Wikipedia talk:Requests for comment/When there is no consensus either way

Status quo, stability

I believe you need to define stability, what is a stable version (WP:STABLEVERSION) and address WP:PRESERVE. I don't think exclusion should be leaned on when there's sufficient stability to counterweigh it, which is relative, and context-dependent. Andre🚐 23:47, 27 December 2023 (UTC)

The story is about a new article, so there is no stable version. WhatamIdoing (talk) 01:22, 28 December 2023 (UTC)

Ah, gotcha. But I was reading new as relative. You could write a new article that hasn't been reviewed for 3 months or even for several years, and if the content was never reviewed or discussed, I guess, it is neither stable nor has consensus, but not new either. I guess I'm raising a not super relevant edge case here, so let me read it more closely. Andre🚐 01:24, 28 December 2023 (UTC)

The story's timeline would have to include both time for an RFC and discussions before the RFC, so it's probably at least five weeks old, and maybe two months, by the point at which "you" are being asked to decide whether no consensus means inclusion or exclusion of the sentence/fact.

As for "reviewing", I doubt that the Wikipedia:Page Curation ("reviewing") status is relevant (it doesn't qualify for speedy deletion, which is all they IMO should care about, and it's both an obviously notable subject and properly sourced, which is what they do care about). If by reviewing you mean "editors looking it over", then all the editors who agree that the entire article is accurate, verifiable, neutrally phrased, etc., must have done that when they formed their opinions on those qualities. WhatamIdoing (talk) 01:32, 28 December 2023 (UTC)

I will make these points explicit, as it's likely to come up. WhatamIdoing (talk) 01:33, 28 December 2023 (UTC)

Reasons

@Blueboar, thanks for your comments. It's too early to really reply (and I might re-write the whole thing before launching it), but I wanted to tell you that I've hesitated to give reasons for inclusion/exclusion, because I'm not sure that we all have the same ideas about those reasons. Consider "It's an interesting detail": Some editors think that interesting details are good (e.g., "will hold readers' attention and encourage them to read more of the article" or "gives you the real flavor of what this is like"), and others think it's bad (e.g., "unencyclopedic" or "just your personal opinion; I don't think it's interesting" or "trivia").

I have similarly resisted the impulse to add a specific subject area, because those can be biased. "Editors are divided on whether to mention the popularity of this plant's use in a traditional herbal remedy" will get a predictable exclusion vote from the Wikipedia:Fringe theories/Noticeboard regulars. "Editors are divided on whether to mention this subject's connection to the Game of Thrones television show" will get a predictable inclusion vote from fans of the show, and a predictable exclusion vote from people who dislike ==In popular culture== sections.

I can tell you that whatever the hypothetical disputed point is, it is not something that can be solved through copyediting. The editors who want it excluded prefer an article that does not mention it at all, but if it's going to be mentioned, then they have no objection to the way that it's written. They already agree that it is cited and neutrally worded. They just don't want it in this article. WhatamIdoing (talk) 02:23, 28 December 2023 (UTC)

But I think this gets into a productive neighborhood. "Reasons for inclusion" have weight. Not just source weight, but relevance weight, encyclopedic is basically relevance. Trivia is defined as something that doesn't have a lot of relevance weight. For a biography, we have some idea what that means. It depends on the topic, though, and each topic has slightly different ideas of what is trivial. There's quite a lot of trivial stuff written about sports or games, but biographies not so much. Andre🚐 02:52, 28 December 2023 (UTC)

Yes, of course, but what's important in this story is: An admin already looked at all the "reasons for inclusion" (as well as the "reasons for exclusion") and already determined that the reasons for and against are exactly balanced. You don't have to re-evaluate the reasons yourself; the hypothetical admin already did that work for you. WhatamIdoing (talk) 05:19, 28 December 2023 (UTC)

I think maybe the story is more powerful if it's 3 stories. Story A, is a story of 0/100. Story B is a 50-50, and finally, a case where the extreme opposite is true. A 50-50 is a rare thing, like a spherical cow. Andre🚐 05:23, 28 December 2023 (UTC)

I am disappointed to discover that there is no article about the Spherical frictionless chicken, a species whose primary habitat seems to be physics classrooms.

I don't think that there is any dispute about what to do when a consensus is formed (the 0–100 and 100–0 cases), so I don't see any point in discussing them. The problem is only when there is no consensus (which, I grant, is more likely to be a 55–45 thing than exactly 50.0000–50.0000), and editors each claim that The True™ Policy says it should go my way, and only the Fake Policy says it should go your way. WhatamIdoing (talk) 05:31, 28 December 2023 (UTC)

Spherical cow is the article Andre🚐 05:32, 28 December 2023 (UTC)

And yeah, you're right, 100/0 and 0/100 are boring, but it's the base case and the inverse base case so important to stake it out cleanly, at least, in your head if not in the actual thing you're creating here. But let's go back to the 55-45 or the 45-55 case. or maybe the interesting cases are 70/30, 30/70, not to make it all horribly mathy and abstract. In my view, the distinction, about consensus, does come down to kind of a weighing of balancing of different heuristics, and we could flesh out the bounds of that a bit more. It's hard to say what someone would do, hypothetically, in an exactly balanced case. An exactly balanced case is more like a superposition of 55-45/45-55 that hasn't yet resolved, and is wavering back and forth wildly between two poles surrounding a nucleus of possible interpretations. Andre🚐 05:41, 28 December 2023 (UTC)

The problem with positing even a 49–51 case is that I expect someone to say that the 51% "wins", merely because it was ever so slightly preferred. It is a way of avoiding making a decision about what to do. I don't want editors to weasel out of this decision by pretending a consensus into existence; I want them to tell me what should happen by default, when there is a realio trulio no consensus. I don't care what they say should be done in such cases (include, exclude, have the closer flip a coin, appoint an editor from each 'side' to play Rock paper scissors via e-mail to ArbCom, with the winner deciding what to do...), but I want them to answer that question. WhatamIdoing (talk) 06:02, 28 December 2023 (UTC)

Is a situation where both groups of editors (the “omitters” and the “retainers”) have equally expressed reasons to support their stance realistic? From my experience, at least one “side” in these disputes does not give reasons… they simply reject other side’s reasons without explaining their own. Blueboar (talk) 13:04, 28 December 2023 (UTC)

And when it's anywhere close to 50–50, both sides declare that their reasons are obviously much stronger than the other side's – at least according to their not-so-neutral personal opinions.

We don't have to worry about this, because in this story, we are explicitly stipulating that there is no consensus and that the arguments made by both sides have already been judged by a neutral admin to be exactly balanced. All we apparently have to worry about is editors trying to answer a different question ("What should editors do when one side is stronger according to whatever metric I care about most, but she keeps insisting that there is no consensus?") instead of the one being asked ("When there is really, truly, absolutely, no-way-around-it no consensus to either include or exclude a given fact, what should editors do by default?").

The resistance to this question continues to surprise me. Consider how often we handle this without a second thought:

• Q: "The admin said there was no consensus at AFD. What happens now?"
  • A: "We keep the article, of course. Even new editors know that."
• Q: "The inclusion of this URL in ==External links== is disputed, and we can't reach consensus. What happens now?"
  • A: "Just read the guideline. It stays out per ELBURDEN."
• Q: "The WP:RM closed as no consensus. What happens now?"
  • A: "Easy: Keep it at the last stable version."
• Q: "Our discussion about the appropriate ENGVAR resulted in no consensus. What happens now?"
  • A: "WP:RETAIN the variety found in the first post-stub, just like the MOS page has said for years and years."

But when I ask:

• Q: "The RFC closed as no consensus about whether to mention this fact. The strength of the arguments advanced by both sides are exactly equally strong. What happens now?"

