Wikipedia talk:Verifiability

Onus

I wonder if the "Verifiability does not guarantee inclusion" needs to be clarified. When it says "The responsibility for achieving consensus for inclusion is on those seeking to include disputed content," I presume this is referring to adding disputed content. I have recently seen an editor remove sourced content and then appealing to WP:ONUS to keep it out. So I propose we clarify by replacing "include" with "add" so that it reads The responsibility for achieving consensus for inclusion is on those seeking to add disputed content. StAnselm (talk) 19:17, 18 September 2023 (UTC)

(ec) I have read the two most recent discussions on WP:ONUS (Wikipedia talk:Verifiability/Archive 68#WP:ONUS vs. WP:QUO and Wikipedia talk:Verifiability/Archive 69#RfC: WP:ONUS) and I think this is by far the most elegant solution to a long-standing problem. StAnselm (talk) 19:28, 18 September 2023 (UTC)

"Verifiability does not guarantee inclusion" is a core content policy. I don't agree that there is an implicit consensus for long-standing content that has not been specifically discussed. ONUS is fine as is. It does not need to be modified so you can win content disputes. Hemiauchenia (talk) 19:25, 18 September 2023 (UTC)

You may not agree, but that is, in fact, what WP:EDITCONSENSUS says: "Wikipedia consensus usually occurs implicitly. An edit has presumed consensus until it is disputed or reverted." StAnselm (talk) 19:31, 18 September 2023 (UTC)

Therefore, because the edit has been disputed there is no consensus, presumed or otherwise. Polygnotus (talk) 19:33, 18 September 2023 (UTC)

Ditto. This is effectively WP:Wikilawyering, if you actually want to resolve this content dispute, go to like NPOVN or something. Hemiauchenia (talk) 19:38, 18 September 2023 (UTC)

Yes it has been controversial, but those discussions have all ended with the current wording staying in place. So they're not a reason to make a change now. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 21:12, 18 September 2023 (UTC)

What you want is not a clarification, but a change. The people removing content always have an advantage over those adding it, which is a good thing because otherwise anyone can write anything and you'd need to spend a lot of time justifying removing it. This is why the rule was made that you have to justify including something; not the other way around. Polygnotus (talk) 19:26, 18 September 2023 (UTC)

Agree with this, it's not a clarification, but a fundamental change in the policy, and one IMO that would ultimately be detrimental to Wikipedia, and Wikipedia talk:Verifiability/Archive 69#RfC: WP:ONUS, was overwhelmingly opposed to effectively the same proposal. Hemiauchenia (talk) 19:32, 18 September 2023 (UTC)

I also agree. It sounds to me like it would create a barrier to editors who want to object to content, and a loophole helping editors who want to add disputed content. --Tryptofish (talk) 19:35, 18 September 2023 (UTC)

• No - The ONUS is on those who wish to add or keep disputed content. There are many reasons why sourced content might be questioned and challenged (perhaps it violates one of our other policies or guidelines, or perhaps it is trivial to the topic of the article, or perhaps it is better presented in some other article, etc). So… Those who wish to keep it must explain why the disputed content should be retained in the article. Remember, no one “owns” an article, and so others are free to come along and (partially or completely) rewrite it (and that rewrite may well omit things that were included in the previous version). Blueboar (talk) 20:42, 18 September 2023 (UTC)
• Regardless of the outcome, during discussion a "best practice" is to tag the disputed text and leave it in place. WP:QUO. Butwhatdoiknow (talk) 00:53, 19 September 2023 (UTC)

Nope; until there is a consensus it is best to get rid of it. We don't have a deadline, so leaving bad stuff in is worse than taking potentially good but disputed stuff out. It can return once there is consensus. WP:QUO is just an essay. Polygnotus (talk) 02:46, 19 September 2023 (UTC)

Yes, QUO is "just an essay." And your comment is the just opinion of one editor. Butwhatdoiknow (talk) 15:52, 19 September 2023 (UTC)

Turning to the substance of your remark, how do you know that the stuff is "bad" as opposed to "potentially bad but disputed" until a discussion takes place? Also, under your approach, how do you prevent edit warring during discussion? - Butwhatdoiknow (talk) 15:59, 19 September 2023 (UTC)

We had long conversations about this which died from exhaustion (or me dropping the ball and not creating a proposal as I said I would) IMO it conflicts with wp:Consensus, is totally out of place (what the heck does it have to do with wp:verifiability?) and probably is only there/ survived to emphasize the "does not guarantee inclusion" that it is coupled with, and doing a poor job at that. IMO we could fix all of these problems (plus a few more) by replacing "does not guarantee" and "onus" with "Verifiability is a requirement for inclusion, not a reason for inclusion" By slightly tipping the balance in two different opposite directions (removing verifiability as a way to coerce inclusion, and removing the wording favoring exclusion) the change would be pretty neutral on the inclusion/exclusion balance, and what gets in would be based more on its merits. Sincerely, North8000 (talk) 03:04, 19 September 2023 (UTC)

That is a solution in search of a problem. Polygnotus (talk) 05:45, 19 September 2023 (UTC)

I think this might be a good idea. I think that ONUS isn't a blank check to remove whatever you want, because there is WP:PRESERVE. However I think ONUS was never meant for removing content that is well-sourced, but adding new content. That doesn't mean content can't be removed, but removing also needs a reason, not just I don't like it/ONUS. Andre🚐 05:29, 19 September 2023 (UTC)

ONUS is not just for newly added content. Any content that is disputed should be re-added only when there is consensus to do so. That is a very low bar and easy to achieve when the content is a good fit for inclusion of course. No one is using ONUS to just randomly remove content without any reason. Polygnotus (talk) 05:44, 19 September 2023 (UTC)

Yes they do! I don't know in what areas you mostly edit Wikipedia but I have encountered this type of reasoning a lot (removing sourced content for various reasons and then arguing that the onus is on those who want to restore it to establish consensus). I would support North8000's rewording. Alaexis_¿question? 06:48, 19 September 2023 (UTC)

@Alaexis: You said "Yes they do!" in response to my "No one is using ONUS to just randomly remove content without any reason" but then you wrote "for various reasons". Which is it? If it is great content that should definitely be included; uncontroversial, neutral, balanced, well-written, non-promotional et cetera then why would someone remove it? There must be a reason. And if there isn't, then establishing consensus takes perhaps 20 seconds. Sounds like there was a reason, but you didn't agree with it, which means the content was disputed, which means its better out than in until there is consensus. Wikipedia would be a giant mess if you could just add whatever, but had to get consensus to remove stuff. Polygnotus (talk) 07:01, 19 September 2023 (UTC)

"No one is using ONUS to just randomly remove content without any reason" – I suppose the truth of this statement depends on whether you mean true randomness, and whether "I personally believe that Wikipedia is improved by containing less information" is a reason. WhatamIdoing (talk) 04:16, 20 September 2023 (UTC)

I always thought the tension between ONUS and PRESERVE was a good thing, depending on the situation either is correct. These aren't laws, they don't have to internally consistent or entirely sensical (although laws aren't always those things).
Bad content should be removed, and any editor wanting to restore it should find consensus to do so. Good content that is flawed should be tagged and discussed before it is removed. What's the difference between good and bad content? Well that's in the eyes of editors.
Ultimately these policies are ideas that have consenus, and how they apply comes out of consesnus as well. If you just have two editors stating ONUS and PRESERVE use WP:3PO to see which one is right. If you disagree with the group consensus of which applies, well move on or take it to a noticeboard. If you can't convince the community you're right, then maybe it's time to accept your wrong.
There is no problem to be solved here, how or when certain policies applies is as defined by consesnus as the policy was defined by consesnus in the first place. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 10:24, 19 September 2023 (UTC)

Please remember that the section immediately after WP:PRESERVE is WP:DON'T PRESERVE. This will shed some light on when ONUS applies. Blueboar (talk) 10:51, 19 September 2023 (UTC)

I've entered a 16 day period with a mix of few or none wiki minutes including 9 days off the grid. I commented here because the thread came up but can't do much more such as a thorough response to the responses to my comment. Wikipedia 95% works and the other 5% is a eternal painful mess that hurts us in many ways. And IMHO policies and guidelines are the cause of and solution for most of the 5%. The wililawyer warrior's heaven of the urban legends that stem from wp:ver (that verifiability is a way to coerce inclusion) that the current section made a failed attempt to address and the wililawyer warrior's heaven of wp:onus are two of the most easily fixed areas regarding this. Folks can always cite the 95% as a reason to not fix the 5%. Sincerely North8000 (talk) 12:12, 19 September 2023 (UTC)

But folks will also always find 5% they disagree with, even if 95% of the time it's fine. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 12:19, 19 September 2023 (UTC)

@ActivelyDisinterested: I don't think "fine" is the right word in this context. Polygnotus (talk) 12:27, 19 September 2023 (UTC)

My point is that that point is what editors disagree upon. So no, other editors won't think fine is the right word. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 12:32, 19 September 2023 (UTC)

@ActivelyDisinterested: Yeah I was being silly, the word "fine" is roughly 5% of that sentence. So I was proving your point by disagreeing with you. Polygnotus (talk) 12:33, 19 September 2023 (UTC)

Well I have to disagr.... ok enough -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 12:42, 19 September 2023 (UTC)

I didn't say there was 5% content which people disagree with. I said that the other 5% is an eternal painful mess. As in filling the drama boards and arbcom with unsolvable problems, unnecessarily losing good editors etc. North8000 (talk) 14:37, 19 September 2023 (UTC)

And I didn't mean content, but that editors don't agree what 5% is the problem. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 16:55, 19 September 2023 (UTC)

Please keep in mind that no one knows what ONUS means. - Butwhatdoiknow (talk) 16:18, 19 September 2023 (UTC)

I agree with the point argued by ActivelyDisinterested above, and I think I disagree with Polygnotus. "ONUS abuse" is a real thing, though it should be discouraged. Andre🚐 17:18, 19 September 2023 (UTC)

Surely I could edit this page with my proposal above and the onus is on those who favor the current wording to achieve consensus for their preferred version... StAnselm (talk) 19:06, 19 September 2023 (UTC)

Except… WP:V does not apply to policy pages, and ONUS is part of the WP:V policy… so, no, you can’t apply ONUS to itself. On policy pages (especially CORE policy pages), we require consensus for ANY change - regardless of whether it is an addition, subtraction, or rewriting of text. Blueboar (talk) 19:23, 19 September 2023 (UTC)

Ah, you've got me there. StAnselm (talk) 19:32, 19 September 2023 (UTC)

Yes, but you need a rationale. A revert citing simply "no consensus" is not valid or at least highly irregular and discouraged. I believe there is a sentence somewhere in policy that specifically says you cannot just revert an otherwise valid bold improvement or addition and say "get consensus for your edit first" absent any other reason. Andre🚐 19:32, 19 September 2023 (UTC)

If there is it could, by its own wording, just be changed. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 20:01, 19 September 2023 (UTC)

@Andrevan, I'm not sure that's actually in a policy, though the principle appears in other forms, e.g., as examples of WP:OWN behavior or in Wikipedia:Tendentious editing#Deleting the pertinent cited additions of others. WhatamIdoing (talk) 04:19, 20 September 2023 (UTC)

Blueboar, I'm not sure exactly what you're trying to say but, to be clear, preapproval not required for changes to policy per wp:PGBOLD. See also wp:DRNC. Butwhatdoiknow (talk) 06:00, 20 September 2023 (UTC)

@StAnselm, if "Verifiability does not guarantee inclusion" feels confusing, do you think it would make more sense to say that "Verifiability is a necessary but not sufficient condition for including material in an article"? They're meant to mean the same thing. WhatamIdoing (talk) 04:21, 20 September 2023 (UTC)

That would be en excellent improvement. You should make a specific proposal. North8000 (talk) 12:12, 20 September 2023 (UTC)

@North8000: Why do you think that would be an improvement? The reason for the subtle change in the wording (which does change the meaning no matter what WhatamIdoing says) is clearly not to make understanding the text easier (despite the claim to the contrary). It seems like the goal of the change to make wikilawyering easier for people who don't want to justify the stuff they've written... Polygnotus (talk) 12:19, 20 September 2023 (UTC)

I'd also be opposed to linking a mathematical concept. Policies aren't logic puzzles that need to be solved. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 12:52, 20 September 2023 (UTC)

Because it's a half way version (1/2 is better than nothing) of the ideal fix which would be to change it all to "Verifiability is a requirement for inclusion, not a reason for inclusion" . I could write a book on all of the problems that that would fix and things that it would improve but unfortunately have very few wikiminutes until Oct 6 and will be off the grid for much of that period. North8000 (talk) 16:55, 20 September 2023 (UTC)

And if the requirement was removed I could write a thousand books on the problems it would cause. Honestly the problem here seems to be editors potentially misusing ONUS. The solution to that is not to water down to nothing an important statement. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 19:13, 20 September 2023 (UTC)

Fair, but what is the way to ensure that ONUS serves the purpose. ONUS is similar to the "burden of proof." Sometimes, your burden is one hand, but there are exceptions. For example, I write an article about a politician and obtain consensus to add something about a prominent scandal in the lead section. Several years later someone comes along and removes that material saying "UNDUE". I restore it, citing that it's verifiable, referenced, relevant, and no reason was offered to remove it according to policy. They remove it again and say "ONUS is on you to obtain consensus to restore it." A discussion on the talk page is joined by 2 other users and nobody emerges to defend the content. We're now memory-holing the content even if there was a well-attended RFC with hundreds of users that came to a strong decision 3 years ago. We need to play the long game and remember that AGF doesn't mean ignore it when the system exhibits a failure mode. Andre🚐 22:38, 20 September 2023 (UTC)

Then you restore it citing PRESERVE, in the discussion you bring up the RFC and post neutral notifications to any relevant boards or projects. If after all that that there is still consensus to remove the content, well consensus can change. Not accepting that consensus can change starts to sound like stonewalling.

As my comment said these issues should be worked out by discussion and consenus, not by removing a policy so that it's easier to win an argument. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 11:23, 21 September 2023 (UTC)

I don't think anyone is seriously proposing to "remove a policy". I think there is some justification in having the onus blurb in a different location (and people are suggesting that), as you say it is really about consensus and discussion (I would add editing, too), not verifiability per se. Selfstudier (talk) 11:31, 21 September 2023 (UTC)

If it's moved to consensus, then a link to WP:CONSENSUS would still be appropriate. If not then ONUS would no longer exist in its current state. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 11:45, 21 September 2023 (UTC)

I just think that the system needs stopgaps to protect good faith contributors from failure modes that are asynchronous. I don't think it's a given that protecting a past consensus against drive-by changes is always stonewalling. My point is that ONUS isn't a blank check, which I think we agree on. To the extent that it affects verifiability: I guess my point is that really all edits need consensus and all edits need proper sourcing and evidence. ONUS says the responsibility is on someone adding content. Why isn't there an equal ONUS on someone removing material (assuming it's otherwise verifiable and policy-abiding)? Why isn't it that a consensus is needed to deviate from the status quo if that status quo has suitable precedent and consensus on its side (not talking about cases where it doesn't) Andre🚐 15:04, 21 September 2023 (UTC)

Well I can answer the question: "Why isn't there an equal ONUS on someone removing material". It is because people post an neverending tsunami of crap on Wikipedia, and having to get consensus for removing even just 10% of it makes keeping Wikipedia usable impossible for those who fix the problems caused by others. Polygnotus (talk) 15:09, 21 September 2023 (UTC)

I'm talking about removing long-standing content which already had a consensus for its inclusion, ie the status quo or the before state of a discussion/RFC Andre🚐 15:09, 21 September 2023 (UTC)

To me its weird to see people defend the status quo; as if everything on Wikipedia is perfect. Spend an hour doing recent changes patrol and you'll hate change. Spend an hour reading existing articles (excluding GA/FA/botgenerated stubs) and doing some factchecking and you'll hate the status quo. Depends on the topic of course; mathematicians are usually excellent Wikipedians. The "consensus" because no one fixed a problem yet comes often from obscurity or our lack of people willing to fix the problems caused by others. Writing stuff is far more "glamorous" than reading stuff other people wrote and comparing the statements in the article to the sources, trying to fix the typos and removing the cruft/puffery/selfpromo/bias/pov/misinformation. I doubt there are people who remove long-standing nonbiased npov reliably sourced perfect-in-every-way content and use ONUS to do it without giving a reason. And I am not sure why people who arent povpushers would pretend such a problem exists. But, I am willing to learn. Polygnotus (talk) 15:17, 21 September 2023 (UTC)

You can see an example of a recent discussion where content was removed and ONUS was cited, and there was a disagreement on the interpretation of sources, see here, while I cited PRESERVE. This does happen periodically. It is definitely the case that there are differences of opinion on some content's relevance, and that some people try to remove certain material citing UNDUE or claims like "this section is overly detailed, too long and needs to be trimmed" etc (not quoting anyone specific). Andre🚐 17:13, 21 September 2023 (UTC)

And again changing policy because you don't like it use in a discussion, is not a good reason to change policy. There was health discussion there and consensus resolved the issue. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 19:41, 21 September 2023 (UTC)

I'm not really saying the policy needs to be changed, but if it is changed, it should be changed to strengthen the status quo and not strengthen the use of ONUS to challenge edits for no otherwise good reason, that is my humble and not at all strong opinion. Andre🚐 20:17, 21 September 2023 (UTC)

But this example is about Trump. Of course its gonna be a battleground. ONUS did nothing to make that situation worse or better. "challenge edits for no otherwise good reason" It was a Trump supporter they will challenge any edit that makes Trump look bad whether its true or not. This is a problem with the fragile American democracy and public school system. It has nothing to do with ONUS. If it was the other way around and you used ONUS to remove content that praised Trump you would like ONUS. ONUS does not give you the right to challenge any edits without a reason. Good or not is in the eye of the beholder of course. Polygnotus (talk) 20:22, 21 September 2023 (UTC)

Without commenting on the views of any particular editors, who at least in theory are supposed to not let their opinions affect editorial decisions, I generally don't think ONUS should be a reason on its own to remove content, even that content was something I didn't agree with or like. I think removing content itself should have a good rationale. And I don't see why that shouldn't be an explicit guideline. PRESERVE, sort of implies this. Andre🚐 20:29, 21 September 2023 (UTC)

You are making the same mistake republicans make. You create a reality in which ONUS is "a reason on its own to remove content" (by repeatedly claiming that it is, or is used as such), even though it is not, and then you complain about something that does not exist. "I think removing content itself should have a good rationale. And I don't see why that shouldn't be an explicit guideline." We call removing content for shits and giggles vandalism, and people who do it get blocked. Wikipedia:Vandalism : "Removing encyclopedic content without any reason, or replacing such content with nonsense" is a type of vandalism. This is policy, a step above a guideline. Polygnotus (talk) 20:33, 21 September 2023 (UTC)

If User:Example is randomly deleting stuff, and gives no rationale whatsoever except perhaps "WP:ONUS" please report them on WP:ANI and they will be dealt with. Polygnotus (talk) 20:43, 21 September 2023 (UTC)

Haha, alright. Point taken. Andre🚐 21:15, 21 September 2023 (UTC)

I think we should distinguish between content that was added a while back and never challenged vs content that was previously challenged. I view that as the difference between implicit consensus and explicit consensus. When anything is added and both not challenged and not removed it has implicit consensus. If it's challenged (or explicitly supported by others) and consensus is established for the content to remain then, in my view, ONUS has been addressed and we now need consensus to change/remove. So I think Andrevan and I are in agreement here. Springee (talk) 16:09, 21 September 2023 (UTC)

Wow Springee, glad to hear we agree, I feel like that doesn't happen too often but I always appreciate it when it does. :-) Andre🚐 17:14, 21 September 2023 (UTC)

• Something else I would add is that text that was immediately challenged and not removed (yet no clear consensus was reached to keep it) ought to lack implicit consensus. This is perhaps slightly counterintuitive (it was examined and somehow managed to stay, after all) but it's important because otherwise we end up encouraging people to edit-war to "prevent" something from gaining implicit consensus. If an editor can say "I object to that, but don't have the time to hash it out now" then that encourages article stability without leading to situations where contested text gains a false sense of consensus just because one side in a dispute became exhausted or gave up. I think WP:SILENCE talks about this - once someone has objected, going silent doesn't necessarily mean their objection is gone; in the absent of a clear-cut consensus the other way, it just puts the dispute on hold. --Aquillion (talk) 20:15, 16 October 2023 (UTC)

  I think that is reasonable. Springee (talk) 01:38, 17 October 2023 (UTC)

ONUS - a different idea

• I think the problem with ONUS is really just the final sentence: “The responsibility for achieving consensus for inclusion is on those seeking to include disputed content”.

I happen to agree with what it says, but I think the WP:V policy is the wrong venue in which to say it. A statement on who is responsible for achieving consensus belongs in the WP:Consensus policy.

The rest of the paragraph (the part I associate with the shortcut VNOT) does fit in the WP:V policy. It is a useful reminder that verifiability isn’t enough to guarantee inclusion.

