Archive 75

Archive 76

Archive 77

Archive 78

Archive 79

Archive 80

So I'm editing Tiutchev (horse) to add citations. But when I insert a citation, it looks like the source I'm using might be interpreted as providing a source for more than it does, when it doesn't. Previously, I would have inserted a ^{[citation needed]} behind the information that had been uncited to let other editors know this, but now obviously I don't want to do that because it'll be interpreted as me challenging very reasonable and likely perfectly true content.

Is there any way for me to signal to other editors that the citation I'm adding isn't actually a source for everything that precedes it, but without indicating I'm challenging perfectly reasonable and likely true content? I've bannered the article for more citations needed, but that still doesn't tell other editors which content is uncited. Valereee (talk) 13:25, 12 November 2022 (UTC)

Use the "reason" parameter in the CN tag? - Donald Albury 18:16, 12 November 2022 (UTC)

@Donald Albury, but that would still be interpreted as me challenging the information, wouldn't it? Valereee (talk) 19:22, 12 November 2022 (UTC)

I think something like {{citation needed|date=November 12, 2022|reason=Preceding content is not supported by the citation at the end of the paragraph.}} would be neutral enough. Or, maybe a hidden note such as <!-- Preceding content is not supported by the citation at the end of the paragraph. --> would work. I know either way would be more work, but if some of us start doing something like that, it might help improve the encyclopedia. Donald Albury 23:09, 12 November 2022 (UTC)

I don't mind the more work (although sometimes, as in this case, I'm not interested in doing an exhaustive search to find sources for every statement) but I really don't want other editors to take my cn as a challenge and remove perfectly reasonable content, and I'm not sure a reason parameter is enough unless I put in something like "reason=THIS IS NOT A CHALLENGE TO THIS CONTENT. THIS IS A SIMPLE NOTE THAT THIS CONTENT IS NOT INCLUDED IN THE NEXT APPEARING SOURCE." Which seems like it would be outside policy, and in any case could simply be ignored. It's a pickle. Valereee (talk) 19:51, 13 November 2022 (UTC)

Break the paragraph so that the material ahead of your addition so that your source is disconnected from the olde info? Carter (Tcr25) (talk) 18:22, 12 November 2022 (UTC)

@Valereee, I think you might want to look at Template:Ref supports and Template:Ref supports2, and for the opposite case,Template:Citation needed span. WhatamIdoing (talk) 19:14, 12 November 2022 (UTC)

@WhatamIdoing, I wouldn't want to use CNspan, as that could be interpreted as a challenge. Maybe Ref supports might work...I'll have to look into how refsuppport is interpreted, thanks! Valereee (talk) 19:20, 12 November 2022 (UTC)

Carter, I tried to do a bit of that, but some of what I was citing was just a person's name -- the source listed all the trainers, so I dropped the citation behind each of their names. Valereee (talk) 19:21, 12 November 2022 (UTC)

I think Template:Additional_citation_needed can be used for it: "The claim is multi-part and the current citation only sources part of it". PaulT2022 (talk) 20:30, 23 November 2022 (UTC)

Hello. I have a question. A few days ago, a discussion happened in the Impact Wrestling roster talk page. The article is about people signed by the pro wrestling promotion Impact Wrestling. Some of the employees, like referees, the source is their twitters, stating that they work in that promotion (or reading the bio) . While some users said that it falls under "published about themselves" other thinks that isn't enough. Also, it is different if the account is verified or not? So, there are any opinions? --HHH Pedrigree (talk) 00:19, 24 November 2022 (UTC)

HHH Pedrigree, I'm not sure that I understand the question. Which of these situations are being considered?

Alice claims to have a job:

• Alice has a Twitter account. (We are reasonably certain that this is Alice's legitimate personal account.)
• Alice, using her Twitter account, tweets that she has recently been hired by this business as a referee.
• A Wikipedia editor wants to add a sentence (e.g., "Alice works for them") to a Wikipedia article, citing Alice's self-published tweet about her own employment.

Coach claims to have hired Alice:

• Coach has a Twitter account. (We are reasonably certain that this is Coach's legitimate account.)
• Coach, using his Twitter account, tweets that this team has recently hired Alice.
• A Wikipedia editor wants to add a sentence (e.g., "Alice works for them") to a Wikipedia article, citing Coach's self-published tweet about Coach's new employee.

Team claims to have hired Alice:

• The team itself has a Twitter account. (We are reasonably certain that this is the official account for the organization.)
• The team's PR person, using the official corporate Twitter account, tweets that they have recently hired Alice.
• A Wikipedia editor wants to add a sentence (e.g., "Alice works for them") to a Wikipedia article, citing the team's self-published tweet about their new employee.

I think the first question for you to ask the other editors is: Do they have any genuine, legitimate concerns that this information might be factually wrong? The approach we take to possible hoaxes is different from the approach we take to suspected trivia. WhatamIdoing (talk) 00:53, 29 November 2022 (UTC)

@WhatamIdoing: thanks for your answer. I think it's case 1, where a person claims to have a job. In this particular case, for example, source for Allison is her Twitter [1] which reads "1⃣2⃣3⃣❗Based in Memphis TN! As seen on Impact, AEW, GCW & Warrior! Ask me about comics & Magic! 🤓" Also, source for Daniel (verified account) [2] "Pro Wrestling Referee at IMPACT Wrestling | Senior Official at OVW | Host of Ringside Podcast | DM for Bookings | Sports Talk | Dad Jokes 🦓" --HHH Pedrigree (talk) 10:58, 29 November 2022 (UTC)

I don't think you should be approaching this from the POV of verifiability. I think the editors at that list should be asking themselves: if there are no independent reliable sources about the referees in these events, then why should Wikipedia mention them at all? WhatamIdoing (talk) 19:50, 29 November 2022 (UTC)

I remember this coming up regarding Krystal Ball's relationship with Kyle Kulinski. Both have confirmed the relationship on their social media accounts, but there is still no RS covering it. According to the latest reliable sources surrounding her personal life (which are several years old), she is still married to Jonathan Dariyanani. But clearly we can't continue to be publishing information that is outdated and untrue. -- King of ♥ ♦ ♣ ♠ 08:05, 1 December 2022 (UTC)

You can use Wikipedia:Editorial discretion to remove information that you have reasons to believe is out of date. WhatamIdoing (talk) 17:33, 1 December 2022 (UTC)

I see the outright removal of the sentence The onus to achieve consensus for inclusion is on those seeking to include disputed content has already been discussed but without any satisfiable outcome. According to the hatnote hidden on the WP:ONUS redirect page Onus is a Latin word for legal concept of "burden" (as in burden of proof). Should or why shouldn't the sentence be reworded (and preferably improved) using the word "burden" and according to the WP:BURDEN policy, to remove ambiguousness? -Vipz (talk) 00:01, 2 December 2022 (UTC)

• I like the fact that we use different words in these two sections. BURDEN is used when the issue under discussion is Verifiability. ONUS, on the other hand, is used when the issue being discussed is something other than Verifiability (be it NOR, NPOV, or even simple consensus). Using two different words helps us to clarify whether the issue is WP:V, or some other policy. Blueboar (talk) 00:24, 2 December 2022 (UTC)

  If that is the case, the relevant sentence needs a lot of improvement (in clarification and difference to "burden"). I've recently been "onus-ed" on my talk page about needing to achieve consensus for removing (not including) completely unsourced content, whereas WP:BURDEN says it is on the editor who adds or restores material. -Vipz (talk) 00:46, 2 December 2022 (UTC)

  Agree with Blueboar; ONUS and BURDEN have different scopes, and complement each other quite nicely. DFlhb (talk) 13:50, 2 December 2022 (UTC)

• The sentence is meaningless and, read broadly, conflicts with wp:NOCON. Unfortunately, there is no consensus regarding how to fix these two problems. - Butwhatdoiknow (talk) 01:17, 2 December 2022 (UTC)

  Can we all agree ONUS is a conduct policy stuck in a content policy? Kolya Butternut (talk) 01:23, 2 December 2022 (UTC)

  And it has nothing to do with Verifiability! It's about consensus. Kolya Butternut (talk) 01:37, 2 December 2022 (UTC)

  It IS about verifiability: it describes the limits of this policy. It seems quite logical to explain on the WP:V page that WP:V alone does not guarantee inclusion, for example, because the text must also comply with NOR, NPOV, BLP, etc. And it seems quite logical to note that even if a proposed text complies with WP:V and all other content policies it may be not suitable for inclusion (for example, because it is off-topic). That is not about conduct, it is about content. Paul Siebert (talk) 02:44, 2 December 2022 (UTC)

  I agree with Paul. I also believe that it's not the optimal wording. @Vipz, if you have a moment to look up one section, I would very much be interested in hearing your reaction to my little hypothetical story, which is about whether to default to inclusion/exclusion of already cited material. WhatamIdoing (talk) 05:34, 2 December 2022 (UTC)

  I also agree that it makes sense to put something in WP:V noting that verifiable or even verified doesn't mean the content should be in the article. This seems to be a common conflict when new content is added to an article then removed as UNDUE. The person who first added it often argues that it should be in the article because it was verified. I do see how ONUS, as debated above, extends beyond just that limited statement. Springee (talk) 05:38, 2 December 2022 (UTC)

  The problem is that people argue that there is some kind of magic, unknown time limit, when WP:UNDUE content (or any other irrelevant content for that matter) after which nebulous time limit, such content no longer needs to be justified for being in the article, rather just "no one caught this problem until too late, so now it has to stay in". When content is under dispute, it is always best for the disputed content to remain out of the article until the dispute is solved; it is better to be silent than wrong. --Jayron32 13:30, 2 December 2022 (UTC)

  Fully agree on this. I'll also add that it's better to be silent than cause harm (a very real possibility, given how influential Wikipedia is). ONUS is a pretty vital piece of Wikipedia policy on that front. DFlhb (talk) 14:07, 2 December 2022 (UTC)

  While it's fine to *note* this under WP:V, it doesn't make sense to cite WP:V to remove verifiable material which causes harm. The last sentence of WP:ONUS is merely reiterating other policies, and it is those other policies which should be cited when removing verifiable material. Kolya Butternut (talk) 19:11, 3 December 2022

• @Butwhatdoiknow: I don't understand why ONUS contradicts to NOCON: the latter is dealing with the actions that do not contradict to other policies. Thus, consider a situation when Alice wants to add a picture A instead of the picture B. Both A and B are under CC, they are neutral, they are based on a single reliable source X, and contain no questionable information about any living person (so they comply with NFCC, V, NOR, NPOV, and BLP). However, Bob prefers the old picture B. This is a situation where NOCON takes effect: if there is no consensus to put A instead of B, B stays.

  Now consider another situation: Alice proposes the picture A, which is not free, whereas the picture B is under CC (or A was made based on some unpublished source, or it is a synthesis of several sources). Can Alice achieve a consensus to add A? No.

  Similarly, if Alice propose the picture A (which is under CC or GPL) instead of the picture B (which is non-free), can Bob object to that citing NOCON? No, WP:V, WP:NPOV and other content policies are non-negotiable, so if there is a serious reason to believe that some content does not comply with one of our core content policies, and some free version exists, the non-free content should be replaced with the free one, and no consensus can overrule it.

  Therefore, if ONUS contradicts to NOCON, that is just a perceived contradiction: NOCON says that if we have a choice between two options, an old and a new texts, and both these options are allowed by our policy then if no consensus is achieved to make a change, the old text stays.

  That means we are allowed to write whatever we want in WP:V, for NOCON regulates the cases that are not covered by WP:V. Paul Siebert (talk) 02:35, 2 December 2022 (UTC)

  I said that, "read broadly," ONUS conflicts with NOCON. Read broadly, ONUS says that when a discussion about existing content results in no consensus, the content is removed. NOCON says the content stays. That looks like a conflict to me. - Butwhatdoiknow (talk) 03:53, 2 December 2022 (UTC)

  Even when read broadly, it depends on the situation. Case in point, Alice adds a verifiable statement, Bob reverts the change. If Alice is unable to attain consensus, both policies support removal. There is no conflict. On the other hand, if Alice removes longstanding material, and Bob restores it, then there could be a conflict if neither can gain consensus. BURDEN would favor Alice's removal, while NOCON would favor Bob's reversion to the old text. --GoneIn60 (talk) 05:36, 2 December 2022 (UTC)

  @Paul Siebert, they conflict only under a fairly narrow set of circumstances:

  • Alice adds cited information to an article a very long time ago.
  • Bob removes that information.
  • There is an intractable, unresolvable dispute that goes through many forms of dispute resolution, and the result is no consensus.
  • Alice says "Ha! I win! NOCON says that if we have a choice between two options, an old and a new texts, and both these options are allowed by our policy (which they are), then if no consensus is achieved to make a change, my old text stays!"
  • Bob objects: "No, I win! ONUS says that you have to achieve consensus for inclusion, and you didn't, so I get to remove it!"

  If you don't have this particular combination (the material is already cited, so BURDEN doesn't apply; the material is old, so QUO can be credibly claimed; the result of discussions is evenly divided, so neither side can claim consensus [this includes compliance with other policies, because there's always a consensus against copyvios and spam and such obvious policy violations]), then the policies get along quite comfortably. I suspect that the rarity of this scenario explains why it's not a frequent source of disputes. WhatamIdoing (talk) 05:42, 2 December 2022 (UTC)

  wp:QUO only applies "during a dispute discussion." - Butwhatdoiknow (talk) 06:02, 2 December 2022 (UTC)

  Well, QUO says that now. It didn't say that back in the day. WhatamIdoing (talk) 06:49, 2 December 2022 (UTC)

In my opinion, if we assume that WP:V is non-negotiable in the same sense as NPOV, there is no conflict between ONUS and NOCON. Indeed, WP:CON is just one policy, and each policy has its own domain, thereby complementing each other. Thus, we cannot keep some content that is verifiable, but not neutral: if some text is not neutral, it cannot stay in Wikipedia, and it doesn't matter whether it is verifiable or not.

Similarly, the CON preamble ("Decisions on Wikipedia are primarily made by consensus") clearly refers to the cases that are not covered by other policies. Clearly, the "decisions" it is talking about are not ALL decisions, but only those decisions that are allowed by other policies. In other words,

• Consensus may be achieved about replacement of the picture A with the pictures B when they both are under some public license. This question is a matter of taste, so if different editors have different opinia, they should negotiate and come to some consensus. However, if a picture B is non-free, and the picture A is under CC (and it is a free equivalent of B), consensus cannot be achieved that usage of B is ok.

I see no difference between ONUS and, e.g. REDFLAG, BURDEN, etc: if we assume ONUS contradicts to NOCON, REDFLAG contradicts to it too. Meanwhile, if I see some statement that stayed in some article for 10 years, and that statement is obviously outstanding, and it is supported by a single obscure source, I have a right to remove it, an no references to NOCON can prevent me from that: our policy prohibits outstanding statements that lack proper support, and, unless convincing evidences have been presented that that statement is not outstanding or it is supported by multiple sources, no references to NOCON can prevent its removal.

To summarise: CON explains how decisions are made in a situation that are not clearly explained in other policies. If some other policy (V, NPOV, BLP etc) clearly define what we should do, we follow what it say. If ONUS says that unless a consensus was achieved to include/restore some information, it should be removed, than we should follow that rule. If NOCON says otherwise, then, clearly, this case is one of exceptions (like BLP violations etc) not covered by NOCON.

Actually, it seems that not ONUS, but NOCON should be modified to bring it in accordance with WP:V (the core content policy). I propose to exclude the word "remove" from the sentence:

"In discussions of proposals to add, modify, or remove material in articles, a lack of consensus commonly results in retaining the version of the article as it was prior to the proposal or bold edit."

If we change it to:

"In discussions of proposals to add or modify material in articles, a lack of consensus commonly results in retaining the version of the article as it was prior to the proposal or bold edit."

there will be no contradiction between NOCON and ONUS. In addition, as I already said, removal of the content should be somewhat easier than addition. That is a necessary condition for maintaining a good quality of Wikipedia texts and their improvement.

--Paul Siebert (talk) 17:33, 2 December 2022 (UTC)

You don't necessarily have a right to remove longstanding material if another editor disagrees with you that the material violates NPOV or any other policy. They would argue that the material had achieved implicit consensus and that the material should stay in pending dispute resolution. Kolya Butternut (talk) 17:51, 2 December 2022 (UTC)

• I think this is something that should be clarified somehow, ONUS implies it should come out whereas in practice, the material tends to stay in until it is forced out. CONLEVEL is also relevant (or should be). Selfstudier (talk) 17:59, 2 December 2022 (UTC)
• IMPLICIT CONSENSUS says "An edit has presumed consensus until it is disputed or reverted" (bold added). Any argument "they" would make to the contrary is not supported by the actual policy. - Butwhatdoiknow (talk) Butwhatdoiknow (talk) 18:08, 2 December 2022 (UTC)

• Yes, once it's challenged you have to figure out whether there's still consensus for it. Wikipedia:Silence and consensus discusses this. Kolya Butternut (talk) 18:14, 2 December 2022 (UTC)

Again, in practice, you get an edit war, multiple reverts and finally an RFC if people dig their heels in, which they frequently do. By CONLEVEL, I mean with reference to the initial introduction of material whenever that might have been, there is a difference between material added with a reasonable consensus and a quiet edit with no active consensus behind it. Selfstudier (talk) 18:16, 2 December 2022 (UTC)

And pending dispute resolution, material with implicit consensus remains in the article, correct? Kolya Butternut (talk) 18:30, 2 December 2022 (UTC)

That's what wp:QUO recommends. - Butwhatdoiknow (talk) 18:38, 2 December 2022 (UTC)

That's what seems to happen in practice although it might vary case to case. Selfstudier (talk) 18:38, 2 December 2022 (UTC)

Quo is certainly what used to happen in practice… but I am not sure that it is still what happens in practice. I think we have shifted somewhat (case by case, but increasingly practice has been shifting towards “omit”). Blueboar (talk) 02:08, 3 December 2022 (UTC)

Sorry, it's so long, I must have lost track...where we did decide to remove The onus to achieve consensus for inclusion is on those seeking to include disputed content.