– well, the replies I receive regularly and consistently (we've been talking about this problem for three years now?) indicate a mental refusal to believe that any RFC could actually end up as 'no consensus' about whether to include a particular fact. Even if I write "The strength of the arguments advanced by both sides are exactly equally strong", I can count on someone to say that editors should go with the side whose argument is stronger, not with a head count (because Wikipedia:Nobody reads the directions, even in RFCs), and when I point out that the question already stipulated both that the arguments are equally strong and that the vote count is exactly equal, they tell me that the story is false, and that any admin who closed an RFC by saying the arguments were equally strong on both sides would be wrong.

Why do you think this question is so difficult for editors to think about? Are we afraid of the idea that there might not be a single right answer? WhatamIdoing (talk) 22:39, 28 December 2023 (UTC)

Perhaps the real “right” answer is: “There is not yet any consensus in either direction… so, keep discussing, keep calling in more people (especially those who might have new things to say on the issue), and (hopefully) a consensus will eventually emerge.” Blueboar (talk) 23:16, 28 December 2023 (UTC)

And in between now and that magical future, what should editors do? WhatamIdoing (talk) 23:23, 28 December 2023 (UTC)

Continue to discuss… sometimes a consensus can take YEARS to form. Blueboar (talk) 23:40, 28 December 2023 (UTC)

The (hypothetical) article exists today. Editors have failed to agree today. What should be done to the article today?

If a consensus forms in the future years, then editors in those future years can implement their agreement. What are editors supposed to do before the agreement exists? Set a bot to re-add and re-remove the disputed fact every 24 hours, so that each side of the discussion gets their due proportion of time? WhatamIdoing (talk) 00:41, 29 December 2023 (UTC)

Why the rush? I am not sure that anything has to happen to the article today. Just leave it as “undecided… keep discussing.” Blueboar (talk) 21:26, 29 December 2023 (UTC)

They've already had normal discussions and an RFC. How many more discussions will you expect them to undertake before someone reverts again because "there's no consensus for your preferred version"? WhatamIdoing (talk) 22:12, 29 December 2023 (UTC)

I think Blueboar's point, and I think ActivelyDisinterested or someone made this point, and I agree somewhat, has to do with the radical eventualist model of wikipedia and collaboration. Let's use an open source project for example, like the Linux kernel, or just pick whatever one you might be familiar with that's a bit smaller.

They have long time horizons, on the order of years. Some patches are reviewed and merged on the order of days or weeks or months, but not always. Some issues, including extremely high priority bugs that impact user experiences and security in a harmful or showstoppy way, may persist for years and exist to this day. Everyone's a volunteer. Sometimes companies pay for work. (WikiEdu, allowed disclosed COIs, similar I guess?) But for smaller projects with no financial support per se, since everyone's a volunteer, that means that instant gratification isn't a thing. Nobody may demand satisfaction, extreme patience is needed, which also means that some problems are WONTFIX, CANTFIX, or just didn't get around to it. (SOFIXIT is what we call some of that. An explicit reference to the software engineering practice of READMEs, TODOs and FIXMEs in code. I've essayed about this on my user page.)

Because Wikipedia embraces fuzzy boundaries and blurry lines and possible contradictions existing in superposition as opposed to clear rules, line-following, and simple numerical metrics to determine things, because they aren't determinative, and this thought experiment is a good illustration of that. Some RFCs, and I'll send you a few to review if you want privately because I don't want to reopen any old contentious issues that are at rest -- Because WP:CCC and there's WP:NODEADLINE, but also WP:NOLIMIT (why doesn't that exist?), I've seen RFCs that have been reopened repeatedly, and the community norm of "that's disruptive" doesn't always re-close the RFC. And then eventually it overturns. And, you might expect, it will be still kind of balanced on a pin and can overturn back. So you would expect if there's a function or a model of the consensus that's really close to 50-50, a timeline like this is possible

1. RFC, no consensus
2. edit war
3. edit war
4. RFC, option A
5. peace
6. RFC, option B
7. edit war
8. RFC, no consensus
9. eventually, option B just stays in the article

Andre🚐 23:57, 29 December 2023 (UTC)

I'm definitely an eventualist, but eventualism doesn't tell us what state the article should be in between today (when there is no consensus) and that magical future (when a consensus is eventually formed). However, unless we delete the article entirely, the article will be in one state or the other between now and its eventual state.

Imagine that the order of events is:

1. RFC closed as no consensus on Monday, resulting in an
2. edit war all day Tuesday because "there's no consensus for your version", resulting in
3. a major ANI thread, during which
4. uninvolved editors at ANI volun-tell you on Wednesday to decide (and inform the warring parties) whether the fact will be included or excluded unless and until there is a consensus to do it the other way.

We know which answer you eventually arrived at: After grumbling about whether there might have been a consensus that the closing admin just didn't notice and how unrealistic the story is, you eventually said that if one really had to decide the thing purely on principle, you'd usually err on the side of including the disputed (but verifiable, properly cited, non-BLP, neutrally worded) content.

I need that bottom-line answer from a lot more editors than you. WhatamIdoing (talk) 21:25, 30 December 2023 (UTC)

I, on the other hand, would lean towards omitting the material if forced to end discussion “today”. Blueboar (talk) 22:49, 30 December 2023 (UTC)

Thanks; remember that for next week/month/whenever I think this is ready for the community to decide.

I really don't know what I want. Sometimes I think that it would be best for Wikipedia to include no-consensus content (except for BLPs). At the moment, my desire to restrict and control (which means excluding) is strong, but it doesn't feel good (in the sense of noble). WhatamIdoing (talk) 00:01, 31 December 2023 (UTC)

@Andrevan and @Blueboar, the RFC is officially open, and I'd be happy to have your thoughts on the official page, whenever/if you feel like posting them. WhatamIdoing (talk) 19:35, 10 January 2024 (UTC)

an answer which poses another set of chained subproblems?

an answer to the problem that poses a dependent set of chained subproblems is: the admin has to weigh a different test or set of factors to determine, in the case of a no consensus outcome, whether the content is retained (status quo) or removed (ONUS trumps stable). The admin should write "there is no consensus to include this content, but, since the status quo has the property of being not a BLP-issue, but rather the discussion did find a consensus that it was NOT a BLP issue but there was still no consensus to have it in, but, we're defaulting to status quo." I guess you would then say, I favor inclusion, and I am at heart, and have always really been, an inclusionist, and QUO should trump ONUS (the problem with ONUS we've been harpin on about occasionally). IFF the content is sourced, NPOV, V, RS, BLP-clean, etc, (hopefully this isn't NP-hard :-)), and the content cannot be impeached for any policy-based reason but it's exactly balanced on the head of a pin? Andre🚐 06:35, 28 December 2023 (UTC)

Which version is the "status quo" for an article that is five weeks old, and whose inclusion of the sentence has been contested from its first hour?

Is that "best edit warriors win" (whichever wrong version is in place at the time the RFC closed wins)? Or "first mover advantage" (inclusion wins)? Or WP:FOLLOWBRD (exclusion wins)? WhatamIdoing (talk) 06:38, 28 December 2023 (UTC)

It's none of those. It's, admin/closer decides based on the reading the paper and text of the case in depth and poring it over. The order of reverts or the state right before the freeze doesn't matter, like you said, a wrongversion. Admin has to read the policy, read the discussion, weigh all the policies and the discussions in a mechanical type way, and then determine in this situation: how good were the edit warriors? Good meaning, did they make edit in favor of policy - and not wrong versioning per se but the content of the edit itself regardless of the revert status. And it's not a WP:SUPERVOTE to weigh policy in consensus. That is what admins are supposed to do, and always has been. Many major decisions on Wikipedia have been made when a numeric majority was against it, but because a sizable minority didn't understand the meaning of a policy according to the empowered closer. Those policies and norms are the reason why wikipedia's the last good website. Tho there are few other good sites out there, but I believe the ossification of admin-weighs-policy is not a good trend. Andre🚐 06:42, 28 December 2023 (UTC)

Again: The admin has already weighed the policies. Weighing the policy-based arguments is how the admin concluded that there was no consensus either way. Both sides have very good – but, unfortunately, exactly equally strong – reasons for their preference. I've re-written the sentence to make that clearer (Both sides had exactly equal arguments based on policies, guidelines, common sense and Wikipedia's values.) but let me know if it needs to be highlighted as well, or if you think that editors will be confused by the idea that Wikipedia has values. (Also, there's no numeric advantage to either side, so there is no numeric majority for the admin to close against.)