So… I would propose that, for now, we simply move the ONUS sentence to WP:CONSENSUS … and then examine how best to say it (or even whether we should say it) within the context of that policy. Blueboar (talk) 12:54, 20 September 2023 (UTC)

Bad idea. It is in the place where it belongs. Polygnotus (talk) 12:55, 20 September 2023 (UTC)

Why? Blueboar (talk) 12:56, 20 September 2023 (UTC)

The reason it is where it is, is because people defended the stuff they wrote by saying it is verifiable. They used WP:V as a magic shield against all criticism. The sentence has little to do with consensus. People who read the verifiability policy need to be told that just because something is verifiable, it does not mean it should therefore be on Wikipedia. What you see as the problem with ONUS is not a problem. It is the solution to a problem. Polygnotus (talk) 12:58, 20 September 2023 (UTC)

need to be told that just because something is verifiable, it does not mean it should therefore be on Wikipedia Then we should just tell them that and not dress it up as something else.Selfstudier (talk) 13:24, 20 September 2023 (UTC)

@Selfstudier: What do you mean? Polygnotus (talk) 13:25, 20 September 2023 (UTC)

What I just said. And there is no need to ping me. Selfstudier (talk) 13:26, 20 September 2023 (UTC)

Who or what is dressing what up as something else? Polygnotus (talk) 13:33, 20 September 2023 (UTC)

The bit I don't understand is: "Then we should just tell them that and not dress it up as something else" Polygnotus (talk) 13:36, 20 September 2023 (UTC)

Replace the current sentence with (something like) what you said, is that clearer? Selfstudier (talk) 13:37, 20 September 2023 (UTC)

We can add "Just because something is verifiable, it does not mean it should therefore be on Wikipedia" to that policy, but it cannot replace the current sentence: "The responsibility for achieving consensus for inclusion is on those seeking to include disputed content" because those sentences have a different meaning. Like I explained above, the sentence that is currently there is a solution to a big problem (people using WP:V (and verifiability in general) as a magic shield against all criticism). Polygnotus (talk) 13:41, 20 September 2023 (UTC)

Part of the existing sentence is in the wrong place, I agree with Blueboar on that. Selfstudier (talk) 13:42, 20 September 2023 (UTC)

Why? Which part? Polygnotus (talk) 13:44, 20 September 2023 (UTC)

The part that I said to replace with your sentence. Selfstudier (talk) 13:45, 20 September 2023 (UTC)

But why? What I wrote is easier to understand, but it also means something completely different. The policies on Wikipedia were forged in fire and written in blood. A lot of time and effort has been spent to get to the point we are now. Polygnotus (talk) 13:47, 20 September 2023 (UTC)

Take a look back at how that sentence got there. You say the problem is that people useV as a shield, so I agree with that, just tell people that they can't do that. Then editing and consensus are something else nothing to do with V. Selfstudier (talk) 13:50, 20 September 2023 (UTC)

It does not make sense to rip that last sentence out of its context. Read the lines previous to it. They are not complete without the last line, and the last line on its own lacks context. Polygnotus (talk) 13:59, 20 September 2023 (UTC)

I am talking about the principle, not the precise wording. As an example of possible wording, another editor at one point suggested "Replace that entire 3 sentence paragraph with "Verifiability is a requirement for inclusion, not a reason for inclusion"". Selfstudier (talk) 14:11, 20 September 2023 (UTC)

It already says: "While information must be verifiable for inclusion in an article, not all verifiable information must be included". And the reason that is not going to happen is that people will use WP:V as a magic shield again. It is really hard to improve something that is so battletested as a Wikipedia policy. Polygnotus (talk) 14:14, 20 September 2023 (UTC)

Well, I have said my piece. Just so we are clear, I agree with Blueboar (and I suspect, some others) and I don't think your objection is valid. Selfstudier (talk) 14:16, 20 September 2023 (UTC)

I do not think ONUS should be moved to CONSENSUS because I think that strengthens ONUS' use as a catchall to say for any given sentence in the article, that it must demonstrate an affirmative consensus. The status quo isn't discussed much in policy. I think the status quo should be even stronger than it is. As the project has grown and become more prominent, we have needed more stability and protection, whether NPP, stable/flagged revisions, patrolled, etc (see my essay WP:DAQ which predates many of the aforementioned technical features and social features of Wikipedia, sorry for the retro-self-plug). Anyway, I think moving ONUS to CONSENSUS makes it seem more like ONUS can just be invoked for IDONTLIKEIT or IDHT. Which IS a real problem, throughout many of the sensitive and contentious topic areas: people removing stuff that is verifiable. The point of ONUS is that verifiability doesn't guarantee inclusion: agreed there. But a BOLD removal of long-standing content that is reverted, for legitimate rationale and in active discussion on the talk page or RFC, defaults to the status quo. So in my view, policy should have more protection for the stable state that had obtained consensus. Consensus is hard to obtain at times in contentious areas, and it's an abuse of process to simply start a new discussion again and again until you get the result you want. Andre🚐 14:33, 20 September 2023 (UTC)

The point of ONUS is that verifiability doesn't guarantee inclusion - no. THAT is the point of VNOT.

I know because I was the person who originally wrote that first part of the paragraph. The ONUS sentence (about who should achieve consensus) was added by someone else… and originally appeared in a completely different part of the policy. Then a third person merged VNOT and ONUS (I assume they did so because both mention the word “consensus”). I have never been happy with that merger. Blueboar (talk) 16:45, 20 September 2023 (UTC)

I forget where it is, there is a diff somewhere showing that the current wording was added with a pretty low conlevel, in which case I find it fascinating that many are arguing quo when, by ONUS' own terms, disputed content should be removed :) Selfstudier (talk) 16:55, 20 September 2023 (UTC)

That's been tried and yet it still stands. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 19:38, 21 September 2023 (UTC)

Btw, the most recent discussion of all this (there have been many and there will continue to be until it is resolved, methinks) was here where we nearly but did not quite manage to agree on a draft RFC. Selfstudier (talk) 14:39, 20 September 2023 (UTC)

There's no reason that the principle expressed in ONUS cannot appear elsewhere in addition to the Verification page. SPECIFICO talk 14:52, 20 September 2023 (UTC)

There is some confusion regarding what principle ONUS expresses. See this discussion. Butwhatdoiknow (talk) 15:24, 20 September 2023 (UTC)

And there have been other discussions on the issue that have found no problem. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 16:20, 20 September 2023 (UTC)

Yes, there is confusion, even among experienced Admins who think it prohibits removal of content. SPECIFICO talk 16:41, 20 September 2023 (UTC)

I wouldn't be against adding the wording to CONSENSUS while maintaining it here. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 16:24, 20 September 2023 (UTC)

Homework

If we are interested in pursuing this _yet again_ : Somebody with programming skills could make a flow chart that identifies the permutations and real world paths in which ONUS is cited, where it works, where it breaks down. E.G.

is content verified? are there multiple high quality RS per WEIGHT? are they news sources? were they contemporaneous sources that are no longer current? Has the content recently been added to the article? Was the content previously challenged? Does the page have more than 50 watchers? Was the content discussed on the article talk page Was there explicit consensus on the talk page Has the content been in the article for 3 months? Has the article been edited by 20 different editors since the content was added?

I'm sure these are not all the right nodes and maybe not good statements of each fork, but we ought to be able to enumerate and state all the relevant factors and break down the areas of agreement and disagreement in operational terms that relate to the actual editing situations that arise. SPECIFICO talk 19:26, 21 September 2023 (UTC)

I don't think we should, the decision should be made page by page by discussion and consenus not here amongst a small subset of editors and imposed project wide. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 19:37, 21 September 2023 (UTC)

In the end it will be only the less than a handful of people who apparently dislike ONUS because someone who they disagreed with used it once (and cannot comprehend its value) who talk among themselves while everyone else gave up. Then they will declare that consensus has been achieved by boring everyone who doesn't care as much. Polygnotus (talk) 20:04, 21 September 2023 (UTC)

If any change were to be made, it'd have to be publicized in an RFC placed prominently on community village pumps, announcement noticeboards, signpost articles, etc etc to make sure it was not just a local consensus from a few wonks and nerds who like arcane stuff. Andre🚐 20:34, 21 September 2023 (UTC)

I think Specifico's point is well-taken that we should find out some examples of when ONUS has been invoked over the years, and what those circumstances were. I've seen ONUS invoked on at least 2 occasions that I thought it was standing in for any actual policy rationale. I'm not sure I've ever seen ONUS used when I thought it was shedding some meaningful light on the situation - does anyone have one? Andre🚐 20:33, 21 September 2023 (UTC)

Thank you, Andre. I think that all of these past discussions -- each of which may have failed in its own special way -- suffer from too abstract an approach that doesn't start from a common understanding of what problem, if any, needs to be solved. A flowchart would identify specific situations for which the language or placement of policy needs to be improved. We'd first need to workshop the nodes. I just threw down examples without much reflection. SPECIFICO talk 22:16, 21 September 2023 (UTC)

• Since someone asked for examples above, here (more discussion here) is another example of how the current wording of ONUS causes misinterpretations that lead to problems. A single editor objected to a sentence that had been longstanding and stable on an extremely high-traffic article for roughly eight years; they edit-warred to remove it repeatedly, constantly citing WP:ONUS (and little else) on talk. Even in the face of multiple editors saying it should stay, they seem to have concluded that it was not sufficient consensus, asserting that they would continue to come back and remove it because of ONUS. Obviously they misinterpreted multiple policies, but this sort of trainwreck is the result of a casual misreading of the final sentence of ONUS giving people the perception that they can just sort of cite it as an "I WIN" policy in any situation where they're trying to remove something from an article. This isn't what the section in question is or was ever intended to be, it doesn't lead to useful or constructive conversations, and it actively derails actual consensus-building by giving one side in disputes the false misapprehension that they have no responsibility to seek any sort of consensus themselves as long as they can cite ONUS over and over. The wording needs to be changed somehow; it reflects neither actual practice nor sound policy. The version of it that people who take the most expansive possible reading of it wish for does not and has never had consensus and the text needs to make that clear. --Aquillion (talk) 09:18, 16 October 2023 (UTC)

Remember that “consensus can change”. Some bit of material might have once had a consensus to include/exclude, and yet not have that consensus NOW. It is perfectly acceptable to question or challenge a previous consensus to see if it still exists. Blueboar (talk) 11:37, 16 October 2023 (UTC)

The post you respond to says the edit warring party was "constantly citing WP:ONUS (and little else)." It is not perfectly acceptable to challenge material with a substance-free "prove it to me" argument. - Butwhatdoiknow (talk) 15:47, 16 October 2023 (UTC)

I concur with Aquillion and Butwhatdoiknow. Andre🚐 15:50, 16 October 2023 (UTC)

If the problem is ONUS being used to edit war, then change it so it only applies after a discussion has run its course, not during the discussion. i.e., follow WP:QUO and WP:BRD until the discussion is done. Then there's no incentive to bring substance-free arguments or be disruptive. DFlhb (talk) 16:02, 16 October 2023 (UTC)

Works for me. Andre🚐 16:16, 16 October 2023 (UTC)

This is another case of an editor edit warring, but such behaviour isn't just an issue for ONUS. I would support mentioning WP:QUO and WP:BRD, but such statements apply for all edit warring. If the example above had been one editor trying to WP:PRESERVE content, and multiple other editors reverting them then the same discussions would apply. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 17:52, 16 October 2023 (UTC)

• But per WP:QUO (and common practice), after a discussion has run its course with no consensus, the longstanding version usually remains, outside of a few situations like WP:BLP or situations where someone can clearly show that a longstanding version lacked even implicit consensus. This has been one of the major disputes over ONUS. I don't disagree with your proposal as long as QUO applies as it always has, but this would completely defang the sweeping interpretation of ONUS that some people want, since it would mean, as a practical matters, that it usually only applies to new additions. Also, for genuinely new additions we usually do leave them out after an objection until things are resolved, although there's no hard-and-fast rule. --Aquillion (talk) 17:58, 16 October 2023 (UTC)

  ONUS shouldn't only apply to only new text, the flaw with that argument is that an editor could put unnoticed text into an article and the claim "silent" consensus later. My comment on PREVERSE wasn't in regard to new content, but about retaining content in opposition to multiple other editors.

  The issue here, and when editors edit war for the opposite, is that it's usually one editor trying to force a situation against many others. The solution would seem to be changing the edit warring wording to discuss such situations. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 18:09, 16 October 2023 (UTC)

  As I said, it is not only new text; the crux of implicit consensus is that the text has been seen (and often edited) by many people, so if they can demonstrate that the text was unnoticed (eg. because the article or that section is low-traffic) then they can show that it has no consensus. But they have a burden to do that, when removing longstanding text - once there's a clear dispute, someone who wants to remove well-established text does have some responsibility to show that it has never had consensus; or, if it has some degree of consensus, they must then demonstrate a consensus to remove. Both these things are to some degree actively occluded by ONUS, which makes people falsely believe that when arguing for removal they can push the entire burden of demonstrating consensus on people who want to retain text, and which falsely leads them to think that removals almost never require consensus. Clearly, they do; once there is a dispute over an edit, everyone must seek consensus - only in the most trivial of cases can you make a significant and clearly-disputed change to an article, including a removal, "by default" in the absent of a consensus. WP:QUO more or less covers this; if there's a dispute over well-established text, and people fail to reach a consensus, the default is the status quo, not removal. (This is, as an aside, part of what constantly baffles me about discussions concerning ONUS and well-established text. An RFC that fails to reach a consensus defaults to the status quo, never to removal. Every established editor knows this fact. The "default to removal" interpretation of ONUS that people argue is not and has never been practice. Hence my explanation for why DFlhb's proposed solution would effectively defang ONUS, because QUO instructs us to default to retaining disputed text.) --Aquillion (talk) 19:34, 16 October 2023 (UTC)

  QUO, by its text, specifically applies "during discussion." Maybe you're thinking of WP:NOCON. - Butwhatdoiknow (talk) 22:53, 16 October 2023 (UTC)

  Getting deja vu here :) Selfstudier (talk) 22:56, 16 October 2023 (UTC)

  Right; I worded that carefully. DFlhb (talk) 02:26, 17 October 2023 (UTC)

• Sure, but in that situation the burden is on the person who wants to make the change or removal to demonstrate that consensus has changed. WP:ONUS definitely doesn't allow you to just assert that a consensus may have changed and demand that other people show that it still exists or you're going to remove something. The heart of the disputes over ONUS are basically - when can someone argue that they have no obligation to demonstrate anything at all, pushing the ball entirely into someone else's court? And for CCC, it's not like that; the person who wants to remove something that previously had consensus has to demonstrate the change in consensus to remove it (or point to something like WP:BLP that shifts things to defaulting to removal, but that requires engaging on the merits and not just repeatedly saying ONUS.) I don't think non-engagement is something we should encourage at all outside of the most trivial circumstances for recent additions or ones that haven't had many eyes on them, but you definitely can't combine ONUS with CCC. --Aquillion (talk) 17:58, 16 October 2023 (UTC)

  Non-engagement is covered by WP:COMMUNICATE, WP:ONUS doesn't give an editor any right to not engage.

  Editors should not have to find consensus to remove obviously wrong content. The issue here, again and again, is edit warring. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 18:14, 16 October 2023 (UTC)

If there is no dispute over whether it is "obviously wrong", then naturally they can remove it. But if there is a dispute, then it is not obviously wrong. In that case they must go to talk, explain why they think it's wrong, and work to obtain consensus for their proposed change. They should not attempt major changes to the article (including removals) without that consensus. It is sometimes possible to demonstrate that the text in question never had consensus in the first place (especially for very new text) but they bear some of the responsibility to do so, especially for more longstanding text - they cannot simply presume that it lacks consensus, nor can they simply say "the ONUS is on you, I'm going to remove this until you show it has consensus." The presumption for longestanding text on high-traffic articles (that is to say, text that we can assume people have regularly seen) is that it has consensus. It is a rebuttable presumption but you must actually work to rebut it, you cannot simply cite ONUS over and over as some people attempt to do. --Aquillion (talk) 19:34, 16 October 2023 (UTC)

Longstanding isn't the point, that just means no one saw it to object to it. With so many articles text could stand for years, silent consensus only has any standing as long as noone objects to it then it has no standing whatsoever. Also minor amounts or major amounts doesn't change the issue, major amounts of trash should be removed as much as minor amounts. Editors shouldn't presume content has consensus.

No, editors shouldn't edit war or refuse to communicate. But your list of editors behaving poorly and using ONUS for the basis, is matched by a list of editors behaving poorly and using PRESERVE to maintain OR, synth, POV, etc.

If an editor removes content because they feel it's problematic they have to say why it's problematic, and ONUS isn't a recognised exception to 3RR. If the situation descends into reverts whether the content should be maintained or removed is something to work out on a case by case basis. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 19:47, 16 October 2023 (UTC)

Established text certainly does sometimes have a consensus, and sometimes does therefore require a clear consensus to remove; the exact threshold and whether or how it applies to specific text is complicated and requires case-by-case adjudication, I agree, but the current wording of ONUS fails to capture that nuance. With that said, would you object to the last sentence of your reply replacing the last sentence of ONUS, then? Something like If there is a dispute over whether an edit has consensus, that should be worked out on a case by case basis to replace The responsibility for achieving consensus for inclusion is on those seeking to include disputed content? That would address my concerns completely. Certainly that sentence of ONUS is not the only badly-written part of policy (I agree that PRESERVE is also sometimes misused, and would be happy to focus on any proposed fixes to it once we've replaced the problematic sentence in ONUS) but that's no reason not to fix it; as it is, it absolutely does encourage both edit-warring and failure to communicate, since it gives some editors the false belief that they can automatically assume that something lacks consensus and that they have no responsibility to engage and demonstrate that fact. If they have a responsibility to engage, and if a dispute resulting from that is best decided by determining consensus and by consensus-building on a case-by-case basis, which both parties have to participate in, then the current last sentence of ONUS doesn't actually mean anything and serves only to lead editors who rely on it astray - you seem to be agreeing that the ONUS in a disputed removal falls on both parties and not just one. --Aquillion (talk) 20:02, 16 October 2023 (UTC)

I wouldn't agree with that exact wording. That content requires consensus needs strong backing. The issue is that two editors can't resolve the PRESERVE / DONTPRESERVE debate, but only in the most contenious can multiple editors fail to resolve the issue.

Something like the following sentences could be added without removing the final sentence.

But whether the content is preserved or excluded until a consensus is formed should be determined on a case by case basis.

This policy is not an accepted exception to rules about edit warring.

This should not be used as a reason to remove content.

With appropriate wordsmithing. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 20:25, 16 October 2023 (UTC)

I'm generally ok with most of the directions this is going in. Most anything would help toward explaining ONUS better. As Aquillion has pointed out, ONUS is a description of how consensus works, and is not supposed to itself be a blank check. The point of ONUS is that "consensus is always required" which is to say that if someone BOLDly does something, and it is reverted, don't edit war: discuss, and reach consensus. How? That's case-dependent, but, don't just keep reverting or keep on trying to sneak your change back in. Discuss. That's the point of ONUS. The burden of someone making a change is to explain why, be that change a removal, or an addition, or a copyedit, if it's disputed by another editor, seek the discussion board to hash it out. In cases where there's no agreement, WP:QUO and WP:PRESERVE tell us that existing, policy-abiding, content, with no consensus to remove or delete it, is protected for inclusion if it had a historical consensus for being included. Historical consensus could have been a bold addition that was kept by many editors implicitly, or it could be a formal consensus found in a process. It doesn't have to always be the same thing. So I think we can explain this practice better. Andre🚐 22:03, 16 October 2023 (UTC)

I disagree completely with this (and with AD's 3rd sentence). 'If it's been there for a long enough time, it stays' is equivalent to deprecating ONUS.

edit: I should clarify: I agree completely with the first half, disagree with the second half. Happy to support suggestions that will get people to properly make their case though DFlhb (talk) 22:27, 16 October 2023 (UTC)

Well I respect your opinion and your edits (I think, forgive me, I sometimes confuse the editors with some characters like yours that don't appear phonetic), why do you disagree that ONUS itself isn't a valid reason to remove content? Isn't that just "IDONTLIKEIT" in another name? "This new content doesn't have a consensus. It should be removed." Invalid. Right? Andre🚐 23:48, 16 October 2023 (UTC)

Right, I don't see it as a valid reason to remove content; actual arguments need to be presented.

ONUS does two things. First it dispels the notion that silent consensus has weight, or, rephrased, that affirmative consensus is needed to remove something added with no consensus and never discussed (I agree with Springee 16:09, 21 September and replies beneath). That helps with stonewalling. When bad edit falls through the cracks, get scrutinised by no one, and when it's caught it's "too late", that's not in keeping with the spirit of collaborative editing. Accessorily, ONUS combined with VNOT is a good policy-based avenue to resist recentism, and the notion of "we can't remove it, it's sourced" which I've heard too many times in non-ECP articles. That scenario's not only about old content; people edit-war to reinstate hours-old content, and in that sense ONUS also helps support BRD.

None of the value I see in ONUS overlaps with "I don't need to make arguments" (I'd look down on that), so I'm sure there's a way to thread this. If Aquillion's diff was representative of the way I saw people use ONUS, I'd dislike it too.

BTW thanks :) DFlhb (talk) 01:51, 17 October 2023 (UTC)

Consider this scenario. 4 editors, all extended confirmed, edit war on an article. After the edit war has stopped simmering, editor A boldly adds material to the article. Editor B reverts it. Editor C reverts editor A and claims that their discussion on the talk page, supports the consensus to retain the content, as it is relevant and encyclopedic. Editor D leaves a comment ambiguous, but probably could be read supporting including the content. Editor B leaves a comment probably supports removing it. The conversation dies down and the content sits in the article for 2 years, let's say. Then a new editor comes along and says "ONUS." Editor B says "yes, ONUS. This content never had a consensus to begin with." Andre🚐 02:31, 17 October 2023 (UTC)

So editors A, C, and (kinda) D support inclusion, and B opposes. That's slim consensus, not overwhelming, but still a consensus; it wasn't "never discussed". DFlhb (talk) 08:03, 17 October 2023 (UTC)

To clarify my third point is that it's not in itself a reason to remove content, but it is a reason to revert the re-inclusion of content.

