User:Tothwolf/rescued essays/NPOV and how to find it

This was written as a post to Wikien-l, but was too long to directly post there, so here it is. This could become an essay in Wikipedia space if that seems warranted, but it would need quite a bit of editing.

The key to understanding this is, first of all, that NPOV isn't a thing, a fixed state, a property of text in itself, it is a balance that represents consensus.

We can measure NPOV by the percentage of editors who agree with a text, and our goal should always be 100%. While we may, in controversial areas, never be able to reach 100%, we should always maintain some level of skepticism that text is truly neutral if there is even a single dissent from a responsible editor. We may have overlooked something, and, if that editor can find *any* support from other responsible editors, we should, as a community (which may require only one of us), examine the reasons for dissent and see if it is possible to address them and either convince the dissenting editors to support the text or, if convinced that the text can be improved to broaden consensus, implement that change.

I have written a draft essay that examines this, and a process that could be followed, at

Let's exclude for the moment what might be called "unreasonable editors." These are true POV-pushers, they will not accept balanced text no matter what, even if, on reflection, they would agree with every statement in it, if that text can be read as contradicting what they believe. I'll return to these editors, and how to deal with them and with the "eternal stream" of new editors who are attached to one side of a dispute, after establishing how NPOV is a matter of consensus.

Further, I need to assume that there is such a thing as truth, and it's an absolute, and that we, as human beings, are capable of recognizing it. I hasten to add that our recognition is imperfect and we can err; however, I have no doubt that I'm sitting at a keyboard, typing this as a message to this list. Do you? Sure, I can posit all kinds of theoretical possibilities, such as that I'm dreaming, I'm lying to you, this was written last year and I set up my computer to automatically send it as a response to a message containing certain keywords, but we don't need to go there. Routinely, we accept things as true, and share this acceptance, even though some abstract possibility of error remains. There are people, plenty of them, who don't believe that truth exists, but these are among those we will exclude from consideration at first.

We can all recognize a statement such as "Israel is an aggressor state" as POV, it involves complex judgment and we know that reasonable people can disagree on this. "Reasonable" doesn't mean neutral and without bias or prejudice, but, indeed, who is free of all these? However, suppose a newspaper, biased or otherwise, call it Middle East Review, reports that Abdullah al-Arab, [a made-up name for a Palestinian nationalist], said, "Israel is an aggressor state," and there is no controversy over the report. Abdullah al-Arab doesn't deny it, his blog states it, etc. Is the text:

"As reported in Middle East Review, Abdullah al-Arab claimed that "Israel is an aggressor state."

Our reasonable editors will agree that this statement is true, and it doesn't matter what side of the dispute they might be on, as to the rest of their opinions. Thus every verifiable text, with proper attribution and context, becomes true. Without that context and attribution it may well be false and is certainly POV.

So how do we find NPOV text in complex and contentious issues? It's actually known how to do it, outside of Wikipedia, but it takes patient and famously tedious process. If we consider NPOV as a percentage figure, a goal, rather than necessarily an absolute, we have 100% NPOV if there is no dissent to the text that is maintained after thorough examination, among "reasonable editors."

So, suppose we have found this magic text, 100% accepted. There is no problem maintaining it, by definition, excepting the situation of new editors.

Now I need to bring in the concept of "Majority POV-pushing," MPOV. There has been a lot of attention to the problem of minority or fringe POV-pushing, but, I claim, this is actually a lesser problem, because there is, practically by definition, no shortage of editors ready to oppose it by reverting contributions or editing them to make them more neutral, or even more to a side that is unfair to the minority opinion. The danger of MPOV pushing which doesn't mean fairly representing the "mainstream" position as mainstream, it means attempting to present the mainstream position as the only position, or the "correct" position, or otherwise to create text that even reasonable editors with minority POV cannot accept as true.

(I will also need to address the difference between "true" and "verifiable," but I'm setting that aside for the moment. Our standard is verifiability, but editors attach themselves to views based on personal understanding of truth.)