Deciding that the policy-based arguments are exactly equally strong, so we're going to do what I want would probably be considered a Wikipedia:Supervote, though we could actually say that this is the perfect situation in which to cast a supervote.

More to the point, if you believe that a statement that is properly cited, neutrally phrased, BLP-clean, etc. – one for which the only identifiable problem is that editors can't agree on whether to have it in the article – should be included (or excluded), then please just own that position! WhatamIdoing (talk) 06:53, 28 December 2023 (UTC)

OK, yeah, I basically guess my position comes down to, if I must be pidgeonholed, "supervote leaning inclusion, with a possible supervote exclusion." It's similar to a 2/3 majority with a line-item veto, like in government and parliament and that sorta stuff. And I DO agree with the values being important. I just don't really think there are a lot of cases where it's a perfect no consensus on all of that stuff, usually, if it's a NOCON on the policy, there's a value, or some intangible thing to weigh. No consensus usually defaults to "a status quo," except in the cases where the admin supervotes it back to "exclude." I don't think, if a new article, which had content that was according to the admin policy-OK but there was just a perfect unclear scenario, then it should remain. Andre🚐 07:07, 28 December 2023 (UTC)

No consensus usually defaults to "a status quo" – assuming that there is one. In this scenario, there is no 'long-standing stable version' to prioritize. WhatamIdoing (talk) 23:02, 28 December 2023 (UTC)

I hope I'm not bugging you with this btw but helping. Let me know if not. I figured out what it is. A tie goes to the runner. But if his foot wasn't on the base, he's not in. Andre🚐 09:53, 28 December 2023 (UTC)

I like the "tie goes to the runner" concept. Thanks for sharing that way of putting it. WhatamIdoing (talk) 23:02, 28 December 2023 (UTC)

Thoughts

Your fine print is saying this dispute falls into one of the editor discretion / decide by consensus areas. I wonder really how often it is realistic to say "Both sides had equally strong arguments based on policies, guidelines, common sense, and Wikipedia's values". Is it not more likely that while both sides cited policy and guidelines as though conclusive for their case, their P&G claims were either utterly wrongheaded or less conclusive than hoped. For example, WP:NOTNEWS as an uppercase shortcut sounds like it would be definitively negative about information sourced to newspapers, and is part of WP:NOT which we'd hope would tell us what not to include. But in fact, the "not" is "Wikipedia is not a newspaper" and is saying that editors need to be mindful that our inclusion and emphasis may differ to that of the press. It doesn't offer an algorithm for working that out other than by discussion towards consensus. A valid argument then wasn't so much "based on" WP:NOTNEWS as "permitted and valued by" it. So would it not be more realistic to say, in the large print bit, that policy and guideline were not definitive and permitted editor consensus which unfortunately did not arise despite good faith efforts.

You ask participants to carefully read any P&G they cite in justification of their solution. But surely your point is that if there were P&G tending to one solution, the closing admin would have cited those to close one way or another. So wouldn't anyone doing that be saying that this RFC is misguided and unnecessary because X. You seem to be saying that earlier discussions did exactly that and you felt they were wrongheaded because you believe the actual P&G text didn't support their claim. You also, in a way that is bound to frustrate, tell participants that "we" know what P&Gs conflict but are not going to tell you what they are. I think you are trying to say that you want an answer from the heart rather than an answer from an editor's interpretation of existing policy. In which case, why not just ban basing one's argument on existing P&G altogether in the hope of fresh argument.

I think there is a possibility that voters just saying "err toward inclusion" or "err towards exclusion" will simply display the inclusionist or exclusionist general tendencies and bias in the editor population and may itself reach no conclusion. After all, if the community was clear in its tendencies, wouldn't we already after all these years have that as a rule (e.g., if it doubt leave it out). Are such votes likely to persuade anyone to change their bias? If not then that's not really "discussion towards consensus" but just "vote with an opinion".

You ask editors to give examples from their topic-based editing experience. But I wonder if there are article-topic or content-specific reasons to be more inclusionist or exclusionist wrt material then surely those reasons are good enough they deserve to be more than a 50:50 tiebreak but actually written down as a general tendency for such content.

I'm concerned that the RFC ends up bigger than it should be. If the decision matters then let's make that tendency clear in P&G and not just for 50:50 edge cases. If the decision doesn't matter, let's not have editors waste their time giving heartfelt encyclopaedic reasons why we should do X, but instead focus on clever arbitrary tiebreak methods that prevent further disputing.

Another concern is about not asking people questions you don't know how they will answer. What happens if you get a definitive inclusionist position, say. Are you really, really, going to get admins citing that when they truly find themselves with a perfectly inconclusive 50:50 where both sides were equally intelligent and cooperative and displayed outstanding qualities of seeking to find compromise and work towards some consensus? Or will your RFC result end up getting cited by admins who find doing that easier than justifying why one side's arguments won? Or worse, your carefully nuanced edge-case tiebreak policy it gets cited during such discussions ("Remove per WP:WHENINDOUBTLEAVEITOUT -- ThoughtfulEditor") as though one side was the obvious natural choice. -- Colin°^Talk 18:58, 30 December 2023 (UTC)

Numbered lists, in case you want to complain about specifics:

1. You ask participants to carefully read any P&G they cite in justification of their solution. But surely your point is that if there were P&G tending to one solution, the closing admin would have cited those to close one way or another. – You've mixed the hypothetical story with the real editors in this (future) RFC. I ask participants in this discussion to read whatever WP:UPPERCASE they cite because I'm losing patience with people who cite some SHORTCUT to a section that is either irrelevant or says the opposite of what they think it means. In particular, if anyone cites WP:QUO as a reason to include the information post-RFC, then you will know that this editor read neither the story (because the article is too new for any of the content to be long-standing) nor QUO (which says that it applies only during the discussion/RFC).
   The hypothetical admin found the arguments equally strong in the story's discussion. Whether you choose to read that as "both sides had equally good, obviously policy-based arguments" or "both sides had equally weak arguments" is up to you, but what matters for consensus is that neither side had an advantage in either numbers or reasons.
   Also, if you're interested in the NOT problem, then you might be interested in the list at User:WhatamIdoing/Sandbox 3#Non-NOTs. I'd hoped that there would be some obvious candidates to send to WP:RFD (either for re-pointing to a section at NOT, or possible for deletion [if they weren't used much]).
2. You also, in a way that is bound to frustrate, tell participants that "we" know what P&Gs conflict but are not going to tell you what they are. I think you are trying to say that you want an answer from the heart rather than an answer from an editor's interpretation of existing policy. – There are three things behind this decision.
   • The first is that I really do want an answer about what The Community™ believes the best practice is, because we need to make the advice pages match reality, rather than matching one or the other advice pages. Whether you understand this as "from your heart" or "from your experience" or "from your best judgement" is immaterial to me; any of those will do.
   • The second is that while there are two policies that are obviously in conflict under certain limited circumstances (e.g., when there is no consensus about whether to retain long-standing content), the problem likely affects other policies and guidelines as well. I don't have a complete list, and the list would likely depend on the outcome (e.g., the Wikipedia:Editing policy might need to be tweaked if editors default to exclusion, but it probably wouldn't if they default to inclusion).
   • The third is that there's a particularly stupid form of argument that happens occasionally with discussions about changing policies, and I think this discussion is uniquely susceptible to it. It sounds like this:
     • Alice: "Hey, guys, this policy says X. I'd like to change it to say usually not-X."
     • Bob: "Oh, no, we can't have this policy say not-X, because this policy says X! Saying not-X would violate the very policy you propose changing!"
   • I'd like to avoid having to explain to Bob that if the policy is changed as proposed, then it won't actually say X any longer, and therefore saying not-X would no longer violate the policy.
3. If the decision matters then let's make that tendency clear: I agree. We currently have a situation in which editors approach these discussions with "my policy says exclude because you didn't get consensus to include" vs "well, my policy says include because you didn't get consensus to exclude". When it is possible for me to tell you today that I get my way because Policy A says no consensus means exclusion, and for me to tell you tomorrow that I get my way because Policy B says no consensus means inclusion, then we have a problem that needs solving. We get about three-quarters of a million registered editors each year, and we have about 10,000 highly active editors at any given point in time. We should not assume that every single editor's character is so sterling as to be completely immune to the temptation to cherry-pick a page that favors my preferred outcome. If that isn't self-evident, then think about the strong likelihood that the 10,000 includes about 200 narcissists and 100 psychopaths, because editors are people. Instead of continuing with conflicting policies, we should figure out what the rule really is and then write it down so that everyone is working from the same ruleset. I have written this story on the assumption that the rule that will be chosen sounds more like "If you truly can't decide, then include/exclude" than "We appoint WhatamIdoing as Her Serene Illustriousness, Princess of the Wiki, and all shall do whatever she wants" (though I am open to that outcome, despite its many obvious flaws, if that's the only way we can get this resolved. My first wish will be to resolve the conflicting advice on this point).
4. voters just saying "err toward inclusion" or "err towards exclusion" will simply display the inclusionist or exclusionist general tendencies and bias in the editor population – It might, but to the extent that said tendencies and bias show up in reality (e.g., in RecentChanges), then documenting that reality could help editors understand what to expect. Our real policy is the things we do everyday, not the words we write in gold ink on purple vellum on a page that says "policy" at the top.
5. I wonder if there are article-topic or content-specific reasons to be more inclusionist or exclusionist wrt material – Finding this out is the purpose of that request. If such reasons appear, then we need to write the policies and guidelines to say that. We already have a rule that says when in doubt, exclude for BLPs; this is why the story specifies that the content is not related to any BLP. Perhaps the result here will be to say, e.g., that potentially unwanted material about the arts and culture get included, but potentially unwanted material about health and science gets excluded. (There's a list of broad categories at Wikipedia:Size of Wikipedia#/media/File:Size of English Wikipedia (1000 vol).svg, if anyone's interested.)
6. Are you really, really, going to get admins citing that – Right now, we get some citing one policy to "prove" that the policies require inclusion, and others citing a different policy to "prove" that the policies require exclusion. Having them all cite the same thing would be an improvement.