For example:

1). Content is removed for reason 'x'

2). Content is restored without resolving 'x'

3). Content is removed per ONUS -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 12:16, 17 October 2023 (UTC)

Makes sense then, I misread that - DFlhb (talk) 12:37, 17 October 2023 (UTC)

After reading your comment I realised the third sentence was ambiguous. Any suggestions on better wording would be welcomed. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 13:00, 17 October 2023 (UTC)

• Another question; above, people asked for examples of where the last sentence of ONUS led to or exacerbated conflicts. But can anyone point to cases where the last sentence, in particular, has been used successfully to resolve conflicts? Not the "verifiability does not guarantee inclusion" part, but cases where someone successfully cited the final sentence of ONUS to argue that someone else had the responsibility to demonstrate consensus and it, for lack of a better word, went well. I have personally never seen that happen - in my experience seeing someone citing that is usually an indication that discussions have already or are about to go to hell. --Aquillion (talk) 20:10, 16 October 2023 (UTC)

  @Springee cited ONUS recently in a way that was somewhat conciliatory and de-escalatory. Believe that one or not! :-) Andre🚐 22:06, 16 October 2023 (UTC)

  Clearly you are thinking of someone else ;) Springee (talk) 01:33, 17 October 2023 (UTC)

  It's thousands of uses within the summary of the third revert by the third different editor of a POV warrior in an uncountable number of articles. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 12:12, 17 October 2023 (UTC)

• Here is another example of an editor abusing WP:ONUS in a way that derails discussions. They WP:BOLDly removed a lead sentence that had been stable in the lead of a high-traffic article for three years, where the lead had been subject to extensive discussion and scrutiny; in the face of three other editors objecting (when they were, so far, the only editor who had advocated removing it), they nonetheless cited WP:ONUS to argue that the sentence in question needed to stay out of the article until, I guess, they were WP:SATISFYed or a formal RFC was held, since the existing objections were clearly not enough for them. They continued to do so even when the number of objections ticked up to five! This is the sort of thing that shows the problems caused by the current wording of ONUS - it leads to situations where individual editors mistakenly believe that they can make changes to longstanding, stable, high-traffic articles and then WP:STONEWALL them in place in defiance of WP:BRD and even basic consensus-building; and it encourages editors to be intransigent about recognizing that their changes lack consensus. --Aquillion (talk) 17:44, 25 October 2023 (UTC)

  That's the exact same guy. Please bring those concerns to the appropriate noticeboards. DFlhb (talk) 18:20, 25 October 2023 (UTC)

  @Aquillion, don't forget to link to Wikipedia:Administrators' noticeboard/Archive352#Block review: Willbb234 if you decide to follow up on this.

  More generally, "if I don't agree, it's not actually consensus" is a problem that can sometimes be solved through education about Wikipedia:How to lose, but it might be more frequently addressed with the block button. WhatamIdoing (talk) 20:42, 26 October 2023 (UTC)

• I don't appreciate this discussion taking place behind my back. I don't think it's very fair. Anyway, in response to the comments regarding the discussion at Talk:Transgender, I will summarise what happened:
  • I removed content for one reason
  • It was reinstated without the issue being resolved
  • I initiated discussion
  • When this stalled, I referred to ONUS as those wishing to include the content need to find consensus.
• This seems very much like the general course of action outlined by ActivelyDisinterested above and very much unlike what Aquillion claimed happened - constantly citing WP:ONUS (and little else) on talk. (I'm unsure where you got that impression). I never used ONUS as the sole justification for removing content. You also failed to mention that consensus ended up being in my favour, but I guess that doesn't meet with your objectives here. You're welcome to disagree with what I have to say, but please don't lie about me and my actions behind my back. Willbb234 23:44, 26 October 2023 (UTC)

  AFAICT, that second bullet point should read "It was reinstated without any other editors agreeing that the thing I dislike is actually a problem". WhatamIdoing (talk) 16:26, 27 October 2023 (UTC)

That's a shame we couldn't have had a constructive discussion here, or that you couldn't have assumed good faith, or that you couldn't have apologised about your actions. Then again, I didn't really expect much else. Willbb234 17:02, 27 October 2023 (UTC)

A new suggestion re ONUS

Would editors agree with the section being reworked so that The responsibility for achieving consensus for inclusion is on those seeking to include disputed content is maintained but:
a). Include a warning that this is not an exception to 3RR.
b). Whether the content is preserved or excluded until a consensus is formed should be determined on a case by case basis.
c). ONUS isn't an initial reason to remove content.
The wording on that last point is poor. As per my earlier example:
1). Content is removed for reason 'x'
2). Content is restored without resolving 'x'
3). Content is removed per ONUS
In the example 'x', the initial reason for removal, can't be ONUS.
This would all be open to proper wording. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 13:55, 17 October 2023 (UTC)

• Anything that might prevent the endless discussions about this, I would support. Selfstudier (talk) 13:59, 17 October 2023 (UTC)
• "The responsibility for achieving consensus for inclusion is on those seeking to include disputed content, though all editors are expected to engage constructively. DFlhb (talk) 14:21, 17 October 2023 (UTC)
• "Ultimately, the responsibility for achieving consensus for inclusion is on those seeking to include disputed content. For a best practice during discussion, see WP:QUO. - Butwhatdoiknow (talk) 14:50, 17 October 2023 (UTC)

A first attempt at a rewrite. Note I dropped the sentence Such information should be omitted or presented instead in a different article as it in part duplicates the "preserved or excluded" part added later.

While information must be verifiable for inclusion in an article, not all verifiable information must be included. The responsibility for achieving consensus for inclusion is on those seeking to include disputed content. Consensus may determine that certain information does not improve an article. Whether the content is preserved or excluded from the article while finding that consensus, should be handled on a case by case basis. Note this is not an allowed exception to the policy on edit warring.

I still have no good way to make the point about initial reasons for removal, which I think is important to solve the cases of ONUS abuse. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 15:55, 17 October 2023 (UTC)

I think this tries to address too many things and is too complicated. Any high-profile policy gets abused. The problem is battleground mindset; a civility reminder addresses that and makes it easier to sanction people if they misuse it. Keep in mind this'll probably go to RfC so the simpler the better. DFlhb (talk) 16:11, 17 October 2023 (UTC)

I'm generally supportive of the proposals. Andre🚐 21:45, 17 October 2023 (UTC)

I was trying to think of good phrasing for this but nothing came to me. In theory ONUS pauses the, if you will, consensus clock. So just as NOCON shouldn't be used as a justification for a first revert of newly added material, ONUS should be used as a justification for material that is otherwise part of the stable article. NOCON does reference QUO so that might be acceptable here though I think ONUS would, logically, favor remove first, restore only if consensus is met. While NOCON suggests returning to the stable version of the article absent consensus, ONUS suggests the opposite if the material has been around for a while (if it's new then returning to status quo gives the same result as ONUS suggests). However, for edit warring reasons going slow is always a good idea but since QUO and ONUS may be seen as conflicting I'm not sure it's a good reference. I was trying to think of a way to deal with the "ONUS as the justification" issue. Something like, "ONUS establishes who should seek consensus for a change but the justification for removal of content should be based on other aspects of VERIFY or other Wikipedia policies such as NPOV, BLP, etc." I'm trying to say "Don't edit war, have a reason for removal other than ONUS, don't be an ass." with more official sounding language. Springee (talk) 02:29, 18 October 2023 (UTC)

Could all this be solved by changing:
The responsibility for achieving consensus for inclusion is on those seeking to include disputed content
to
The responsibility for achieving consensus for inclusion is on those seeking to restore disputed content?

Maybe also linking "omitted" in Such information should be omitted or presented instead in a different article. to WP:DONTPRESERVE, which discusses reasons why content might be removed. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 13:15, 20 October 2023 (UTC)

No, that doesn't solve the underlying issue. To me, the core problem is exemplified in the misuse of ONUS I just linked above - an individual editor, with no reason to think they have consensus, should not feel that policy supports them in removing a longstanding sentence from the lead of a high-traffic, high-scrutiny article in the face of multiple editors objecting. That is a WP:BOLD edit and WP:BRD applies; it ought to be restored at the first objection, and remain in the article during discussion, because the lead of an article like that (where every sentence has had people going over it with a fine-toothed comb) is presumed to have consensus. These are the sorts of things that I see constantly - an individual editor makes an edit to a high-traffic lead that is the result of extensive discussion and consensus-building; because these things are complex, it's hard to cite a specific discussion for any individual sentence, so they cite WP:ONUS, and continue to hammer it until / unless an actual RFC is held, regardless of the level of objections. That interpretation of ONUS is unworkable because it doesn't align with the way that most consensus is formed. In fact, the issue above (where the editor continued to cite ONUS even after roughly five people argued for restoration and only they were arguing for removal), while extreme, is an inevitable result of that, because the subtext of that interpretation of ONUS is that only the most unambiguous forms of consensus-building actually "count". "It was longstanding text in the lead of a high-traffic article" is weak consensus; a quick nose-count in a discussion is weak consensus. But those weak consensuses still need to actually have force (enough force to overcome ONUS) or we end up in situations where any editor (this includes, like in the discussion above, well-meaning ones who simply think that they're right) who misinterprets ONUS the way some people above have can effectively cause any discussion where they're arguing for a removal to drag out tortuously long unless someone is willing to just bluntly tell them their interpretation of ONUS is wrong and ignore it. Once even a weak consensus is demonstrated, the ONUS shifts to the people arguing for removal to dispel that weak consensus, show that it has changed, or demonstrate a stronger one for removal; the current wording of ONUS (while it doesn't contradict this) has this one-sided vibe that makes people think they can ignore that. That's what needs to be corrected. --Aquillion (talk) 17:52, 25 October 2023 (UTC)

All this would suggest result in fringe/pov warriors being able to maintain their text, and endless discussions/RFCs that waster editors time to get rid of it. You've shown a few editors who are abusing ONUS, but seemingly ignore the many benefits it has. Only some kind of comprising is going to bridge the gap. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 18:33, 25 October 2023 (UTC)

So, I agree with Aquillion that this is a classic example of "ONUS abuse" and this is far from the only instance. There's value in ONUS as well. Maybe a constructive approach is to game out a few scenarios in explanatory essay form of good and bad usage of ONUS. Also, I think there's an aspect to ONUS that could be explored more. Sentences in Wikipedia are similar to logical premises, which is that they are arguments that rest on an evidentiary basis, which is similar to the burden of proof. We could explain that burden better, I think, and solve the ONUS-abuse problem as well. Andre🚐 20:25, 25 October 2023 (UTC)

I like that the best; it fights the misuse and "ball in the other court" problem, without enshrining "silent consensus" as having inherent weight, which poses the exact same "ball in the other court" problem in the opposite direction. highly-visible text in established high-traffic articles is too open to gaming, and wouldn't solve the misbehaviour problem DFlhb (talk) 21:36, 25 October 2023 (UTC)

Collapsed own comment

Reframing ONUS to only apply after a discussion has run its course, would largely solve this, and might be the most likely to pass proposal here. I've (repeatedly) seen ONUS misused with regard to new material too (where there is majority support for inclusion yet ONUS is asserted and debated), not just for older material; the issue's the same: contentious procedural arguments about what we should do with the content, distracting from substantive discussion about the content. In practice that means taking what WP:NOCON says about non-BLPs, flipping it around to suggest (not require) exclusion rather than inclusion, and then changing ONUS to match that new NOCON; since NOCON is focused on what happens after discussions end. I also noticed a big error in my earlier proposal on the NOCON talk page; it should have said add or remove, not add, modify, or remove. The wording will need to be subtle and robustly-discussed, because we don't want, for example, people blanking articles after "no consensus" AfD closes. DFlhb (talk) 09:28, 31 October 2023 (UTC)

Collapsed; that would actually weaken WP:QUO, since it would lead to new additions staying in during discussion, rather than kept out temporarily per WP:QUO. I increasingly think that since the problem is wikilawyering, misusing its wording without regard to its spirit, the solution would be to explain the spirit explicitly, or give examples of good/bad use in some explanatory page. DFlhb (talk) 18:08, 31 October 2023 (UTC)

I mentioned some of this in A New Suggestion b) and c). The other part is wikilawyering, and this isn't the only 'ALLUPPERCASESTATEMENT' that suffers from that. How we stop editors from trying to misusing ONUS as an "I win" trump card I don't know, but the same problems exist with edit warring the removal of content because "ISAYBLPRESTOREAPPLIES".
Maybe some statement that if other editors disagree you're probably wrong would help. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 20:24, 31 October 2023 (UTC)

• I mean, that's why I'm saying things like "in the lead of a high-traffic, high-scrutiny article." Certainly ONUS does not and has never protected removals like that, and someone who tries to cite it for such a removal is misreading it or misusing it. If something has been there for an extended period of time, in a place where it can be reasonably concluded that it's been seen and reviewed by many different editors, and you want to assert that it is nonetheless WP:FRINGE oder WP:POV, that's clearly something you have to demonstrate and get consensus for; after all, removals (especially of longstanding text that an article has been written or balanced around) can also lead to an article becoming WP:POV or tilting it towards undue emphasis on WP:FRINGE viewpoints; removal of longstanding, highly-visible text in established high-traffic articles is every bit as dangerous as additions and requires the same degree of scrutiny and consensus-building. (The longstanding, highly-visible text in established high-traffic articles bit there is the compromise you're asking for.) We determine these things by discussion and consensus-building, there's no silver-bullet policy; so ideal policies should encourage everyone to come to the table, with the spectrum of consensus left, to an extent, deliberately vague and hard limits on any policy that an editor could use to try and say "nah the ball is completely in your court." Definitely there's a spectrum of consensus; my problem is that many people read ONUS as not recognizing this at all. I'd be happy to have a discussion of what constitutes implicit consensus, but I'd like to see at least some feedback on the definitions I've given. --Aquillion (talk) 20:44, 25 October 2023 (UTC)
• Also, as far as compromises go, I think that that is my outer limit - a sentence saying that ONUS does not apply to longstanding, highly-visible text in established high-traffic articles, which are presumed to have consensus. That reflects both actual practice and our existing policies on consensus-building. --Aquillion (talk) 20:46, 25 October 2023 (UTC)

  Support that Andre🚐 21:30, 25 October 2023 (UTC)

About the "reason to remove content": Since the community doesn't have a 'thing' that they call WP:Reasons to remove content, then telling an editor that "____ is not a reason to remove content" (or the opposite, e.g., "verifiability is not a reason to include content") is not likely to be understood. You'd have to introduce that kind of idea with a lot of education. The shortest possible explanation that I can think of now is:

• There are many reasons to include material, and many reasons to exclude material. For example, material may be removed because you can't find reliable source that supports it after a reasonable effort, because it is non-neutral, or because it is inappropriate for a Wikipedia article. See WP:PRESERVE and WP:DONTPRESERVE for some other reasons. However, by themselves, "There's no consensus" or "per ONUS" are invalid reasons, even when they are true. Instead, give the reason for the lack of consensus to include the material, e.g., "There is no consensus to include this because editors felt it was misleading" or "Per the talk page discussion, this material belongs in [[other article]]".

I don't think it will help our friend with the 1RR sanction and the moderately long block log, but it might eventually nudge some editors towards clearer explanations. WhatamIdoing (talk) 20:52, 26 October 2023 (UTC)

A novel use of ONUS?

Hmmm. this? (Summary - A no consensus AfD entitles an editor to redirect/delete the page because by ONUS it has no consensus) Selfstudier (talk) 16:54, 31 October 2023 (UTC)

I've commented at the thread. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 17:42, 31 October 2023 (UTC)

ONUS gives revert ninjas the upper hand

WP:ONUS unfairly gives revert ninjas the upper hand in any content dispute because any Ninja editor can delete content (regardless if it's recently added) and say "go get a consensus for inclusion". Is this a good thing? I can't see how that would be the case, given that there must be a clear consensus to delete articles (see WP:NOCONSENSUS). It's beyond me why the same standard would not apply to any content within articles that passes WP:V. Where is Matt? (talk) 22:48, 8 December 2023 (UTC)

You are saying that as if its a bad thing. Actually, it is by design that those who protect the encyclopedia from nonsense have the upper hand over those who want to include nonsense. Read Brandolini's law. If you patrol recent changes for a while you will understand. Polygnotus (talk) 22:52, 8 December 2023 (UTC)

If it passes WP:V it wouldn't be nonsense. Where is Matt? (talk) 22:56, 8 December 2023 (UTC)

Lots of nonsense is verifiably true; its just not fit for inclusion in an encyclopedia. Polygnotus (talk) 23:07, 8 December 2023 (UTC)

If the disputed content is truly not encyclopedic ("nonsense" per your vocabulary), there would easily be consensus to exclude it from the article. Where is Matt? (talk) 23:11, 8 December 2023 (UTC)

So what you are proposing is that a povpusher or marketing/PR person could just add nonsense to Wikipedia (no consensus required), and the people who actually create and improve the encyclopedia have to get consensus to remove it. Do you see the problem? Polygnotus (talk) 23:15, 8 December 2023 (UTC)

Any content not in compliance with Wikipedia's policies (in your example, WP:NPOV) can be edited, up to, and including deletion. In your example, if content is opined to violate WP:NPOV, the onus would obviously be on the person who wants to add the content to show that it's not in violation of WP:NPOV.

The way WP:ONUS currently reads, revert ninjas can delete any content by citing WP:ONUS. Where is Matt? (talk) 23:21, 8 December 2023 (UTC)

What you call revert ninjas are the volunteers who spend hours of their finite life on earth to protect the encyclopedia many people worked so hard on against those who would damage it. Polygnotus (talk) 23:28, 8 December 2023 (UTC)

That is the exact opposite of what I see. What I see, is a revert ninja hitting the revert button (only takes a few seconds) to wipe away the hard work of the editor who goes through the effort of finding information and documenting it. Of course, not all information is encyclopedic, but the burden should be on the editor who wants to delete the information, just like the burden on deletion of entire articles. Where is Matt? (talk) 23:44, 8 December 2023 (UTC)

"If the disputed content is truly not encyclopedic ("nonsense" per your vocabulary), there would easily be consensus to exclude it" Now flip that around. If the content is so great, getting consensus to include it should be easy right? But it wasn't, because it isn't. Polygnotus (talk) 23:47, 8 December 2023 (UTC)

I think you're missing the point. The proposed change to WP:ONUS impacts whenever there is no consensus either way. If there is consensus one way or the other, WP:ONUS does not apply. — Preceding unsigned comment added by Where is Matt? (talk • contribs)

I am not. Polygnotus (talk) 00:04, 9 December 2023 (UTC)

If someone is citing ONUS, but removing something solely because it hasn't gained consensus without giving a legitimate reason (e.g., WP:UNDUE), then that can be viewed as disruptive editing and WP:GAMING. That is what separates typical editing with this policy vs. the scenario you're bringing up. Not that it doesn't happen, but this policy does not allow for that either. The other comments here a very on point about actual practice with the policy. KoA (talk) 23:43, 8 December 2023 (UTC)

There are plenty of pieces of information that are verifiable in reliable sources, but are otherwise nonsense or even outright false. Passing V is a very low bar to meet, which is why we have additional policies and guidelines like WP:BLP, WP:FRINGE, and WP:MEDRS. When an addition has been contested, it is right that consensus must be sought for its inclusion, not exclusion. Sideswipe9th (talk) 23:18, 8 December 2023 (UTC)

Verifiable and false are contradictory. If material appears in WP:RS it is presumed to be correct, unless contradicted by other WP:RS. Where is Matt? (talk) 23:26, 8 December 2023 (UTC)

That is not true. Polygnotus (talk) 23:30, 8 December 2023 (UTC)

Verifiable and false are contradictory. No, that is not correct. Both mis and disinformation are well known concepts, and reliable sources that ordinarily have a reputation for fact-checking have fallen prey and been known to publish both. Additionally, reliable sources are written by people, and people make mistakes. Sideswipe9th (talk) 23:56, 8 December 2023 (UTC)

You need to read both sentences. In the next sentence, I concede that WP:RS can make mistakes, but the presumption is that they don't. Where is Matt? (talk) 00:02, 9 December 2023 (UTC)

You can safely assume Sideswipe9th has read the entire comment Sideswipe9th replied to. Polygnotus (talk) 00:07, 9 December 2023 (UTC)

Lots of people think ONUS allows that, in practice it is not so easy to get away with only claiming ONUS. Selfstudier (talk) 22:53, 8 December 2023 (UTC)

I would change the following sentence:
From: The responsibility for achieving consensus for inclusion is on those seeking to include disputed content.
To: The responsibility for achieving consensus for removal of content that passes WP:V, and is not in violation of WP:BLP, is on those seeking to remove the disputed content.
Where is Matt? (talk) 22:55, 8 December 2023 (UTC)

Terrible idea. And its not going to happen. Polygnotus (talk) 22:56, 8 December 2023 (UTC)

A classic WP:IDONTLIKEIT. Where is Matt? (talk) 22:57, 8 December 2023 (UTC)

Except its not an argument. And not in a deletion discussion. Polygnotus (talk) 22:58, 8 December 2023 (UTC)

It's an argument against making the change, and here lies the point: what's good at deletion discussions should apply for any dispute. Afterall, a deletion discussion is a dispute whether a topic merits an article on Wikipedia. Where is Matt? (talk) 23:02, 8 December 2023 (UTC)

Yeah no. Sigh. Polygnotus (talk) 23:03, 8 December 2023 (UTC)

Where is Matt? is probably looking for Wikipedia:Arguments to avoid on discussion pages, which also addresses purely subjective arguments. But Polygnotus is correct that this being a "terrible idea" is not an argument. It's an observation, borne out by the WP:SNOW result that instantly formed, and all the similar discussions of the same idea before. This should probably just be listed at WP:PERENNIAL. Any time a user who has been here for only 4 months thinks policy is broken and that they have the fix for it, they are making a mistake.  — SMcCandlish ☏ ¢ 😼  07:47, 9 December 2023 (UTC)

ONUS, part 7 incoming :) Selfstudier (talk) 22:57, 8 December 2023 (UTC)

A terrible idea for all the reason stated in the dozens of previous discussions. If someone is editing disruptively take them to ANI, don't try to change policy so you win. -- LCU ActivelyDisinterested «@» °∆t° 23:38, 8 December 2023 (UTC)

Oppose. This is self-serving attempt to change a policy in a way to make poor editing easier, and would do incalculable harm to the project. The request is based on no evidence of any problem to address at all (and much to the contrary; see below). If Where is Matt? were actually meeting with unreasoning, PoV-laden resistance when he tried to add neutral and demonstrably factual material in an article or category of articles, with impeccable sources for his edits, but no valid rationales for undoing them, then this would be a behavioral matter to address at a noticeboard.