Because I'm still not comfortable with that. valereee (talk) 23:16, 22 July 2022 (UTC)

As far as I can tell, there's not a consensus for removing it, there's been a long inconclusive discussion about it thats's been going on this talkpage for months now. I am not convinced that the relatively small number of participants to the discussion here are reflective of the broader Wikipedia community. I don't see this being resolved without a well-publicised RfC with a simple binary choice to keep or remove it. Hemiauchenia (talk) 23:20, 22 July 2022 (UTC)

Thanks, @Hemiauchenia. I was going back and forth on the diffs and I was like, 'what did I miss?' valereee (talk) 23:24, 22 July 2022 (UTC)

(edit conflict) Indeed. With this much of a flurry of edits going on, I'd expect blocks based on seeing similar or even lesser occurrences on other policy pages.

I've gone ahead and restored the policy section to the last status quo [3], and any additional changes really need consensus at this point. Propose changes here, but don't go making them to the policy itself like an article. KoA (talk) 23:27, 22 July 2022 (UTC)

Do normal editing policies apply to policy pages? If so, #The_meaning_of_the_CURRENT_text_at_ONUS_(introduction) meaning #13 states "it's your job to prove there is an active, positive agreement to include your material in that article." The alternative meanings for #15 seem to indication that the onus sentence is redundant with "Consensus may determine that certain information does not improve an article", which would mean it could be removed. Kolya Butternut (talk) 23:35, 22 July 2022 (UTC)

Policy pages are even more stringent. Especially for anything contested like this, and where all editors on the talk page would already know the change is controversial, there is really no excuse to be editing the policy directly like what happened recently. KoA (talk) 04:04, 23 July 2022 (UTC)

KoA, where can I find the "policy pages are even more stringent" provision? wp:PGBOLD would suggest there isn't one. (That said, wp:QUO tells us the best practice during discussion is to leave the status quo version in place.) - Butwhatdoiknow (talk) 04:13, 23 July 2022 (UTC)

@Butwhatdoiknow, to me WP:PGCHANGE (which incorporates PGBOLD) says exactly that policy pages are more stringent. It says talking first is usual, that editing boldly is permitted (rather than encouraged), and that folks should follow 1RR or 0RR. That's more stringent than BRD. valereee (talk) 12:47, 23 July 2022 (UTC)

Thank you. Perhaps the problem is the use of the word "stringent" (which does not appear at wp:PGCHANGE). While PGCHANGE encourages caution when editing polices and guidelines, it does not set forth a rule. And, to my mind, recommendations cannot be "stringent." - Butwhatdoiknow (talk) 15:16, 23 July 2022 (UTC)

Okay, the guidelines for making changes to policy pages specify using more caution than the guidelines for making changes to articles? valereee (talk) 15:25, 23 July 2022 (UTC)

Well, now we're splitting hairs. In my mind "encourage" equals "recommend" and "specify" equals "requires." - Butwhatdoiknow (talk) 18:39, 23 July 2022 (UTC)

I do not think it's splitting hairs to point out that there's stronger language re: policy pages. valereee (talk) 18:48, 23 July 2022 (UTC)

• My big problem with all the text above is nobody has said what 'disputed' means. I have no real problem with people trying to achieve consensus over some disputed text, but unfortunately in Wikipedia we have loads of people who say something is 'disputed' because it is different from what was there before, or they just say they dispute it or some other non-reason. If we could get people to state a reason that can be true or false and can be contended to an objective result that would be an immense step forward. This I think could help greatly especially when ogres just dismiss some newbie's contribution with a list of WP:TLAs that each points to thousands of K of text. NadVolum (talk) 07:29, 23 July 2022 (UTC)

This is something I touched on above - I think that a lot of the problems could be resolved by making it clear that once a credible source has been produced, further challenges require an actual policy-based objection related to article content (rather than a procedural one.) It doesn't have to be a strong one or a good one, that's something we settle through discussions or RFCs; "rv, WP:UNDUE" is notionally fine, although you'd be expected to explain in more detail eventually if there was an ongoing dispute. It's also not necessary to continuously repeat your objection once it's been stated (although it doesn't hurt to mention it in the edit summary, just to ensure everyone is on the same page.) But "rv, get consensus per WP:ONUS" when nobody has expressed anything that could remotely be construed as a policy-based objection is inappropriate, and I feel that one of the issues with ONUS is that some people have interpreted it as allowing that. This is part of what I mean when I say that policies that are unnecessarily strongly-worded can impede consensus-building by making it harder to bring everyone to the table. --Aquillion (talk) 09:54, 23 July 2022 (UTC)

That's more an issue of taking things out-of-context: All you have to do is comply with Note 3, and if we want we can repeat Note 3 at the end of ONUS. Alanscottwalker (talk) 11:32, 23 July 2022 (UTC)

I might be getting a bit ahead of things here, don't we have enough in all these discussions to ask in an RFC whether that ONUS sentence should stay? Selfstudier (talk) 11:34, 23 July 2022 (UTC)

Before we have an RFC… I want to suggest an option: rather than simply removing the ONUS sentence, we should consider moving it over to WP:Consensus. I don’t really object to what the ONUS sentence says, but I have never understood why we say it in this policy.

I understand why this policy includes the VNOT sentences (they are a useful reminder that there is more to inclusion than just WP:V). But why does our Verifiability policy outline who must gain a consensus when the dispute is about something other than Verifiability? Shouldn’t that be outlined at WP:Consensus? Blueboar (talk) 13:32, 23 July 2022 (UTC)

That is an idea with which I do not fundamentally disagree and if the thing is to be argued about, then let it be argued about there rather than here. Likely still needs an RFC though. Selfstudier (talk) 13:52, 23 July 2022 (UTC)

@Blueboar, because newer editors often need to see these two policies (V and ONUS) presented together so that they understand that the simple fact we can verify something doesn't mean we include it, and that the person who is insisting we include some trivial fact is the one who needs to gain that consensus. These two policies are interrelated, and if we don't present them together, that's often difficult for newer editors, who are still on a steep learning curve, to see. valereee (talk) 13:53, 23 July 2022 (UTC)

ONUS by itself is just a sentence, not a policy, pretty troublesome one at that. Selfstudier (talk) 13:56, 23 July 2022 (UTC)

Indeed, valeree, and the V policy is unique, in that its focus is on one source, it has practically no regard for context, broad research, or subject matter knowledge. And it's not the case that V has nothing to do with NPOV and NOR, indeed they must (we say, although this is what fails too often) they must be read together. Moreover, the burden of V is very light, "any source [the editor] believes, in good faith, to be sufficient." That's not nearly enough for vouchsafing good content. Alanscottwalker (talk) 14:09, 23 July 2022 (UTC)

The short version of how onus works in the wikipedia system is that it puts a finger on the scale towards exclusion of material when there is a debate. It's also a too-vague version of reinforcing the point of the paragraph that it is in. Which more directly would be: "Verifiability is a requirement for inclusion, not a reason for inclusion". If we want to get rid of it, it would take a big RFC, and to do that right we'd need to do our homework on formulating a good plausible RFC and then supporting it. Which also means dealing with the two current effects of onus. IMO the "finger on the scale towards exclusion" isn't by itself needed or even good, and some would say conflicts with another policy, but IMO the latter reminder is much needed. Even with the (weak) reminder we have the persistent urban legend is that merely being sourced is an argument for inclusion. My idea for an RFC (which IMO would fix everything in one fell swoop) would be:

Replace: "The onus to achieve consensus for inclusion is on those seeking to include disputed content" with "Verifiability is a requirement for inclusion, not a reason for inclusion"

Which, BTW, is pretty neutral on the overall "inclusionist vs. exclusionist" balance, which many folks care about. Sincerely, North8000 (talk) 13:10, 23 July 2022 (UTC)

It's not about exclusion, its what is needed for inclusion. Ignoring everything else, Verifiability is subject to consensus. We don't want content in which verifiability is in substantial doubt, nor do we want content where a NPOV is in substantial doubt, nor OR, nor etc. Alanscottwalker (talk) 13:36, 23 July 2022 (UTC)

But isn’t all that covered by the VNOT part of the paragraph? Blueboar (talk) 15:02, 23 July 2022 (UTC)

If an RFC on this matter is held? Would the Village Pump be the proper place for it? GoodDay (talk) 14:08, 23 July 2022 (UTC)

It's not a new policy as such, just an amendment to an existing one (or two if we are to transfer the subject to another). Wouldn't here + WP:CENT be sufficient? Selfstudier (talk) 14:45, 23 July 2022 (UTC)

For something like this, high participation is more important than the specific venue… So wherever we hold it, we should put very prominent notices in as many different locations as we can… maximize the number of participants and minimize the number of people who will later complain that they never knew about it (of course, some will… no matter what we do). Blueboar (talk) 14:59, 23 July 2022 (UTC)

So we have something like this so far?

Should the sentence The onus to achieve consensus for inclusion is on those seeking to include disputed content

1) Be replaced with the sentence "Verifiability is a requirement for inclusion, not a reason for inclusion" or

2) Be moved to an appropriate place at WP:CONSENSUS or

3) Stay as is. Selfstudier (talk) 15:40, 23 July 2022 (UTC)

In this proposed RfC are options 1 and 2 mutually exclusive? You could in theory modify the text here to include the "requirement for inclusion" rewording, and put the current onus text in consensus. If they are not mutually exclusive, then I agree with what North8000 says below in that it would complicate the RfC and the closer's job. Sideswipe9th (talk) 16:10, 23 July 2022 (UTC)

We may want to strike "appropriate" as non-neutral phrasing. --GoneIn60 (talk) 16:27, 23 July 2022 (UTC)

I think "appropriate" in this context means "an appropriate place within the text of WP:CONSENSUS" and not "WP:CONSENSUS is the appropriate place to put this text". Sideswipe9th (talk) 16:36, 23 July 2022 (UTC)

For me there's a fourth possibility -- if ONUS redirects to CONSENSUS, keep language (with a link to the new location) at V. For me that would (mostly) addresses the issue of needing this language to be at V. valereee (talk) 16:52, 23 July 2022 (UTC)

Can we move the ONUS wikilink to WP:CONACHIEVE? Nothing about onus is stated there, but it's implied you have to engage in the consensus-building process? Kolya Butternut (talk) 18:21, 23 July 2022 (UTC)

Having three or more choices makes it pretty complicated. IMO we should hammer out a proposed change here and then do a "support/oppose" RFC. And IMO, an RFC at the place involved (here) plus listing at centralized discussion would be OK & best. North8000 (talk) 16:09, 23 July 2022 (UTC)

Concerning the proposed RFC. You'll find it would become messy, if we ignore the fact, that it's mostly the majority who establishes consensus in content disputes. If enough editors argue that Orange is Green? the result will be Orange is Green. GoodDay (talk) 16:16, 23 July 2022 (UTC)

Is that concern not inherently covered by the overall WP:CONSENSUS policy and specifically WP:DETCON? DETCON already states that it's not the majority, but that it's the quality of the arguments given on the various sides of an issue, as viewed through the lens of Wikipedia policy which determines the consensus outcome of a discussion. I'm not sure we need to duplicate that here. Sideswipe9th (talk) 16:20, 23 July 2022 (UTC)

Those are the words, yes. But putting them into practice is usually messy, if the position of the minority is chosen. GoodDay (talk) 16:23, 23 July 2022 (UTC)

As written, DETCON is arrant nonsense that nobody actually believes, since it would give the self-selected closer to impose their own view of the "quality of the arguments" over the community's. If I went around closing discussions based on DETCON I would expect to get desysopped right quick, and rightly so. At best, it is a guide to resolving rare corner cases involving e.g. manipulation or other suspected bad-faith conduct. Outside of those situations, if you need to ask yourself how to determine if there's consensus, there probably isn't one. -- Visviva (talk) 21:21, 24 July 2022 (UTC)

I'm not sure if folks were responding to my "makes it pretty complicated" but what I meant is that with having 3 or more choices can divide up the support of persons with similar views. Also that they usually a blend of 2 or more questions. North8000 (talk) 19:40, 23 July 2022 (UTC)

• I think this may need more than one RFC… but the initial question could be asked with just 4 choices:

1. Cut the ONUS sentence
2. Move the ONUS sentence to another page (such as WP:Consensus)
3. Edit the ONUS sentence
4. Leave as is.

Those are the “basic” options we have been discussing. Blueboar (talk) 21:39, 23 July 2022 (UTC)

5. Leave the ONUS sentence and move the ONUS shortcut to WP:Consensus.

Kolya Butternut (talk) 15:44, 24 July 2022 (UTC)

§ Removing ONUS...

The right question to ask here isn't "where we did decide to remove The onus to achieve consensus for inclusion is on those seeking to include disputed content." but rather where did we decide to add it?

As discussed above the current statement under dispute: The onus to achieve consensus for inclusion is on those seeking to include disputed content.

is materially identical to the statement which was BOLDLY added 13 August 2014: The onus is on those seeking to include disputed content, to achieve consensus for its inclusion.

Clearly from this extended discussion The onus to achieve consensus for inclusion is on those seeking to include disputed content. is itself disputed content and therefore the onus to achieve consensus for inclusion is on those seeking to include this disputed content.

So any multiple choice RfC on this matter is out-of-order and premature. We need a binary-choice RfC with yes/no or keep/remove the disputed statement The onus to achieve consensus for inclusion is on those seeking to include disputed content. which is necessary per the policy to achieve consensus for its inclusion.

Questions about moving or editing the sentence should only happen after the binary-choice RfC finds a consensus to keep it at all. – wbm1058 (talk) 03:14, 24 July 2022 (UTC)

Good point, but for several reasons I don't think that that will work. First, many would not accept that retaining something that has prominently been around that long requires a consensus. Also, folks like me think that it does a (albeit weakly and poorly) does an important job for that paragraph and so simple removal would have another headwind there. I think that it would more useful to strongly make/remember the point that it was added without consensus. Sincerely, North8000 (talk) 14:40, 24 July 2022 (UTC)

WP:CCC (silent or otherwise) and I think we are paddling around in that pool at this point. The question is how to make progress? I get the desire to have a yes/no scenario, tend to agree that might be insufficient, then again, four/five choices might well result in 25/20% each so how to arrange it? Selfstudier (talk) 15:07, 24 July 2022 (UTC)

How to move forward. The problem is that there are multiple questions (content and location) and infinite possibilities. Let's start by ruling out the unlikely and unfeasible one which is moving it to another policy. That would mean changing two policies at once and getting a different policy to accept our problem-child phrase. :-) So now we're down to potential changes at this policy... The three choices there are status quo, delete or change. And under "change" we still have infinite possibilities. So, let's discuss to decide best one or 2 to offer under "change". And then offer the three choices, but ASK EVERYBODY TO GIVE THEIR OPINION ON EACH ONE OF THEM. This will avoid the math problem associated with having 3 or more choices. North8000 (talk) 20:36, 24 July 2022 (UTC)

The easiest solution isn't necessarily the correct one. I think this debate is really about IMPLICITCONSENSUS, and we just need to make clear that the onus sentence reiterates CONSENSUS. This could be achieved through option #5, and then our discussion would move on to IMPLICITCONSENSUS. Kolya Butternut (talk) 20:54, 24 July 2022 (UTC)

It could be the case that there is a relatively silent majority out there thinking that there should be no change. Perhaps we ought to first test that hypothesis. That is a straight choice between no change and some change (to be determined if there is a consensus for it). Selfstudier (talk) 21:40, 24 July 2022 (UTC)

@Selfstudier, that doesn't usually go well. People feel like they're being asked to buy a pig in a poke. They fear that a vote to change it will later be interpreted as a vote to change it to something they disagree with. WhatamIdoing (talk) 19:55, 25 July 2022 (UTC)

I wouldn't feel that way but maybe I'm all alone in that. Anyway, just putting it out there, trying to pin down what an RFC should look like if it is not to result in nocon (=no progress, come back later). Tbh, I was hoping some silent folk might comment, guess not. Selfstudier (talk) 21:23, 25 July 2022 (UTC)

It's not the "easiest" way, it's just A way which is pretty good and following in general what people think should happen. Vs. it being eternally heading towards no resolution process. North8000 (talk) 22:09, 24 July 2022 (UTC)

OK I might boldly put that structure forward in a week. For that we need two change proposals. I have one that I suggested. Does somebody have another 1 or 2 to float? North8000 (talk) 11:37, 25 July 2022 (UTC)

Arbitrary break for discussion of draft

Could you show an outline draft of the RFC (with a space for the missing change proposal). Pretty please. Selfstudier (talk) 11:57, 25 July 2022 (UTC)

Responding to @Selfstudier: here is a draft of the RFC, proposing putting it up on August 2nd:

This RFC is about the existing sentence in wp:verifiability which reads: "The onus to achieve consensus for inclusion is on those seeking to include disputed content" Which is often called "onus". The discussion has already been lengthy, here are 5 possibilities. To avoid math problems associated with more than two choices, please give your input on EACH of the 5 possibilities where you have an opinion.