The current form of this story was developed based on discussions that have gone approximately like this:

• Me: "So, in principle, if editors really, truly can't come to any sort of agreement, then what should they do?"
• Editors: "The disputed material can be removed per WP:BURDEN in a WP:CHALLENGE."
• Me: "It's already cited, so you can't challenge it."
• Editors: "Oh. Then they should have an RFC, or otherwise follow relevant dispute resolution processes."
• Me: "Right. So they followed all of them, and they discussed until they were blue in the face, and it's still no consensus."
• Editors: "They should go with the long-standing version of the article until a consensus emerges."
• Me: "The article is brand-new. There is no long-standing version."
• Editors: "Then they should go with whatever side has the better argument."
• Me: "If one side had materially better arguments, then we'd have had a consensus, right? No consensus == neither side had stronger arguments, so we can assume the arguments were equally strong."
• Editors: "Sure, but surely one of the sides was at least a tiny bit stronger."
• Me: "A tiny difference still wouldn't be enough to make a difference in whether consensus was found, but if it would help you understand that no consensus means no consensus, then we'll say that the arguments from each side were carefully reviewed and judged be exactly, precisely, completely, absolutely equal in strength."
• Editors: "How about they find a compromise?"
• Me: "You can't compromise on whether something is mentioned at all in an article."
• Editors: "Well, maybe they should go with whichever side did a better job of canvassing, I mean, whichever side had more supporters".
• Me: "Vote counting is against the Wikipedia:Consensus policy, and besides, the votes were equal."
• Editors: "Well, this is all just completely unrealistic. There's always one side that wins."
• Me: "So which side is it? Which way do these directly conflicting policies get changed to?"
• Editors: "Um, they only directly conflict when there's no consensus by either strength of argument or by vote-counting, and we can't find a way to block or ban enough people from one side that a re-vote would have a different result, right? Well, that hardly ever happens, so who cares? Besides, the only thing that really matters to me is that you can't take away the policy I quote when I tell newbies that no consensus means that I win and they lose."

WhatamIdoing (talk) 22:59, 30 December 2023 (UTC)

I don't think I have the story and the RFC confused as they do interact. The story can't be "true" (that an admin has no golden policy they can cite to say what to do for "no consensus", as the existing policies are either inadequate or contradictory) if there is a golden policy that a participant on this RFC could cite that would convince you/everyone that there is in fact an existing solution. Perhaps the best that anyone can do when citing existing policies is to say which they favour and which they disfavour, and we rewrite or remove the least favoured. Your Alice and Bob scenario is indeed a problem but would be improved if instead of saying "X is policy" as their justification, said "I think having X as policy is great, and helps with A, B and C. And I think Y as policy is awful and leads to P, Q and R." If that's what you are wanting people to do with existing policy (which ultimately we may be altering/removing) then I don't know why you won't mention them and ask people's opinions on which they prefer (or if they can be tweaked to apply to completely distinct scenarios and thus not conflict). And alternatively, one could ask people just to explain in their own words what rule they would like, without reference to existing policy at all.

Looking at, *cough*, a couple of policies in this area, I can see that one is written to value the encyclopaedic content that any reader sees and to entirely ignore history and past contributions and contributors. In this policy we have decided that it is better if readers don't see / do see contentious content, and what Wikipedia displayed last month is unimportant and all the polishing that editors did three years ago when it became a GA is unimportant. And the other is written to value history and stability (and thus the contribution of editors) and does not care one bit about the encyclopaedic content (except in a few clear content areas, like BLP, copyright and EL). In this policy we have chosen to arbitrarily value what previous editors did even if they were idiots while hoping they were wise. So there are different philosophies and values going on here. Is that worth exploring? It is possible that valuing one thing has clear community consensus and the other can be dropped. Or can we address the concerns of one group in a different way? For example, I don't think many of our readers care one bit about stability or what the article said last week and would instead value polished content they are actually reading. But if the editors writing that material got disheartened that it was too easy to wreck their longstanding prose with either additions of cruft and tittle tattle or removal of carefully selected nuggets of information, then we no longer have a project. Can we fix that without having contradictory policies?

I'm glad you included that exchange above, as it has saved me from repeating it! I think you've purified your example in order to address some of the "obvious" solutions you feel are not universally adequate. But in doing that you end up creating a story you are asking participants to decide upon that will end up being used in lots more common stories that you didn't ask them to think about. It is like pretending, for now, the existing conflicting policies are only in fact a problem in this pure scenario and could have been resolved already in all the other more common ones. While those existing policies do get cited in closing remarks in an RFC, they are mostly cited outside of that. For example:

• Bob adds some material. Alice removes it and cites policy X to say Bob needs to gain consensus for his addition. Bob and Alice argue and wonder why they keep bumping into each other on policy pages since they clearly hate each other. They fail to reach agreement and Bob is forced to give up. Bob is fed up that it is hard to add good content when there are fools like Alice on the project and never any other adults around when you need them.
• Alice wrote some material years ago. Bob comes along and removes it for the nonsense it clearly is. Alice restores it and they argue. Alice and Bob's friends join in and the discussion wanders about the page for a bit till everyone gets fed up. After two days of no further comments, someone says policy Y says the old content, that Alice wrote, should stay. Nobody even bothers to add numbers on one side or another, for it is clear that there are two significant camps and they do not see eye to eye. Bob is fed up that it is really hard to remove nonsense from Wikipedia, especially nonsense that has been embarrassing the project for years.