But I seriously doubt this is the case anyway. In looking over several weeks of edits by this editor that have been reverted, I see: unsourced and probably WP:OR claims; attempts to WP:COATRACK an entire section of material into one article that is already covered at another (already linked at the first) and which does not provably relate strongly to the first article anyway; lots of unhelpful and long-winded rewriting of headings; red-linking things that are never likely to become articles; incorrect link-piping; circular linking back to the same article the link is in; mis-capitalization of descriptive phrases as if they are proper names; adding unnecessary subheadings above very small bits of content; really insistent and WP:IDHT-leaning attempts to change various essays (to make them longer but no clearer, and often misapplicable in various ways after the meaning changes); terrible "make it longwinded for no reason" writing, like changing "swiftly" to "in a swift manner"; improper changing of disambiguation pages to move the WP:PRIMARYTOPIC meaning out of the lead and into the list of other meanings; experimental changes to markup in a page being saved as tests, instead of using "Show preview"; extremely confusing heading changes like "Texas" above material entirely about political matters in that state being changed to "Tax returns" (incidental subject of the Texas legislative activity, which is like similar activity in other jurisdictions); repeatedly changing an article back to having a large list of commercial examples, over objections that it is spammy or will attract more spam, yet zero participation by this editor in the talk page thread already open about the matter; deleting sourced information simply because he doesn't understand the difference between two syntactically similar but factually different claims; adding speculative claims to a BLP with nothing but primary-source news stories, despite him already being aware that the claims contradict reliably sourced events.

The further back I look, the worse it gets. Much of this editor's activity consists of edit-warring to reinstate his changes after it has been reverted for very clear reasons. And I find zero evidence of any kind anywhere that one or more editors are reverting him improperly, as he alleges above and wants to change the policy to prevent. So, in short, this is basically a bunch of noise, by someone who does not yet understand how the encyclopedic works, and who is treating it like his personal blog and getting frustrated that he doesn't have total editorial control over it. This is nothing to devote any more attention to.  — SMcCandlish ☏ ¢ 😼  07:47, 9 December 2023 (UTC)

Graham's Hierarchy of Disagreement

Attacking me, rather than the proposal, constitutes a classic Ad hominem attack that ought to be rejected out of hand. The current policy gives revert ninjas the ability to say WP:IDONTLIKEIT and say "go get a consensus in the talk page". We have even seen this in this discussion. The idea is to reject the use of WP:IDONTLIKEIT in all content disputes, not just deletion discussions. In my opinion, removal of verifiable content should be based on policies, not on WP:IDONTLIKEIT (and yes, it's redundant to say "verifiable content", since WP:V is one of the Wikipedia:Five pillars of Wikipedia -- I purposefully added the word "verifiable" for emphasis).. Where is Matt? (talk) 22:06, 9 December 2023 (UTC)

I'm not seeing anything in your argument that hasn't been discussed to death. The policy works fine, other than a few bumps with disruptive editors but the issue there is the editors not policy. -- LCU ActivelyDisinterested «@» °∆t° 22:25, 9 December 2023 (UTC)

I have had and continue to have strong disagreements with User:SMcCandlish about several other issues (especially the Wikipedia:Manual of Style). But on this issue, I fully concur in User:SMcCandlish's excellent analysis. WP:ONUS is fine as is. --Coolcaesar (talk) 17:09, 10 December 2023 (UTC)

FWIW, I have no recollection of particular disagreements with you. I don't pay much attention to user names or hold grudges. :-)  — SMcCandlish ☏ ¢ 😼  17:42, 10 December 2023 (UTC)

I'm not "attacking you", Where is Matt?, I'm analyzing your activity as it pertains to the nature of this proposal, and finding a great deal of it objectively unconstructive in ways that would probably be further enabled by this proposal succeeeding (which has no chance at all), and finding no evidence to support your claims about PoV-pushing "revert ninjas", anywhere. Anyone who goes into weird lecturing like "it's redundant to say 'verifiable content', since WP:V is one of the Wikipedia:Five pillars of Wikipedia", toward people who have been here far longer than the erstwhile lecturer and who know orders of magnitude more about the system and how its parts work and interrelate (most of the regulars at WP policy talk pages), is simply in the throes of the Dunning–Kruger effect, and it shows in the very nature of this proposal. You'd do well to get a feel for how WP operates the way it does and why, instead of railing against the way it operates and refusing to understand why no matter how many times it is patiently explained to you.  — SMcCandlish ☏ ¢ 😼  17:42, 10 December 2023 (UTC)

Also note that the Ad hominem article explicitly states: "But it also may be a sound argument, if the premises are correct and the bias is relevant to the argument.". In this case the bias is relevant to the argument. Polygnotus (talk) 18:09, 10 December 2023 (UTC)

To put this simply… when there is disagreement between editors over whether some bit of information should be included or not, we don’t ask editors to prove a negative (ie “this doesn’t belong”) we ask them to prove the positive (ie “this does belong”). This stance makes our articles better. Blueboar (talk) 22:22, 9 December 2023 (UTC)

And that's exactly what I am challenging. If an editor went through the trouble of adding content that satisfies all of Wikipedia's policy requirements, the onus should be on the revert ninja to remove the content. Where is Matt? (talk) 22:32, 9 December 2023 (UTC)

OK… then we disagree on what is best for the project. I suppose you are entitled to your opinion, as I am entitled to mine. Good luck trying to convince others. You haven’t convinced me. Blueboar (talk) 22:43, 9 December 2023 (UTC)

And saying "revert ninjas" over and over again is not making your case for you. It just reinforces that you think everyone who reverts you is wrong, while a close examination of you being reverted shows that in virtually every case the reverts were somewhere between well-justified and downright necessary.  — SMcCandlish ☏ ¢ 😼  17:42, 10 December 2023 (UTC)

Many people have explained the same thing. Maybe its time to move on. Polygnotus (talk) 22:33, 9 December 2023 (UTC)

In your opinion your addition has met all requirements, that doesn't mean everyone has to agree with you. Another editor can in good faith disagree with you, with out being a ninja of your imagination. If it happens I suggest following WP:BRD and using WP:Dispute resolution as necessary. -- LCU ActivelyDisinterested «@» °∆t° 22:37, 9 December 2023 (UTC)

I agree with you. The way WP:ONUS currently reads, the revert ninja does not have to provide any argument for removal of content, other than to point to WP:CONSENSUS and say, "go get consensus for inclusion". That's how revert ninjas operate. It ought to be the other way around, just like it is in WP:AfD discussions. Where is Matt? (talk) 22:46, 9 December 2023 (UTC)

There are no revert ninjas. If editors are behaving in a disruptive manner the correct venue is WP:ANI, changing policy won't change their behaviour they will just find something else to be disruptive with. No policy, or part of a policy, is a trump card. Discussion is the preferred method of solving disputes. -- LCU ActivelyDisinterested «@» °∆t° 23:01, 9 December 2023 (UTC)

Given how WP:ONUS is written, the revert ninja doesn't do anything wrong by removing content with the only justification being "there no consensus for inclusion". WP:ANI does not handle content disputes; it only handles violations of policies. Where is Matt? (talk) 23:18, 9 December 2023 (UTC)

Someone who continues to revert without joining a discussion is WP:STONEWALLING, a behavioural issue that should be handled at WP:ANI. If they continue to revert then it may come under WP:3RR, which is better suited to WP:AN3.
Having no discussion and trying some other method to WP:WIN could also be seen as disruptive. -- LCU ActivelyDisinterested «@» °∆t° 23:56, 9 December 2023 (UTC)

The only thing I would say is that I am of the view that ONUS is more of a consensus/editing issue than it is verifiability so I think it is sitting in the wrong policy but that's just me. Selfstudier (talk) 17:26, 10 December 2023 (UTC)

I agree that ONUS is in the wrong policy. I support what it says, but not where it says it. Blueboar (talk) 12:56, 12 December 2023 (UTC)

Moving it to consensus but leaving mention of it here seems a solution. -- LCU ActivelyDisinterested «@» °∆t° 14:23, 12 December 2023 (UTC)

How revert ninjas weaponize ONUS

I guess I did not do a good job describing how revert ninjas weaponize ONUS, so I'll try to do that here.
Since we are talking about content that passes WP:V, which can be objectively determined, the revert ninjas have to use subjective reasoning such as WP:UNDUE or WP:TRIVIA to delete the content that they want to delete (it's a matter of opinion whether something is undue or trivial). The revert ninjas then use WP:ONUS to put the responsibility on the editor who added the content to get WP:CONSENSUS that the content (per the examples listed) is not undue / not trivial.
In my opinion, the responsibility to form a consensus based on such subjective arguments should rest with the revert ninja, and not with the editor who adds the verifiable content. Where is Matt? (talk) 05:42, 12 December 2023 (UTC)

I think you're confusing failing to describe your concerns with failing to do so in a manner that other editors find persuasive. You may want to consider taking a look at WP:STICK. DonIago (talk) 06:07, 12 December 2023 (UTC)

Don't worry I fully understand you position, I just don't agree with it. -- LCU ActivelyDisinterested «@» °∆t° 12:42, 12 December 2023 (UTC)

@Where is Matt?, please immediately stop calling editors who disagree with you about including something "revert ninjas". Valereee (talk) 13:02, 12 December 2023 (UTC)

In spite of the ad hominem attacks that have been directed towards me in the discussion, this is not about me. Where is Matt? (talk) 14:11, 12 December 2023 (UTC)

You started this discussion in a way that was very negative about anyone who might disagree with you, and have continued in that same vein. Framing your arguments in this way won't win you any converts. -- LCU ActivelyDisinterested «@» °∆t° 14:22, 12 December 2023 (UTC)

@Where is Matt?, the discussion of namecalling is about your behavior, and I'm asking you to stop doing it. Valereee (talk) 14:55, 12 December 2023 (UTC)

How could I possibly be engaging in namecalling if I'm not calling anybody by any names? I did not write the essay WP:NINJA, and I did not accuse any specific editor of being a revert ninja. Where is Matt? (talk) 15:10, 12 December 2023 (UTC)

Taking this to your talk. Valereee (talk) 15:41, 12 December 2023 (UTC)

I agree that ONUS says the excluder (who may or may not be the reverter) wins. Also, this appears to be intentional and wanted by most of the community.

I encourage you to not think of UNDUE and TRIVIA as "subjective", which some people will read as a smear word. I encourage you to think of it as requiring "good editorial judgement", which is something that Wikipedia can't operate without.

On the broader subject, meeting the minimum burden of verifiability is just not enough to get something in a given article. If someone shows up at Cancer (disease) with well-cited information about Cancer (astrology), you really don't want the excluder to have to prove consensus for its removal, when it's obviously in the wrong article ("UNDUE", in our jargon). Any modification to the rule would still have to account for honest mistakes like that. WhatamIdoing (talk) 15:47, 12 December 2023 (UTC)

It should be noted that any deletion of content is a revert, even if performed years after the content was written. Someone had to have written the content for the content to be removed, reverting the original writing (even if it would not count in the WP:3RR).

I disagree that the word "subjective" could reasonably be interpreted as a "smear word".

As for the examples of Cancer (disease) vs. Cancer (astrology), any reasonable editor who adds legitimate content into the wrong article should be more than happy to add it to the correct article. Where is Matt? (talk) 02:45, 13 December 2023 (UTC)

Some policies routinely get used contrary to their intent including by wiki-clever editors I think that it's a valid concern for discussion. I do have a proposal in mind to resolve this and other onus-related problems. North8000 (talk) 13:51, 12 December 2023 (UTC)

Looking forward to seeing your proposal. Where is Matt? (talk) 14:11, 12 December 2023 (UTC)

Mass deletion proposal of Wikipedia:Village_pump_(idea_lab)#Unreferenced_articles based on WP:VER

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

---

An editor has declared that unreferenced articles violate WP:VER, at Wikipedia:Village_pump_(idea_lab)#Unreferenced_articles and has proposed that all 100,000+ of them be summarily deleted through a change to the prod system making deletion mandatory if references are not provided by a drop dead date. My question is: Is the verifiability policy being correctly interpreted over there?    — The Transhumanist   23:29, 1 November 2023 (UTC)

Verification only requires that sources exist, but the discussion is about changing policy. So it isn't a valid question. If policy discussions had to follow existing policy then you couldn't even start a discussion about changing policy.

The original proposal to template new articles without any form of referencing could have got some support, but it somehow became PROD all existing articles (which immediately failed at RFC). I don't think asking established editors to a least include a single reference in the article upon creation is asking much, and it does only apply to established editors as new/IP editors don't get away with not doing that. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 01:44, 2 November 2023 (UTC)

@The Transhumanist, please see Wikipedia:Village pump (proposals)/Archive 207#Request for comment: Unreferenced PROD for the rest of the story.

This, and many other conversations like it, show two points of confusion:

1. Some editors believe that content is never WP:Glossary#verifiable unless it is already WP:Glossary#cited.
2. Some editors believe that this policy (specifically this policy) already requires the addition of at least one source to every page in the mainspace (except WP:DAB pages).

We probably could fix these problems, thought the glacially slow speed of information spread in an environment in which Wikipedia:Nobody reads the directions and in which editors are incentivized to overstate the minimum requirements means that it would take a while to have any noticeable effect. That would likely require re-wording a few things in the policy, e.g.,:

• A clearer definition up top, such as:

  −

  '''verifiability''' means other people using the encyclopedia can check that the information comes from a reliable source

  +

  '''verifiability''' means other people using the encyclopedia can check that the information matches what is stated in a reliable source, through any means, including finding a source themselves that hasn't been cited in the article yet

• Adding a direct statement about whether this policy always requires citations (it's always a good idea, but the confusion is about whether it's always a requirement):

  −

  +

  If an article contains none of the [[WP:MINREF|four types of material that are required to have inline citations]], then ''this'' policy does not require that article to have any sources at all. Citing sources in excess of the minimum requirements is encouraged but optional.

• A clearer statement about BURDEN:

  −

  Any material lacking an inline citation to a reliable source that directly supports the material may be removed

  +

  Editors are permitted, but not required, to remove any material that lacks an inline citation to a reliable source that directly supports those claims

On the middle point, I suspect that if editors were given a choice, they'd actually prefer something like "Every article except DAB pages and non-BLP navigational lists (which can be distinguished from other stand-alone lists by permitting only blue-linked entries with brief descriptions) must contain at least one independent source, even if none of the material on the page technically requires an inline citation. (This source does not need to be formatted as an inline citation.)" WhatamIdoing (talk) 04:32, 2 November 2023 (UTC)

@WhatamIdoing: Changing WP:VER is problematic. I tried to "simplify the convoluted wording" of the policy many years ago, believing that verifiability meant referenced, and got straightened out in short order. Reading the discussion again after all these years, I find it rather amusing. Fortunately, they were very patient with the newb, and took it easy on him, and explained the whole thing in plain English. It set me on the road from cluelessness toward being politically informed. The gist of the policy page is that its wording is a carefully balanced compromise between the deletionist faction and the inclusionists, both groups perpetually closely watching the core of the policy for drift one way or the other. If you so much as add a comma, it'll be inspected by who knows how many editors. That being said, I have noticed a (glacially) slow shift favoring deletionism. Well, that's my $00.02. Sincerely,    — The Transhumanist   19:35, 2 November 2023 (UTC)

Based on your comment believing that verifiability meant referenced, perhaps we could add a sentence that says Verifiable does not mean cited. Verifiable means that it is possible to cite it. Do you think that would have clarified things for you back in the day?

We'd probably have to add yet another sentence about citations sometimes (i.e., frequently) being required, and we can predict opposition from the folks who believe that accurately describing the current policy is lowering standards (or at least making it difficult for them to claim that inline citations are required at the end of every sentence), but if we leave the practicalities aside, do you think that would have helped you back then? WhatamIdoing (talk) 20:15, 2 November 2023 (UTC)

Yes, but I don't think it would have flown, per Blueboar's explanation. I like this wording a little more: Verifiable does not mean cited. Verifiable means that accessible sources exist (out there, somewhere). :)    — The Transhumanist   20:52, 2 November 2023 (UTC)

I’m glad I said something useful, but don’t remember what it was. Could you link to my explanation? Blueboar (talk) 21:07, 2 November 2023 (UTC)

@Blueboar: You shared your wisdom about WP:VER in Wikipedia talk:Verifiability/Archive 24#Proposal: change the name of this policy to "Verification is required". Here are some excerpts:

I think the ambiguity is somewhat intentional. The ambiguity is a nice compromise between deletionists and inclusionists. The idea is that if something is verifiable (ie it is likely that it could be verified, but hasn't been verified yet) it does not completely violate this policy... it simply needs to be verified. It covers the grey zone between deleting unsourced statements immediately, and leaving them tagged with a tags forever. Blueboar (talk) 5:27 am, 7 February 2008, Thursday (15 years, 8 months, 29 days ago) (UTC−8)

There is a difference between the "some sort of random speculative 'I heard it somewhere' pseudo information" that Jimbo is talking about, and information that is likely to be verifiable, but isn't yet verified. Obviously we should remove nonsense and challenge statements that are questionable... but, as Carl states, not everything needs to be given an inline citation - and some statements are so basic that they do not need citation at all (these are frequently referred to as "Water is wet" or "The sky is blue" statements... statements of uncontroverted and obvious fact). There is a ballance here between "get rid of it if it is not cited" and "let people say what ever they want". Blueboar (talk) 8:06 am, 7 February 2008, Thursday (15 years, 8 months, 29 days ago) (UTC−8)

Transhumanist seems to be a hard liner where it comes to verifiability... that is OK. Carl seems to be a soft liner (if that is a term)... That is OK too. Me... I fall somewhere in the middle. That is also OK. Indeed, the entire point of the current language is that it accomodates all points of view on verification. If any of us sees something that we think needs a citation, we can challenge it. If a citation is not provided after a reasonable time we may delete the unsourced statement. The question then becomes: what is a reasonable amount of time?... and that depends on the nature of the statement and what kind of article it appears in... In some cases (such as an unsubstatiated negative comment a BLP) we should delete immediately. In others we can wait several months. We use our good judgement. If there is a disagreement over whether the challenge is realistic or not, we discuss and reach a compromise. All that fits with the current language... which is why I prefer it. Blueboar (talk) 9:32 am, 7 February 2008, Thursday (15 years, 8 months, 29 days ago) (UTC−8)

   — The Transhumanist   21:40, 2 November 2023‎ (UTC)

Thanks… (wow, that Blueboar guy sure said some smart things… on occasion) Blueboar (talk) 22:02, 2 November 2023 (UTC)

@Blueboar: Thank you. I'm not a hardliner, by the way. Your quotes inspired me to write this:

It is enforced. But it is verifiability and not verification that is being enforced. Citations are not required by this policy -- that's a misinterpretation that is very easy to make, and a great many people do. Only the potential of sourcing is required, that is, sources must be out there. If something doesn't have sources out there somewhere, then it isn't verifiable and that item is subject to being deleted. It's a subtle distinction, yes, but very carefully maintained by those who watch over this policy. The wording of the policy is misleading, perhaps intentionally so, and because of this it provides the image that Wikipedia seeks reliability by requiring that everything in it be referenced. But it is just an appearance - because verifiability is much different than actual verification. So people are allowed to jump to this conclusion, as no clarification is provided in the policy. It's a political balancing act, between being largely unreferenced and wanting to be fully referenced, and between verification and the wholesale deletion of a huge proportion of Wikipedia. The reality is that most of Wikipedia is unreferenced. We can't allow that material to be deleted, because that would be too damaging and costly. So, if someone goes on an enforcement rampage to delete every article they come across that lacks sources, they will likely be the focus of an RfC, and shut down. Therefore, the policy is not useless, and serves an extremely important purpose. The Verifiability policy is the fulcrum point of a check and balances system. It is a compromise between deletionism and inclusionism, and is at the very center of Wikipedia's highly political structure. Once you understand this and embrace it, you will have achieved Wiki-Zen.  :) Become one with the wiki. Good luck. Have fun. The Transhumanist 11:51 pm, 17 June 2008, Tuesday (15 years, 4 months, 20 days ago) (UTC−7) (In response to: Wikipedia talk:Verifiability/Archive 29#Useless if we don't apply it)

   — The Transhumanist   22:20, 2 November 2023‎ (UTC)

Yes, but we are at Wikipedia, not "out there" in the world. We have an obligation, clearly explained here: "Responsibility for providing citations". Readers and other editors should have the evidence, right there in the text, of which reliable source backs up the content. No editor has a right to add content (that is not "sky is blue", IOW that might be challenged) with a "trust me" attitude. Hell no. We are skeptics by nature. We demand evidence, and for our content, that evidence is in the form of sources. We must demonstrate that the source of the content is not from our own brains, but from outside sources that are verifiable. If we can't name them in the form of inline citations, then other editors have a right to assume we are pushing OR BS and delete the content. -- Valjean (talk) (PING me) 21:02, 2 November 2023 (UTC)

Maybe I'm missing something, but I don't see mention of the subject in the section "Responsibility for providing citations". Nor do I see emphasis on "All quotations, and any material whose verifiability has been challenged or is likely to be challenged, must include an inline citation to a reliable source that directly supports the material." (bolding added)

We are talking about content that is not "the sky is blue". In theory, it's possible to write an article without any inline citations because every word is about a "sky is blue" fact. Outside of that bizarre possibility, editors must not write content with a "trust me, it's documentable" attitude. No, they have a responsibility to provide inline sourcing from RS for any content "whose verifiability has been challenged or is likely to be challenged". We must not minimize these requirements. -- Valjean (talk) (PING me) 20:26, 2 November 2023 (UTC)

@Valjean, the policy says that the following types of content "must include an inline citation":

1. All quotations,
2. any material whose verifiability has been challenged, and
3. any material whose verifiability is likely to be challenged.

