict of opinion over the order of 2 or more Mayors, that the right opinion was followed?[***]
If I may be allowed to answer my own question, I think there is a way to allow original research into Wikipedia without also allowing its problems in. One way would be to present the conclusions of orignial research overtly as one, non-exclusive POV. To do this requires the use of some despised weasel-words like "based on this evidence, it appears that" or "it may be true that" or "this suggests". After all, even if the evidence screams that a given conclusion is the only conceivable one, further research or peer review may show that some of the evidence can be explained away or be shown to support an entirely different conclusion.
Another approach -- & these are not exclusive -- is to admit that even though we are Neutral Point of View, sometimes because no article on Wikipedia is every truly complete we are only able to present one of many POVs. That is, the original editor adds her/his research with the full knowledge that another editor may drastically rewrite it to add other material. In a way, this is identical to current policy -- well, as I understand current policy; but this not only reinforces the idea that no one Wikipedian owns an article (avoiding issue #1 above), but also it may be useful for everyone to occasionally step back in a dispute & determine whether our arguments are based on a myopic & enthusiastic embrace of our own POV (avoiding issue #2 above).
This is all meant as more material to chew on. The [[No original research]] policy helps in a lot of different ways, but clearly it has a few problems that require some tweaking with it.
Geoff
[*] I selected this example not necessarily because I believe it is true, QED, but because I live with 3 cats who would shred me into hamburger were I to argue the opposite opinion. Let's not debate this topic.
[**] Just a selection of fiction writers from the shelf next to my desk, nothing more.
[***] I've encountered a similar problem in compiling a list of the Patriarchs of Antioch -- an article I've likely left unfinished, & hampered by countless cases of umarked POV content, but at the end I grew tired of the subject, & decided it was time to leave the article for someone else to work on.