None of these even reached RFC let alone closing remarks by a third party. And how many admins would really be comfortable declaring that e.g., the Brexit vote of 48% vs 52% represented anything other than "no consensus". So my point is that whatever changes are made here, will influence debate and outcome at a much earlier stage and in more casual scenario than described by the story. Is that a fair setup? -- Colin°^Talk 16:40, 31 December 2023 (UTC)

Your second story reminds me of WP:WikiSpeak#consensus:

consensus n.

One of the three states that can be reached at the end of a discussion after all parties have become thoroughly fed up with it; the alternatives are no consensus or for pity's sake, I wish I'd never gotten involved in this.

As for whether it's fair, having conflicting policies (our current state) is unfair, and resolving that should improve fairness. Of course, it's possible that we won't be able to resolve that; the conflicting policies might represent a real divide in the community's views. WhatamIdoing (talk) 22:28, 31 December 2023 (UTC)

I still think the two value-systems is useful. Consider the "encyclopaedic content is all that matters" view. In this case your good-sized RFC generated good reasons to keep and good reasons to remove but importantly, all our policies in this regard mean participants have to focus on the article content, its sources, the topic literature, etc. Nobody can argue "But Bob worked on that for days and will be very upset to see it go". Or "Alice wrote that so I wouldn't be at all surprised if it is nonsense". Therefore if such a scenario reaches a stalemate then it doesn't matter, for the encyclopaedic content, if our rule is inclusionist or exclusionist or if our rule values the new or values the old. Any choice is equally as good or as bad as the other. But this scenario is not nearly as common as the two stories above and others. Such as where the rule about what happens in the absence of consensus is invoked even by the reverting editor before they even begin discussion. Or is invoked during discussion or after informal small scale discussion. These are the most common.

Lots of edits are either kept or reverted and stay reverted without discussion. So, for the edits that merit discussion:

• We'd want to be exclusionist if most of these discussed edits add bad stuff or most edits removed existing bad stuff.
• We'd want to be inclusionist if most of these discussed edits added good stuff or removed good stuff.

And generally, for edits that add or remove or reword:

• We'd want to value retaining longstanding material if there is evidence that the older the text the better it is and new discussed edits generally made things worse.
• We'd want to value retaining discussed edits if there is evidence that edits improve the project and that there is a lot of bad old material that needs improving or removing.

There may be other arguments for or against an approach and maybe the RFC is the place to discuss them in detail rather than here. But off the top of my head, I'm starting to wonder why we have a rule that tends towards keeping existing text in a project where we ask everyone out there to mercilessly edit our shit if they think they can make it better. Anyway, I think some framework for thinking about our answers is better than just saying, here's a story now tell me how it should end. Which may just get people voting. -- Colin°^Talk 21:02, 1 January 2024 (UTC)

A Conceptual framework? Our conceptual framework is approximately:

• Comply with policies and guidelines. If you follow the rules, you'll almost always get a consensus that supports your choices.
• If there are no relevant rules, or if they're contradictory or competing, then (seriously) just get consensus, through any legitimate method and on any rational basis you can achieve it.
• If consensus is not possible, then if it's AFD, keep it; if it's a possible copyvio, then delete it; if it's about a BLP, then remove it; if it's an admin action, then un-restrict it; if it's article content – well, it depends on which page you choose to quote, so that's another thing that you're going to fail to find a consensus on today.

WhatamIdoing (talk) 05:55, 2 January 2024 (UTC)

I don't think that's a framework for thinking about a problem. It's just a series of bullet points, some of them dubious. I would guess that most edits and most results occur in the absence of explicit consensus. And in those times where we expect editors to get consensus, when challenged, it isn't really that there are "no relevant rules". It is that that relevant rule is "discuss towards a consensus". Rules are not algorithms. The rules of the road don't tell you how to drive from London to Glasgow, so why should we expect all rules to tell us how to write and polish content?

I'm thinking more that our decision of what is best for the project has to meet several aims. We want the encyclopaedic content to be great and to get better. We also want editors (both new and old) to feel their contributions are valued. We want this to be as easy and quick as possible because this is a wiki and because nobody is getting paid. Unique among significant publications, we don't have a closed and carefully selected group of expert writers, copyeditors, editors and publishing authority who creates the content for others to read. Instead we expect our readers to become involved themselves and primarily allow anyone to edit anything to any extent. Protecting content is regarded as against our culture and we do it only when we feel we have no other choice. We have no hierarchy of editors to escalate the content decision making process up to. These sorts of things might shape our thinking about what is the best result (and others are probably better than me at expressing this). -- Colin°^Talk 10:02, 2 January 2024 (UTC)

Since we're starting with a dispute, explicit consensus feels more salient than implicit consensus (aka "normal editing").

When I wrote "no relevant rules", I was thinking of us having no rules that are relevant to the content, rather than no rules about editor conduct. For example: Shall we mention an unimportant but not unreasonable detail (e.g., the name of a non-notable family member, the location of the business's original office building, )? Shall we choose this picture of the subject, or that similar picture of the subject? We have a rule that says not to add multiple similar pictures, but we don't have a rule telling us which one of the similar options to prefer.

When there is a relevant rule, editors should normally follow it. For example: Given a choice between a newspaper article that says smoking tobacco causes lung cancer and a recent review article in a reputable medical journal, they should pick the MEDRS-ideal source. When the relevant rule(s) for content exist, are clear, and are not contradictory, you will (normally) achieve consensus.

When there is no relevant rule, the goal is still consensus.

The comment Protecting content is regarded as against our culture makes me wonder if you are leaning towards neither including nor excluding potentially unwanted content, but towards an anti-status-quo approach: if the article said for all of last year that Bob's Big Business's first office was in his garage, its second was on Main Street, and the present headquarters is at One Widget Way, and the dispute is over whether to mention the Main Street location, then perhaps you'd be inclined to remove it, because protecting content is not one of our values? And if it wasn't there last year, you'd be inclined to keep it? WhatamIdoing (talk) 18:57, 2 January 2024 (UTC)

I'm trying very hard not to "lean" in any direction, since the RFC isn't live, but just to think about what values/pressures/aims might encourage me or anyone else to !vote a certain way. I'm currently not sure it matters at all which rule is picked if we had 12 good editors who had exactly a 50:50 split. As I said, I think their discussion should ideally have focused on the value or otherwise of the editorial content, and not on whether it was there yesterday or last year or who added it. If they truly are balanced on the value of the content, then it truly doesn't matter which way we go, for content purposes. Perhaps you can think of a reason why, for content purposes, we should err one way or another in such a case, cause I can't. For project/community purposes, right now I can see arguments for retaining what someone or collaboratively was worked on and created in the past and I can see arguments for valuing what someone/some people new has/have contributed (whether by adding or removing or revising). Right now, I think our explicit values would appear to indicate we encourage the new and perhaps it is old established editors who are being over protective of existing content. But I'm open to hearing if there are good arguments to suggest that's not a workable mindset. That's the sort of discussion I'd hope to appear in an RFC, so why I'm trying to think of some framework or even just "think about X" or "think about Y" or "what is important to you" thought provokers, so that we don't just get "Here's an RFC, read it over and then type your opinion into the box".

I think there is a danger with invented scenarios. They can be helpful but also they can push someone to give an opinion based on the random choice of scenario rather than thinking more abstract. For example, in the absence of any information, one might think including all business locations was reasonable rather than arbitrarily picking two out of three. There must be a reason why not only one editor but half the participating editors feel the middle location isn't worth mentioning or perhaps more likely is misleading to mention according to some criteria not valued by the other editors.