Now, what does the policy say about inline citations for material that is:

1. not a quotation,
2. whose verifiability has never been challenged, and
3. whose verifiability is not likely to be challenged?

WhatamIdoing (talk) 22:23, 2 November 2023 (UTC)

I'm not sure how the policy covers that, but I describe it above when I write about "sky is blue" content. Such content is unlikely to be challenged, so it doesn't need a source. It's also uncontroversial content. If it gets challenged (because of ignorance or cultural differences), a consensus can decide whether to source it.

The initial obligation of the editor adding content is to always add sourcing for content that might be challenged. But what if they don't think it will be challenged (because in their mind it's a "sky is blue" statement), so they don't include a source? That's okay. They were AGF, but other editors didn't agree and they challenged it. Okay, so now it's challenged and needs a source. A source is found and added. We do this all the time. -- Valjean (talk) (PING me) 23:39, 2 November 2023 (UTC)

I agree. Or, as I suggested above, "If an article contains none of the four types of material that are required to have inline citations, then this policy does not require that article to have any sources at all." WhatamIdoing (talk) 15:55, 3 November 2023 (UTC)

I don't disagree with that statement you make above. It covers the 0.001% of the time no references are needed. Because it covers an exceptionally rare situation, can it be made shorter? Come to think of it, do you know of any examples of articles that justifiably do not contain any sources? Can you even describe such an article? -- Valjean (talk) (PING me) 16:11, 3 November 2023 (UTC)

Different people have different ideas about how WP:LIKELY a challenge is. Some editors prefer to think that since someone actually did fact-tag a statement about how many fingers humans have, then everything will get challenged at some point before the WP:DEADLINE. Other editors prefer to think that since only a very few percentage of claims actually get challenged, then probably most uncited content won't be. The first editor might say that this will happen only to 1% of material and basically no articles; the second editor might say that this will happen to 80% of material.

As for articles that justifiably do not contain any sources, the second editor will generally find them in short stubs. (The more content you add, the more likely it is that something will need a source.) Imagine, e.g., a substub that says "Italian Renaissance sculpture is sculpture made in Italy during the Renaissance". You might decide that it's a pointless article, but it would be difficult to imagine someone seriously wondering whether Italian Renaissance sculpture might instead be, e.g., modern Japanese paintings.

If you'd like to look at a "whole" page (i.e., something longer than a single sentence), look at User:WhatamIdoing/Christmas candy and think about whether there is a single statement in there that you genuinely doubt. Look at the first version of Breast cancer, and ask yourself whether any of that is something that you genuinely doubt our ability to source. Don't evaluate it according to whether it would pass FAC; just look at it and classify each claim as either "I could probably find a source for that, if I put a little time and effort into it" or "I have no confidence that I (personally) could find a source for that, even if I spent hours and hours searching for them". WhatamIdoing (talk) 16:36, 3 November 2023 (UTC)

That should be downgraded to an essay not elevated any further, I'm certainly not going to give it much weight. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 16:59, 3 November 2023 (UTC)

Do you see anything in any content policy that explicitly requires an inline citation for any other type of material? For example, if there were a sentence in an content policy that said something like "All statistics or similar numeric descriptions must be supported by an inline citation", then that would indicate an error of omission in WP:MINREF. WhatamIdoing (talk) 18:53, 3 November 2023 (UTC)

The omission in that essay is that it doesn't deal with why content could be removed and the responsibility of editors to reference material if it is challenged. By that omission it gives the impression (that is wrong and doesn't have community support) that content that is challenged doesn't need referencing because it doesn't fall into those for categories. This is what was discussed the last time it was brought up. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 20:10, 3 November 2023 (UTC)

"omission in that essay"? ActivelyDisinterested, what essay is that? -- Valjean (talk) (PING me) 23:05, 3 November 2023 (UTC)

Wikipedia:Inline citation it's an essay about how inline citation should be setup that had extra content added to it. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 23:45, 3 November 2023 (UTC)

MINREF says – this is a copy-and-paste quotation –

"Wikipedia policy requires an inline citation to a reliable source specifically for the following four types of statements:...Any statement that has been challenged (e.g., by being removed, questioned on the talk page, or tagged with {{citation needed}}, or any similar tag)"

What is it about the words "Wikipedia policy requires an inline citation...for...Any statement that has been challenged" that makes you say "it gives the impression...that content that is challenged doesn't need referencing"? WhatamIdoing (talk) 23:06, 3 November 2023 (UTC)

A reminder: Verifiability is truth. "It is commonly cited that the minimum condition for inclusion is verifiability, not truth. However, for Wikipedia's purposes, verifiability is truth. We can't call something true without evidence, and our standard of evidence is verifiability from reliable, published sources." (Copied from Maddy from Celeste) -- Valjean (talk) (PING me) 20:46, 2 November 2023 (UTC)

• One of this things about what editors decide doesn't need referencing is that it's based on the biases of whichever socio-cultural background they come from and serves to reinforce those ideas. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 21:13, 2 November 2023 (UTC)

  I agree, and this can lead to some unfortunate unfairness, in which editors will demand a citation for, e.g., a statement about that a famous Indian Bollywood actor really is an actor, but would not make the same demand for a fairly minor American actor.

  But uneven application of WP:LIKELY doesn't actually help us understand what it means for an uncited statement to be verifiable. For example, the policy talks about "material whose verifiability has been challenged". We don't "challenge" or have disputes over whether material is cited; anyone can look at a paragraph and see whether there are any little blue clicky numbers there. So what, exactly, are we doing when we "challenge its verifiability"?

  I think we're saying that we believe it's impossible for other people to verify that the information matches what the reliable sources say. I believe that to dispute the verifiability of material is to say (a) that verifiable means that sources exist (out there, somewhere), and (b) that we believe that no sources actually exist out there, anywhere for this particular material.

  I believe this distinction between "sources exist (out there, somewhere)" – verifiability – and "there's a little blue clicky number in the article" – cited – is important. WhatamIdoing (talk) 22:49, 2 November 2023 (UTC)

  We are at Wikipedia, not "out there", so we bring the source "out there" and put it "here" as an inline citation. Here the two become one. We should not add content based on what's "in our head". It should be based on "a source we know" and will provide on the spot.

  Readers should not have to go on a hunt for sources. They should never have to "just trust" the unknown editor of the content, especially because we know that editors are fallible, and some are even malicious. Unsourced content that is challenged can be removed. If it's rather benign and seems possibly true (and not a BLP issue), we can tag it with a "citation needed" tag and be patient. If it's been there for a while, and no source has been provided, we can try to find one or we can just delete the content. Standard procedure. The policy works fine as is. -- Valjean (talk) (PING me)

  That's fine, as long as 1) it's not done automatically, by bot, which is what that proposal was leading up to, and 2) deletion is not made mandatory, such as with a drop-dead date, which the proposal included. That proposal disregarded WP:VER and if put in place would have overrode it.    — The Transhumanist   08:17, 3 November 2023 (UTC)

  As I said above the idea was bad, but policy discussions don't need to be policy compliant. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 13:30, 3 November 2023 (UTC)

  Meaning what? That the statements in the discussion weren't verifiable?    — The Transhumanist   13:53, 3 November 2023 (UTC)

  No that a discussion about changing WP:VER cannot be said to be disregarding WP:VER. If that was the case it would be impossible to discuss any changes to WP:VER. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 17:01, 3 November 2023 (UTC)

  What does "LCU" mean? -- Valjean (talk) (PING me) 17:07, 3 November 2023 (UTC)

  See User talk:ActivelyDisinterested/Archive 4#LCU. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 17:10, 3 November 2023 (UTC)

  I agree with The Transhumanist and object to the proposal that is the subject of this thread. It should not be a fully automated process, but something should be done, and automation will be necessary. In theory, it's possible for there to be an article that justifiably does not have sources, but it's such a rare likelihood that we should assume, in the absence of evidence to the contrary, that an article without references is in violation of policy. Just as we should "just assume Trump's always lying and fact check him backwards".[1]

  What could be done is automate a process of notification of the article creator or the last few involved and active editors, or those who watchlist the article. If they don't respond within a specified time limit, then delete the article. That way automation does not control the whole process. A possibility for human intervention is opened, and if involved editors don't think it's important enough to look at the article and add sources (or defend the lack of sources), then delete it. I suspect most of these articles are obscure or lack real notability, and can be created again if someone feels they really are notable. -- Valjean (talk) (PING me) 16:29, 3 November 2023 (UTC)

  This back and forth between claims of what the policy means versus what it actually says, has been going on for decades. We aren't covering any new ground here. WP:VER doesn't specify that an article is in violation of the policy if it is unreferenced. It clearly states when material that is unreferenced can be removed. Here's the nutshell, verbatim: Readers must be able to check that any of the information within Wikipedia articles is not just made up. This means all material must be attributable to reliable, published sources. Additionally, quotations and any material challenged or likely to be challenged must be supported by inline citations. It states "attributable", not "attributed". In the body of the policy it states All quotations, and any material whose verifiability has been challenged or is likely to be challenged, must include an inline citation to a reliable source that directly supports the material. Any material that needs an inline citation but does not have one may be removed. Nobody gets in trouble for not referencing material, and material is not breaking the rule if it hasn't been referenced or removed. The material just becomes eligible for removal. The policy's wording is the result of decades of consensus building between editors of the various wiki philosophies, and is quite deliberate in its meaning — and, it has been written to be as non-confrontational as possible. There is no monster that is condemned to be slain.    — The Transhumanist   17:31, 3 November 2023 (UTC)

  The Transhumanist, the wording "attributable", not "attributed", should be changed, because it contradicts the very first section, which details how editors have the "Responsibility for providing citations". What's "out there" must be linked to by an inline citation. Anything else is OR. Making that change will clear this up. The strength of Wikipedia is that we cite sources. We should not degrade that in any manner. If anything, we should tend to strengthen that. Without that, our credibility suffers. Changing ONE word will resolve this. Just change "attributable" to "attributed". -- Valjean (talk) (PING me) 17:52, 3 November 2023 (UTC)

  @Valjean, are you saying that material is not verifiable unless it is already cited? The policy requires all material to be verifiable. WhatamIdoing (talk) 17:57, 3 November 2023 (UTC)

  Yes. It is not verifiable "here". (What's "out there" is effing irrelevant unless the editor adding content cites it "here".) Editors have no right to add content and force readers to "just trust me because I have read the sources". No, editors must provide the source as an inline citation. Readers should be able to verify content, not by themselves doing the research, but by clicking the inline citation, going there, and double-checking/verifying that the content is backed up by the source provided as an inline citation in our text. Editors are supposed to do the homework of finding sources, and they should "show their work", as our teachers in grade school taught us. Our teachers could thus easily verify that we were not cheating. They could see how we arrived at our solution because we wrote down our reasoning process and calculations. Here we read things in sources and then describe that information in our articles. We are all humans and can make mistakes. Readers should be able to easily check to see if we have interpreted a source correctly. They shouldn't have to ask us where we "got that idea". -- Valjean (talk) (PING me) 18:11, 3 November 2023 (UTC)

  Why are you talking about a "right to add content" at all? Is it somehow an imposition on your own "rights" to have someone attempt to improve Wikipedia, unless they do it perfectly? We might hold certain aspirations, but talking about excluding content because other people having "no right" to participate unless they can do it correctly on the first try and without any collaborative assistance does not sound like the wiki way. WhatamIdoing (talk) 18:24, 3 November 2023 (UTC)

  And it's not what we advertise. This is what we advertise:

     — The Transhumanist   18:33, 3 November 2023 (UTC)

• You are right that all have a right to edit, but no one has a right to expect that such an edit will be saved or later built upon. Our invitation is very simplistic, IOW encouragement to BE BOLD. Here I'm speaking to us in this conversation, so maybe I should have written "experienced editors have no right" to expect us to "just trust me". I'm speaking about what our policies should say. I am not speaking to newbies and drop by editors who are not expected to know all these things. But when they decide to dig deeper and learn why their first edits were not accepted, they should be able to find in our policies why their OR ideas not based on any RS were not accepted. They will then learn to not add what they think is better content. They will learn to only add what a RS tells them. (This obviously has little to do with stylistic, MOS, grammatical, and framing issues.) -- Valjean (talk) (PING me) 18:44, 3 November 2023 (UTC)

  The vast majority of editors are new and inexperienced, and therefore, the vast majority of new edits are unsourced. Seasoned editors have been perpetually trying to catch up with backlogs in pretty much every department, let alone fixing the lack of citations.    — The Transhumanist   19:26, 3 November 2023 (UTC)

  What editors believe has sources "out there somewhere" is a matter of socio-cultural bias. What areas editors believe don't need referencing is a judgment they make based on the ideas and values of their societal background. When content is challenged it is, in part, that bias that is challenged, which may explain some of the backlash it sometimes receives.
  As to what it means to "challenge its verifiability", it's simply a believe that no sources backs the content, a challenged only answerable by supplying referencing. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 00:07, 3 November 2023 (UTC)

  Bingo! What is "common knowledge/sky is blue" varies across cultures and languages. -- Valjean (talk) (PING me) 00:12, 3 November 2023 (UTC)

Too right you are. And on the other side what is obvious to someone with knowledge of a subject can be just impenetetrable to someone without. I recently had the unpleasant experience of arguing with someone who said a citation said something and there were a whole lot of cites that copied it, all with zero knowledge of how it works and just one letter so not a reliable source on the original saying it was codswallop. Notable enough to be due so in it goes. Oh well. NadVolum (talk) 11:58, 3 November 2023 (UTC)

I don't think the "cultural" argument is especially relevant.

Here's the question we're trying to address:

• Is it possible for uncited material to be verifiable?

Imagine that you personally happen to know that the uncited statement is accurate (e.g., you personally read about this exact thing in the news this morning). But it is definitely and indisputably uncited in the Wikipedia article. Is that unverifiable, and thus in violation of the policy's requirement that "All material in Wikipedia mainspace, including everything in articles, lists, and captions, must be verifiable", or is that merely uncited, in which case it is necessary to determine whether it is one of the four types of material required to have an inline citation before you know whether a citation is required (NB: not "a very good idea", which of course it is)? WhatamIdoing (talk) 16:01, 3 November 2023 (UTC)

What you don't think is "relevant" is based on your socio-cultural biases, that's the point. Saying you don't think it's relevant shows it is relevant, not that it's not. Because what you will or will not challenge is based on your preconceived ideas and biases. "How could you not know that? (When everyone I know knows that)", "That's just general knowledge! (In the culture I live in)", etc.

And if someone challenged something I knew to be corrected, I would supply referencing as required by policy. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 17:07, 3 November 2023 (UTC)

Yes, of course: Once someone actually CHALLENGEs the verifiability of the material, then the policy requires that a citation be added. But unless and until someone challenges the material, the policy does not require that a citation be added (assuming it's not a quotation).

What I mean above about it being irrelevant is that I don't think your argument (basically: What's obvious to me isn't obvious to everyone; therefore I think it's best for everything to be pre-emptively sourced even if the policy doesn't require that material to be cited yet) has anything to do with whether verifiable is a synonym for cited. WhatamIdoing (talk) 17:55, 3 November 2023 (UTC)

Of course verifiable is not a synonym for cited, so we should resolve that by changing that one word to verified, or "verifiable by citing the source". Then they are synonyms. Problem solved. Like I wrote above, "changing ONE word will resolve this. Just change "attributable" to "attributed"." -- Valjean (talk) (PING me) 18:22, 3 November 2023 (UTC)

Okay, @Valjean, it's time for you to make up your mind. Above, I asked "are you saying that material is not verifiable unless it is already cited?" and you said yes. Here, you say no, material can be verifiable if it's not cited.

Could I please have exactly one answer to this question – just this one, single, isolated question – and ideally an answer that you're planning to stick with for the next month or two? WhatamIdoing (talk) 18:26, 3 November 2023 (UTC)

Where do I say "no, material can be verifiable if it's not cited."? (It's hard to keep up because of the fast pace and edit conflicts.) -- Valjean (talk) (PING me) 18:51, 3 November 2023 (UTC)

Me:  "Are you saying that material is not verifiable unless it is already cited?"

You:  "Yes. It is not verifiable "here"." WhatamIdoing (talk) 18:55, 3 November 2023 (UTC)

Valjean, I tried to make that very change 15 years ago. And it did not fly. Over the years, many others have tried. And their efforts died. The reason is, that the policy is in a stalemate between deletionists and inclusionists. The current wording is a form of détente, and has been that way for decades. If you look at the talk page archives, you will see this debate run over and over again ad nauseam. The core of the policy has not budged, regardless of how many times editors like you or me or WhatamIdoing have argued on either side or the other of the debate. It's a perpetual tug-of-war between equally matched teams. If you don't believe me, edit the policy, or create a proposal to change it, and see what happens. (It'll just be another waste of time, like this debate we are in right now). See Blueboar's explanations above for clarification. He hit the nail right on the head.    — The Transhumanist   18:56, 3 November 2023 (UTC)

You're right that it "is a form of détente", and that's untenable. We should finally resolve it, but this thread is really about something else and we're deeply into policy discussions. This isn't good. -- Valjean (talk) (PING me) 18:54, 3 November 2023 (UTC)

Somehow, we coexist on Wikipedia in relative harmony, which is as close to tenable as we get.    — The Transhumanist   19:03, 3 November 2023 (UTC)

You keep on sidestepping the central point, but sidestepping not won't make it go away. What is or is not required to be referenced is based on socio-cultural bias. Policy follows what is actually done yes? So that is policy. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 20:10, 3 November 2023 (UTC)

Just looked at that proposal and I strongly disagree with it, but it demonstrates something which has been evolving on WP. I file it under the bureaucracy tendency. One thing I did not see explained either here or in that Village Pump discussion is that Wikipedia is in a constant state of drafting, and never finished. I suppose most articles went through phases where they were badly sourced. If you keep deleting such pages, then we'll never be able to do what we are best known for doing. We would switch over into a temple for tweaking old articles made by people in the good old wild days. If we don't want that then we have to remember the practicalities of our strange way of starting and then building up articles, which was historically often fuelled by the annoyance editors (as opposed to bureaucrats) have at poorly sourced articles. Are all articles supposed to be born in the perfect finished form like Aphrodite?--Andrew Lancaster (talk) 17:34, 3 November 2023 (UTC)

The answer to your last question is a resounding "NO"! But that doesn't mean that as an article is being built from a stub it should not be based on cited RS. It should. Each building block should be cited. -- Valjean (talk) (PING me) 18:00, 3 November 2023 (UTC)

m:Eventualism is my own approach, but it's not everyone's.

My concern about the trend in that area is that it seems to be based on an anti-collaborative style, rather like opposing lawyers in court. Everyone tries to be polite/formal about it, but there's a lot of standing on my inalienable right to not have to do anything except complain about how you didn't do it perfectly.

The story we want to tell is like this:

• Bob added something to an article, but he didn't add any sources. Alice noticed his mistake and fixed it for him. Wikipedia is improved!