1. No change. Keep the status quo.
2. Simply delete: "The onus to achieve consensus for inclusion is on those seeking to include disputed content"
3. Make the following change: Replace: "The onus to achieve consensus for inclusion is on those seeking to include disputed content" with "Verifiability is a requirement for inclusion, not a reason for inclusion"
4. Make the following change: (additional change proposal to be worked out by discussion here)
5. No immediate change. Have an additional RFC to consider other change ideas

Sincerely, North8000 (talk) (7/26/22)

Being a lil bit bureaucrat, what are we going to call RFCbefore (for newcomers), it's most of this page, ha, from "1." on down? Selfstudier (talk) 18:44, 26 July 2022 (UTC)

For the blank, how about Replace "Consensus may determine that certain information does not improve an article. Such information should be omitted or presented instead in a different article. The onus to achieve consensus for inclusion is on those seeking to include disputed content." with "Verifiability is a requirement for inclusion, not a reason for inclusion" (the thing I most dislike about that little para is the segue from V to talking about consensus). Selfstudier (talk) 18:52, 26 July 2022 (UTC)

Sounds good to me. North8000 (talk) 20:20, 26 July 2022 (UTC)

You can add it as an option, but it would be the only option I would actively oppose. It cuts a key part of VNOT. A consensus that X does not improve an article and should be omitted (even though X is verifiable) is one of the key reasons why “Verifiability does not guarantee inclusion”. It directly relates to the section (unlike the ONUS sentence, which is tangential). Blueboar (talk) 00:27, 27 July 2022 (UTC)

Side note… the above just made me realize why I have always found the ONUS sentence so jarring: the VNOT sentences all discuss what might happen DUE TO a consensus to omit verifiable information (ie when there IS a consensus) … ONUS suddenly shifts to discussing what happens when there ISN’T a consensus. Blueboar (talk) 00:57, 27 July 2022 (UTC)

Those additional two existing sentences weren't on my worry list. "Consensus may determine that certain information does not improve an article" is vague and structurally says little-to-nothing (a consensus against an addition can keep the material out...duh) but in the fuzzy Wikipedia world it does add emphasis to the fact that that it also has to be a good idea to make an addition. IMO "Such information should be omitted or presented instead in a different article." is silly but is harmless. (That if there is a consensus against inclusion of material the rejected material should go into another article.) But it never gets used so it's harmless. But we're here to hammer out an additional proposal, and your discussion is a part of that. Sincerely, North8000 (talk) 01:39, 27 July 2022 (UTC)

I'm not married to my suggestion for the blank space, if there is a better something to put in there, fine by me. Selfstudier (talk) 13:07, 27 July 2022 (UTC)

@Selfstudier: Would you like to have your idea listed in the RFC? Please ping me with an answer if so. Sincerely, North8000 (talk) 18:19, 8 August 2022 (UTC)

@North8000: I leave it to you to decide, I can always refer to it in my comments if needs be. Selfstudier (talk) 18:33, 8 August 2022 (UTC)

That's what we have so far. As noted on 7/26, I plan to put the RFC up on August 2nd. North8000 (talk) 16:28, 28 July 2022 (UTC)

I think an RfC should clearly define the problems and how the proposals would address them. Kolya Butternut (talk) 18:02, 28 July 2022 (UTC)

Surely the RFC in whatever form it takes is an attempt to do just that. The raft of RFCbefore on the page would indicate that even though editors seem to be in some agreement that there is a problem, they are not able to agree on what the problem is exactly. With a bit of luck, some progress can be made. Selfstudier (talk) 18:16, 28 July 2022 (UTC)

Kolya Butternut What you are asking for would make the RFC non-neutral. But I would plan to do that when I weigh in as a participant. North8000 (talk) 13:27, 29 July 2022 (UTC)

I do not think this RFC, as written, will be productive. I think it will generate more heat than light. I do not think you should start this RFC next week. I think you should consider alternative questions, and I think you should consider asking for help with drafting a question at Wikipedia talk:Requests for comment.

As one example of a different question to ask (one that will be informative, but not directly and immediately settle the question about what to do with this sentence), we could ask:

"When editors are at an impasse over sourced content (e.g., some editors do not believe it belongs in this article, and others do), should the disputed material be retained or omitted?"

Also, I'd like us to spend a while with the point raised by @Blueboar: Maybe we need to add something to this subsection to explain why this is here. Consider, e.g., the old wording:

While information must be verifiable for inclusion in an article, not all verifiable information must be included. Consensus may determine that certain information does not improve an article. Such information should be omitted or presented instead in a different article. The onus to achieve consensus for inclusion is on those seeking to include disputed content.

versus this:

While information must be verifiable for inclusion in an article, not all verifiable information must be included. Sometimes editors will agree that certain information should be omitted or moved to a different article. Other times, editors are not able to agree. The editors who want to include the disputed material are responsible for forming a consensus to include the disputed content. If they are unable to reach an agreement to include it, the disputed material should be omitted from the article. It is not enough to have a reliable source for the material; you must also have other editors' agreement that the material should be included.

I know that some editors prefer QUO to this rule, but at least this suggested version is clear about what the rule is and how it is relevant to VNOT. WhatamIdoing (talk) 21:00, 29 July 2022 (UTC)

Regarding the RFC, I'll answer from a couple of different angles. The framework is designed to handle multiple ideas while avoiding the math problems that usually come from having 3 or more choices. Also, I try for the "middle of the road" between something that is not really going to go anywhere and being too rushed, the happy medium between zero help making this happen and going too far and fast. IMO it would be a good idea to delay my August 2nd date about another week until August 9 to see if another proposal gels that at least a couple of people like and then if so we put that one into it as well. Sincerely, North8000 (talk) 12:51, 2 August 2022 (UTC)

Be cautious of RFCs with more then three options. Those tend to end in 'no consensus'. Anyways, looking forward to participating in it. I'm assuming it'll be held at this talkpage. GoodDay (talk) 13:39, 2 August 2022 (UTC)

The focus of North's proposed RFC is "make a decision abut wording". I wonder if we need to "gather information" or "make a decision about the best practice" instead.

Part of the reason that I am interested in other approaches is because we have a problem. This problem is not named/described in the proposed RFC question, which makes it difficult for editors to figure out what we're trying to achieve with these five options. Also, if the most popular result is #1, then we will still have the same problem.

Separately, North's line that "Verifiability is a requirement for inclusion, not a reason for inclusion" sounds like a slogan, not a rule. I agree with the sentiment, but if you want more editors to understand it, then I'd suggest wording like "Verifiability is a necessary but not sufficient requirement for including any material in an article. The fact that material is verifiable does not mean that the material should be included. All material in all articles must also be relevant to the subject of the article, appropriate for an encyclopedia, and support an overall neutral point of view for the article." (I suggest that last phrase because each individual sentence does not need to be fully NPOV in isolation. For example "Alice says X" is 100% focused on the POV of one source, but it can contribute to an NPOV article when it is combined with "Bob says Y".) WhatamIdoing (talk) 16:13, 3 August 2022 (UTC)

@WhatamIdoing: IMO it just stops a widespread mis-impression which launches from wp:ver. (IMO, from it's location, something which wp:Onus (badly) attempted to do). The sentence doesn't without trying to say what should happen.....that gets determined by Wikipedia:How_editing_decisions_are_made which is based on multiple policies, guidelines. IMO you are trying to describe that overall process in one section of the Verifiability policy which to me is impossible to do well. That said, you've mentioned 2 ideas. Maybe a shorter broader version of what you are getting at would be good. Is there on which you would like to have listed in the RFC? North8000 (talk) 18:22, 8 August 2022 (UTC)

I think the wording Verifiability is a necessary but not sufficient requirement for inclusion would be better for option 3. Kolya Butternut (talk) 14:00, 12 August 2022 (UTC)

I'll put that in as one of the choices. Regarding other choices, @Selfstudier: said they were neutral about whether or not I put their idea in. Given that, and that their idea involves broadening the question to eliminate additional text that we're not been discussing, I don't plan to put it in. I pinged @WhatamIdoing: (see one post up) to see if they have one to put in; if so I'll put that one in also. Again, asking people to give thier opinion on all of them will reduce the math problem of having more than two choices. North8000 (talk) 16:03, 12 August 2022 (UTC)

You feel "not a reason for inclusion" is better than "is necessary but not sufficient"? It would be best to have fewer options, but I know you've been advocating for that language for a long time. We could provide a rank-choice analysis for the closer, like I did for the lead image of Woman at Talk:Woman/Archive_19#Instant-runoff_method. Kolya Butternut (talk) 16:48, 12 August 2022 (UTC)

IMO ranked choice is pretty clearly voting, which sort of goes against the "not a vote" concept. IMO saying "give your opinion on every item" gets the important benefits of ranked choice without have that issue. North8000 (talk) 19:31, 12 August 2022 (UTC)

I think ranking is no more voting than writing "support" or "oppose"; it's just a way to indicate relative support, and this ranking should not be given without arguments. Kolya Butternut (talk) 19:58, 12 August 2022 (UTC)

I think that people might be interested in adding one of the not-a-reason/necessary-but-not-sufficient sentences without removing the only "what to do if someone stuffs in verifiable trivia" sentence. Consequently, Option 3 asks two separate questions. To support that option, you have to support two things:

• Remove this, and
• Add that

The "Remove this" question is redundant to Option 2. It should probably be left out of Option 3, which would become (only) "Add that".

I think you could re-organize this list (which I still think is premature, because people don't know what the actual problem is) as these options:

1. Do you think we should remove the ONUS sentence?
2. Do you think we should add <North's new slogan>?
3. Do you think we should add <WhatamIdoing's long thing>?
4. Do you have other concerns or ideas about this section that you would like to share?

By asking four separate questions, we lose the need for a "no change" option, as "no change" can be achieved automatically/implicitly by answering all of the questions "no". It also merges the original 4 ("make another change") and 5 (have another discussion first), because both of those views can be reported as "other concerns or ideas".

Also, each of the first three questions can be evaluated separately. There will be no need for ranked-choice voting or anything else, except to the extent that someone might say "first choice" or "second choice" for the middle questions.

It would probably help to provide people with the entire (four-sentence) ONUS paragraph, because most people won't click over to read it themselves. It would definitely help if there were a description or story about the problem that needs to be solved, which is that ONUS gives the advantage to article-blankers and NOCON gives the advantage to change-opposers, and this is not always the same editor. WhatamIdoing (talk) 19:53, 12 August 2022 (UTC)

That sounds like starting the discussion that we've already been having for the last ~6 months. Plus it wanders away form the "what to do with the onus" question. Those other ideas dealt with it because they are in essence "replace the onus sentence with...." How about you write up the WhatAmIDoing idea (to substitute for the ONUS sentence) and we put that in the RFC? North8000 (talk) 20:56, 12 August 2022 (UTC)

• Do you want to remove A?
• How about we remove A and insert B?
• How about we remove A and insert C?

What do you expect people to do, if they want the paragraph to include A+B+C? What will you think if someone says "No, don't remove A, but yes, do remove A and insert B"? Editors who want A retained will say that pair of responses should be thrown out as illogical nonsense. Editors who want A removed will say that this pair of responses means that A should be removed. Framing the questions that way will lead to avoidable disputes. WhatamIdoing (talk) 22:48, 13 August 2022 (UTC)

I plan to put the above RFC up soon. Including asking people to respond on all options that they have an opinion on to avoid math issues from having more than two options. Beside the "keep as is" option, it has one for "no change now but consider more ideas". I've been asking for a while, but if anyone has another specific option to include, please provide it. North8000 (talk) 19:33, 16 August 2022 (UTC)

Muse...delete the whole lot, everything under "Verifiability does not guarantee inclusion" :) Like, yeah, we know that, don't we? Might as well put it up, see what happens. Worst case is we have to do another one. Selfstudier (talk) 21:48, 16 August 2022 (UTC)

So, what does the RfC draft look like now? I like "necessary but not sufficient", but not the rest of WhatamIdoing's "long thing". Kolya Butternut (talk) 23:45, 17 August 2022 (UTC)

Sorry so slow. Got buried in real life. Basically, same as the original except putting in @WhatamIdoing:'s & @Kolya Butternut:'s ideas in as additional choices. I'll put a final draft up tomorrow. Sincerely, North8000 (talk) 15:47, 2 September 2022 (UTC)

@North8000, I'm not paying close attention to my watch right now, would you please ping me from wherever you put it up? Thanks! Valereee (talk) 17:04, 2 September 2022 (UTC)

Update on ONUS RFC

There was an immense discussion on this important point and I think it would be good to bring it to a conclusion. Also, I think that it would be difficult to make a good RFC which gives several good choices without incurring the usual math problems that normally come with those. I've been saying that I would get one done but real life came and buried me from several directions. Now, I'm going top be completely off the grid (not even cell coverage) for 14 of the next 23 days and so I would do a bad job on following through on anything that I presented. I'd still be happy to do it but would/will need a few weeks. Sincerely, North8000 (talk) 19:04, 26 September 2022 (UTC)

I agree about the math problems, RFC that are too complicated rarely resolve anything, so how about something more reductive in the interim as was previously suggested, with the idea that it may require more than one RFC to finalize?

Maybe something very simple to start with, as in "Should the ONUS sentence remain as is?" which will either eliminate the need for further RFCs or lead to another. Selfstudier (talk) 14:58, 27 September 2022 (UTC)

I think that two stage RFC's are a good idea. They are unusual in Wikipedia so somebody has to promote doing an "unusual" RFC. Also at first glance they look like a more lengthy process although if they are well structured and bring the question to a decisive conclusion then that's a lot faster than eternal conversations. If we did your idea we'd need to clarify that if "change it" is chosen, then the next step would be a careful process and a second RFC. Otherwise "change it" might seem like a blank check to do who knows what. North8000 (talk) 18:08, 27 September 2022 (UTC)

I've been trying to think of a positive example of a two-stage RFC for a "small" question (e.g., something smaller than RFA reform), and I haven't been able to come up with anything.

Wikipedia:Straw polls may be an under-utilized option on wiki. The question to ask IMO is not "Should it remain?" but an informal question like "Just looking for a quick show of hands: Do you think this as unambiguous and actionable as it should be?" with a note afterwards that say "If people think this is already completely clear and functional, then we don't want to do any unnecessary work. But if you've ever found yourself wondering how to apply it, speak up, and we'll propose an RFC about improving it. If we need an RFC, we'll ping everyone who responds here when it opens. (Feel free to provide any thoughts you have now, if you want.)" WhatamIdoing (talk) 20:47, 27 September 2022 (UTC)

The second stage didn't end up being needed, but I think this is a positive example of a two-stage RFC for a "small" question. BilledMammal (talk) 23:44, 27 September 2022 (UTC)

Straw polls can be good pre-RFC guidance, but I also view the discussion (and these was an immense one here) as good pre-RFC guidance. North8000 (talk) 21:57, 27 September 2022 (UTC)

Ask the folks who say it is unambiguous to indicate what it unambiguously says. Past results suggest you'll get more than one answer. - Butwhatdoiknow (talk) 06:08, 28 September 2022 (UTC)

Perhaps instead of asking them what it unambiguously says, we could ask them what is says is to be done with long-standing, adequately sourced text that someone wants to keep and someone else wants to remove. Then half of them can say "It unambiguously says to remove" and the other half can say "It unambiguously says to keep it", and perhaps they will conclude on their own that it is not so unambiguous after all. WhatamIdoing (talk) 01:13, 29 September 2022 (UTC)

Now that the relevant RFC's are closed, two real cases involving onus v quo are Talk:Jerusalem#Should the infobox contain this flag and emblem? (removal of material present since 2010) and Talk:State of Palestine#RFC about de jure (removal of material present since 2016). Onus "won" in both cases but an RFC was necessary (ie consensus). Selfstudier (talk) 09:01, 29 September 2022 (UTC)

If ya'll are gonna have an RFC on this topic? Limit its scope to three options, or you'll never get a consensus on anything. GoodDay (talk) 22:12, 27 September 2022 (UTC)

What's wrong with this RfC? Crossroads ^-talk- 22:37, 27 September 2022 (UTC)

Is there something wrong with it? Doesn't that mean we can't move it to consensus? Which is anyway a different policy and we would have to get it approved somewhere else than here. That's still an option, I guess, but there are other options besides that, right? Selfstudier (talk) 22:48, 27 September 2022 (UTC)

This has been repeatedly explained. In Archive 74 I said that the difference with that RFC was that that proposal was for changes to "longstanding content". Kolya Butternut (talk) 03:08, 10 July 2022 (UTC) Kolya Butternut (talk) 09:24, 28 September 2022 (UTC)

IMO it has the same structural flaw as the current one (trying to dictate what happens based on only one consideration) and also lacks the large amount of good that could be accomplished by one of the discussed ideas. North8000 (talk) 13:14, 28 September 2022 (UTC)