There are two axes. One is inclusion vs exclusion and one is change vs status quo. And these are also what are causing conflict in our policies. If one is thinking about whether we tend towards change or tend towards status quo, in the presence of no-consensus, then does it matter if that change is addition, removal or editing it in some way. Perhaps then it helps to not say what kind of edit it was. Or do you think removal changes are different in some way to addition changes that might make one value change over status quo or status quo over change? May be they are? But from an encyclopaedic content point of view, it doesn't matter what the change is. The editor debate was on the merits of mentioning one business location, not whether an editor has proposed adding it or added it and got reverted, or an editor is proposing removing it or removed it and got reverted. What are the merits in taking a purist "the encyclopaedic content is all that matters" approach vs a "It matters what the content was and what the content would be" approach? Is one of these axis always inferior to the other, or can we modify them both so they interact in a way that isn't contradictory? -- Colin°^Talk 08:39, 3 January 2024 (UTC)

But, just to keep this in mind, I think our decision might in fact have a bigger impact on all the casual editor interactions and much less on the "We had a big RFC and still couldn't decide" cases. -- Colin°^Talk 08:40, 3 January 2024 (UTC)

I tend to approach this from the mindset of Clinical equipoise, which often feels like "Antibiotic A or B? Eh, it just doesn't matter than much. Either will work. Flip a coin". But clinical equipoise also happens in other situations, where emotions run much higher: "She'll die if you follow his treatment plan! You must try mine!" vs "No, don't listen to that quack! There's more evidence for my approach!"

Editors could be, and have been, evenly divided over questions like whether to mention genocide or war crimes in articles. You could have Twelve Good Editors evenly divided and thinking that it just isn't that important what decision is made, but you could also have twelve good editors convinced that if the other side 'wins', it will be an unmitigated disaster for the article and well worth edit-warring to protect readers from the m:Wrong Version.

I like your idea about "think about what is important to you". That's what I'd been trying to suggest with the line about our values. Maybe I'll try to do more with that. WhatamIdoing (talk) 18:19, 3 January 2024 (UTC)

Yes but if six of them thinks it will be a disaster if the content is kept (say) and six of them think it is a disaster if the content is removed, there isn't a content-based solution, and you are always going to upset half. It neither matters if the six editors are equally "meh" about in/out or equally "I'll quit otherwise" about in/out. A longer term solution might be to foster a content-based inclination such as we already have for BLP and for which MEDRS is often used, to keep dubious stuff out, or on some other area perhaps, to keep wonderful nuggets in. But, as with BLP, those would then be exercised far far more often than when there is an explicit no-compromise stalemate that will not benefit from further eyes and further discussion. It get exercised early and casually.

The conflict in our policies seem to stem from one set of policy editors valuing content (with a mindset of keeping the crap out, say) and another set of policy editors valuing existing editors (with a mindset of not losing people to all these newbies who do nothing but wreck our precious work). So if the problem has arisen from competing (and possibly wrongheaded) values, then that's where the solution may lie. -- Colin°^Talk 08:28, 4 January 2024 (UTC)

If the editors are 'meh', the lack of clarity about whether no-consensus content should be included or excluded does not usually result in serious behavioral problems. When the editors are upset, the lack of clarity results in edit warring, ANI posts, AE actions, and blocks. Consequently, I think that the emotional temperature of the discussion matters.

I believe you are right about the rule getting invoked early. The existing, contradictory (under certain circumstances) rules are being invoked early. I'd like to reach a point at which the rules that are being invoked, whether they are invoked early or late, are not contradictory. WhatamIdoing (talk) 18:17, 4 January 2024 (UTC)

About the rule getting invoked early: That's happening now. Right now, this is possible:

• Alice: We'll just have to agree to disagree.
• Bob: Yeah, and Policy A says that if we don't agree, I get to exclude your content.

and

• Alice: We'll just have to agree to disagree.
• Bob: Yeah, and Policy B says that if we don't agree, I get to include my content.

I'd like to limit 'Bob' to one answer or the other. WhatamIdoing (talk) 18:48, 4 January 2024 (UTC)

I get you are trying to remove the contradiction, and that is important. But you are also trying to have policy state "what to do" not only in the RFC scenario of "we tried everything and still disagree" but also the "before you even begin arguing with me on the talk page, I'm going to reset/set it to what policy says is the "no consensus" position" or the "I shouldn't have to recruit help on WT:MED every time I want to keep crap out of this article" or "Can we not have an RFC to make every edit in this topic please" areas. And that's not the question you are asking. So you may get a "meh" I don't care, in your scenario, which they pick. Or you may get a variety of votes based on how editors wanted some recent dispute to go. You might not get votes where they consider these scenarios. You could get a result that is a bit random or itself not particularly indicative of a consensus.

If you want to stick with the "we tried everything" story, in order to stop people asking "well, have you tried this", then perhaps at least add. Ok, now you've thought about this pure case, consider that whatever policy says, its going to be, possibly wrongly but still very frequently, invoked at very early stages in dispute and we don't want to make such a bad choice that editors then feel they have to take every little disagreement through an RFC or recruit all their mates every time in order to achieve their goals, which might have been easier before. Ideally we'd actually make it easier to reach the inevitable outcome without there having to be a big discussion. -- Colin°^Talk 08:29, 5 January 2024 (UTC)

• you are also trying to have policy state "what to do" – Not necessarily. The contradiction could be resolved by removing the 'wrong' statement(s) that currently exist, without providing any new statements about what to do. Or by removing all of the contradictory statements, and thus providing no information whatsoever about what to do.
• If it's bad to invoke policy in early discussions (what I'm guessing is your overall feeling?), then why wouldn't it be at least as bad for editors to cherry-pick the policy that favors their view? People are currently cherry-picking the policy that matches their view. Why are you concerned that in the future, the editors who are currently playing games with the rules might be constrained to quoting only policies that match the broader community's view? Would this not be an improvement in fairness?

WhatamIdoing (talk) 02:36, 6 January 2024 (UTC)

I don't really know whether it is bad. It happens and it does have the effect of short cutting what could be a lot of tedious argument that doesn't achieve much. Whether it is fair might matter less than that. I totally agree the current contradiction is undesirable. I'm concerned how the RFC is worded, not whether having one is a good idea. But I also think that if editors here agree, possibly based on them thinking only about pure and well participated stalemates, that Wikipedia's new default opinion, whenever any two editors find themselves in disagreement, is that the material gets included, then you've made a whole big problem that will probably result in either mass resignations or your policy change being really really short lived. -- Colin°^Talk 21:54, 8 January 2024 (UTC)

Or perhaps it would trigger mass participation by editors who post "Whatever it is, I'm against it", to make sure that any borderline dispute ends up as exclusion.

So far, my question seems distinctly unpopular, though. I may not get enough responses to let me know what the community's overall view is. WhatamIdoing (talk) 22:10, 8 January 2024 (UTC)

Possibly the intro is too long and discouraging. We could have a fight and you take me to AN/I for an indef block. Then you'd get eyes on it. -- Colin°^Talk 18:38, 9 January 2024 (UTC)

I appreciate the creative proposal, but ideally this should result in fewer fights, not more. ;-) WhatamIdoing (talk) 19:22, 9 January 2024 (UTC)

To be honest I don't think this RFC is working. We have two editors who think the main reason for excluding "facts" or "encyclopaedic information" is IDONTLIKEIT, which very much isn't the case. Others have rightly spotted that it really doesn't matter which way you pick for your example and so like some coin tossed their leaning is pretty arbitrary and not based on any kind of "if this was the default, we'd get all these arguments happening that didn't happen before". And others have spotted one of the two policies and gone with that, or have failed to mention either of the existing policies and don't think we should have any policy on no-consensus, that they should argue more but better this time. And as you said already, there aren't the numbers here to write or remove longstanding policies. Maybe some will disagree with that summary, but I don't think it is far off the mark.

Maybe another approach is to examine cases where either of the two policies have been cited and see which tend to produce what could be regarded as the suboptimal outcome. Is one of them actively harming Wikipedia? Perhaps editors are mostly picking whichever one is actually the better, in that case, for the encyclopaedia, and the fact that they contradict is only something for policy nerds to fret about? Or maybe in each case it isn't possibly to fault editors picking either, that they clearly picked one because it helped their own case, but if the choice wasn't terribly important, then as long as picking something stopped the arguing... -- Colin°^Talk 10:06, 16 January 2024 (UTC)

Well, you might be right, but I think we're getting an interesting theme right now with @Aquillion and other recent commenters recommending against any default. The solution might be to remove all such advice, so that we really have no policy claiming default at all. That would mean removing ONUS, which says the default is that I get to remove your content unless you produce a consensus in favor of it, plus the line in NOCON that says the default is that I (usually) get to keep my content unless you produce a consensus against it, assuming I can claim that my version is the 'prior' version.