The interactions we get from the ones screaming about how all the burden is on YOU NOT ME!!! is more like the story in Wikipedia:Bring me a rock. In older, more judgemental language, "they bind heavy burdens and grievous to be borne, and lay them on men's shoulders; but they themselves will not move them with one of their fingers". We don't want editors to dump dubious text and then say "you source it, if you want a source; I can't be bothered". But we also don't want editors to blank everything that doesn't have an inline citation associated with it, and say "I challenge every uncited statement from here to infinity, including 'The capital of France is Paris' even though WP:NOR directly says the statement "the capital of France is Paris" does not require a source to be cited". We are looking for a happy balance between the two extremes. Perhaps the path there is to specify in BURDEN that this is what's true in a dispute, and not, e.g., when someone mentions the capital of France. WhatamIdoing (talk) 18:19, 3 November 2023 (UTC)

You seem to be trying to squirm out from under the clear policy statement:

"All content must be verifiable. The burden to demonstrate verifiability lies with the editor who adds or restores material, and it is satisfied by providing an inline citation to a reliable source that directly supports the contribution." (bold original)

We are supposed to provide sources in two cases, either before or after a "challenge". We should provide sources when adding content to prevent challenges that we might reasonably foresee. This is where the imperfections of human cultural differences come into play, so we AGF and provide the source if the content is later challenged. We should try to provide sources for new content, even an updating of the score in a soccer game. This is one of the most frequent types of subtle vandalism. Scores and numbers get changed all the time without the editor providing a source. That's not right. -- Valjean (talk) (PING me) 18:34, 3 November 2023 (UTC)

No, I'm trying to set up the preconditions for an intelligible conversation, namely that when we each use a word, we are using it to mean the same thing. For example, if I encountered "the capital of France is Paris" (sans little blue clicky number) in an article, I would describe that as:

• verifiable,
• uncited, and
• not requiring a citation.

I'm not sure how you would describe that. Maybe verifiable or unverifiable? Uncited, of course. Maybe requiring a citation (because the burden to demonstrate verifiability always lies with anyone except you)? Maybe not, since a core sourcing policy explicitly says that it doesn't? I don't know. But I'd like us to have a shared vocabulary, so that we can understand whether that text already violates existing policy ("must be verifiable") or whether requiring a citation for that statement would require a change to the existing policy. WhatamIdoing (talk) 18:44, 3 November 2023 (UTC)

And the fact that you know Paris is the capital of France, and the reason that you think everyone would know that Paris is the capital of France, are products of your socio-cultural background. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 20:15, 3 November 2023 (UTC)

But it's not relevant to the question of which bits of wikijargon we should use to describe that statement. The statement itself could be anything. The question I've asked Valjean is about which words we should use (e.g., in this policy) to describe an uncited statement for which the editor who is talking about the uncited statement is perfectly certain that a suitable reliable source could be trivially produced. WhatamIdoing (talk) 23:10, 3 November 2023 (UTC)

@Valjean: that is just not dealing with the reality of how WP works, as opposed to how you think it should work. WP as we know it would not exist if everyone had to do every edit that way even when starting to or tweaking new, low quality articles. Since the beginning of WP there has been an idealistic idea that it just isn't fair that WP works without being more strict. Various attempts were made to create wikis which were more strict. They did not work. The only way of achieving what you are demanding is by filtering out the best of WP, which is of course also an old idea that has been attempted various times.--Andrew Lancaster (talk) 19:03, 3 November 2023 (UTC)

Due to the high volume of this discussion, its fast pace, and edit conflicts, I'll just post my view and give this thread a pause:

We are supposed to (according to Jimbo) document the "sum of all human knowledge", and we do that by finding it in RS "out there" and then describing that, with its source(s), "here" in our articles. That's how we build content. All content should be "verifiable in cited" sources "here", not in "citable" sources "out there". -- Valjean (talk) (PING me) 19:16, 3 November 2023 (UTC)

I'm still trying to figure out whether you think "verifiable in citable but uncited sources" is an oxymoron. I understand that you wish it were; I'm trying to figure out whether you believe it already is. WhatamIdoing (talk) 23:21, 3 November 2023 (UTC)

WhatamIdoing, I suspect that's a good question if I can even parse it correctly. Even more worrying is whether I can explain myself well enough that I don't trip over my own awkward attempt to clarify things. Let me make a feeble attempt, and, as usual, you will help me refine my thoughts, for which I am always grateful! I appreciate your clear thinking. That's the blessing of writing. Speech doesn't quite do that. Writings can be tweaked, while speech is gone with the wind.

Let me illustrate. If I have a fatal illness, but you tell me there is a cure, it does me no good if you don't tell me where it's located. That's what an inline citation does. It tells me where the cure is located. When in doubt (if it's "likely to be challenged") you must tell me the location. Our policy tells us that. If content is likely to be challenged, then an inline citation must be included. Is that clear as mud? -- Valjean (talk) (PING me) 03:35, 4 November 2023 (UTC)

Murky as mud. Because what if the editor does not respond well to that type of motivation? They make an edit, and then walk away. You let them know that their work could be removed, but they don't care, and work on 1 or 10 or maybe even 100 other articles that week instead. If you follow them around, deleting everything they write, they'll either report you for stalking or move on to another project or another account. There are millions of people around the world contributing to Wikipedia, and most of them know very little about the policies. My guess is that the vast majority of citing is done by our army of silent but effective (and overworked) wikignomes. If you can figure out how to recruit or train more of those, maybe you could get a handle on the backlog.    — The Transhumanist   07:12, 4 November 2023 (UTC)

@Valjean, I think what would help is if you could imagine a statement that is currently not required by this policy to have any citations whatsoever. Like, none anywhere on the page. ("Christmas candy is candy associated with Christmas"? "Red is a color"? "The human hand normally has four fingers and a thumb"?) Think of a sentence that could be in an article, and that, if someone dropped a fact tag on it, you'd think "What a troll" instead of "Okay, I can imagine why someone might want a citation for this".

Then tell me: if that sentence is uncited, is that sentence still (currently, without the citation) verifiable? (For clarity: The answer to this question can be a single word, either "yes" or "no".) WhatamIdoing (talk) 17:15, 4 November 2023 (UTC)

In this scenario, the request for the citation is unreasonable, and I suspect the following discussion would result in a consensus to not add a citation. A wikilink would probably suffice for the person who doesn't understand what "candy" or "red" mean.

Otherwise, with a reasonable request, yes, it is unverifiable here (the only thing of relevance to the reader), and that lack of a citation requires the reader to perform their own OR to find it out there. We do not do that. We provide the source inline, right here. We prove to the reader that the editor who wrote that content didn't just write their own opinion or get the idea from their own head. The reader knows the author based the content on an external reliable source. Wikipedia is better than many other sources because we provide the documentation. Our current policy is fine. -- Valjean (talk) (PING me) 18:11, 4 November 2023 (UTC)

The policy says All material...must be verifiable. It does not say that material must be verifiable unless it's unreasonable to ask for a citation. Do you think that those statements (e.g., "Red is a color"), when uncited, are 100% completely verifiable within the meaning of that policy requirement? WhatamIdoing (talk) 01:17, 5 November 2023 (UTC)

Our policy should be: "All content that is likely to be challenged must be verifiable in an inline cited source, not just in an undefined citable source. That source must be included." (or something like that) -- Valjean (talk) (PING me) 19:16, 3 November 2023 (UTC)

The present policy requires it to be WP:LIKELY, not just conceivably possible ("might be", as in "pigs might fly"). WhatamIdoing (talk) 23:12, 3 November 2023 (UTC)

Good catch. I'll fix that above. -- Valjean (talk) (PING me) 23:24, 3 November 2023 (UTC)

The existing policy wording is "All...material whose verifiability...is likely to be challenged, must include an inline citation...". The existing statement says "must", and here you are proposing "should". This is inconsistent with what you have advocated for, so I assume that you would prefer to stick with the existing wording. WhatamIdoing (talk) 02:25, 4 November 2023 (UTC)

Yes. Fixed. -- Valjean (talk) (PING me) 03:17, 4 November 2023 (UTC)

My focus above is the key wording: "verifiable in an inline cited source, not just in an undefined citable source." -- Valjean (talk) (PING me) 03:22, 4 November 2023 (UTC)

To my eyes, that looks like a wordier or more redundant way of putting it, but it doesn't change the meaning. If it's LIKELY, the current policy says it "must" include an inline citation; if it's LIKELY, you propose saying that it must "be verifiable in" an inline citation, and you specify the obvious, which is that an uncited source is not the same thing as an inline citation. These have no content difference, and they are not the source of the disputes.

The source of the disputes is editors reading "likely" and then replacing that with their own views. Imagine the editor who would say something like "Paris is the capital of France? Oh, no, people only know that if they've had more than a couple of years of education, so it is extremely LIKELY to be challenged!" or "That fish might be shorter than your finger, but saying it's 'small' is a 'scientific classification' and all scientific classifications are extremely LIKELY to be challenged!" That editor reads the policy (or doesn't, because Wikipedia:Nobody reads the directions) and then makes up non-existent rules, like "WP:V says that every single article must have an inline citation to at least one source". This might be a highly desirable thing for WP:V (or WP:NOT) to say, but the fact is that those words aren't actually in this policy (yet). And we will never get them into this policy unless editors stop twisting the definitions and making up stuff about what's really here.

If editors would stop saying that WP:V already requires inline citations on practically everything, because (ITNSHO) practically everything is LIKELY, and that therefore WP:V certainly requires at least one source per article, then we might be able to get other editors to say things like "What a shocking oversight – Did you know that verifiable only means you could find a source if you really had to, and not that it's verifiable specifically in the source named in the little blue clicky number at the end of the sentence? According to our policy, uncited text could be verifiable even if there's no source in the article and no free-to-read reliable source online for that material! And did you know that, while the policy says that everything must be verifiable/able-to-find-a-source-yourself-with-effort, if there's no MINREF material, we could actually, legally, legitimately have a whole Wikipedia article that isn't required by any policy to cite a single source in it? We should fix this terrible oversight, and require more than just 'verifiable'. We should require a citation to at least one source in every single article".

But we can't get there so long as we have a couple of editors who go around saying things like "WP:V already requires this – why, no, I can't actually find a sentence that plainly says that each article must have at least one inline citation anywhere in the policy, but I can tell you that if (but, sadly, only if) someone needs to cite something, then the person who needs to do that work is not-me, and that's practically the same thing" or "verifiable means the same thing as cited; unverifiable means the same as uncited". If we are ever to get a statement into this policy that says a minimum of one source is required no matter what, then you really need to stop claiming that the policy already says this. WhatamIdoing (talk) 06:57, 4 November 2023 (UTC)

"you really need to stop claiming that the policy already says this". I didn't say that.

We already use (at least) two basic types of inline "citations". We have refs/notes and wikilinks. The first are used for things likely to be challenged, such as quotes and controversial stuff. We try to provide them when we write the content. (Failing that, we just add them later when requested.) We also use wikilinks for the uncontroversial stuff written in wikivoice. They obviate requests a lot of the time. Without them, we'd be getting lots of challenges over trivial "sky is blue" stuff and single words from people with other cultural and language backgrounds.

Do we ever mention "wikilinks as a form of inline citation" in this policy? Maybe we should. When we get a seemingly unreasonable request from someone who lacks basic English language skills or a rudimentary knowledge of English or American history, our first solution should just be to wikilink it. Maybe that will resolve the problem, and after all, isn't that what this is all about? We just want to inform people and make it easy for them to verify stuff. For controversial stuff, we provide outside sources. For trivial stuff, we provide wikilinks to whole articles where sources are provided as inline citations.

For your theoretical "article with no inline citations", I imagine it would still include wikilinks. We should make that a rule. "An article must contain essential wikilinks, especially if there is no obvious need for other inline citations." For trivial stuff, wikilinks obviate the need for other inline citations. Maybe that should be added as a fifth item to WP:MINREF. -- Valjean (talk) (PING me) 18:37, 4 November 2023 (UTC)

A wikilink is not an inline citation, and a Wikipedia article is never a reliable source per Wikipedia:Verifiability#Wikipedia and sources that mirror or use it.

In practice, they are useful as a source of information if an editor is wondering whether it would be reasonable to CHALLENGE something (e.g., in a List of Michelin 3-star restaurants: If you click through to the linked article and see that the article begins with 'Ooh La La is a French restaurant best known for winning three stars in the annual Michelin Guide', then only trolls, idiots, and lazy people would seriously wonder whether that item belongs in the list), but they are expressly disclaimed for the purpose of complying with this policy. WhatamIdoing (talk) 01:23, 5 November 2023 (UTC)

I think we agree. -- Valjean (talk) (PING me) 04:37, 5 November 2023 (UTC)

The core of wp:ver says that sourcing is required if challenged or likely to be challenged. "Likely to be challenged" is non-explicit guidance but it's easy to resolve that by merely challenging it. (I've advocated that saying that a challenge should include a good faith expression of concern about the verifiability of the material but so far we don't have that, and IMO the current policy in this area is perfect for Wikipedia other than that)  :-) ) Based on this, the premise of the assertion that the OP brought to light is based on misreading wp:ver. That said, what more can be said? North8000 (talk) 00:00, 4 November 2023 (UTC)

When experienced editors misread a policy – when, for example, someone says that the four-item table in WP:MINREF is omitting or downplaying policy requirements, but does not notice that it is extremely similar to (but written more forcefully than) the four-item bullet list in WP:BURDEN – then it may be appropriate to re-write the policy to be clearer. Sometimes, of course, clarity isn't desired, especially by people who dislike the current policy requirements. WhatamIdoing (talk) 02:33, 4 November 2023 (UTC)

To the extent that any of the above "we should rewrite the policy" ideas by Valjean et al. could have any effect that amounts to making it no longer valid to cite hard-to-get-at but extant and otherwise reliable sources, I'm absolutely opposed to such a change. I've gone out of my way to obtain (sometimes at considerable expense) various out-of-print (but not out-of-copyright and thus not on Internet Archive, etc.) and otherwise scarce books by subject-matter experts in various topics, and they are perfectly valid as sources (often among the best sources when it comes to quality of information), but not easy for some random reader to get via the Internet without spending considerable money (probably not even through inter-library loan in the US; copies in US libraries, if any, would probably be non-circulating). And the vast majority of the scientific and other academic material we cite is paywalled. The fact that it takes at least an undergrad at a univerisity with institutional access to verify that content is not a WP:V problem, and can't become one. This is the encyclopedia anyone can edit, by slogan, and the encyclopedia anyone can verifify something in if they have the time and the means, by policy. It is not the encyclopedia anyone can verify every fact in with no legwork or investment, and it would fall apart rapidly if it became that.  — SMcCandlish ☏ ¢ 😼  22:29, 4 November 2023 (UTC)

@SMcCandlish, there is no need to worry about this. Valjean is hoping to make people add more sources, not worse ones. WhatamIdoing (talk) 01:26, 5 November 2023 (UTC)

Exactly. I'm an inclusionist, and especially favor the inclusion of sources as inline citations more than is currently the case. We can't expect newbies to understand all this stuff, but they cannot learn it if it's not specified more clearly in our PAG. As noted above by The Transhumanist, "The reason is, that the policy is in a stalemate between deletionists and inclusionists. The current wording is a form of détente, and has been that way for decades." I believe this is an untenable situation. Our PAG should avoid ambiguity. We should finally resolve this. IIRC, there are only two words that would need to be changed to fix this. -- Valjean (talk) (PING me) 04:48, 5 November 2023 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

EXCEPTIONAL tweak to close wikilawyering loophole

WP:EXCEPTIONAL presently says: Any exceptional claim requires multiple high-quality sources.

This is a system-gaming/wikilawyering loophole because "two" matches the definition of "multiple". I believe this should be changed to "numerous", "many", "a majority of", or some other formulation that actually gets at what we mean here. Some disputation about wikilawyering related to this has come up at Wikipedia:Manual of Style/Film, and involved a guideline change from Describing a film with superlatives such as "critically acclaimed" or "box-office bomb" is loaded language and an exceptional claim that must be attributed to multiple high-quality sources. to use many; instead of multiple. It was reverted on the grounds that the wording has to match WP:EXCEPTIONAL, so here we are.

The issue is that if anyone can find a bare two sources that say the PoV that editor wants to push (even if the majority of sources disagree), the mulitple wording gives them all the wikilawyering leverage they need to do it (and seems to be backed up by guidelines parroting the same wording). This is clearly not the intent of WP:EXCEPTIONAL, and is inimical to properly encyclopedic writing. This is just one of those WP:Policy writing is hard matters, and needs some wordsmithing. But I'm not absolutely certain what wording would be best here.  — SMcCandlish ☏ ¢ 😼  08:49, 4 November 2023 (UTC)

Perhaps "a proportional number of"? If the topic has thousands of sources, but only dozens support this exceptional claim, it's probably not appropriate to include - but if there are only ten sources on a topic, then two would warrant including the claim. BilledMammal (talk) 08:59, 4 November 2023 (UTC)

Not sure of the wording but it's something we need to change. Doug Weller talk 09:39, 4 November 2023 (UTC)

In the bulleted list of red flags, this concern is already covered: "Claims contradicted by the prevailing view within the relevant community...". I think it is generally understood that multiple high-quality sources (more than one) is needed for an exceptional claim, but exceptional claims must still adhere to WP:DUE for inclusion and prevalence.

In this specific case at MOS:FILM, the concern was raised that editors would find what they believe to be significant coverage on a loaded language claim like "critically acclaimed", but since the claim was exceptional and typically gets flagged by MOS:WTW, you really needed more than just a passing mention in multiple reviews. High-quality sources help solidify the case for inclusion when challenged. --GoneIn60 (talk) 11:29, 4 November 2023 (UTC)

• As with everything, policies and guidelines must be approached with a degree of common sense. The problem I have with this proposal, is that it changes the emphasis from requiring that exceptional claims are well corroborated to requiring that there is widespread reporting of an exceptional claim. I think this would be a mistake, given the fact that an exceptional claim—by its nature—may not be widely reported. I do not believe we should impose that requirement either, but it is important that an exceptional claim is corroborated by high-quality sourcing. You are correct that a literal interpretation of the policy may mean that an exceptional claim may satisfy the requirement if just two sources report it. Indeed, given how "exceptional" or provocative the claim is, whether it relates to a living person or not, and whether it is libelous or not, two sources may be sufficient. Editors may take the view that given the degree of exceptionalism then three or four sources should be the bar. Of course, there is a direct relationship between the quality of the sources, and the number that editors consider sufficient, but that should be determined on a case-by-case basis.

I think the proposal is unnecessarily replicating WP:DUE where the weight of opinion should be taken into consideration (e.g. film criticism, climate change etc) and imposes a "weighting" condition in cases where it is arguably not appropriate. WP:DUE takes care of the number or proportion of sources required when that is a consideration, but WP:EXCEPTIONAL should concern itself with how well corroborated a claim is.

Betty Logan (talk) 12:29, 4 November 2023 (UTC)

I agree that we should remain flexible on this. A lot depends on HOW exceptional the claim is, given the topic in question. The more exceptional it is, the more we need to support it by citing sources. The number of sources can be determined on a case by case basis. Blueboar (talk) 14:19, 4 November 2023 (UTC)

I agree with Betty and Blueboar. The quality of the sources makes a huge difference in how many are needed, so this is a case-by-case matter. -- Valjean (talk) (PING me) 16:28, 4 November 2023 (UTC)

A problem I came across wasthat the original source was quoted in multiple other reliable sources and they kept saying therefore that there were multiple reliable sources. None of them would have checked it since they were just doing things like reviewing the book and quoted the author. Unfortunately that was no other source which described anything like what was described or any effect that one might expect from it but inferring anything from that is classed as OR. I tried to get third opinion from a Wikiproject that one might expect to have some expertise but nobody there contributed anything. NadVolum (talk) 17:09, 4 November 2023 (UTC)

I would take the republication/repetition of the idea in the original source by multiple other RS as a meaningful confirmation and approval of their thinking, enough that citing a few of them would be good enough. (News aggregators do not count.) A great idea starts somewhere, and when other great thinkers start repeating it, the weight of their approval means something. We'd then cite the original source and also cite some of those who approvingly repeat their idea, often along with some of their added commentary as they will usually do more than just repeat it. -- Valjean (talk) (PING me) 18:51, 4 November 2023 (UTC)

And there you have the problem. A person goes and writes something he almost certainly got the wrong end of in a book. The book reviews go and quote it and the newspapers too. All with absolutely no evidence. And those quotes attributed to him are then taken as extra reliable confirmatory sources overriding exceptional and we go quoting it too attributing it as well - but by your reasoning we've checked it as well and found what the author said is probably right. NadVolum (talk) 20:03, 4 November 2023 (UTC)

I think this goes to the heart of the point I was trying to make. Putting aside considerations such as WP:FRINGE or issues where there are counter-claims for the moment, when a claim is duplicated by high quality sources ordinarily WP:DUE would tip the scales in favor of inclusion. However, with exceptional claims, I think we should be seeking a certain level of independent corroboration. It is very difficult to address this through hypotheticals, but to take this story as an example, it was quite widely repeated but there was no independent corroboration of the central claims. Betty Logan (talk) 23:31, 4 November 2023 (UTC)

I think part of the problem here is that we're talking about a story ("A funny thing happened on my way to work today..."), rather than a universal fact ("Wonderpam reduces the risk of heart attack, stroke, and hemorrhoids").