Recapping my OP, I'll be here only sporadically for the next 3 weeks and then back full strength. Sincerely, North8000 (talk) 17:40, 29 September 2022 (UTC)

We going to let this lie? Selfstudier (talk) 10:00, 27 October 2022 (UTC)

I left off that I'd formulate an RFC. Sorry, after I got back on the grid I got real-world buried. Since we spent months and a lot of work talking about this maybe we should try to take it to a conclusion with an RFC? Or should we just let it die? If the former, I'd include the proposal listed above, the status quo and for 2 weeks invite others to create a choice (in final form) for inclusion. Then I'd create an RFC which asks respondents to weigh in on EVERY ONE of them in order to avoid the usual problem of an RFC with more than two choices. What do y'all think? Sincerely, North8000 (talk) 20:15, 29 November 2022 (UTC)

I'm concerned that there isn't even any agreement among participants here about what to do, so I don't have much hope for an RfC. Kolya Butternut (talk) 02:59, 30 November 2022 (UTC)

So much time has passed everyone has forgotten the discussion, I would rather go back to KISS and just ask the question whether something/anything should be done about the ONUS sentence and see what that throws up before trying a detailed choice type RFC. Selfstudier (talk) 11:00, 30 November 2022 (UTC)

I vote 'let it die'; I see no benefit in a change and no reason to go over this again. Crossroads ^-talk- 05:58, 1 December 2022 (UTC)

To be clear, I vote against let it die. Selfstudier probably has the right idea. Kolya Butternut (talk) 11:04, 1 December 2022 (UTC)

As it stands there is a clear contradiction between policies and clear disagreement over what WP:ONUS means, so an RFC seems necessary. There isn't a current agreed-upon interpretation of ONUS to fall back to as the default, so we can't really let it die unless we want to have this argument again every single time anyone cites ONUS in any nontrivial context. --Aquillion (talk) 11:57, 1 December 2022 (UTC)

I think we need to spend less time worrying about the interpretation of existing wording, and more time worrying about what editors think should be the normal default in practice. I think I'll start a straw poll on that in a minute. I think it would be informative about the general views of the community, even though it will not translate into immediate approval/disapproval of specific wording in a policy. WhatamIdoing (talk) 17:32, 1 December 2022 (UTC)

OK, I plan to start an RFC in about 2 weeks per the above. I plan to tidy up the one idea above, and also provide my rationale which would be a response to the RFC not a part of the RFC. I'll get that done within a week. Then we'll go another week....other are invited to write other choices. The we'll ask people to weigh in on every choice to avoid the issues. North8000 (talk) 16:48, 2 December 2022 (UTC)

Is there any chance we can instead discuss adding clarifying language to WP:IMPLICITCONSENSUS and leaving WP:ONUS alone? Or maybe a better question is: do we all agree that the root of the problem is how implicit consensus works? Kolya Butternut (talk) 17:14, 2 December 2022 (UTC)

IMO this wording has multiple problems. I was planning to detail in my rationale within 1 week but briefly It prima-facie conflicts with another policy, it fails to do what it is supposed to do, and arbitrarily weighs in towards exclusion.North8000 (talk) 19:37, 2 December 2022 (UTC)

What about changing "include" to "add": The onus to achieve consensus for inclusion is on those seeking to include add disputed content. Kolya Butternut (talk) 20:30, 2 December 2022 (UTC)

It won't solve the problem. It requires only novice-level wikilawyering to turn "add" into "your reversion 'added' the information". WhatamIdoing (talk) 20:58, 2 December 2022 (UTC)

But isn't it an improvement? The word "add" implies a new addition. We can also add clarifying language to WP:IMPLICITCONSENSUS. Kolya Butternut (talk) 22:52, 2 December 2022 (UTC)

@Kolya Butternut asked: do we all agree that the root of the problem is how implicit consensus works?

Kolya, I feel like I want to agree with this, but I don't think I understand what you intend in enough detail to be able to agree with you. WhatamIdoing (talk) 20:59, 2 December 2022 (UTC)

I don't think there's any disagreement over what happens for new additions; the disagreement is over what happens when someone removes longstanding text which lacks explicit consensus. Kolya Butternut (talk) 22:52, 2 December 2022 (UTC)

Even if the long-standing text had explicit consensus from back in the day, WP:Consensus can change. If someone objects to it now, then we still have to re-demonstrate that there is consensus for it back in the day. Can you imagine someone saying that an RFC from 15 years ago requires us to preserve out-of-date information, and besides, it's now "longstanding text", so it's doubly protected? That's not how it works. WhatamIdoing (talk) 17:07, 3 December 2022 (UTC)

I don't see how what you're talking about is related to the current dispute. So, the question remains whether you agree with what I stated. Kolya Butternut (talk) 17:20, 3 December 2022 (UTC)

Arbitrary break, continuation of prior discussion

• Regardless of how it was added originally, strong consensus was found for it at this Village Pump RfC. It was proposed to move WP:ONUS to WP:CON, and redefine it as about removals too - as about any change - and it was rejected. The closure states, There is an overwhelming consensus against this proposed change, primarily due to concerns that it would undermine verifiability and efforts to remove poorly sourced information. Crossroads ^-talk- 16:12, 25 July 2022 (UTC)

  Even North opposed that, lol. That's not really where we are at atm, though. Selfstudier (talk) 16:19, 25 July 2022 (UTC)

  I don't think the rejection of that RFC amounts to an endorsement of the status quo. Levivich (talk) 16:40, 25 July 2022 (UTC)

  • That RFC tried to do two things at once … move ONUS and change what it said. It seems that most of the opposition focused on the language change, although a few did oppose the move. Blueboar (talk) 13:05, 26 July 2022 (UTC)

  I wonder whether it would make more sense to ask people what they think should happen in certain scenarios, instead of asking them what words they think should be used. WhatamIdoing (talk) 20:20, 25 July 2022 (UTC)

Quick side request

Twice now, I have attempted to insert a simple paragraph break between the VNOT sentences, and the ONUS sentence - so that those newly arriving at this discussion can better visualize and understand what we are discussing. Unfortunately, on both occasions my simple edit was immediately followed by a more substantive edit, and then a revert back to the original (one paragraph) version. It may be that my attempts to insert a break simply got caught up in objections to what followed… but I am not sure. So… let me ask: does anyone object to my re-inserting a paragraph break (with no change to the text)? Blueboar (talk) 19:07, 23 July 2022 (UTC)

It is a single sentence, a sentence is not a paragraph and the paragraph logically places the sentences together. Seperating words from context in this policy is especially problematic, since this policy tends to get atomization of of analysis on the singularly focused but also important source-content relationship, especially since the burden is so light. Alanscottwalker (talk) 19:18, 23 July 2022 (UTC)

But that’s part of the problem… VNOT and ONUS are talking about completely different things. They don’t logically go together. The only reason I can see for placing them in the same paragraph is that both contain the word “consensus”. Blueboar (talk) 20:13, 23 July 2022 (UTC)

They are not two completely separate things, each sentence explicates and why and how verifiability does not mean inclusion. So, the grounds for your proposal is substantive, not clarifying, it is to bolster taking words and concepts out of context, through the odd means of creating a non-paragraph of one sentence. Alanscottwalker (talk) Alanscottwalker (talk) 20:41, 23 July 2022 (UTC)

Ok… I disagree, but at least I know that there is a thought out objection to my paragraph break. I can let it rest. Thanks. Blueboar (talk) 21:02, 23 July 2022 (UTC)

(edit conflict) If it helps, I'll echo one of my comments in an earlier section that there isn't really consensus that there's a dispute in policy or that there is an issue needing changes here. At the time, others were trying to continue the dispute despite not gaining any traction by adding a tag after significant talk discussion. This would fall into similar territory, regardless of intent. This is the point where editing the policy page itself really needs to stop, and consensus is needed for any change at this point. KoA (talk) 21:04, 23 July 2022 (UTC)

@KoA, I think we agreed last round that there is a dispute in policy, but that it only rarely appears in practice. Consider this scenario:

• Alice creates an article.
• Bob adds some content one hour later.
• Carol objects to Bob's addition one hour after Bob added it. Specifically, Carol does not believe that this content belongs in this article.
• The ensuing discussion attracts many experienced editors. At the end of the discussion, the following facts are presented to you:
  1. All editors unanimously agree that the disputed content is properly supported by an inline citation to an appropriate reliable source.
  2. All editors unanimously agree that the disputed content does not involve anything about BLPs.
  3. All editors unanimously agree that the disputed content does not involve a copyright violation, legal concern, or any other characteristic that could result in either speedy deletion or any other automatic decision.
  4. All editors unanimously agree that it's beyond silly to talk about a "status quo version" for an article that was only one hour old at the time the content was added and two hours old when the addition was disputed.
  5. All editors unanimously agree that, on the question of whether to include or exclude the material, they have produced the most perfect tie you could imagine. Regardless of whether you are counting votes, evaluating the strength of arguments, or even considering the apparent level of passion individuals have about the dispute, there is no consensus for the article to contain this content and equally no consensus for the article to omit this content. The community's view is unanimously declared to be exactly balanced.

What do you think editors should do now? Alternatively, what additional information would you need to make a recommendation to editors about what they should/shouldn't do about the disputed content? WhatamIdoing (talk) 20:19, 25 July 2022 (UTC)

ONUS reiterates Consensus. Per NOCON the material stays out. Kolya Butternut (talk) 21:14, 25 July 2022 (UTC)

What makes you say that? "Per NOCON", why isn't Bob's bold edit removed? WhatamIdoing (talk) 21:32, 25 July 2022 (UTC)

What's the difference? What do you think happens? Kolya Butternut (talk) 22:00, 25 July 2022 (UTC)

NOCON says – noting that this is disputed, but here's what it says – "In discussions of proposals to add, modify, or remove material in articles, a lack of consensus commonly results in retaining the version of the article as it was prior to the proposal or bold edit."

I'm curious why you pick Bob's bold edit as "the version of the article as it was prior to the proposal or bold edit" instead of picking Alice's edit. WhatamIdoing (talk) 22:54, 25 July 2022 (UTC)

I don't understand how you got that from what I said. I said "the material stays out." Bob's edit is the only material in question. Kolya Butternut (talk) 23:06, 25 July 2022 (UTC)

I'm sorry, I misread your comment. But the same question applies: Why isn't Carol's "proposal to...remove material" the one that's reverted, and why don't you want to be "retaining the version of the article as it was prior to the proposal"? WhatamIdoing (talk) 02:48, 26 July 2022 (UTC)

The dispute is over Bob's bold edit. Kolya Butternut (talk) 03:17, 26 July 2022 (UTC)

But Bob says the dispute is over Carol's bold blanking of well-sourced content. WhatamIdoing (talk) 21:01, 29 July 2022 (UTC)

ONUS is a great tool for brigading and wargaming, but for building knowledge not so much. There's no effort to distinguish between what is excluded for being questionable on its reliability, and what is questionable on its importance. Hyperbolick (talk) 09:21, 24 July 2022 (UTC)

I'm too old, weary, tired etc. Hope ya'll will find a solution with WP:ONUS. GoodDay (talk) 23:47, 24 July 2022 (UTC)

Don't hold your breath :) Selfstudier (talk) 10:28, 25 July 2022 (UTC)

This is mostly a response to WhatamIdoing's post (20:19, 25 July 2022 (UTC)). I haven't read the whole discussion in full, and I am apologizing in advance if I raise some questions that have already been concerned.

In my opinion, your question describes two totally different cases:

• First, Bob's addition is in agreement with WP:V, AND WP:NOR, WP:NPOV, BLP and other non-negotiable policies. In that case, Carol's objections may or may not be justified. The reason for the objection may be that the Bob's text is only marginally relevant to the article's subject (if relevant at all), or it is stylistically poor, etc. This type issues can and should be resolved by consensus, and it is Bob's task to convince others that his addition improves the article. Since this type issues are mostly a matter of taste, some soft consensus may be sufficient for new text's acceptance.
• Second, Bob's addition is in agreement with WP:V, but it violates either WP:NOR or WP:NPOV (or both). According to the policy's text, these two policies are non-negotiable, so the discussion of the Bob's text should focus not on the merits of the proposed text, but on its compliance with these two content policies. If during this discussion at least one editor raises a legitimate concern about a violation of NPOV, NOR, BLP etc, and other editors fail to properly address it - the proposed text must be rejected.

Currently, the ONUS is somewhat misleading, because it mixes these two (totally different) cases, and it directly contradicts, for example, to the WP:NPOV's preamble, where it directly says it is non-negotiable. It creates a false impression that inclusion of ANY verifiable information is a subject of negotiations among editors, whereas NOR and NPOV say quite the opposite.

I propose to make it more specific by adding a reference to other two policies (and, probably, BLP and NFC). Something like:

Verifiability of the information, as well as its compliance with WP:NOR and WP:NPOV are necessary criteria for its inclusion in an article. However, not all verifiable information must be included. Consensus may determine that, although certain information meets all criteria described in our core content policies, is does not improve an article.

--Paul Siebert (talk) 21:15, 28 July 2022 (UTC)

In addition, with regard to the "one hour" issue, the time should be measured not in days or hours, but in views and edits. If some article is being viewed just one time per day, and is being watchlisted just by one editor, and it had not been edited during last year, then the edit made one year ago may be considered as pretty recent, and it may be senseless to speak about any stable version at all. In contrast, consider some article that is being viewed 100,000 times every day, and watchlisted by 10,000 users. If some edit was made just one hour ago, but after that 100 other edits have been made, we may have some ground to conclude this edit in not as recent as it seems...--Paul Siebert (talk) 21:48, 28 July 2022 (UTC)

I can agree to your proposed text, and completely agree with your comment on views/edits. Blueboar (talk) 21:56, 28 July 2022 (UTC)

If you look at the scenario, the article is two hours old at the time the dispute appeared, and there have only been three edits to the article ever. I think we can assume that the maximum number of watchlists involved is three. Notice, however, that "The ensuing discussion attracts many experienced editors", so the number of watchlists is not relevant. Imagine an CENT-listed RFC with 100 editors, if that helps. Also, notice that the "many experienced editors" unanimously agreed that it's silly to claim that a STATUSQUO version exists when a dispute arises in a two-hour-old article. WhatamIdoing (talk) 21:24, 29 July 2022 (UTC)

@Paul Siebert, you've missed the point. There is only one case here: Editors cannot decide whether Bob's addition violates NPOV. Consensus didn't determine that the disputed material violated NPOV; consensus determined that half the editors believe it's a NPOV problem (with a reasonably convincing argument) and half the editors believe there is no NPOV problem (with an exactly equally convincing argument).

I agree that "This type issues can and should be resolved by consensus" – usually. But the fact is that sometimes we cannot (or at least do not) form a consensus on every single issue. WhatamIdoing (talk) 21:21, 29 July 2022 (UTC)

If I were to come across the above scenario… before determining who should “win”, I would want to see if there has been any attempt at compromise? Did Bob (or one of Bob’s supporters) suggest reworded versions of what they want to add, and we’re those attempts to compromise discussed or just stonewalled? Did Alice (or one of her supporters) make a similar effort? This would tell me a LOT about whether there was actually a good faith effort to reach a consensus or not. Remember that sometimes, holding a !vote can be the wrong approach, and once the involved editors move beyond binary !vote mode, a consensus is achievable after all. Blueboar (talk) 20:56, 16 August 2022 (UTC)

@Blueboar, I don't know if you'll remember this, from forever ago, but there was a real-world dispute about whether someone's farm was in active operation. (I was under the impression that the government could take over abandoned farms or force a sale in that country, so this was a bigger deal than it sounds like.) Commons had a photo of the farm. More precisely, Commons had a photo of a cow or two standing in front of a house that was believed to be at that particular farm. There were concerns (not in sources, if I recall correctly) that the photo might have been produced as part of a disinformation campaign, since you could borrow a cow from a neighbor long enough to take a few pictures. There's no way to re-word a picture, and no way to include a picture halfway. So either we include the only photo we have, which is normal, or we exclude it, because we think it's faked. Including it implies that Wikipedia has decided the allegations were false. Excluding it (at least if you happen to know that this picture exists) implies that Wikipedia has decided that the allegations are true.

I agree that consensus is usually achievable, and that compromise is usually possible, but sometimes you can't achieve consensus (e.g., this year) and sometimes you can't achieve a compromise. WhatamIdoing (talk) 22:31, 19 August 2022 (UTC)

No, ONUS should not selectively regurgitate the other core content policies; they are best handled at their own pages, because the form of regurgitation would unnecessarily distract policy/discussion. Altanner1991 (talk) 13:09, 14 September 2022 (UTC)

Hi. There is a discussion of disputed content in Talk:Tudor Dixon, under the thread "Restoration of contentious info in place of neutral language ". For some reason even if it has no consensus, it's been questioned and is disputed it remains in place. It's an WP:ONUS issue. If you have the time and the interest, it would be nice if you share your input there. Thanks! Thinker78 (talk) 23:04, 3 December 2022 (UTC)

It has been argued to me on the talk page of the List of Impact Wrestling personnel that Twitter can be used as a primary source to simply confirm an individual's position with a company. My argument is that better sources are needed but I was informed that this is not needed if the account is verified. Another set of eyes may be needed on this but I am not sure which is why I am here instead of in the admin stream. The section concerned is referees. My main gripe is that none of those people listed are notable, which means that the section shouldn't be there at all, but I will leave that to any fresh eyes that wish to look at it. Addicted4517 (talk) 02:26, 8 December 2022 (UTC)

I think this might be better directed to WP:RSN than here, but otherwise I'll leave it to editors with more expertise in such matters. DonIago (talk) 05:22, 8 December 2022 (UTC)

In fact @Addicted4517:, I made the same question here. #Twitter for jobs --HHH Pedrigree (talk) 10:01, 8 December 2022 (UTC)

@HHH Pedrigree:, I think a report to WP:RSN in light of that note of yours may be in order, but I can't do it right now. I hate to pass the buck but could you do it? This needs to be sorted by others so certain people (you know who I'm talking about) can't claim consensus. Addicted4517 (talk) 00:00, 9 December 2022 (UTC)

Footnote 3 contains the following wording:

Once an editor has provided any source they believe, in good faith, to be sufficient, then any editor who later removes the material must articulate specific problems that would justify its exclusion from Wikipedia (e.g. why the source is unreliable; the source does not support the claim; undue emphasis; unencyclopedic content; etc.). If necessary, all editors are then expected to help achieve consensus, and any problems with the text or sourcing should be fixed before the material is added back.