(For a brand-new article like the one in my story here, you will, of course, declare that the correct long-standing version is the one before 'my proposal to add it', and I will insist equally loudly that the correct long-standing version is the one before 'your proposal to remove it', but that can probably be settled with little more than one or two ugly sections at ANI.) WhatamIdoing (talk) 17:57, 16 January 2024 (UTC)

Well if we are considering that path, the above analysis would seem even more important. How many cases were there where those rules were cited and it was a good thing. Would Wikipedia have improved if we forced the editors to keep fighting or to escalate?

The two policies do have a synergy that means it is a little harder to add some random fact of disputable merit to Wikipedia but also that it is harder to remove that random fact of disputable merit once it has been in for a while. I'm not sure that synergy is healthy but it's there. If we just removed ONUS but keep NOCON, say, then it is possible that Wikipedia fills up with shit and also makes it hard to remove that shit. So be careful if we go down that route. I'm not sure what detrimental effect there would be if we removed the current advice at NOCON but kept ONUS. -- Colin°^Talk 08:40, 17 January 2024 (UTC)

I believe the concern expressed by other editors is that keeping ONUS but removing NOCON would make it harder for them to protect their existing content (e.g., from anyone who thought their existing content was inappropriate). WhatamIdoing (talk) 02:16, 19 January 2024 (UTC)

But isn't it telling that you say "their content" here. Isn't that ownwership? Surely, those exact same editors go to an article written by someone else, like some long ago FA for example, and look with despair at all the shit it has accumulated over the years, and say "this needs a scythe taken to it". I am beginning to think there is faulty thinking going on here, where we just remember how stupid the other people are, but don't consider that we are other people to someone else, too. -- Colin°^Talk 09:14, 19 January 2024 (UTC)

(outdent)

I do think that some of these disputes, and the resulting cherry-picking of which policy to name in support of "my" preference, arise from an WP:OWN mentality. Additionally, we also firmly believe that we make better decisions that average. This mental mistake is the same as the one that makes us each believe, when we get in the car, that we are above-average drivers. Since we firmly believe that we're better than most editors, that our opinions are more neutral, that our search for sources was more skillful, etc., when the discussion doesn't immediately go "my" way, then of course I'm going to do my best to get "my" outcome. WhatamIdoing (talk) 14:54, 19 January 2024 (UTC)

I'm sure believing one is among the better 50% of drivers or Wikipedians is no bad thing, for one's own mental health. But it certainly skews thinks if that influences whether the no-consensus default should be to "retain my text" vs "accept the other guy's changes" from the viewpoint of an editor watchlisting text they or their wikifriends wrote. I'm wondering why we should remember conflict where we were the editor retaining our existing text rather than the editor modifying other's existing text.

1. Watchlists give us far more opportunity, without even making any effort, to come into conflict with editors who are editing articles we once edited (or discussed with others). Watchlists don't randomly select an edit that needs some eyes on it, where we had no previous investment.
2. As text ages, the editors who created it fall off the project or unwatch. So a given edit may only be observed by the editor making a change rather than the editor who wrote the text being replaced. That makes that edit appear easier from the editor's point of view, and the point of view of the original writer is lost, as they have gone.
3. Many editors find themselves doing less new-text editing over time, and more defending/discussing text as they age on the project. Maybe the ones that are still really productive on writing new stuff really are also in the top quartile of edit quality and so really do find their work retained without fuss. One can become intolerant of other editors who are are also improving the project, even if they are making lower quality edits and don't get the citations right, or at all, or make MOS errors that needs fixing. So there's a demographic difference where older defensive editors who know how it should be done but maybe aren't doing are making rules that would have prevented their younger creative selves from getting started. -- Colin°^Talk 15:33, 21 January 2024 (UTC)

The hidden question

Should we deprecate “No Consensus”?… Because, essentially, that is what this RFC is asking. Removing the option to close as “no consensus” is the only way this scenario can be resolved. We are told: we need a resolution, one way or the other… and “no consensus” leaves it unresolved. Blueboar (talk) 14:09, 1 January 2024 (UTC)

I don't think that's what I'm asking. I think I'm asking to change from the current system:

• Editors: Argue, argue, argue, RFC!
• Closing summary: No consensus.
• Editors supporting exclusion: "Since it's 'no consensus', policy A says that we must exclude!"
• Editors supporting inclusion: "Since it's 'no consensus', policy B says that we must include!"
• Editors: "I will edit war to the death to enforce the One True™ Policy!" (Optional response from the admins: Block, block, block!)

to a new system:

• Editors: Argue, argue, argue, RFC!
• Closing summary: No consensus.
• Everyone: "Since it's 'no consensus', all policies say that we must {include|exclude} (unless and until a true consensus emerges)."

We have no consensus at AFDs without anyone wondering whether that should normally result in the article being kept or deleted. We all know that no consensus = keep. We should be able to have a no-consensus outcome on article/content RFCs with the same level of agreement about what should happen when there is no consensus about inclusion/exclusion. WhatamIdoing (talk) 16:38, 1 January 2024 (UTC)

I think it is valuable to have a rule for what to do when "no consensus". It is a common scenario in life as well as on wiki. The problem is conflicting policies, not that we have to demand or pretend there was a consensus. The question is whether the current rule(s) on what to do in the absences of consensus should all be kept.-- Colin°^Talk 20:37, 1 January 2024 (UTC)

But… a rule for “what to do if no consensus” essentially negates the entire point of “no consensus”. A finding of “no consensus” means “we don’t agree on how resolve this” - and (apparently) that is no longer acceptable. So… why not simply say: “Closer, sorry but “no consensus” is not an option. you need to make a decision one way or the other - so there is a resolution.” Blueboar (talk) 22:06, 1 January 2024 (UTC)

Ask the closers to supervote in such cases? IMO there are worse options.

We have to have an option for declaring no consensus; when there really is no consensus – and editors realistically will not always come to an agreement – then claiming there is a consensus would be counterfactual. But once that's been truthfully declared, we could set the rules to do anything we wanted in response: supervote, flip a coin, compromise by making the content visible on desktop but not mobile (this is technically possible, for discrete material [e.g., a sentence, paragraph, or image]) – anything. WhatamIdoing (talk) 06:02, 2 January 2024 (UTC)

Blueboar, it isn't just the big story on the current proposed RFC where we presumably have a decent participation by some representative subset of the community. If Bob removed something and Alice added it back and they had a brief discussion and failed to agree... that's "no consensus". Are you saying that somehow we should, on a volunteer project, force Bob and Alice to argue in perpetuity or recruit enough friends to swing it every time this happens? The "what do we do if we can't agree" is a valid thing. In the real world we often have a hierarchy with someone further up the chain or considered more experienced or whatever who gets to make a final decision. And very often "status quo" is an appropriate default in life, but possibly it isn't for a wiki that should be easy to edit. After all, the default on Wikipedia is not "Here's a great encyclopaedia we built already. If you think you can make it better, please start a discussion on the talk page and gain consensus for your edit". It very much is that if you think you can make it better, please just go ahead and change it. The "having to gain consensus for" bit is exceptional. And I don't know the stats for how much that happens for when it would have been better to keep/remove or retain/revert. Anyone got any ideas? -- Colin°^Talk 09:47, 2 January 2024 (UTC)

I’m not actually advocating that we deprecate “No consensus”… I am saying that the (perhaps unintended) consequence of this RFC will be to do so. It will replace “No consensus” with a “Default consensus”. Blueboar (talk) 13:10, 2 January 2024 (UTC)