In the first instance, sources will repeat stories that they find entertaining or illustrative of someone's character. The point isn't really whether it happened, so they're not trying to assemble evidence that would prove that it happened. They're trying to give an impression, so they're telling stories that illustrate the broader outlines (e.g., "Here's a modern-day Robin Hood, who commits crimes to thwart the powerful and help the oppressed"). Independent confirmation is usually unrealistic and may be presented that way (e.g., "When I was alone in the hospital room with him just before he died, he said..."), and I think the solution for such instances is to follow their lead: "He claimed that Assange confessed to committing computer crimes in Canada" rather than "Wow, what a super-nice dude who's been proven in multiple sources to have help out those oppressed brown people!"

In the second instance, independent confirmation is usually realistic and desirable. Cold fusion that only works in one lab isn't proven, drugs are going to be evaluated by government agencies, etc. One drug company's story being repeated in a dozen news publications is not sufficient evidence. WhatamIdoing (talk) 01:46, 5 November 2023 (UTC)

Well we know Assamge was convicted for that twenty years previously. And it makes a good story so is repeated. But should we be repeating stuff like that without the slightest other evidence of at least somebody having done any hacking of anything vaguely relevant to Egypt's telecoms at the time? After all he is facing stiff charges from america for hacking. I'm sure some of both the robbed and the poor given money by Robin Hood would say something. I don't want to go into that business - I think it would require technical assistance - but stories about people can have consequences. Do we just say we don't care, attribute it and we're done? NadVolum (talk) 11:59, 5 November 2023 (UTC)

I think this may be based on a false premise. Regarding "the fact that an exceptional claim—by its nature—may not be widely reported", one might argue that the opposite is the case, where reasonable claims are not widely reported (because they are banal, because they are obvious, because so many different, boring, sources assert them, for whatever reason) while exceptional claims may be very widely reported. See "Jewish space lasers" for example. The focus, rather, should be on independent sources, as raised in several comments below. Mathglot (talk) 16:28, 5 November 2023 (UTC)

Hence the "may". An exceptional claim may or may not be widely reported. A reasonable claim may or may be widely reported. I was arguing against the suggestion that exceptional claims should be supported by many sources, on the basis that an exceptional claim may be strongly corroborated but not widely reported. Am I not partly making the case that the emphasis should be on independent sources, rather than the quantity of sources, by arguing that WP:EXCEPTIONAL should concern itself with corroboration? Sorry if I am misunderstanding your point, but you appear to be suggesting that I am arguing the opposite of what I am. Betty Logan (talk) 17:00, 5 November 2023 (UTC)

That is clear, but I'm not sure how to square this idea with the principle that we give due (if any) weight based on prevalence in reliable sources. If some claim that is exceptional is barely reported and most sources say something different, then we might not include it at all, even if various editors are convinced it is correct. We'd wait until the preponderance of current source material told us that the real-world consensus (in whatever the applicable field is) has become that it is correct.  — SMcCandlish ☏ ¢ 😼  17:26, 5 November 2023 (UTC)

Things like WP:DUE, WP:AGEMATTERS and possibly WP:FRINGE would presumably play a role in those case. The scenario you describe commonly arises for ordinary claims all the time. The purpose of this policy is (at least as I interpret it) to establish a threshold/precondition for exceptional claims. Betty Logan (talk) 00:08, 6 November 2023 (UTC)

• No, I don't think this is workable. There are people who start to complain if you put more than three sources on a statement, even. And changing it to "multiple" opens the door to wikilawyering in the other direction - declaring that a statement is EXCEPTIONAL is already, to some degree, a value judgment, so if we were ambiguous about how many sources are required then someone could declare any statement they personally find objectionable to be exceptional, and continuously refuse to accept any number of sources as sufficient for it. This can even be done in good faith! If someone genuinely believes that something is simply wrong regardless of what the sources say, they're going to find it exceptional and no reasonable number of sources is ever really going to WP:SATISFY them. In terms of how we want disputes of that nature to go - I think that once multiple high-quality independent sources have been presented, we want people to move on to examining and evaluating those sources, checking to see if there are other sources that dispute them, arguing over WP:DUE and so on. What we don't want is someone repeatedly citing EXCEPTIONAL alone as a reason for removal with no further engagement, which is something that I'd fear could be invited if the sourcing requirement is raised; I don't think it's good for the wiki or for dispute-resolution to have policies that people can beat repeatedly like a drum - outside of trivial disputes, both sides need to engage, which means that once there's multiple sources you have to examine them or start doing the serious legwork of digging up sources of your own, you can't just keep asking for more and more and more forever. However, we could perhaps add a word and have it say multiple high-quality independent sources - citing two sources that aren't independent of each other isn't really better than citing one (and citing a source that isn't independent from the organizations or people being discussed isn't any good at all.) --Aquillion (talk) 19:48, 4 November 2023 (UTC)

  Extraordinary statements will normally be due if they're true! And unfortunately if something is not true you can have zero sources that say anything to deny it no matter how silly it is. NadVolum (talk) 20:14, 4 November 2023 (UTC)

  That's completely extraneous, Aquillion. This has nothing to do with how many citations are put into the article at the same spot, or at all. It's only to do with what proportion of the available reliable sources, that editors examine in determining WP:WEIGHT, actually support the claim when someone cries "I found two sources, thus multiple sources, saying this, so I can make Wikipedia say it".  — SMcCandlish ☏ ¢ 😼  22:19, 4 November 2023 (UTC)

• Has anyone ever been able to use the exceptional clause in circumstances other than WP:FRINGE when a reliable source says something silly but a contributor argues it should be included? NadVolum (talk) 22:56, 4 November 2023 (UTC)

  I think this comes up mostly when it comes to terms covered by WP:WTW, but I can also see it happening when various generally reliable sources have not aged well on certain points, and are contradicted by newer and presently-better-accepted research (think topics like Race (human categorization), where there is a never-ending stream of randos trying to make our articles say things that are considered scientifically obsolete ideas).  — SMcCandlish ☏ ¢ 😼  23:30, 4 November 2023 (UTC)

  I wonder whether the WP:ECREE section was written for the Cold fusion disputes. @Blueboar, do you happen to remember? WhatamIdoing (talk) 01:31, 5 November 2023 (UTC)

The common sense answer would be the more exceptional and questioned the claim is, the stronger the sourcing must be. (and vice versa) And implicitly, the strength of the the sourcing can arise from various factors, with quantity of sources being (only) one of them. If we could come up with something which incorporates some of those principles, IMO that would be good. North8000 (talk) 01:30, 5 November 2023 (UTC)

We could just stick that in. It is quite succinct in cutting through to the heart of the issue. Betty Logan (talk) 03:47, 5 November 2023 (UTC)

I'd like it to say independent sources. But I'm not sure we can agree what an independent source is if people think having a second reliable source quoting the same source is confirmation that it is true. NadVolum (talk) 11:23, 5 November 2023 (UTC)

I think that independence of the source is one of many factors that that go into gauging the strength of the sourcing. Others might include the expertise, knowledge and objectivity of the source with respect to the item which cited it and the extent that the source has been established to have those qualities. But I that is too much and too universal to try to put into this change. North8000 (talk) 14:35, 5 November 2023 (UTC)

Jytdog used to pound on the idea of intellectual independence. I've seen this abused (e.g., if the company is famous for its financial situation, then it's not notable, because all financial information ultimately traces back to the company/someone affiliated with the company, and that's 'not independent' – see also churches known for their attendance, list of largest employers, etc.), and I'm not sure that it's a good idea in the first place (because taking information from other sources is an important form of endorsing and elevating what's most important – it's impossible to do a systematic review or a meta-analysis that is, in this strict sense, fully 'intellectually independent' of the sources you're citing), but that would be the language people have used to describe this idea. Again, I have my doubts about the utility, practicality, and desirability of this concept; I only mention it for the name that's been given to it. WhatamIdoing (talk) 15:53, 5 November 2023 (UTC)

Are people saying company reports, church attendance figures, or largest employers would be cases of dispute under EXCEPTIONAL? They don't sound like the things mentioned there. NadVolum (talk) 16:07, 5 November 2023 (UTC)

I don't think that these are commonly taken to be extraordinary claims, though given that editors are people living in their own little filter bubbles, we do get the occasional odd remark. People who don't attend church and don't know many people are occasionally surprised to discover that a sizeable minority of Americans do, for example. WhatamIdoing (talk) 01:10, 6 November 2023 (UTC)

Sure, and I like where this is going. I'm often not too particular in exactly how a loophole gets patched, as long as it does (without creating new ones in the process). But a simple sentence along these lines, which is consonant with other policies and guidelines, seems to be a good way to go.  — SMcCandlish ☏ ¢ 😼  11:46, 5 November 2023 (UTC)

I'm not sure that we really have a loophole. GNG says "multiple", but that's doesn't mean "exactly two gives you the right to demand inclusion". WhatamIdoing (talk) 15:54, 5 November 2023 (UTC)

Hmm, except I see that argument all the time at AfD. Something has one cite that seems to qualify as non-trivial coverage, but it's not enough by itself. Someone adds another, then !votes keep because "multiple" has been satisfied. If we mean to say "lots" in one wording or another, then we need to find some way to say it instead of the ambiguous/gameable "multiple" – especially when it comes to the WP:V matter, where the credibility stakes are much higher than whether we do or don't have an article on some random band or actor or squash player.  — SMcCandlish ☏ ¢ 😼  16:12, 5 November 2023 (UTC)

Except that we don't need "lots" of high-quality sources. Chances are that if two high-quality sources have come to the same determination on a particular claim, then there will be a significant number of lower-quality sources parroting this claim as well. Are there a lot of real-world situations where that's not the case? Would it just be easier to inject a reminder that EXCEPTIONAL claims must still adhere to DUE? GoneIn60 (talk) 19:53, 5 November 2023 (UTC)

If we don't have sources opposing it is very easy for something exceptional to be due. It is quite easy for mistakes to arise. I've noticed lots in newspapers nearly always it isn't worth bothering about. Even in maths you can have mistakes, I've just noticed in the article on Abraham de Moivre in the section on De Moivre's formula it says Euler proved the formula for any real n. Which is wrong as described in De Moivre's formula#Failure for non-integer powers, and generalization and the talk page is filled with arguments about this as it is a high school formula and the mistake was in a widely used textbook. And of course there aren't textbooks saying it is wrong! NadVolum (talk) 21:09, 5 November 2023 (UTC)

Sure, people sometimes advocate for that at AFD (but only when I think it should be kept, and never when I think the subject should be excluded). That doesn't mean that their argument is accepted. WhatamIdoing (talk) 01:13, 6 November 2023 (UTC)

On these tricky areas where Wikipedia "usually works", I've always thought that if you put the actual "common sense" process that occurs for "usually works" into words that you have a good plan. Nearly always this involve multiple people weighing multiple variables and deciding. (Wikipedia:How editing decisions are made) While individual policies and guidelines usually have "binary flow chart / combinational logic" provisions that are counter to this they also includes softer language which allows the Wikipedia to work at the larger scale level. My post here was just suggesting using this concept for this particular question. Sincerely, North8000 (talk) 01:11, 6 November 2023 (UTC)

Common sense bah! There's editors around who investigate every last thing about the topic and are determined to stick everything in. If a reliable source writes up something it has to go in even if they are totally naive and ignorant about what they say. If an expert makes a mistake it has to go in because nobody has written anything saying they made a mistake. Maybe we are all missing some common sense but it seems to fly out the window on Wikipedia when confronted by a reliable source and EXCEPTIONAL does not currently seem to be any defence except in the FRINGE project which I think sometimes acts like a flash mob of canvassed like minded editors. We need something better in policy rather than something so desperate. NadVolum (talk) 12:09, 26 November 2023 (UTC)

Remember that “Verifiability does not guarantee inclusion”. Per WP:VNOT, Consensus can determine that some bit of information should be omitted, even when that bit of information is not considered extraordinary.

So, when some Wikilawyer argues “I have multiple sources for this - it must be included” they are simply wrong… while having lots of sources makes it more likely that there will be a consensus for inclusion, there is no special threshold of sources that mandates inclusion. Blueboar (talk) 14:58, 26 November 2023 (UTC)

Do you think I'd eb bothering here if EXCEPTIONAL was redundant like you seem to be saying? It is not redundant. An extraordinary thing is quite liable to be interesting and worthy of inclusion in an article if it is actually true. The problem is that they very often are not and we need a better definition of what is exceptional and requires some better verification than the same original source being duplicated. Even just attributing something can be too much - often these extra cites are just extra attributions rather than sources saying it in their own voice and there's no reason for us to give extra weight by doing that as well. NadVolum (talk) 18:59, 26 November 2023 (UTC)

This behavior is one that makes me think we need a policy called Wikipedia:Write an encyclopedia article. "I have multiple sources" – "So? Including that would be a violation of the 'Write an encyclopedia article' policy". If we could have conversations instead of throwing WP:UPPERCASE at each other, it wouldn't ever be necessary or useful, but... WhatamIdoing (talk) 20:42, 26 November 2023 (UTC)

Sounds like a good essay at the least.  — SMcCandlish ☏ ¢ 😼  08:03, 9 December 2023 (UTC)

Specific revision idea

From the above discussion, I distilled this:

Any exceptional claim requires multiple high-quality independent, reliable sources. The more exceptional the claim, the stronger the sourcing must be.

What do we think about this? It's not a huge change from Any exceptional claim requires multiple high-quality sources, and I don't think it's actually a substantive change (it's just drawing on extant policy), but would have a strong clarifying effect, and make the "multiple" issue I opened with effectively moot. I think this is more important to say than several other things already in the section, like the mention of conspiracy theories, and of statements that seem out of character.  — SMcCandlish ☏ ¢ 😼  08:03, 9 December 2023 (UTC)

The 'independent' is what's really good there, where hopefully independent means the original sources are independent like two reporters at an event rather than two independent newspapers reporting the same reporter at the event! A lot of editors have been arguing that since two different newspapers report the same original source that's good enough. NadVolum (talk) 12:50, 9 December 2023 (UTC)

You CAN prove a negative

Both of these tasks are equally practical:

• Verifying that source X does say statement Y.
• Verifying that source X does not say statement Z.

Consider a citation of X, which simply lists the name of the work; in order to find statement Y in work X, the reader would potentially have to read through the entire work. Similarly, the reader could read through the entire work to verify that it does not say statement Z; it is just as neutral, factual, and verifiable to say each of these statements:

• (positive) X says Y.
• (negative) X does not say Z.

Of course, the negative statement is frequently not useful; there are an infinite number of things that X does not say. However, sometimes, it is useful to point out a negative statement, particularly as it relates to a positive statement. For instance, consider this text:

In 2022, Fields Medalist David Cohen–Steinberg published a paper that said "Goldbach's conjecture is herein proven to be false without a shadow of a doubt."^[11] However, his paper neither provides a proof nor cites one.

[…]

11. ^^{a b c} Cohen–Steinberg, David. "Fool's Goldbach". MIT Department of Mathematics. November 21, 2022. Retrieved November 23, 2022.

Here, we have 2 statements:

• (positive) Reference [11] says "Goldbach's conjecture is herein proven to be false without a shadow of a doubt."
• (negative) Reference [11] does not contain a proof or even cite one.

It would be worthwhile to make an explicit policy regarding such negative statements; specifically, it should be permissible to say that work X does not say Z, especially when Z relates to some other statement Y that is positively cited as being in X. Um… I think I said that right… 24.118.62.152 (talk) 04:02, 26 November 2023 (UTC)

Why? Why not just remove the other statement Y that is incorrectly representing the contents of X, or find a third source A, which actually does support statement Y? WhatamIdoing (talk) 05:22, 26 November 2023 (UTC)

• Y is not incorrectly representing the contents of X.
• Finding a third source to say something (at all) is precisely what is not necessary. This is what needs to be made clear to editors; the one source is good enough.
• I will add that I do not think your reply is particularly responsive to the issue raised; it seems to miss the point, which I say not to call you out, but rather to indicate to readers of this thread that I do not think we are actually talking about the same thing here.

24.118.62.152 (talk) 06:59, 26 November 2023 (UTC)

We are not restricted to using only one reliable source for a statement. In fact, we should search for multiple reliable sources for any given content to be sure that we are representing in Wikipeida the consensus view of experts on the field. If multiple reliable sources agree on a fact/position/interpretation, then we only need to cite one of the sources. If one source contradicts multiple other reliable sources, then the policy at Wikipedia:Neutral point of view#Due and undue weight applies, and the editors contributing to the article can reach a consensus to exclude the fringe source (we should always exclude non-reliable sources). If there are multiple reliable sources supporting each of two or more different facts/positions/interpretations of something, then the policy at NPOV:Due and undue weight again applies, and the article can reflect the different positions/interpretations roughly proportional to the number and quality of reliable sources supporting each fact/position/interpretation. I have dealt with cases when a fact/position/interpretation is found in only one reliable source, but I think the said fact/position/interpretation is not compatible with related facts/positions/interpretations in other reliable sources. I cannot let my doubts about the fact/position/interpretation in question affect how it is presented in Wikipedia, as that would not be sticking to the source. - Donald Albury 16:19, 26 November 2023 (UTC)

What you say would fit better in the previous discussion about EXTRAORDINARY. It sounds like you would not admit any change in reliable source requirements even if you think something is extraordinary. NadVolum (talk) 19:07, 26 November 2023 (UTC)

Personally I'm all for this, I don't know of any policy which covers it, WP:CALC seems to me to be the closest. Wikipedia does need better protection from untrue statements. The immediately preceding discussion #EXCEPTIONAL tweak to close wikilawyering loophole covers another aspect of this problem. Basically people are not all that interested in writing citable errata or don't have the qualifications to do so - at most we get some random response in a comments section of an article saying it is wrong or rubbish or stupid but we can't cite that. NadVolum (talk) 11:53, 26 November 2023 (UTC)

Why? The IP apparently wants to write sentences like "Reference [11] does not contain a proof or even cite one." To which I have to ask: Who cares about Reference [11]?

Imagine that someone has used a completely invalid source for an otherwise true statement, like:

• According to The Bible, the winner of the 2022 Academy Award for Best Actor was Brendan Fraser for his role in The Whale.

The IP wants editors to read the cited source, discover that it {{failed verification}}, and then add a second sentence that says something like:

• The Bible does not contain any information about the Oscars.

This would not be a valuable contribution. A valuable contribution would look like removing the incorrect in-text attribution to the Bible and adding a citation to a reliable source that actually contains the relevant information, e.g., https://www.oscars.org/oscars/ceremonies/2023. There's no encyclopedic purpose at all in saying that the Bible doesn't talk about the Oscars, or that any incorrectly cited source doesn't say whatever it's claimed to say.

Additionally, if we move in that direction, we'll end up with POV pushing: This medical journal article claims to prove that HIV causes AIDS, but I don't think it's sufficient proof, so I'll just write "Reference [11] does not contain proof that HIV causes AIDS or cite a source that proves HIV causes AIDS." Just before posting this, the IP was edit warring over cannabis products to say that the FDA didn't cite studies proving (to the IP's satisfaction) that the product had side effects. The FAQ at the top of this page says that reliable sources don't have to cite their own sources, but the IP's goal is to be able to discredit reliable sources if they don't cite sources. WhatamIdoing (talk) 21:00, 26 November 2023 (UTC)

You missed out an indirection in the question. It is more like 'Famous person' says 'source' proves 'something'^[1] - and ^[1] is reliable. But 'source' doesn't say aything much about 'something'. For instance Lady Tartagua says the Bible has a list of Oscar winners^[1], but no Oscar winner appears in the Bible. The rule followers will say that searching the Bible is OR and lots of editors are quite happy to have daft things in no matter how stupid provided they are attributed like this is to Lady Tartagua. EXTRAORDINARY in the discussion before this is more like 'reliable source' says 'somethng', and the 'something' sounds very unlikely. NadVolum (talk) 21:43, 26 November 2023 (UTC)

As to what the IP said in the article - any decision here would not support the edits they made. The FDA itself said the statement in its own voice - and that sort of thing if it is to be countered without sources would come under EXTRAORDINARY in the previous discussion. Or for a medical article WP:MEDRS.NadVolum (talk) 22:07, 26 November 2023 (UTC)

• If source X claims that source Y includes a proof of Z… and source Y does not, in fact, include a proof of Z… then we must at least question the reliability of source X. Blueboar (talk) 19:17, 26 November 2023 (UTC)

I'm not sure that I follow your reasoning. If a statement X (e.g., Goldbach's conjecture is proven in this paper) is challenged by an editor, then the best course of action would be to find another source which says that X is not true (e.g., "Goldbach's conjecture is not proven in that paper" OR "Goldback conjecture is not proven").

Of course, it will not always be possible to find such a source. Failing that, the editor should establish consensus that the source does not support the claim and remove it from the article. Alaexis_¿question? 21:10, 26 November 2023 (UTC)

What the IP apparently wants to do is add "However, the FDA does not cite any studies to back up these claims" after the statement that "In 2022, the FDA stated that 'scientific studies show possible harm to the male reproductive system...'^{[FDA source]}".

To use the sentence @NadVolum wrote above, instead of 'Famous person' says 'source' proves 'something', it's an attempt to discredit 'US FDA' says 'scientific studies' prove 'something' – and the IP doesn't want people to believe that 'something' might actually be scientifically proven. So instead of finding MEDRS-compliant medical review articles like "The effects of cannabidiol on male reproductive system: A literature review" oder "The endocannabinoid system, cannabis, and cannabidiol: Implications in urology and men's health" oder "Review of the oral toxicity of cannabidiol (CBD)" (and, from a brief review, 100% of other medical journal articles addressing this particular subject), the IP wants instead to say "Well, the exact cited source didn't cite sources that convince me, so let me scatter some disbelief over this widely acknowledged fact." WhatamIdoing (talk) 22:30, 26 November 2023 (UTC)

So in this case there is no consensus that a source does not support this claim, therefore the source should stay and no statements should be appended to it. Alaexis_¿question? 22:41, 26 November 2023 (UTC)

Tagging without challenging

We talked about this a while ago, but I can't find the discussion. I ran across Glossary of communication disorders today. It's 100% uncited, and from my spotcheck, also 100% verifiable. That is, the only problem with the content is that nobody has yet spent several hours spamming in a bunch of little blue clicky numbers.