Is my reading that it's supposed to provide grounds for removal of verifiable content unless an explicit consensus for its restorationthat the text doesn't have any problems exists correct and it extends past the scope of the WP:V/BURDEN intentionally (as the edit comment seems to suggest)?

I think it contradicts the spirit (if not the word) of WP:NPOVHOW and WP:PRESERVE, which give an opposite advice to improve by editing rather than removal.

No opinion on what it should be, but feels odd that something fairly significant with scope extending the scope of the policy is hidden in a footnote. PaulT2022 (talk) 21:58, 14 December 2022 (UTC)

Quick note… Verifiability issues are covered by WP:DON'T PRESERVE (the next section … immediately after WP:PRESERVE). Blueboar (talk) 22:37, 14 December 2022 (UTC)

Thanks! To be clear, I didn't raise it because of verifiability issues (for which there's unambiguous guidance in the main text: "Any material lacking an inline citation to a reliable source that directly supports the material may be removed and should not be restored without an inline citation to a reliable source.") but because of the overreach into NPOV (undue emphasis) and vague open-endedness of etc and any problems with the text... should be fixed before the material is added back, as well as unclear reasons why it's in a footnote rather than the main text. PaulT2022 (talk) 22:42, 14 December 2022 (UTC)

That footnote more or less describes general editing practice as well as being ONUS related. If it is possible to improve via editing, that's preferable to removal unless there is a decent argument for it. Just claiming ONUS and saying something like biased, undue, blah to justify removal won't do, specifics are necessary. Talk page discussion should follow a la BRD but that does not actually prevent reinsertion following removal and talk page discussion will likely follow that as well (or competing reverts until...). It's the usual WP circularity and connects up the different policies afaics. Is the footnote strictly necessary, maybe not but does it do any harm? Selfstudier (talk) 08:53, 15 December 2022 (UTC)

I've seen two related problems in editing:

• sometimes editors may claim to remove material because of sourcing/verifiability, whereas actual issue discussed was NPOV. So I think some abuse of scope of BURDEN does happen, and there may be confusion whether BURDEN or ONUS applies, although I don't remember this footnote being referenced in discussions specifically.
• more significantly, I think policies are not sufficiently clear that the practice is to remove material until the consensus is reached that the problems with it are fixed. Do they say it directly anywhere other than this footnote? I've seen edit wars around this, as well as done a few unnecessary reverts myself due to misunderstanding of QUO/BRD/NOCON.

I came here trying to figure out seeming contradictions between burden, onus and consensus that were discussed multiple times. After reading archives I think the confusion stems from some of the essays rather than policies, but finding something major like this tucked in a footnote of BURDEN felt strange.

To me, it'd make more sense to have it as non-footnote text and possibly moved to ONUS if it's intended to explain editing practice related to ONUS.

Also, I think "policy-based problems" or simply "the problems" to make it clear that it refers to the articulated problems justified by policy (as required by preceding sentence) would make more sense than "any problems". (As it's what it means anyway?) PaulT2022 (talk) 12:41, 15 December 2022 (UTC)

policies are not sufficiently clear that the practice is to remove material until the consensus is reached that the problems with it are fixed sometimes it is not so clear and material may stay in pending formal DR.

This might well be worth raising in the ONUS RFC (if we ever start one, lol). Selfstudier (talk) 13:53, 15 December 2022 (UTC)

That's a good point, I should've said "not to re-add material". Agree that best not to touch it without RfC too. Thank you for the feedback! PaulT2022 (talk) 23:41, 15 December 2022 (UTC)

@PaulT2022, I have been thinking about writing Wikipedia:Bring me a rock for years. Let me know if you think it would be helpful to add a section that explains the connection between this footnote and the story. WhatamIdoing (talk) 01:14, 20 December 2022 (UTC)

Love the writing! Could be useful if there's a non-obvious one. PaulT2022 (talk) 07:08, 20 December 2022 (UTC)

I think it's obvious, but since I wrote both the footnote and the essay, maybe it's only obvious to me: The reason this footnote specifies "any source they believe, in good faith, to be sufficient" (originally "one source", if memory serves) is so that editors can't claim protection from BURDEN and CHALLENGE to send others on an endless errand to find sources. The reason the footnote lists several reasons (beyond "Bring me a source") is to help the reverting editor remember that WP:BURDEN isn't the only policy-based reason for removing content.

It might be time to move this out of the footnote. WhatamIdoing (talk) 16:25, 20 December 2022 (UTC)

Thank you for this explanation, that's a very strong argument for leaving the footnote in BURDEN actually. I did focus more on the final sentence of it, which kind of negates the main point of the footnote almost encouraging the Wikipedia:Bring_me_a_rock.

I do get that it happens not because of the footnote, please don't get me wrong. It's just that the footnote in the current form somewhat excuses it, framing such behaviour as an effort to maintain verifiability. PaulT2022 (talk) 02:51, 21 December 2022 (UTC)

I love hearing all the ways that policies don't quite make sense to other editors, so you don't need to worry about my reaction. If you weren't trying to help, you wouldn't have bothered starting this discussion.

If someone like you finds this confusing and contradictory, then someone else will be even more confused by it. I very much appreciate you bringing this up and sharing how it landed with you. WhatamIdoing (talk) 06:02, 21 December 2022 (UTC)

• Perhaps it helps to realize that the fundamental substructure of main content policy issues (V, NPOV, NOR, (and BLP)) are sourcing issues, interconnected by the issues of quantity of sourcing and the quality of sourcing, leading to the article aimed at fair representation of the source isolate and the sources in topical context. The policies overlap because they are cooperating aspects of good articles not hermetically sealed totems. Alanscottwalker (talk) 11:59, 21 December 2022 (UTC)

Often a key RS like the NYTimes will publish a major story, and we will see other sources present that in summary, usually referencing the NYTimes story with language like "According to...". Do we have advice that encourages editors to use the original source when this is clearly obvious, and particularly when the original source is generally seen as a more reliable work than the summarizing source? Should we have this type of advice? Masem (t) 14:34, 6 December 2022 (UTC)

To some extent, all sources are representation of other sources, primary becoming secondary en route. I have noticed that newsorgs use this "according to.." more frequently these days. (Putting AFP/AP up in the corner is also popular). Happens the other way too, the "better" source will credit the "lesser". I think it is a matter of editorial judgement for the most part, subject the usual guidelines for source reliability. Selfstudier (talk) 14:49, 6 December 2022 (UTC)

I don't know that there is a universal way we should handle such a situation. It is going to be a case-by-case basis for these things. Sometimes, it is better to cite the New York Times, because they "according to" is the primary source itself. "According to Press Secretary Karine Jean-Pierre..." seems like something we would want to cite to the Times and not directly to Jean-Pierre's notes, for example. Sometimes, the Times is repeating, essentially verbatim, what another secondary source says. In those cases, we'd want to bypass the Times and go to the other source. There's probably another several dozen nuanced ways in which a sentence in the Times could say "According to..." and each one would require a different approach at Wikipedia. No need to itemize each one. --Jayron32 14:53, 6 December 2022 (UTC)

• Both types of sources are important to writing a good article. The original report has more depth and is thus better for Verifiability - while the “according to” repetition is better for establishing notability and due weight. Blueboar (talk) 14:59, 6 December 2022 (UTC)

Authors should use the source they can access. So if they can't get hold of the original NYT article, using the information as presented in the Phoenix Sun (or wherever) and sourcing it to that is fine. If you check a source and see that it's rereporting information from another outlet and can get a hold of the original, then, after verifying the information, changing the info to the more detailed source makes sense. That said, checking sources is a good practice; sometimes what an academic paper or book says a journal article or other sources says isn't always correct. — Carter (Tcr25) (talk) 16:48, 6 December 2022 (UTC)

Sometimes it's better to go to the original (e.g., when it is authoritative), and sometimes it is better to use the second-hand (which is not necessarily secondary) source (e.g., when it is a stronger source in some respect; when it provides context or analysis). Whichever source you choose to use, you should WP:SAYWHEREYOUGOTIT. WhatamIdoing (talk) 22:49, 7 December 2022 (UTC)

Even in the case where the reusing source offers additional analysis (for example, Vox often does this), it should still be desirable to ref the originating source in addition to the reusing source. Masem (t) 23:14, 7 December 2022 (UTC)

Not necessarily. We should never write things like "According to Anthony Fauci, HIV causes AIDS", much less "According to Anthony Fauci, as reported in Vox, HIV causes AIDS." WhatamIdoing (talk) 00:18, 9 December 2022 (UTC)

I don't think that's what's being suggested; more a matter of whether or not there should be ^[NYT][Vox] or just ^[Vox] when the information in the article is being sourced to Vox but the Vox report draws on (or builds on) New York Times's reporting, or if it should be "According to The New York Times, ..."^[Vox] — Carter (Tcr25) (talk) 01:31, 9 December 2022 (UTC)

If you didn't read the NYT article yourself, then you shouldn't cite the NYT article. You can (see the example in WP:SAYWHEREYOUGOTIT) write a citation that says "NYT article, as cited in Vox article", but you shouldn't have a separate ^[NYT] footnote. WhatamIdoing (talk) 20:44, 9 December 2022 (UTC)

re " Do we have advice that encourages editors to use the original source when this is clearly obvious, and particularly when the original source is generally seen as a more reliable work than the summarizing source?", we already have this:

The greater the degree of scrutiny given to these issues, the more reliable the source.

IMO, that answers your question. If some source just quotes another source, it is hardly more reliable than the original source. If a source A provides detailed analysis of some event, and the source B just quotes the source A, the former is more reliable. However, it seems that the policy has no direct advise to chose A in this situation. It says:

"Base articles on reliable, independent, published sources with a reputation for fact-checking and accuracy. "

and that may be a problem, for both A and B are reliable. In connection to that, if we modify this text as follows:

Base articles on reliable, independent, published sources with a reputation for fact-checking and accuracy. Best quality sources are preferable. "

If we do that, it will be a direct advice to chose A instead of B is both of them are available. Paul Siebert (talk) 21:21, 9 December 2022 (UTC)

I don't think articles are really meant to be WP:Based upon any single source, though. WhatamIdoing (talk) 20:23, 19 December 2022 (UTC)

• What's usually happening in such a case is churnalism. Once one reputable news outlet has published a story, others will then repeat it. See research, for example. Such recycled news would usually be less reliable than the original because rewording may distort or change the meaning. And, if no additional research is done, then no value is added. So, it's best to find the first source in the chain. Andrew🐉(talk) 23:10, 9 December 2022 (UTC)

Sources that quote other sources

Consider a little bit different situation. Some reliable source "A" quotes another source "B", but provides no analysis or discussion (it is just a verbatim quote). Should we use "A" instead of "B" when "A" meets RS standards, but "B" does not? --Paul Siebert (talk) 19:52, 19 December 2022 (UTC)

Usually, that implies either endorsement or disapproval, does anyone quote with zero context? A quote just hanging in the air? Do you have an example? Selfstudier (talk) 20:20, 19 December 2022 (UTC)

What does it mean for "B" to not meet RS standards? If you check the FAQ at the top of the page, all sources are reliable for some purpose. Every published source is reliable for a statement that says nothing more than "Source B said <whatever the source actually said>". WhatamIdoing (talk) 20:24, 19 December 2022 (UTC)

Under "does not meet RS standards", I mean "it is reliable only for the author's opinion".

Ok, I'll ask it more concretely. In my example, the source "B" is an SPS, the source "A" is a peer reviewed publication. However, "A" provides no analysis of the source "B", and does not discuss it. It just quotes what "B" says. Paul Siebert (talk) 21:16, 19 December 2022 (UTC)

I'm genuinely trying not to pick your example to pieces, but it kind of matters whether this is being used for quotation-y purposes (e.g., a literary quotation sprinkled in for fun) or if it's a factual claim. Consider:

• "Toto, I've got a feeling we're not in Kansas anymore." – Source A, as quoted in Source B, to set a humorous tone in advance of a tricky explanation.
• 'Gale's report said that Oz was "not in Kansas".' – Source A, as quoted in Source B, to report seriously on a geographical question.

In the first, I'd usually consider Source A to be more authoritative, because misquotations and misattributions happen, and if you can get back to the original, you can avoid those problems. In the second, I'd usually prefer source B, as it provides not only evidence that the information came from some source, but also an indication (however vague) that this source is worth taking notice of. There could be dozens or hundreds of self-published sources, and only Source B gives us a reason to prefer this one over the others. WhatamIdoing (talk) 01:32, 20 December 2022 (UTC)

Well, I can give you a concrete example. It is just the most recent one, and most impressive. The Holodomor and Soviet famine of 1932-33 article claim that during the Soviet famine of 1932-33, Soviet authorities printed a poster "To eat your own children is a barbarian act". That looks absolutely weird, because Stalin's authorities were trying to conceal the facts of cannibalism, and any posters of that kind would have a Streisand effect. The information about that poster was found in the article by Vardi&Vardi, but these authors just quote (without any critical analysis) another article: it is the article by some Tapay, which was published by Hungarian association, which is just an emigre club. I was not able to find any information on Tapay, and his credentials are totally unknown. I looked through the text of his article, and I found that he used five sources: two chapter of The Black Book of Communism (authored by Werth, who is a very reliable source), a book by a political writer Medvedev, a book by a playwright Radzinsky, and his own article about Khruschev. That's it. Obviously, Medvedev and Radzinsky are hardly reliable for such a statement, reliability of Tapay himself is totally unclear (there is no information about his credentials in google), but Werth is reliable. Tapay took a lot of information from him, and many paragraphs are just a translation from Werth (the Tapay's article is written in Hungarian). However, Werth's book contains no information about the poster. Interestingly, Werth himself took information about cannibalism from Conquest's "Harvest of sorrow" (and he cites that book). That means, we have the following chain of sources.

1. In his "Harvest of sorrow", Conquest shows a dramatic picture of cannibalism during the Soviet famine of 1932-33.

2. Werth includes this information into the Black Book of Communism. He supplements it with an inline citation to Conquest, so it is clear where the information was taken from.

3. Tapay writes the article for "Proceedings of Hungarian association", where he provides no inline citations, but almost verbatim similarity of hist text with Werth makes the origin of information quite clear. To this text, he adds a statement about the poster, and this information is not found in Werth.

4. Vardi&Vardi quote Tapay without any analysis and commentaries.

5. This information is now in Wikipedia.

Let me reiterate: there is absolutely no evidences that Tapay is a professional historian, and that he is doing his own research. I was unable to find any publications authored by him in google.scholar. Therefore, all facts presented in his article are supposed to be found in the literature he cited. Werth does not say that. Medvedev is a political writer, and his book was written in 1973, when he had no access to archives. Radzinsky is not a historian, he is a playwright. Therefore, the information about that poster is not more reliable than any other information that can be found in an average SPS.

Does the fact that that information was quoted in "East European Quarterly" make it reliable? "East European Quarterly" is a journal with no impact factor, and the authors do not analyze the Tapay's statement, they just quote it.

In my opinion, this information is very questionable, and it is not a surprise that we have no independent confirmation of it.

Does our policy provide any mechanism to protect Wikipedia from this type low quality information? --03:06, 20 December 2022 (UTC) Paul Siebert (talk) 03:06, 20 December 2022 (UTC)

@Paul Siebert,

What an informative illustration. I have two approaches to suggest. The first is that we use the movement's size as a strength, and ask some Hungarian speakers. User:Botulitisz, User:Syp, and User:Korovioff have all contributed to the Holodomor article at the Hungarian Wikipedia. Syp isn't very active these days, but the other two might be around and might be able to provide more information. (Perhaps, for example, Tapay's name is normally transliterated into a different spelling, or perhaps they'll know someone who can advise us on the history of that particular journal.)