I don't think you'd call it "default consensus" as consensus literally means the general agreement among a group of people, which didn't occur. I think the term would be "fallback position" or "fallback procedure" which might be "status quo" or "Ask Jimbo to decide for us". If we (not going to happen, I know) decided that our fallback was "Ask Jimbo" then nobody would say "Oh we decided to go with keeping X because the default consensus was to keep it". We'd say "The community couldn't agree so we asked Jimbo". Similarly for "There was no community consensus on that, so the fallback position was to remove it". -- Colin°^Talk 13:57, 2 January 2024 (UTC)

Meh… I’m sure we could fit a few more angels on that pin head. Blueboar (talk) 15:10, 2 January 2024 (UTC)

I'm not sure this warrants a rude reply implying the reason I have given that you are wrong is tediously trivia of no value. The reason you are wrong is that the word "consensus" means a certain thing, and the community would be most upset if people went around saying we had decided something by "consensus" but clearly hadn't cause there wasn't. There's simply no such thing as a "default consensus". Further I don't think WAID's proposal would be to remove NOCON but to potentially amend the procedure/position that it recommends. -- Colin°^Talk 15:37, 2 January 2024 (UTC)

Colin, most edits don't get reverted, so presumably most edits are positive contributions. However, relatively few edits add or remove a significant amount of material. In the visual editor, about 20% of edits during the last hour added ref tags, so that suggests they added more than a few words of content.

In the 2010 wikitext editor (which is what the Undo button leads to), contributions tend to be more minor, especially from high-volume editors – either undoing (which could be an addition or a removal) or non-content edits (e.g., 90% of the edits to Science by press conference between late 2016 and early 2020), or small copyedits.

I don't know if that feels like a solid answer to your question, but it's possible that we could get statistics for reversion rates at Wikipedia:Request a query if you wanted more answers. WhatamIdoing (talk) 20:39, 2 January 2024 (UTC)

The lack of a swift revert doesn't mean the contribution was viewed as positive, or even, for that matter, viewed (by someone who might know one way or another). Witness the amount of student contribution that sits for years until someone spots the source doesn't back it up or it is plagiarism. And some challenges don't fit into your "ultimately up to editor consensus on whether the article is improved" model. A WP:V challenge always follows the "it's removed unless there is consensus there are reliable sources for this" and plagiarism or blatant copyvio isn't arguable and we'd have to exclude BLP where that's already a given. Someone adding ref tags is unlikely to get disputed unless those refs are dishonest or mistaken. I think it would be pretty hard to just analyse article text who's worth/value/appropriateness/etc was disputed. It may vary from subject to subject. -- Colin°^Talk 13:07, 12 January 2024 (UTC)

I guess the point is that if text is disputed then at least one person thinks the text was good to add and at least one thinks it is bad to add. This isn't true of the majority of edits, as you point out. -- Colin°^Talk 13:08, 12 January 2024 (UTC)

I must be wrong here...

But WP:NOCON appears to very specifically state that When discussions of proposals to add, modify, or remove material in articles end without consensus, the common result is to retain the version of the article as it was prior to the proposal or bold edit, which appears to both give a clear answer to the dillema and render the argument that WP:STATUSQUO is inapplicable moot. I assumed that this section was one of the two policies referred to in the sentence One policy says content without consensus is usually kept, and the other says content without consensus should be removed., but, after reading the text of the RfC multiple times, it appears that the two conflicting policies are never stated. Obviously @WhatamIdoing read the text of NOCON before creating this, so there must be something I'm missing, but I cannot forumulate a response to the RfC without clarification. Mach61 (talk) 03:58, 22 January 2024 (UTC)

@Mach61, I would very much prefer that you answered the question from the viewpoint of what you personally believe is best for Wikipedia. We are dealing with a WP:PGCONFLICT situation involving two core policies (one sentence in each of them, both sentences boldly added years ago without evidence of support for their addition in discussions at the time).

Our policy on that problem says If policy and/or guideline pages directly conflict, one or more pages need to be revised to resolve the conflict so all the conflicting pages accurately reflect the community's actual practices and best advice. I'm asking what your best advice is, and what the community's actual practices are. WhatamIdoing (talk) 05:47, 22 January 2024 (UTC)

@WhatamIdoing: Could you please just state precisely what sentences in what policies are in conflict? If they're clearly contradictory it will be obvious that citing one of them in the discussion is pointless. Mach61 (talk) 06:27, 22 January 2024 (UTC)

Unless and until the best practice is identified, we don't actually know which ones might need to be updated. For example, if editors declare that having less information is better than having more information, then the second sentence of Wikipedia:Editing policy#Adding information to Wikipedia might need to be changed. If they decide that more information is better than less information, then that sentence will not need to be changed.

The main conflict is:

• one core policy says to exclude information unless and until there is a consensus to include it, and
• another core policy says that when there is no consensus either way, then follow the "status quo".

If (and only if):

• the "status quo" version includes the information, AND
• there really is no consensus about whether to include or exclude it,

then the first policy says to remove it and the second says to retain it. When both of those conditions are true, it is not possible to comply with both policies. You have to pick one or the other. WhatamIdoing (talk) 19:13, 24 January 2024 (UTC)

The problem is, we can’t say whether to have less or more information without knowing what kind of information we are talking about. In the abstract, I think most of us would say “less” when it comes to trivia, but “more” when it comes to substantive facts. However, editors might well disagree on whether a specific bit of material actually is trivia or substantive. All we can do is leave the question open and call in more people to help us decide. Blueboar (talk) 19:31, 24 January 2024 (UTC)

Mach61, as noted in the above discussion, WP:ONUS is the main conflict with WP:NOCON, and says While information must be verifiable for inclusion in an article, not all verifiable information must be included. Consensus may determine that certain information does not improve an article. Such information should be omitted or presented instead in a different article. The responsibility for achieving consensus for inclusion is on those seeking to include disputed content. It is saying that being verifiable is a necessary but insufficient condition for inclusion and consensus is needed to include. So a lack of consensus would be to exclude This is slightly different than WP:BURDEN, which also says you need consensus to include but here it is consensus that the text is verifiable, not consensus that it meets all sorts of other requirements like WP:WEIGHT or WP:NOTNEWS.

These two policies conflict but they are operating on different axes. One is inclusion vs exclusion. The other is old vs new. I think there are examples where picking one or the other as a default will please some people and disappoint some people, will make some things easier or make some things harder. I think WAID is hoping you can give an opinion from experience rather than citing policy. -- Colin°^Talk 09:32, 25 January 2024 (UTC)

Subpages note

Just a note to express reservations about putting high-stakes RfCs on subpages. When a discussion happens on VPP, for example, it comes to editors' attention not just at the single edit of announcing the RfC, but with every subsequent edit that boosts its visibility to the top of 3,700 watchlists. I don't think many of us regularly click on VPP and scroll through the whole page to look for announcements. We scan recent edits/discussions as time allows, and so recent edits to the section in question are more easily noticeable. If putting it on a subpage, I'd encourage whomever does so to take great pains to advertise (perhaps even multiple times). In this case, it doesn't look like it was mentioned at VPP or WT:CONSENSUS. More than two weeks into a big question about a core policy and only 18 users have participated, and the page has fewer than 30 watchers. FWIW. — Rhododendrites ^talk \\ 17:22, 24 January 2024 (UTC)

Feel free to post links to this RFC anywhere you'd like (as long as it complies with Wikipedia:Canvassing, of course). WhatamIdoing (talk) 19:05, 24 January 2024 (UTC)

In case anyone wonders, in addition to being listed at Wikipedia:Requests for comment/Wikipedia policies and guidelines, this page is linked to at:

• Wikipedia:Village pump (miscellaneous)/Archive 76
• Wikipedia talk:Verifiability
• Wikipedia talk:Requests for comment
• Wikipedia talk:Featured articles

(plus one editor's User_talk: page). I'm not aware of any other efforts to advertise it (e.g., on Wikipedia:Discord). WhatamIdoing (talk) 06:12, 11 February 2024 (UTC)