I'm wishing that there was a way for me to spam in {{citation needed}} at the end of every entry, so that it would end up in Wikipedia:Citation Hunt. However, under our present system, there's no way for editors to distinguish between "she didn't want to do all that work herself, so she set it up to work for the citation-adding game" and "This is a serious WP:CHALLENGE to the contents, and if a source is not supplied within a reasonable length of time, the material should be removed". WhatamIdoing (talk) 17:52, 7 December 2023 (UTC)

The {{unreferenced}} banner template is appropriate. Even though WP:GLOSSARIES can be even more of an oddball than lists, they still need verifiability, but on something that's an established field and verifiable in any single textbook like this, one could just pick up (or download) a single mainstream grad-level textbook, go to the index, and tick the boxes. I'd suggest you don't cite each entry to their page number in the textbook index unless you're actually verifying, word-for-word, that the description is accurate.

The concerning bit is the handful of entries that might be there are not in the standard textbook. You would probably find those in the edit history after the bulk of the glossary was written. Those are the ones to individually tag with {{cn}}, or else to google-scholar quickly to ensure they are established in the field.

You could of course flag WP:WikiProject Medicine to do this, but it seems to me an easy enough job if you can get a relatively recent textbook. SamuelRiv (talk) 18:23, 7 December 2023 (UTC)

@SamuelRiv, I probably should have been more explicit, instead of just giving links. This policy says that all material must be verifiable. As I said, the list is 100% verifiable. That means, from the Wikipedia:Glossary, that all material in the page must be something "that someone (although not necessarily you) could, with enough effort and expense, determine has been published in a reliable source, even if no source is provided in the article."

This page "has verifiability". What it doesn't have is "cited sources". "Cited" means little blue clicky numbers on the page this minute. "Verifiable" means "an editor realistically could find a source to cite and add a little blue clicky number". In the language of WP:UPPERCASE, we have WP:V; we don't have WP:CITE. I would like to have both, and I don't want to do the work myself.

As a way of drawing attention to this page by people who actually do enjoy doing that work, I would like to get it into Wikipedia:Citation Hunt. However, it's important to me that this be done without anyone (e.g., an editor who is unclear on the distinction between verifiABLE and citED) being able to use that as an excuse to gut the article later. This is presently not possible in our current system. We don't have a system to say "It'd be super sweet of someone to volunteer to add a source here, and I am absolutely not issuing a CHALLENGE because I know perfectly well that this is verifiable content". WhatamIdoing (talk) 18:41, 7 December 2023 (UTC)

This again? Anyone can put a cn tag if they want to and delete it after a time if one isn't forthcoming, no obligation to go hunting for one. Selfstudier (talk) 18:52, 7 December 2023 (UTC)

Sure, I could add a tag, wait a few weeks, and then remove it, but I'd like these entries to remain as opportunities in Citation Hunt until they are cited. WhatamIdoing (talk) 19:15, 7 December 2023 (UTC)

Ok, I understood what you were saying about the article (and maybe I didn't communicate that well), but I think I misunderstood the point of this thread (which isn't meant to be about any particular article).

I've found something similar in several old academic-centric articles, where a great bulk of material may lack references of any kind but be 100% verifiable (or in the case of math, simply verifiably true). One knows enough to know that supporting references definitely exist, but one lacks the time/energy to get those references, and one wants a tag that says to other editors "citation needed, but it's definitely WP:V and the reference material is out there, so please don't nuke this text in 6 months like you would with ordinary cn stuff." Do I have that right?

So it sounds like this is just a matter of creating a new inline template, and getting some consensus to put it in WP:V and other meta pages so people use it? You also need a good title for the template -- and I'm sorry but I can't think of one. SamuelRiv (talk) 19:06, 7 December 2023 (UTC)

Yes, I think we're talking about the same thing now. A new template, or a "known good" parameter for the existing one might work. A new template (unfortunately) might be safer, as changing from {{cn}} to {{optional but it would be ideal to have a little blue clicky number here}} would be more noticeable in the diffs. WhatamIdoing (talk) 19:18, 7 December 2023 (UTC)

There is {{sources exist}}, which is a header template that could certainly be improved a bit. But it's not so popular that it is heavily restricted. Adding some customizable text fields like |reason= would help. SamuelRiv (talk) 19:57, 7 December 2023 (UTC)

What's the copyright status of www.nidcd.nih.gov? The entries on this list appear to be very close or simple copy paste of NIH entries. I know certain US governmental publication are free to use, but don't know the complete details.

. -- LCU ActivelyDisinterested «@» °∆t° 21:08, 7 December 2023 (UTC)

Everything created by an employee of the US federal government in the course of their official duties is public domain. On official websites, it's not unusual to find that the text is public domain, but there are exceptions, and photos are more likely to be copyrighted.

https://www.nidcd.nih.gov/glossary is a glossary there. The page has no statement about copyright, and https://www.nidcd.nih.gov/policies#d says "Unless otherwise stated, the information on this site is not copyrighted and is in the public domain." It appears therefore to be in the public domain.

Our page goes back to 2002 (=21 years). The first functional copy of the NIDCD page that I can find is https://web.archive.org/web/20160405050019/https://www.nidcd.nih.gov/glossary in 2016. Many descriptions match, but the list of entries is quite different (e.g., under "G", they have Genetic counselor, Geneticist, Global Aphasia, and Gustation, while we have Grammar, Gustation, and Glaucoma). WhatamIdoing (talk) 22:42, 7 December 2023 (UTC)

Thanks I could find how they declare non-free content but https://www.nidcd.nih.gov/policies#d is clear. -- LCU ActivelyDisinterested «@» °∆t° 00:51, 8 December 2023 (UTC)

A little geeky "I'm good at whipping up templates" part of me at first (for like two seconds) liked the "just a matter of creating a new inline template" idea, but in the end it would simply be abused by every PoV pusher under the sun to assert that their nonsense is somehow sourceable out there and shouldn't be removed. The only way to demonstrate that something is verifiable is to make it verified.  — SMcCandlish ☏ ¢ 😼  01:14, 8 December 2023 (UTC)

Can you make template trustmebro? Sources exist but I am not gonna tell you where.User:Polygnotus/trustmebro Polygnotus (talk) 12:13, 8 December 2023 (UTC)

On a more serious note, that {{Sources exist}} template is a joke and should be deleted. I randomly picked an article where it was used, "Technical decision". Literally no sources. I compared the Wikipedia article with a boxing dictionary.

Wikipedia: "A technical decision is a term used in boxing when a fight has to be stopped because of a headbutt."

Boxing dictionary: "Technical Decision: When a fight is stopped early due to a cut, disqualification, or any situation when the bout is stopped and the scorecards are tallied."

So the Wikipedia article is not just unsourced, it also uses a wrong definition of its subject. Instead of using {{Sources exist}} one should simply post a bunch of sources on the talkpage. Polygnotus (talk) 12:28, 8 December 2023 (UTC)

I do exactly that at various topics (in a thread I always create and call "Additional sources" for consistency, and sometimes cross-reference them between related topics). It saves me the trouble of trying to learn some complicated bibliography app to keep track of stuff I find that might be of use, and with me putting the sources (usually already pre-formatted in citation templates for easy copy-paste) in the talk pages, someone else may beat me to the punch in actually using them to improve the article. Plus they're a line of AfD defense.  — SMcCandlish ☏ ¢ 😼  12:53, 8 December 2023 (UTC)

Not to get sidetracked, but the source you posted isn't an official rulebook or summary or something like that. When I looked up other sources, they were mixed depending on I guess how technical they were -- the Indian Boxing Council site seemed to indicated the TD can technically come about from any cut or similar injury, from any legal collision that is not a punch, but it seems from summaries like that of Sports Pundit like it's almost always a head collision or headbutt.

So our article does seem indeed to be verifiable by sources, but it's certainly not gotten at the heart of the issue that we want for any decent WP article -- what is the technical definition in every circumstance in every major jurisdiction. I'd say the {{sources exist}} tag is accurate, appropriate, and flags this article as one needing significant attention. SamuelRiv (talk) 12:58, 8 December 2023 (UTC)

But when you dig a bit deeper that just turns out to be wrong, e.g. https://www.boxingbase.com/boxing-results-explained/ Cuts are one of many potential reasons for technical decisions, and the definition in the wikipedia article is incorrect. It would be much much better if the driveby tagger posted some sources on the talkpage instead of slapping on a sources exist tag. The template gives people an easy way out, which is great for lazy people but bad for the encyclopedia. Polygnotus (talk) 13:06, 8 December 2023 (UTC)

That link says exactly what I said. The fouls in question are almost always headbutts and collisions resulting in cuts, and half the google results refer to such exclusively. "Dig deeper" means doing the article right and getting strong RS, not this or that website. A driveby tagger may have simply known or believed, as a person familiar to the sport, that a TD is in practice always a head collision. They tagged the article as problematic. Now here we are discussing a problematic article. Had such a tag not existed, perhaps the tagger may have considered, as is the point of OP's thread here, that the article were sufficiently verifiable as to not warrant tagging that it "needs sources for verification", and moved on. The lack of a suitable tag would not magically motivate said tagger to spend 5--10 extra minutes on a single page to list a bunch of sources (which they probably would not themselves have vetted, because that's another 15 minutes at least extra commitment you're asking of them) on the Talk page.

If the point of having quick straightforward tagging templates is that problematic pages are tagged for later review by those with the time and energy, which is exactly what is happening right now with the TD article, then that is the system working as intended.

If a new tag with more granular semantics will mean more tagging of problematic articles with more precision, as OP is suggesting, then I'm all for it. SamuelRiv (talk) 13:21, 8 December 2023 (UTC)

This is also one of the great properties of the trustmebro template. When Wikipedians see User:Polygnotus/trustmebro sprinkled throughout articles they may end up talking about those articles. So if I understand you correctly you think its a great idea if I add User:Polygnotus/trustmebro to all unsourced statements, which will then lead to discussions that mention the articles (despite the fact that mentioning an article does nothing to improve it). Being in favour of a template because people complain how useless it is is a bit weird. In response to the OP, if someone is too lazy to use sources they are too lazy to add content to Wikipedia and every edit they make should be reverted. Polygnotus (talk) 14:01, 8 December 2023 (UTC)

@SamuelRiv: Of course I was thinking about fixing the Technical decision article, but after reading your comment I decided I am too lazy. And when you are done, you will have only 1969 articles to go. Writing articles Wikipedia:BACKWARD is not to be encouraged. Imagine if we WP:TNT all that unsourced content. Have we lost anything of value? Pulling out the weeds gives the other plants more space and sunlight and nutrients to grow. Polygnotus (talk) 14:16, 8 December 2023 (UTC)

I am thinking of Template:Sources exist as an intermediate point between Template:Notability and an article that has cited sources proving notability. I'm thinking of it as "I'm not questioning notability, because I did a BEFORE search".

I looked at a dozen articles tagged with this template. They ranged from Greasy Kid Stuff, with only an WP:ELOFFICIAL link, to Dragon (Dungeons & Dragons), which cites 73 sources. WhatamIdoing (talk) 19:08, 8 December 2023 (UTC)

@WhatamIdoing: See below. Polygnotus (talk) 21:03, 8 December 2023 (UTC)

• We have a ”dubious” tag for uncited statements we are challenging… It sounds like we need a “Likely” tag for uncited statements we think are ok, but would benefit from being cited. Blueboar (talk) 13:01, 9 December 2023 (UTC)

  Sounds like a good idea. -- LCU ActivelyDisinterested «@» °∆t° 14:46, 9 December 2023 (UTC)

Template:Sources exist

The first 50 articles that contain the "Sources exist" template and the amount of times the string "<ref" occurs in them

Number Article Count 1 Genosha 31 2 Latveria 17 3 Zerg 25 4 The Black Pits of Luna 5 5 Mara Jade 15 6 Everett, Massachusetts 41 7 The Best... Album in the World...Ever! 0 8 List of Taoists 0 9 Dead Prez 4 10 Terminal ballistics 3 11 Ş 0 12 Little John 4 13 Greasy Kid Stuff 0 14 Tongji (spirit medium) 0 15 Trinovantes 2 16 Rabbi ben Ezra 2 17 Elisa Maza 14 18 Mimic (comics) 37 19 Charles Strite 5 20 Dragon (Dungeons & Dragons) 115 21 Travis McGee 12 22 Witches (Discworld) 9 23 Lucius II: The Prophecy 5 24 Department of Computer Science and Technology, University of Cambridge 16 25 What Is Enlightenment? 2 26 Caribbean Disaster Emergency Management Agency 18 27 Belisarius series 7 28 Covenant (Halo) 48 29 Artistdirect 10 30 Michael Bivins 1 31 Lala Mara 0 32 G'Kar 3 33 Omaha's Henry Doorly Zoo and Aquarium 53 34 Knockout (1971 comic) 0 35 Dhalsim 25 36 Starscream 50 37 Murder of Kristen French 8 38 Greenfield land 3 39 Initial D (video game series) 8 40 Liquid Snake 25 41 Anointed One (Buffy the Vampire Slayer) 0 42 Nebuchadnezzar (The Matrix) 11 43 Painkiller (magazine) 6 44 Master Mold 10 45 HK-47 30 46 John Scolvus 14 47 Season of the Sakura 1 48 Fantastic Five 2 49 Raúl Díaz Arce 0 50 Sally Baldwin 7

At what point should we remove {{Sources exist}} from an article? For example, the city of Everett has a population of 50.000 people, there is no doubt that it is notable.

I propose we remove the template from all articles that contain 5 or more references the string "<ref" 5 times or more. Polygnotus (talk) 21:08, 8 December 2023 (UTC)

Any editor can remove a template of they believe that the point raised is no longer an issue. However doing things en masse requires care, "<ref" 5 times could be one reference using a refname 5 times and no evaluation of the quality of that source would be taking place. -- LCU ActivelyDisinterested «@» °∆t° 23:46, 8 December 2023 (UTC)

OK, agreed, I just wrote it like that to avoid 2 extra lines of code. I changed it to 5 or more references. I don't think its fair to ask for a evaluation of the quality of the sources. Since the template has been slapped on articles with very little thought behind it it should be easy to remove. Polygnotus (talk) 01:28, 9 December 2023 (UTC)

Since the template has been slapped on articles with very little thought has it? I'm sure in many cases, as with other tags, it might have been. But to classify ever use of it as thoughtless seems a stretch. -- LCU ActivelyDisinterested «@» °∆t° 14:48, 9 December 2023 (UTC)

classify ever use of it as thoughtless I didn't. Polygnotus (talk) 15:01, 9 December 2023 (UTC)

What is the purpose of removing a template, designed to flag an article to have its verifiability reviewed by an editor, if you're not actually going to review its sources for verifiability? By my accounting references fail verification on spot check nearly half the time (although in fairness that's biased toward me only spot-checking hot, controversial, stale, and questionable articles). SamuelRiv (talk) 01:41, 9 December 2023 (UTC)

@SamuelRiv: Why do you insist on being part of a discussion about a template when you do not know what the text on the template says? The template is for notability, not verifiability. Please read what the template actually says. Also read Template:Sources_exist#When_not_to_use. If you want to make a new template that says "someone should compare the sources with the statements in the article" you could, and then you could put it on basically every article (except maybe the 0.000000000001% FA/GA). Almost no-one is actually comparing sources with claims made in the article, and this template is not telling people to do that. I honestly think a large majority of people are unable to write a decent article. What is the purpose of having a template that is not helpful to the reader or editor on hundreds of pages? Cool, some rando thought sources existed to establish notability. That is no reason to put a giant banner at the top. I have been working on Category:Unreferenced_BLPs_from_March_2020; it would be cool if you could do Technical decision. Polygnotus (talk) 01:45, 9 December 2023 (UTC)

One of the articles I looked at the other day (with this tag on it) had multiple refs, but almost all of them were non-independent. I don't remember the subject, but imagine an article about a band, citing different pages of the band's own website. While I think that filtering the list for 5+ refs would be a good starting point for manual review, I also think that manual review would be helpful.

@Polygnotus, do you think this banner should be on the talk page? WhatamIdoing (talk) 17:47, 9 December 2023 (UTC)

I think this diff by BoomboxTestarossa is an important example. They replaced the {{notability}} tag with the {{sources exist}} tag. In short, it's being used to clear up bigger problems (should this get deleted?) and prevent duplication of effort (essentially meaning "I already did a BEFORE search, and I decided it was unlikely to get deleted at AFD"). WhatamIdoing (talk) 17:52, 9 December 2023 (UTC)

I think there is a presumption of non-notability whenever an article is created. Someone has to prove notability, by describing what the article is about (e.g. Everett) or by including sources. New Page Patrollers will make sure it gets deleted if it is not notable. But if an article exists for years, and/or has a bunch of sources, we can safely presume that it is probably a notable topic. So it would make sense to me to say, after x month or x references we can automatically remove Template:Sources exist.

If we look at The Black Pits of Luna it was nominated, the template got added, and then the AfD was closed as keep, but the template was not removed. If an AfD is closed as keep, does that not mean the article is notable and the template can be removed?

Another option is to add a parameter to the Notability template that changes the first sentence of that template from:

The topic of this article may not meet Wikipedia's general notability guideline. Please help to demonstrate the notability of the topic by citing reliable secondary sources that are independent of the topic and provide significant coverage of it beyond a mere trivial mention. If notability cannot be shown, the article is likely to be merged, redirected, or deleted.

to:

An editor has performed a search and found that sufficient sources exist to establish the subject's notability. Please help to demonstrate the notability of the topic by citing reliable secondary sources that are independent of the topic and provide significant coverage of it beyond a mere trivial mention. If notability cannot be shown, the article is likely to be merged, redirected, or deleted.

If someone did a WP:BEFORE search and concluded that a topic is notable, is that simple fact reason to put a giant ugly banner on an article for years? The information that someone believes the topic is notable is only potentially relevant in a deletion discussion, so sticking it on the talkpage would make far more sense. The readers of Wikipedia articles don't want to see a giant ugly banner. And they won't be converted from readers to editors because we warn them that the article is low quality. Polygnotus (talk) 18:11, 9 December 2023 (UTC)

I'm not sure whether there is a presumption non-notability – you might well be right about that – but according to the guidelines, there's supposed to be a presumption of notability if the real world (NB: not the article) has a handful of decent sources.

I like your suggestion of merging the two templates. The {{sources exist}} template could become a wrapper template (essentially, a template that auto-adds the new parameter). WhatamIdoing (talk) 19:28, 9 December 2023 (UTC)

How do we get this ball rolling? There are probably quite a few ways to achieve this on a technical level and I am not sure what the best way is. Polygnotus (talk) 21:05, 9 December 2023 (UTC)

On a technical level, I don't think it would be difficult. Compare Template:Original research section and Template:Original research; the former passes a parameter to the latter that changes the displayed text. WhatamIdoing (talk) 21:46, 10 December 2023 (UTC)

If sources exist why not use {{Refideas}} instead, that would include details of what you found. -- LCU ActivelyDisinterested «@» °∆t° 21:00, 9 December 2023 (UTC)

Exactly, the template is just facilitating drive-by tagging. No effort has been made to clean up the mess. You can use refideas or just stick some sources on the talkpage. In a deletion discussion the moderator will certainly check the talkpage before making a decision. Polygnotus (talk) 21:05, 9 December 2023 (UTC)

Spot checking some of the articles it's used on, I wonder if some of these are editors adding the wrong template. As far as I can tell[2] there are less than two thousand articles with this template, that's in the realms manual cleanup. -- LCU ActivelyDisinterested «@» °∆t° 21:18, 9 December 2023 (UTC)

I have to wholeheartedly agree with "use {{Refideas}} instead" and "[{{sources exist}}] is just facilitating drive-by tagging" by either PoV pushers or cluebags.  — SMcCandlish ☏ ¢ 😼  02:56, 11 December 2023 (UTC)

I would say editors using the incorrect template, rather than cluebags, but yeah Refideas sounds more useful. -- LCU ActivelyDisinterested «@» °∆t° 15:26, 11 December 2023 (UTC)

Yeah, well, see other discussion going on at this page. >;-)  — SMcCandlish ☏ ¢ 😼  21:48, 11 December 2023 (UTC)

Thinking about this structurally the subject tag is a few things:

1. A useful piece of information that sources probably exist to establish wp:notability. And an implied statement that those sources are not in the article. "Probably" because it's only as good as the reliability and objectivity of the person who placed the tag, but for usage of this tag both are probably likely
2. Ramdom automated commentary, not the subject of the post that the article might have wp:Ver, wp:or and completeness problems. Maybe true, but random automated commentary that is not the subject of the post is not very informative. It's also "off target" of the post which is about wp:notability, not wp:ver, wp:nor or completeness.

I have no strong opinion on what should happen, but thought these notes might be useful.

Sincerely, North8000 (talk) 23:28, 10 December 2023 (UTC)