Second, I suggest that when we have a single source for a colorful detail about a very large subject, that it doesn't belong in the article anyway. Consequently, it doesn't really matter whether it's true or whether it's verifiable; it only matters that this detail – one of many, many thousands of details from a multi-year horror – should not take priority over the details that multiple sources deemed worthy of mentioning. WhatamIdoing (talk) 16:39, 20 December 2022 (UTC)

Thank you. My goal here is not to draw attention to this concrete text. I am asking if this policy has some mechanism to prevent this type incidents. If it does not, we should discuss if this type cases are frequent. If they are (in my opinion, yes), we need to propose some mechanism for protection of Wikipedia from this type spam. It is necessary to note that this type cases are teh most frequent in the topics that are beyond the scope of a broad community, and are interesting to just a small group of users, so development of a local consensus that goes against the norms or our policy is quite likely.--Paul Siebert (talk) 17:23, 20 December 2022 (UTC)

I think the main mechanism is what I describe in my second paragraph. It becomes a question for NPOV/DUE/BALASP rather than a question for WP:V. WhatamIdoing (talk) 00:00, 21 December 2022 (UTC)

I had a situation with the Bill Mullahey article when it was nominated for a DYK item. A reviewer objected to a statement that was cited to a self-published work (ref #5). That source met WP:USINGSPS, but the relevant point was quoting from an older WP:RS published book that I couldn't locate a copy of to confirm the details. In the end, the citation used credited first the original work but noted that it was "as quoted in" the self-published work. For the situation you've outlined, I believe that's a reasonable solution. If you can confirm the accuracy of the original work, the source that alone; if you can't, then source the original, but note that it is filtered through a different source. — Carter (Tcr25) (talk) 20:43, 19 December 2022 (UTC)

I'd give the opposite advice for Wikipedia:Biomedical information, because we put a premium on recent sources. If a recent source approves of a 20-year-old source, then we tend to cite the recent source, to indicate that this information isn't outdated. WhatamIdoing (talk) 01:17, 20 December 2022 (UTC)

Yeah, the context matters. — Carter (Tcr25) (talk) 02:21, 20 December 2022 (UTC)

WP:RS/QUOTE: To ensure accuracy, the text of quoted material is best taken from (and cited to) the original source being quoted. So "B" in this context doesn't just meet RS standards, it is required by them if it is available, regardless whether some WP:RSN conversation said "generally" unreliable for facts. I recently had a difficulty because the quoted source (Breitbart) had been placed on a spam blacklist, and worked around it by citing but not putting the http: stuff in the url section of the cite. Peter Gulutzan (talk) 17:28, 20 December 2022 (UTC)

When that page says that citing the original "is best", it doesn't mean that citing the original is required or the sole acceptable option. WhatamIdoing (talk) 23:59, 20 December 2022 (UTC)

As I hinted, and as the next sentence in the paragraph makes clear, if it's impossible to quote then one can use secondary. Why anybody otherwise would think that what's best isn't what's needed or made necessary, I don't know, but of course following guidelines is up to participants in particular situations. Peter Gulutzan (talk) 13:46, 21 December 2022 (UTC)

What's needed and required is to use a source that meets or exceeds the rock-bottom minimum requirements. What's best is hopefully better than the bare minimum. We don't want editors (especially editors reviewing other editors' work) to be making the perfect be the enemy of the good.

(Also, quoting someone doesn't make your source secondary. It makes the quotation second-hand, but that's not the same as secondary.) WhatamIdoing (talk) 17:16, 21 December 2022 (UTC)

Surely using what the guideline says is "best" is by definition "better". (Also, I used the word secondary when describing an exception because that's the word the guideline uses.) Peter Gulutzan (talk) 18:17, 21 December 2022 (UTC)

Thanks for the reminder. I fixed the guideline; anyone who's confused about why should find WP:LINKSINACHAIN helpful. WhatamIdoing (talk) 19:22, 21 December 2022 (UTC)

Editors have been talking for months about the interaction of various policies and procedures around cited material whose inclusion is disputed on grounds related to whether the material is verifiable. Here's the scenario:

Alice and Bob decided to create an article together. They were editing collaboratively until they ran into a serious disagreement. One of them wants to briefly mention a related subject, and the other doesn't think that related subject should be mentioned at all (in this particular article). They are unable to agree, and when you look on the talk page, you find a discussion between multiple editors, which is very evenly divided. These are the facts:

1. The material in question is fully verifiable, factually correct, well cited to two high-quality reliable sources, and not about a BLP.
2. There is no consensus to include the material and equally no consensus to exclude the material. There are good arguments on both sides.
3. It is not possible to compromise half-way, because there is no way to write an article so that it both briefly mentions, and also does not mention at all, the related subject.

You edit in this area regularly, so these editors have asked you personally for your advice on the dispute. Specifically, they ask you to tell them whether you believe that it is generally better for Wikipedia to include or exclude material in this type of situation.

Your job is to finish this sentence: "When editors have been unable to reach an agreement about whether to include a piece of verifiable, cited material, I believe it is usually best for Wikipedia to ______". 18:08, 1 December 2022 (UTC)

Your answers will help us resolve a possible conflict between some policies, guidelines, and essays. Expected answers include: default to inclusion, default to exclusion, default to the older version (assuming any exists), default to whichever version is in the article today, depends on the subject matter – but please surprise us with new ideas, and give an example or a reason. Reasons based in common sense would be particularly valuable (e.g., readers are best served by longer/shorter articles).

Thank you for your help. WhatamIdoing (talk) 18:08, 1 December 2022 (UTC)

Answers

If you object to the poll altogether, please comment below. — Preceding unsigned comment added by Butwhatdoiknow (talk • contribs) 20:55, 1 December 2022 (UTC)

• "When editors have been unable to reach an agreement about whether to include a piece of verifiable, cited material, I believe it is usually best for Wikipedia to remove it".

Rationale A continuous process of adding/copyedit/removal can be described as a random walk. Its trajectory is not limited from the top (texts can be infinitely long), but it has a natural limit at the bottom (no text can be shorter than zero symbols). A random walk that is limited from one side always drifts to the opposite side. If we make addition and removal equally easy, Wikipedia texts will start to infinitely inflate, and dilution of good texts with questionable additions will become unstoppable.

We are amateurs, and the only condition that allows us to create reasonably good texts is the fact that our texts are much easier to remove than to add. Paul Siebert (talk) 18:46, 1 December 2022 (UTC)

• When editors have been unable to reach an agreement....yet. This is ONUS again, albeit dressed in different clothes... "..Such information should be omitted or presented instead in a different article. The onus to achieve consensus for inclusion is on those seeking to include disputed content." No consensus, no include. Selfstudier (talk) 19:00, 1 December 2022 (UTC)
• Wikipedia already has guidance on how to proceed, and has had such guidance for quite a good many years; it's guidance I agree with. While information must be verifiable for inclusion in an article, not all verifiable information must be included. Consensus may determine that certain information does not improve an article. Such information should be omitted or presented instead in a different article. The onus to achieve consensus for inclusion is on those seeking to include disputed content. (bold mine). If there is not a clear consensus to include something (or at least a lack of objection), then it stays out until there is consensus to include. If Bob wants it in, and Alice wants it out, and neither budges, Wikipedia's best practices are clear and unambiguous on the matter: It stays out. Bob has the responsibility to convince a preponderance of people that the content belongs in. Alice's objections, so long as reasonable and in good faith, are enough to keep the information out. If Alice's objections are not reasonable, other people will say so. If Bob needs help finding people to say so, WP:DR is a thing. The principle is simple: It is better to be silent than wrong. If we don't know whether something is proper to include or not, we default to silence. --Jayron32 20:30, 1 December 2022 (UTC)

  This may work well when it's an article that gets a lot of eyes and opinions, but more often than not it seems to come down to an article that two or three editors are watching and then it becomes an WP:OWN question and you either "let the Wookie win" or take a trip to ANI, neither of which is an ideal solution. — Carter (Tcr25) (talk) 20:42, 1 December 2022 (UTC)

  Nope. You have MANY steps before ANI, which won't get involved in content disputes anyways. You can try WP:3O, you can go to WP:DRN, you can start a WP:RFC. You can ask from input from relevant Wikiprojects. There's tons of ways to get people's attention in neutral, non-WP:CANVAS ways, you just have to care enough to try. And if no one ever comments, that's fine. The information stays out of the article unless and until someone does. I am 100% okay with that every time, all the time, with no qualms or problems. --Jayron32 20:51, 1 December 2022 (UTC)

  I opened an RFC at a major guideline earlier this year. It followed up on some serious drama at ANI and weeks of discussions about how to write the RFC question. Exactly one (1) editor responded to the RFC within the usual 30 days. You can't always get people's attention. Sometimes they just don't care (any longer). WhatamIdoing (talk) 05:46, 2 December 2022 (UTC)

  In response to both your reply and Carter's: your problem-set tells us that there is a talk-page discussion about the matter, with no consensus. I don't see a problem-there, and don't think WP:OWN applies; it's not the end of the world if content is excluded for a while; and it won't be "lost to time" since it's immortalized in easily-searchable archives. That discussion can be resurrected at any point in the future, and if it has any merit, it likely will be.

  Taking a step back: the real issue I see, echoing the long discussion from earlier, is with (a) low-attention pages, where (b) large amounts of unsourced content is deleted, which (c) is lost to time since it never hits the talk page and no one ends up checking the revision history, where (d) the deleted content is due, and could be sourced to old newspaper/magazines if you look deeply, but (e) equally crucially, you'd only find those sources if you know what you're looking for; that material would never surface in a simple Google or Google Books search, therefore (f) anyone trying to rebuild the article based on proper sources, not being aware of the earlier deletions, would miss a lot of the deleted-but-due-and-factually-correct information due to FUTON bias. This is a minor and mostly hypothetical scenario, which I very intentionally define narrowly. History of the Macintosh is one article where this narrow series of events may occur.

  But you know what? I also firmly believe WP:PRESERVE already addresses "my" issue satisfactorily, and I trust editors' good faith and judgment. I'd much rather my issue occur, than us weakening BURDEN to address these edge cases. Wikipedia's credibility, and its efforts to avoid causing real-world harm with false content are far, far more vital than comprehensiveness in obscure articles. I endorse Jayron32's reasoning; it's already covered, and we default to remove; the current guidance is excellent, and I'd strenuously oppose weakening either BURDEN or ONUS. DFlhb (talk) 15:45, 2 December 2022 (UTC)

• You wrote an example which is logically cleaner than most but still left out variables which I think should affect the result. So lets say that I nail those down by saying it has a medium-direct degree of WP:Relevance and a medium degree of importance regarding the topic. Then my answer would be "include". But in no way should this be considered to be an application of our three main prima-facie-conflicting policies regarding this. Instead it might contribute to an effort to fix them. Wp:undue never was usable and post-Walter Cronkite era now produces bad results. WP:Onus wording is a misfire at something that is really needed in that paragraph and wp:consensus is a good well-written and useful guide. North8000 (talk) 20:44, 1 December 2022 (UTC)
• "When, after a good faith discussion, editors have been unable to reach an agreement about whether to include a piece of verifiable, cited material, I believe it is usually best for Wikipedia to remove it or, if appropriate, move it to another article". wp:NOT. - Butwhatdoiknow (talk) 21:08, 1 December 2022 (UTC)
• Leave the new addition out. Per WP:NOCON, In discussions of proposals to add, modify, or remove material in articles, a lack of consensus commonly results in retaining the version of the article as it was prior to the proposal or bold edit. However, this has nothing to do with the months long ONUS discussion, so either this RfC should be closed or that line should be struck. Kolya Butternut (talk) 22:53, 1 December 2022 (UTC)

  What is "the new addition" that you mention? If you mean the disputed material, the scenario above does not say that there is any "new addition". The disputed material could have been present in the very first edit that created the page. Then what? WhatamIdoing (talk) 05:19, 2 December 2022 (UTC)

  Your example states that Alice and Bob were editing collaboratively until they ran into a serious disagreement, i.e., there was a "new addition". That's how everyone's going to interpret your question, so if that's not your intention you should stop the RfC. Kolya Butternut (talk) 08:01, 2 December 2022 (UTC)

  The serious disagreement could also have been the one of them, upon further reflection, wanted to remove something that was in the first revision. It could even have been removed by the person who originally added it. WhatamIdoing (talk) 15:53, 2 December 2022 (UTC)

• Since this is a hypothetical I will assume there are no policy related issues that might favor include or exclude and it's 100% editor judgement. In such a case, I think you exclude because CONSENSUS says we need the consensus to include, not exclude. To address, WhatamIdoing's question to KB, I would still say exclude. If the content has never been discussed thus we can't show it had more than an implicit consensus to include, again, leave it out as it doesn't have consensus to include. Springee (talk) 05:31, 2 December 2022 (UTC)

  Ah, but that strikes at the heart of the issue. NOCON speaks of any proposal to "add, modify, or remove" needing consensus, otherwise the text reverts back to its longstanding version. If that longstanding version happens to contain the disputed material that has nothing more than implicit consensus, NOCON inadvertently opposes removal and favors keeping it when the editor who tried to remove it is unable to attain consensus. --GoneIn60 (talk) 05:56, 2 December 2022 (UTC)

  Springee's assumption is fair, if we restrict ourselves to content-related policies and guidelines. Other policies, such as Wikipedia:Edit warring, always apply, but we are going to pretend that isn't relevant for my story. WhatamIdoing (talk) 06:45, 2 December 2022 (UTC)

  I'm confused. NOCON is a part of CONSENSUS, yes? Then the correlation is that CONSENSUS isn't only about consensus to include. It also says we need consensus to exclude when dealing with a change to longstanding content. Note: I happen to agree with Springee, but I caution pointing to CONSENSUS as a basis for that opinion. --GoneIn60 (talk) 08:18, 2 December 2022 (UTC)

  "Longstanding content" is just a euphemism for "Things that are wrong that no one noticed for a long time". Why do wrong things that no one noticed get privileged status? --Jayron32 12:03, 2 December 2022 (UTC)

  Still, there is definitely confusion, even among experienced editors, on the QUO/ONUS thing, see this recent RFC before, which is QUO (not noticed for 12 years) and ONUS (it's wrong because reasons X and Y) and led to a lengthy RFC during which the disputed material stayed in. Selfstudier (talk) 14:43, 2 December 2022 (UTC)

  wp:QUO only applies "during a dispute discussion." - Butwhatdoiknow (talk) 15:14, 2 December 2022 (UTC)

  Uhuh, it was a dispute discussion (later called pre RFC) and the QUO/ONUS was further discussed in the actual RFC. Closer comment is of interest "Additionally, the consensus to include those symbols in those articles has been reached by a more limited discussion than this one or previous ones about the Jerusalem status. Thus, I believe that (see WP:CONLEVEL), they cannot override wider community consensus on the Israel/Palestine issue or how NPOV applies to infoboxes of political entities." and the subsequent close review is also worth a look.Selfstudier (talk) 15:39, 2 December 2022 (UTC)

  Jayron32: I completely agree with your sentiment; it's the head-scratcher in NOCON that I think needs to be addressed. See my response to the straw poll below. --GoneIn60 (talk) 20:13, 2 December 2022 (UTC)

  (edit conflict) To be clear, I'm also fine with the guidance of WP:QUO as well. There are really two different questions at play here 1) What is best for the text of a Wikipedia article? and 2) What kinds of behavior should people engage in to allow for harmonious editing? The answer to question 1) is that disputed content should be left out of an article, because it is better to be silent on a matter than to be wrong on it; and it is best for the article text if disputed content is left out, no matter how long it has been there, until the dispute is resolved. The answer to question 2) is leave the fucking article alone until the dispute is resolved. Even if that violates #1, if there is a dispute ongoing, people involved in the dispute (or people who intend to get themselves involved in the dispute) should leave the article alone at all costs, even if someone else ISN'T doing that. Let's say, for example, that we say that per WP:ONUS, some content is removed until the dispute is resolved. If someone comes along and adds it back, saying "I just don't agree", well, then from a behavioral point of view, let them do it. Don't take it back out again. Wrongness is not binary, it's graded on a curve, and if we're going to get in an edit war over the matter, it's just not worth it. So yes, the material should remain out. But you shouldn't be the one to remove it again and again if it means there is to be an edit war. My most important behavioral guideline has always been "let other people be assholes". If you yourself do the same behavior, you're just making it worse. Yes, there is some inconsistency there. No, there isn't a magic pill that will resolve it. If there was, we'd have fixed it by now. --Jayron32 20:17, 2 December 2022 (UTC)

  The heart of the issue is related to implicit consensus, but this RfC question doesn't address that. If two editors are creating an article together it makes no sense for one of them to cite implicit consensus. Kolya Butternut (talk) 08:12, 2 December 2022 (UTC)

• exclude it – When an editor (or group of editors) are unable to show consensus for inclusion, the default outcome should be exclusion. But consensus is never final, and discussion doesn't have to end there. Having a slightly higher bar for inclusion leads to better quality articles at the possible expense of being incomplete in the short term, a reward that is worth the risk IMO. Current loopholes like NOCON's favoritism for longstanding content should be modified (or at least clarified) to prevent conflicts. Implicit consensus is weak and should have no bearing in a discussion where disputed material is sufficiently challenged. --GoneIn60 (talk) 08:18, 2 December 2022 (UTC)

  This RfC question isn't about longstanding content or implicit consensus. Kolya Butternut (talk) 08:39, 2 December 2022 (UTC)

  Thanks, I realize. The last two sentences are preemptively addressing the elephant(s) in the room. Feel free to ignore them if you need my response to seem more applicable. --GoneIn60 (talk) 20:03, 2 December 2022 (UTC)

• Exclude it, per GoneIn60. Nikkimaria (talk) 00:47, 3 December 2022 (UTC)
• Default to inclusion, if it really does just come down to editorial judgement. Benjamin (talk) 01:52, 3 December 2022 (UTC)

Objections

• This doesn't accurately represent the conflict, so I would ask that you close this poll. The actual conflict has more to do with how implicit consensus works. Kolya Butternut (talk) 19:12, December 2022 (UTC)

  I am not trying to directly resolve any conflicts. I am trying to gather information that I would find informative. I particularly hope to discover what the community of experienced editors thinks. It is very easy for me, at least, to incorrectly assume that my thoughts on a matter are "of course" what everyone else thinks. If I am lucky, I will learn that my view isn't necessarily the dominant one. I will not learn this if people do not share their honest views. I encourage you to add your advice to Alice and Bob above. You might also be interested in the FAQ at the top of WT:RFC. WhatamIdoing (talk) 20:16, 1 December 2022 (UTC)

  You're misrepresenting this RfC as being about what Editors have been talking for months about, but this scenario absolutely does not represent the discussion. For that reason I would ask that you end the RfC. If you think we should have one, let's discuss it first. Kolya Butternut (talk) 22:48, 1 December 2022 (UTC)

• Something can't be simultaneously a straw poll and an RfC. Either it's binding or it isn't. I believe the RfC tag should be removed. Crossroads ^-talk- 19:59, 1 December 2022 (UTC)

  RFCs aren't binding. Per WP:RFC itself: "RfCs are a way to attract more attention to a discussion". I want more attention; therefore I use an RFC. If you want a binding decision, then the RFC page seems to suggest ArbCom is your only option. I would value having your view of my question recorded here; I have no doubt that your view would be both well thought out and interesting to both me and other editors. WhatamIdoing (talk) 20:11, 1 December 2022 (UTC)

  RFC is "binding" if a closer judges there is a consensus to be found in the discussion. The chances of a closer finding a consensus for this RFC as worded are not that great imo. An RFC is not a straw poll, either. Selfstudier (talk) 22:11, 1 December 2022 (UTC)

  An RFC is not binding, per many discussions. If you disagree with the outcome of the discussion, you can just have another discussion and see what folks think this time. I can give you two specific examples of that happening: it took a series of RFCs to remove a nude picture from the lead of Pregnancy, and the other was a revolt at WT:MEDRS against an RFC summary that claimed the country of origin for sources must be ignored, despite peer-reviewed academic journal articles saying that publications about a particular subject, from a particular country, were extremely biased and should not be trusted. In other cases, the results of an RFC get overtaken by events, and everyone tacitly agrees to ignore it. You are not allowed to edit against consensus, but it's consensus, not the RFC, that matters.