MPOV pushing is more seriously harmful than the other kind because it creates text that seems true to the majority, and can maintain it against modification, and, in our context, this means removing or unfairly framing reliably-sourced text and refusing to balance the article to reflect the weight of what is found in reliable sources in order to extend consensus beyond mere majority. Hence MPOV pushing can, if unrestrained, set up conditions for extended edit warring and continual discontent among a significant minority of editors. New editors arriving will attack the text, while dissatisfied experienced editors will cheer them on, and the battle continues indefinitely without resolution, wasting editorial effort on all sides without ever settling on true consensus, only "rough consensus," which pretty much means, in practice, some kind of majority decision.

Now, by some miracle or through the intervention of inspired editors or even by pure chance, a text is found, through detailed discussion, that is sufficiently neutral that all reasonable editors sign on to it, and agree to support it. They understand that if the text is altered from this, and the change is allowed, it will destabilize the article. For simplicity, I will assume two POVs, MPOV and mPOV, for majority and minority POV. I'm also going to assume that mPOV is fringe, say, on the order of 10% of editors, reflecting real-world distribution of opinion.

Suppose the change shifts the text toward mPOV. The MPOV editors can easily revert it and keep it out, and, since there are more of them, they are most likely to see the change first, so, let's say, one of them reverts it. The new editor, who now thinks, as we can expect, that the truth is being suppressed, reverts back. What happens? Here is where we would start to deviate from existing practice. The exact response would depend on many variables, but I'll give an extreme response that would remain legitimate: the new editor is blocked, but not just by any administrator, by one affiliated with the mPOV. Or if by a neutral administrator, an mPOV editor takes on the task of educating the newcomer. This experienced editor points the newcomer toward the discussion, which has been refactored into a consensus document showing all arguments presented and the process by which consensus was found, and invites the newcomer to find problems with it. Perhaps some relevant assertion wasn't included. This discussion can be far, far more detailed than the article, and can include, as assertions considered, any proposal made by any editor, as long as another editor seconded it and was willing to share responsibility for it. Perhaps the new editor can think of something not there. The editor proposes this for discussion. Can the editor find any mPOV editor, or any other editor, MPOV or neutral, to support this as being worthy of inclusion in what has become a FAQ on the existing consensus text?

If not, the new editor is faced with a reality: not even those who agree with him or her on mPOV are willing to accept the assertion as relevant and worthy of consideration. If so, however, the new argument is put up and considered and may end up making some change to the text, and what has happened is that consensus has been broadened, and 100% NPOV is maintained.

In the reverse direction, if a new editor shifts the article text toward MPOV, those editors maintaining the article will revert it, because they do not want to waste more time battling with mPOV editors. One of them, again, takes this editor aside and explains why the text is neutral and verifiable, and this is done sympathetically, as with the mPOV editor case. "I agree, I can understand why you think the text unfairly favors [mPOV], but we have discussed this thoroughly and every editor agreed that the text was fair and balanced, and if there was some mistake made in the process, you are welcome to enhance it, please see the process page...."

The process page is a consensus document, every statement in it is NPOV because of attribution, what's been eliminated there is our ordinary notability requirement, which has been relaxed to the extent of requiring only the assent of two editors to include an argument. It's a maintained page and has a Talk subpage associated with it. The process page itself, because of its nature, would be semiprotected (by consensus request) to reduce random vandalism and useless effort; a new editor would need to obtain the support of an autoconfirmed editor to post to the process page, and frivolous posting to process pages would be considered disruption. But the new editor could post freely to the talk subpage.

This requires far more detailed and tedious process than we currently see. However, it's not as hard as it might seem to those who don't have experience with real-world consensus process. Wherever there are organizations that value unity, rules are often created that require complete consensus, and, yes, bypasses exist to deal with unreasonable hold-outs, but they are loathe to invoke them because they damage unity, they will go to extraordinary lengths to accommodate minority opinion. When consensus becomes an absolute rule, however, the result is a kind of minority rule, whenever the status quo favors the minority.