  Rarely, ArbCom will declare a moratorium on a subject, which effectively makes the most recent discussion binding for a period of time (e.g., six months), but never permanently. Wikipedia:Consensus can change, even if there was a previous RFC on the subject. WhatamIdoing (talk) 05:17, 2 December 2022 (UTC)

  That's why I put "binding" in quotes, its valid until there is a new consensus. As you say sometimes it can be set for a period, the Jerusalem RFC was set at 3 years in force. Selfstudier (talk) 10:02, 2 December 2022 (UTC)

  I'd say it this way: RFCs aren't binding, but consensus is (at least until the consensus changes, and then the new consensus is binding). WhatamIdoing (talk) 21:17, 2 December 2022 (UTC)

• This seems to have immediately gone off the rails, with multiple editors trying to determine what is actually under discussion. The original question seems to obfuscate the issue previously bunder discussion. Close this and start again with a much clearly question, that directly identifies the policies at hand. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords°
  • This discussion is collecting information, rather than voting on any change. Editors are invited to share the advice they would give in a situation like the story I've told above. I realize that finding out what people think is the best practice will seem rather old-fashioned, especially to editors whose only experience with RFCs is voting on simple support/oppose questions, but so far, I think it's working. I'm learning that my guess on the balance of views in the community was not entirely correct. I hope you will add your own advice, so that I'll know what your view is, too. WhatamIdoing (talk) 21:12, 2 December 2022 (UTC)

  My view is that you've used a very narrow question to bring the focus of the discussion to exactly what you would like it to be. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 19:55, 9 December 2022 (UTC)

  Thank you for the compliment, unintended though it may have been. I am trying to learn more about the community's view on a very narrow point, without distractions like "except if it's a BLP" and "except if it's been there for ten years" and "except if one of them is a primary contributor" and "except if it claims that homeopathy works" and so forth, so I'm glad that you agree that I've managed to achieve my goal.

  I also hope that you might be able to believe that I'm at least potentially smart enough not to turn the responses about a dispute between two active contributors over well-sourced, non-BLP content in a brand-new article into a statement about what should happen during a dispute with a drive-by POV pushing newbie adding unsourced BLP information to a high-traffic FA.

  I do still want to know whether you would, if it really does just come down to editorial judgment (as two editors described it above), default to inclusion or exclusion, or if there is missing information that you would need before you could answer that question. WhatamIdoing (talk) 20:40, 9 December 2022 (UTC)

  No. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 21:37, 9 December 2022 (UTC)

I had occasion recently to look up WP:SPS and saw these two clauses, which I think might need some clarification:

Self-published expert sources may be considered reliable when produced by an established subject-matter expert, whose work in the relevant field has previously been published by reliable, independent publications. [...] Never use self-published sources as third-party sources about living people, even if the author is an expert, well-known professional researcher, or writer.

The latter sentence seems to leave open the possibility that SPSs might exceptionally be used as "first-party" sources; that is, sources about their creators themselves, even if it's a bit hard to imagine someone being recognized by independent publication as a subject-matter expert on him/herself.

But supposing someone took the last sentence as exceptio probat regulam, what I'm wondering is, are there standards for verifying that the SPS was actually written by the purported author?

(In case anyone's wondering, the reason for my question is an exchange at Harvey Friedman. Friedman, or someone impersonating him, apparently took a provocative position on Quora, and an IP editor thought this should be mentioned in Friedman's bio, with the seeming subtext being that it should count against Friedman's academic trustworthiness. To be fair, it's not unknown for Harvey to take provocative positions, and I personally don't think the impersonator scenario is the most likely case. I think there are several other reasons the Quora post should not be mentioned in Friedman's bio unless picked up by a secondary source, and possibly not even then, but I'm still curious about the general case.) --Trovatore (talk) 08:15, 5 January 2023 (UTC)

Twitter or blogs are fine for about self statements. If a person posts about themselves, see WP:ABOUTSELF instead of WP:SPS. The Quora should be kept out of the article because of 4. there is no reasonable doubt as to its authenticity; as per ABOUTSELF. Any reasonable doubt is enough that the source should be discounted. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 12:27, 5 January 2023 (UTC)

Hmm -- that's enough for the moment, for the specific issue at Harvey Friedman. But suppose the attribution were solid? This is another thing I think might need clarifying.

It seems to me that the statement attributed to Friedman, on its face at least, is not about Friedman but about the election. If it's about the election, then ABOUTSELF doesn't apply, and Friedman is certainly not an established subject-matter expert on elections.

But then what does it cover? Just things like "my favorite pizza is quattro formaggi"? If it's only for trivia, then I wonder whether it's worth having at all. --Trovatore (talk) 18:41, 5 January 2023 (UTC)

If the subject is stating a matter of opinion, then there is no reason to not state that they said it. You shouldn't use it in an article about a topic they are not experts in (SPS would apply), but in their own article it should be fine (obviously due, false balance, OR, synth, etc still apply). -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 19:36, 5 January 2023 (UTC)

Well, I think that would need to be clarified. ABOUTSELF refers to "information about themselves". If Joe Bloggs claims the Moon is made of green cheese, this is not a claim that Bloggs had made about Bloggs, but about the Moon. It is a fact about Bloggs that Bloggs made the claim, but the source does not report that Bloggs made the claim; it simply reports the claim itself.

So I think the statement is ambiguous. Personally I would lean towards at least discouraging the use of SPS to report claims that bio subjects make about things other than themselves. It seems to be too easy to use it tendentiously, as in the dispute at the Friedman bio. I would be interested to see cases where you think it's actually valuable and encyclopedic to include such material. -Trovatore (talk) 20:45, 5 January 2023 (UTC)

cases where you think it's actually valuable and encyclopedic to include such material. Someone stating their birthday or alma mater on Twitter, announcing the birth of their child or the wedding on Facebook, a personal blog relating a relevant detail about their childhood or creative process, etc., are examples of ABOUTSELF that may not pop up in a WP:RS, but might be able to be used under ABOUTSELF. It's not necessarily trivial, but it might be the sort detail that fills out an biography or infobox but isn't part of a RS article. — Carter (Tcr25) (talk) 21:13, 5 January 2023 (UTC)

Right, those are the things that seem to be covered by the unambiguous part of ABOUTSELF. I'm not sure whether they're really valuable enough to justify ABOUTSELF, but at least they're clearly covered by it.

The direct reporting of controversial claims made by the bio subject, not about him/herself but about something else, is another matter; it isn't clear from the text of ABOUTSELF whether these are covered or not. I tend to think they should not be (that is, I tend to think that at least in general we should not report such claims), but in any case the language should be clarified. --Trovatore (talk) 21:19, 5 January 2023 (UTC)

We can state that their opinion is that the moon is made of cheese, we can't state in wikivoice that the moon is made of cheese and use them as a reference. ABOUTSELF covers the former, SPS wouldn't allow the later. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 22:26, 5 January 2023 (UTC)

ABOUTSELF as currently written does not clearly cover the former. If it should cover it, the language needs to be clarified. My preference would be that it should not cover it, and that should also be clarified. --Trovatore (talk) 23:00, 5 January 2023 (UTC)

Sorry I don't see it.
1. the material is neither unduly self-serving nor an exceptional claim; To state that someone has an opinion is not a exception claim. Note they may have a exceptional opinion, but this isn't about verifying if that opinion is correct only that the subject has it. We can't use that self claim to say that opinion is true.
2. it does not involve claims about third parties; The opinions a subject holds is about the subject. The subject may believe that a third party is a poo poo head but we're not verifying that the third party is a poo poo head, only that the subject has that opinion.
3. it does not involve claims about events not directly related to the source; Again the opinions held are about the subject. The subject may state elections are controlled by bees but we are not verifying if that is true, only that the subject said that.
The verification that someone holds an opinion is covered by ABOUTSELF, whether that opinion should be included in the article is covered by other policies. Verification of whether any of those opinions are true must use reliable third party sources.-- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 23:37, 5 January 2023 (UTC)

The key point is that ABOUTSELF covers only material subjects write about themselves. If Joe Bloggs posts that the Moon is made of green cheese, he is posting about the Moon, not about Joe Bloggs. Therefore ABOUTSELF as currently written does not unambiguously apply to that statement. --Trovatore (talk) 23:52, 5 January 2023 (UTC)

Is your point that if Joe Bloggs states "I believe the Moon is made of green cheese," then it is an ABOUTSELF-covered statement; however, if he says "The Moon is made of green cheese," then that cannot under ABOUTSELF be used to state that "Bloggs believes the Moon is made of green cheese"? — Carter (Tcr25) (talk) 01:23, 6 January 2023 (UTC)

Well, of course if that is the effect of the current language, we surely don't want that; it would be kind of silly. Nevertheless it is a possible reading.

I also note that clause 3 says it does not involve claims about events not directly related to the source. That would seem to rule out using Friedman's Quora posts about the election, but oddly, not the green cheese, because that is not an "event". Maybe clause 3 should be broadened beyond "events"; say, to "facts"?

Anyway, we don't want editors to have to be lawyers, and making them dig for a close reading of clause 3 is clearly not optimal. I think Tcr25's earlier construction is probably close to the intent of ABOUTSELF; it's so we can share personal data that subjects have shared (birthday, alma mater, etc), not their opinions on facts. I would support rewording in the early going of the text of ABOUTSELF to make that more clear, in such a way that readers understand that this is the intent without having to dig through five clauses of sub-rules. --Trovatore (talk) 01:35, 6 January 2023 (UTC)

But it's not a claim about the event; it's a claim about the subject's belief or statements. As ActivelyDisinterested noted, ABOUTSELF here is verifying the person's stated opinion or belief, not the veracity of that opinion or belief. Sourcing a tweet and using ABOUTSELF to justify a statement like "Joe Bloggs supported efforts to introduce wolverines to Detroit city parks," should be fine ... if it is germane to the article about him. In an article on efforts to introduce wolverines to city parks, it might be acceptable to use if Bloggs's opinion is relevant, so long as it's presented as his opinion, but it could not be used to source a statement like "Releasing wolverines in city parks will help reduce Detroit's chipmunk problem," even if that's what Blogg's tweet said. That statement, in Wikivoice, makes it about the event, not the source's opinion. I think that's a pretty clear reading and doesn't seem inappropriate or silly to me or in conflict with clause 3. The claim is about the person's belief/statement/opinion, not the event or action itself. Other policies govern whether or not an edit adding that sort of claim is appropriate, but, if it is relevant, then the tweet should be an acceptable source for documenting someone's opinion. — Carter (Tcr25) (talk) 13:35, 6 January 2023 (UTC)

It says it doesn't "involve" such claims. "Involve" is quite a broad word. Reporting claims by the subject does "involve" such claims, even though they're made by the subject, not by Wikipedia.

It is at least a possible reading that a subject's opinions are not "facts about the subject" in the sense of the section. If they are meant to be taken as such, then that should be clarified. But I personally don't think such a reading is the better one. --Trovatore (talk) 16:51, 6 January 2023 (UTC)

Sorry, that seems a rather tortured reading of "involved". So long as it is clear the statement is describing the person's opinion then what's involved is the opinion about the subject, not facts about the subject. There may be other policy reasons to exclude such a statement, assuming it's relevant in the first place, and it may be necessary to rephrase a statement to make clear that it's the person's opinion, but ABOUTSELF seems a pretty clear, acceptable standard for sourcing something someone has said or supported. — Carter (Tcr25) (talk) 17:21, 6 January 2023 (UTC)

I don't think it's a tortured reading at all. Look, I understand your view. I don't think it's clearly expressed by the existing language, or that it's even necessarily everyone's understanding of the current intent.

If that is the view we want to take, then the language should be clarified. But first we should make sure we're in agreement about what the policy should be or currently is. --Trovatore (talk) 17:30, 6 January 2023 (UTC)

I think I understand what you want: ABOUTSELF-supported statements to be things unique to a person (family, school, work, self), but not what they believe about something external to their life. Is that correct? So how would you clarify it?

Current: 3. it does not involve claims about events not directly related to the source

What would you change to make it fit your preferred understanding? How would that differ to cover a (relevant) statement a person makes about their beliefs or opinions? Carter (Tcr25) (talk) 17:52, 6 January 2023 (UTC)

Fair question. I'll try to get back to it after my work day. --Trovatore (talk) 18:35, 6 January 2023 (UTC)

Almost all wiki rules leave some things open to interpretation. This is a necessity, because "interpretation" means taking other things into consideration rather than just the one rule dictating the result. And keep in mind that sourcing not only relates to satisfying the wp:ver rule, it also affects inclusion/inclusion under other rules and considerations. Editors might require a more thorough authenticity check if their Twitter feed says "I admire Hitler" than if it says "I went to the Bears game on July 15th" . If they ostensibly make an outlandish claim that makes them look stupid, it might be a matter of interpretation whether they meant it or meant it as a joke and editors might demand a secondary source to allow it in as a statement of their belief. Editors might decide on a stricter interpretation of the wp:ver rule on an edge case on the above discussed criteria if the material looks worthless for the article. North8000 (talk) 21:29, 6 January 2023 (UTC)

I've sent the article that prompted this to Wikipedia:Requests for page protection, because the way to resolve a dispute with an IP who isn't showing any signs of hair-splitting about the policy wording is to make it impossible for them to screw up that article, rather than tweaking the wording in the policy. Wikipedia:Nobody reads the directions. WhatamIdoing (talk) 06:56, 7 January 2023 (UTC)

One concern I've had about ABOUTSELF in the past is with editors taking things from subjects' social media feeds and other verified, WP:ABOUTSELF-able sources for things that could potentially cast the writer in a negative light (often place like Twitter or Facebook, where huge amounts of content are posted), then highlighting it on Wikipedia. An extreme example would be eg. taking a quote of someone using a racial slur or expressing an opinion that is generally considered abhorrent, cited solely via ABOUTSELF with no secondary sourcing; another example might just be a statement that is wildly at odds with the subject's usual positions. I think that that's clearly not allowed (it's WP:OR / WP:SYNTH in a way that violates BLP) but it could be worth making it more clear in ABOUTSELF itself. It already forbids exceptional claims, but by my reading that is just for when the subject themselves is making some exceptional claim, not for eg. quoting them saying something stupid or reprehensible or something that makes them look like a hypocrite. So I think we should have an additional clause for statements with exceptional or BLP-sensitive implications about the subject. --Aquillion (talk) 06:42, 4 January 2023 (UTC)

Aquillion, I think that it would be useful to have at least two or three real world examples, so that editors commenting can see the alleged problem in context. Cullen328 (talk) 06:51, 4 January 2023 (UTC)

The section I just added below seems not unrelated. I read this section after posting. --Trovatore (talk) 08:47, 5 January 2023 (UTC)

I think that's a DUE problem, not a WP:V problem. WhatamIdoing (talk) 06:35, 7 January 2023 (UTC)