What real-world experience with consensus process shows is that the attention paid to maximizing consensus minimizes disruption and division. In the end, it would be, here, more efficient than the roll-the-boulder-back-up-the-hill maintenance of MPOV text against wave after wave of mPOV pushers. That there are such waves is a strong sign for us that we have not yet found true NPOV. Yet, because, to an MPOV editor, MPOV text can look balanced and neutral, we easily identify these mPOV editors as disruptive and damaging and "against NPOV."

From this line of thinking, if we truly want to demand absolute NPOV, we should not include any text that is not unanimously supported; essentially, we'd have an extreme exclusionist policy. But this is not what I'd recommend at all. Absolute consensus should be a goal, not a policy.

If an editor can be identified as "unreasonable," i.e., unwilling to accept text that is clearly true and verifiable, typically because of the *implications* of the text, and the editor is willing to fight against such text by reverting it or unbalancing it, that editor should be excluded from the body of editors who must agree in order to show consensus, at least within that topic, and possibly site-wide. This is not necessarily the same as a topic ban, though it might be associated with one.

However, I do not recommend consensus as the standard for inclusion, though flagging text as disputed would be fine when it is only supported by a majority (even if that's a supermajority), and is opposed by at least *two* "reasonable" editors, or perhaps by a new editor and one reasonable editor.

Nor am I suggesting that we set up some bureaucracy to enforce standards; rather, I'm suggesting a cultural shift, a shift toward the understanding of the value of 100% consensus. It may not be achievable, but to the extent that we do not achieve it, to that extent we cannot be certain that our text is NPOV. We need to recognize that NPOV is not a quality of a text, but of the community's relationship to text, that finding it requires process (when there is controversy), and that, for efficiency, the process must be documented and made easy to access and build. In other words, we need to build, not only an encyclopedia, but a penumbra of pages, far more detailed, that explain why the text is the way it is, when controversy has been encountered, and that then becomes an avenue for new editors to challenge consensus non-disruptively. Thus our consensus becomes a living process, not a static and immovable object.

It is not necessary for all editors to pay attention to this process, and, indeed, it's probably enough that only a few do it. But articles which are maintained as battlegrounds need this. It does not require policy changes to build this structure, but simply a few editors to pioneer it and demonstrate it, and more editors to support it. I'm personally working, at this point, with Cold fusion, because, there, investigating behavioral issues, I needed to research the field and, reading both supportive and skeptical sources, extensively, personally found that ordinary MPOV on this was false, and easily demonstrated to be so, providing that the MPOV editor is willing to sit through the process of examining the basis for the MPOV. Indeed, it's actually preposterous, which is why the article has been so frustrating for editors arriving with knowledge of the field.

(Not that we should treat cold fusion as real, there is no clear and convincing reliable source on that, and MPOV among the general scientific community is probably still that cold fusion is rejected, dead, fringe; yet, there is, instead, evidence in reliable source that there is something going on, and we don't know what it is, and that cold fusion is one among a number of reasonable hypotheses, but the basic hypothesis most easily supported is that "There is an anomalous effect here, and we don't know what this is." This is a field where there is an abundance of reliable source for facts that are nowhere in the encyclopedia, and that material is being excluded by the majority of involved editors on the basis of "undue weight," which should always be a suspicious argument: if a fact is supported by reliable source, it's, by definition, notable enough for inclusion somewhere, yet, because of strong anti-fringe bias ("anti-fringe" is by definition MPOV), material is being excluded because what would ordinarily be sufficiently reliable sources are tagged as "fringe," and efforts to create other articles where sufficient detail could be presented without imbalance have been prevented as "POV forks.")

I will be working on the process I have roughly described here, with that article. If I'm not topic-banned! To do this, however, I don't need to be able to edit the article, because I am, indeed, seeking consensus, and if I can facilitate consensus, I don't need to be the one to implement it. And the process will be such that any editor can ignore it unless consensus is found among those participating, in which case the editor can then, still, reject article edits but might find that increasingly difficult to maintain.