I think this applies to most of these situations, yes ABOUTSELF means they are verified but that doesn't mean they should be included. If an article subject posts something stupid or contentious, so what. People have been upset about it on social media, so what. Unless it is picked up by reliy sources it doesn't seem DUE in the article. Also there seems to be situations where quantative judgements are being made when using the subjects own posting, e.g. the subject posts the moon is made of cheese and an editor adds "subject is a loony" referencing the subjects post. Well that fails either WP:OR or WP:SYNTH, as well as verification. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 13:46, 7 January 2023 (UTC)

I decided to start this section because I've just realized that, whereas WP:NPOV page says

"This policy is non-negotiable, and the principles upon which it is based cannot be superseded by other policies or guidelines, nor by editor consensus."

the WP:V, as well as WP:NOR, WP:BLP etc does not say that. Does it mean that, in contrast to NPOV, this policy IS negotiable, and its principles may be superseded by editor consensus? The answer to this question may be important for resolving the possible contradiction between ONUS and NOCON.--Paul Siebert (talk) 04:33, 2 December 2022 (UTC)

WP:V, WP:OR, and WP:NPOV all have the same status - see Wikipedia:Core content policies. BilledMammal (talk) 04:35, 2 December 2022 (UTC)

...but only NPOV is explicitly "non-negotiable". I am asking again, if they have the same status, why that clause was not added to V/NOR? Maybe, it makes sense to add it? Paul Siebert (talk) 04:39, 2 December 2022 (UTC)

It should go without saying that all policies are non-negotiable. The addition of that text in one of them was likely a user error. Chris Troutman (talk) 04:41, 2 December 2022 (UTC)

Actually all policies are to varying degrees. Wikipedia:The rules are principles. NPOV just happens to be one with very few exceptions or edge cases. Andre🚐 04:55, 2 December 2022 (UTC)

NPOV was labeled as non-negotiable because there was a good soundbite about it in the early days. I wouldn't read too much into that. WhatamIdoing (talk) 05:03, 2 December 2022 (UTC)

I know that a lot of editors really enjoy nitpicking about wording. But in my view, our three core content policies, WP:OR and WP:NPOV and WP:V are non-negotiable. Also, compliance with WP:BLP and WP:NLT, policies with legal implications, should always be non-negotiable. In all of these examples, debates about edge cases are permissible, as long as nobody argues that the policy itself should be set aside. But no argument to the effect that "this topic is so interesting and relevant and useful that we can ignore verifiability" can ever be accepted as legitimate in any way. Cullen328 (talk) 06:10, 2 December 2022 (UTC)

• There is considerable fuzziness around the difference between negotiation and interpretation. A policy can be non-negotiable and still people can legitimately interpret a specific application of that policy in different ways. --Jayron32 12:05, 2 December 2022 (UTC)

I think this is the best answer so far. The "principle" or "theory" is non-negotiable, but this allows a lot of space where necessary. We have 3 core content policies which all need to be considered. There is also WP:NOT, and of course WP:IAR. We should keep all those principles in mind.--Andrew Lancaster (talk) 17:10, 22 December 2022 (UTC)

• NPOV was a founding principle, and I seem to recall that a similar page was once hosted here. The idea maybe that policies and guidelines were all negotiable—make it up as you go along—but that the founding principles were non-negotiable. That might be the origin of the wording in NPOV that is missing in the others. fiveby(zero) 12:58, 2 December 2022 (UTC)

  It's perfectly fine. The idea that we strive for neutrality is indeed not negotiable, but it's also not a big deal. If someone is arguing that, sometimes, we want to be deliberately biased, that's the only thing I can think that objecting to the non-negotiability of NPOV could mean, then I don't know what to say about that. --Jayron32 13:27, 2 December 2022 (UTC)

  I wasn't saying that it was not fine, but that the wording might be an artifact of the good old days of ignore all rules and product, process, policy ([4],[5]) Reading This policy is non-negotiable and the lack of that wording in V and NOR might be misleading. A better statement might be: The NPOV principle is non-negotiable, but go ahead and change the NPOV policy if it leads to better product. fiveby(zero) 14:29, 2 December 2022 (UTC)

  You seem to be saying that the phrase "non-negotiable" is preventing people from editing the policy... Evidence is clear that you are just wrong about that. --Jayron32 14:40, 2 December 2022 (UTC)

  No, i am trying to help answer Paul Siebert's question: I am asking again, if they have the same status, why that clause was not added to V/NOR?. fiveby(zero) 14:52, 2 December 2022 (UTC)

  ¯\_(ツ)_/¯. Why does it need to be? --Jayron32 15:21, 2 December 2022 (UTC)

  Don't ask me, as a reader i see the policies and guidelines often leading to bad article content, so dump them all and reboot. As someone who has to maintain the encyclopedia and control the "anyone can edit" you probably see things differently. fiveby(zero) 15:32, 2 December 2022 (UTC)

  I'm not sure where an anarchic free-for-all with no regulations at all would make for a better written product...--Jayron32 15:36, 2 December 2022 (UTC)

• As someone else wrote, I think that the difference in wording is more of a historical accident than a deliberate distinction. Concerning negotiation, don't forget WP:CONLEVEL: "Consensus among a limited group of editors, at one place and time, cannot override community consensus on a wider scale." It means that although a talk page discussion can establish a consensus that some text is in conformity with WP:V or not, a consensus to commit a violation of WP:V is null and void. Zero^talk 03:48, 3 December 2022 (UTC)

NPOV is just more comprehensive in focus. Underlying NPOV is V in singular and V in conjunction other other V (while V's focus is singular). NOR is a special case related to NPOV and V, it is delimiting a tempting means of editor bias that departs from V and NPOV. -- Alanscottwalker (talk) 11:45, 22 December 2022 (UTC)

"Original research" means the editors made stuff up on their own. Making stuff up in a strictly impartial, unbiased way is still making stuff up. I'm not sure that it's really related to bias per se, although it will overlap in practice.

One of the formative early moments for the core content policies was dealing with a crackpot very enthusiastic editor who believe he had single-handedly disproven Einstein's theory of relativity. Thus we get DUE ("No, your crackpot ideas don't deserve to be mentioned in this particular article") and NOR ("No, your crackpot ideas don't deserve to be mentioned anywhere in Wikipedia unless a source published them outside of Wikipedia") and especially SYNTH ("No, your crackpot ideas aren't acceptable, even if you claim that it's 'obvious' that the combined meaning of these sources agrees with you, even though none of them says anything even remotely like 'we agree with that crackpot from USENET'."). WhatamIdoing (talk) 23:01, 22 December 2022 (UTC)

An "impartial way of making stuff up"? If you are making stuff up (and committing it to writing) there is no way you are impartial to it. You are biased in its favor, explicitly so. -- Alanscottwalker (talk) 08:40, 23 December 2022 (UTC)

I don't think so. I could write, for example, that several large vehicles drove past me an hour ago, or that I'm wearing a red shirt, or that I'm typing on a laptop. All of these facts are unverifiable; putting any of them in an article would be a violation of NOR. They constitute "material—such as facts, allegations, and ideas—for which no reliable, published sources exist." But they are still impartial, unbiased, objectively true facts. You would never say "Oh, WhatamIdoing, you're so biased: those vehicles weren't truly 'large', your shirt isn't actually 'red', and you're not 'typing on a laptop'. If you looked at the situation objectively and impartially, you'd know that you were interpreting everything in a biased way." WhatamIdoing (talk) 21:41, 27 December 2022 (UTC)

I would say you're biased: in at least two ways, you are biased by your personal experience, and biased to put your personal experience in an encyclopedia article, leading to any article that is biased to convey your personal observation. (BTW, It seems odd that you switched from "making stuff up" to personal observation, but that does not matter, either way.) Alanscottwalker (talk) 23:26, 27 December 2022 (UTC)

Do you think that an unbiased person would have different facts?

NPOV would care about whether to include one person's personal experience. But when I take the exact definition of OR out of the policy, it doesn't have anything at all to do with bias.

Compare and contrast:

• "NOR is...is delimiting a tempting means of editor bias"
• "OR is...material—such as facts, allegations, and ideas—for which no reliable, published sources exist."

Wikipedia:No original research does not include the word bias anywhere. I can't imagine why you think it is about bias, instead of about stuff that nobody else has published before. WhatamIdoing (talk) 05:44, 28 December 2022 (UTC)

In your scenario, no one else here has any facts, none they can examine, all they have your say so. They literally have to take your word for it, or not at all (and in Wikipedia, it is not at all.)

If it's stuff that nobody published before, it has to come from somewhere, that somewhere and is the editor, and the editor is putting their editorial bias into the article. The editor is conveying their own thoughts, feelings, or analysis, biasing the article to convey their own thoughts, feelings, or analysis - editorial bias. -- Alanscottwalker (talk) 08:54, 28 December 2022 (UTC)

You're absolutely right: all you have is my self-report. That's why I say it's "material—such as facts, allegations, and ideas—for which no reliable, published sources exist", aka "original research".

You, on the other hand, think that original research means is about editor's bias, rather than unsourceable content. But where's the actual bias in saying my shirt is red? Where are my own "thoughts, feelings, and analysis" about the color of my shirt? Color is an objective fact. The dominant wavelength reflected by an object can be measured with a smartphone. (There's an app for that.) Bias, according to the article, is "a disproportionate weight in favor of or against an idea or thing, usually in a way that is closed-minded, prejudicial, or unfair" (italics in the original, though they probably shouldn't be). Do you think I'm being "close-minded, prejudicial, or unfair" when I say my shirt is red? It's 100% solid crayon red. Do you think I'm giving "disproportionate weight in favor of or against" the color when I accurately report the color? I don't, and if you saw the shirt, I suspect you wouldn't, either.

I don't think that NOR is about bias at all. NOR gives a definition of OR that doesn't mention bias. NPOV isn't mentioned in NOR except to say that it is one of the other core content policies. I am beginning to feel like NOR is not well understood (perhaps similar to the way that WP:QUO doesn't say what its proponents claim, that WP:BRD doesn't say what certain reverters claim, etc.). We seem to project into NOR whatever gap we feel is being overlooked. In this case, editorial bias is covered by NPOV, beginning with the very first sentence. It does not happen to be part of NOR. WhatamIdoing (talk) 03:11, 29 December 2022 (UTC)

OR is closely related to SYNTH. OR is indeed about reflecting what the sources say, but you can write OR without it being unourced. It is about following and not leading the sources. How it's related at times to having a bias or an axe to grind is when some novel crank comes to Wiki promoting their fringe theory. But garden variety original research could simply be citing material from a primary source to make a conclusion. But NOR is not just a special case of NPOV. I agree that NOR isn't about bias or NPOV, but it's not only about being well-referenced either. It's about not introducing a new idea or interpretation, even of a valid reference. Andre🚐 03:58, 29 December 2022 (UTC)

When someone writes 'my shirt is red' into an encyclopedia article, they are not just making an assertion about their shirt, they are making the assertions that they are a valid source and that their observation of their shirt matters to the topic. They are biasing the article to these assertions. You are close-minded and in favor concerning this fact as encyclopedic (and you want the reader to be, too).

When someone writes in an encyclopedia article that in effect, 'this conclusion is drawn from this fact, based on my say so' they are biasing the article to identify the relationship and to state their conclusion. They are close minded and in favor of these assertions (and they want the reader to be, too).

When someone writes in an article in effect, 'This theory, I created matters', they are biasing the article to their assertions. They are close minded and in favor of it (and they want the reader to be, too).

When someone writes in an article that in effect, 'I need you to know, that this source for 2, added to this source for 2, means 5' or any other synth, they are biasing the article to these assertions. They are close-minded and in favor of it (and they want the reader to be, too).

These are all ways the editor decides (closes their mind, and yours too, since they want you to accept what they write) to prize their 'facts, allegations, and theories' above others (the source material that is published).

But let me add, I think it is rather odd (including WAID'S now claim to superior exegesis, indeed WAID seems to assert there is some kind of canonical exegesis), that WAID so much wants to quibble the very short statement, I originally made, since there is no dispute on the effective rationale and outcome: viz, 'there are many ways to do it, but don't stray from the sources.' (the temptation and indeed intellectual predilection (bias) to write in our pedia about ourselves, viz our own facts, our own allegations, our own theories instead, of what we need to be doing, is strong, and policy is useful against it.)Alanscottwalker (talk) 15:01, 29 December 2022 (UTC)

I agree with your statement that adding any content to any article is implicitly asserting that the content is relevant to the article. I don't agree that adding content is "biasing the article to these assertions".

You've probably never been to Mulberry, Kansas. (I haven't either.) Imagine that there were no existing Wikipedia article for the town. Then consider this story:

• I drive through the town. I notice that it's small. I create a Wikipedia article that says say "Mulberry is a small town in Kansas."

Am I biased? No. Is the article biased, now that it says the place exists and is a small town? Also no.

Compare it against this story:

• I find the US census report from 2020. It says the population is 409. I create a a Wikipedia article that says say "Mulberry is a small town in Kansas."

There's no difference in the article content, but you've been saying that the first is biased and the second isn't. WhatamIdoing (talk) 00:16, 1 January 2023 (UTC)

I think you're touching on an important point about how we write articles now versus how many articles were written back in pre-2005. Back then, you might start an article about a topic that you knew about from original research and fill in the references later. I would still argue that "Mulberry is a small town in Kansas." is a perfectly fine stub, but what would be original research would be to write, "Mulberry's most popular sandwich shop is Sam's Deli." To do proper research you'd need to find a source for what is the most popular sandwich shop, and it may not be what you thought it was. I wouldn't call that bias though. I would call that an educated guess or an assumption based on background knowledge or personal experience. And it's not necessarily wrong to let those assumptions guide you as long as you provide appropriate sources. The problem is when you start to look for sources to support a preconceived assumption and selectively interpret. Andre🚐 00:26, 1 January 2023 (UTC)

Given the size of the town, and a back-of-the-envelope estimate that you could expect one restaurant per thousand residents, it may not have any restaurants at all, much less the possibility of having more than one sandwich shop, but I agree with you that "Mulberry is a small town in Kansas" would make a perfectly fine, NPOV-compliant stub. WhatamIdoing (talk) 06:53, 7 January 2023 (UTC)

Yes, you are biased and so is the article. If you write in an article "Mulberry is a small town in Kansas," just based on you driving through it, you are asserting the legal identity of this place, basing the article only on your conclusory bias of what you think a town should be. There are other things this place could be besides a town, it could be a village or have some other form or no legal form at all, but you don't know, and your reader has no idea, except the one you constructed for them, because all the reader has is the bias of your personal observation. -- Alanscottwalker (talk) 22:09, 3 January 2023 (UTC)

I think I need you to tell me what definition of bias you're using.

For example, I could say that your analysis is biased, i.e., that it is exhibiting "disproportionate weight in favor of or against an idea or thing, usually in a way that is closed-minded, prejudicial, or unfair" and "personal and sometimes unreasoned judgment" in suggesting that the English word town can only be applied to places with a particular legal identity.

But I'm not at all sure that it is biased of me to use the word town when I describe any "a compactly settled area as distinguished from surrounding rural territory". WhatamIdoing (talk) 01:19, 7 January 2023 (UTC)

I think he is referring more to cognitive bias or maybe confirmation bias, either way though, it would appear to be an unusual usage or unique coinage to describe the concept he is describing. Andre🚐 07:11, 7 January 2023 (UTC)

I think your problem is, here, not accounting for the context you are writing in (in short, our work is simply, not about you). As readers of your articles, no one cares or wants to know one whit about your use of the word 'town' in you personal life, nor do they want to know what you see from your car, nor do they care what you think about while driving around (in short, none of what you are writing has any weight).

This, here, is not about your personal life, it is not your journal (it is the encyclopedia's articles), thus not the place for you to write about yourself (your ideas, your experiences). In the examples you put forward of writing in our articles - your various personal views from you car - you not only greatly value your experiences above others (the sources and the readers) by putting them in an encyclopedia article for others, you literally base the articles on your point of view. Since what you are writing has no weight, the weight you do give it, is more than disproportionate (indeed its only measure of value is your personal bias, you chose it, and what you chose is about you -- you chose to value your personal experience). -- Alanscottwalker (talk) 15:15, 8 January 2023 (UTC)

So, standard dictionary definitions are about my personal life? Are editors not allowed to Wikipedia:Use our own words? NOR should stop telling editors that to be "Rewriting source material in your own words while retaining the substance"?

I don't think that I've based this hypothetical single-sentence article on my own point of view. I think I've based it on these points:

• Settled places (e.g., towns) are considered notable on wiki.
• It's widely accepted among editors that there should be a separate article for every separate such place.
• The two most common and obviously encyclopedic facts about any settled place are (1) its location and (2) its size – which is the only information I've included.

What I'd like to explore more with you is that you say that if I write this sentence based on seeing the town in person, then the sentence is biased, but if I write this sentence based on a book, then the sentence is unbiased. It does not seem to me that this is actually possible. A sentence can certainly be biased, but the method I use to construct the sentence does not seem to me to be especially relevant to determining whether the result is biased. WhatamIdoing (talk) 01:12, 9 January 2023 (UTC)

You changing your story about why you wrote the article and what you base it on, just highlights the dishonesty. Had you been honest with your readers, what you would have written is 'I was driving in Kansas and I came across this small group of houses and buildings, let's call it the Town of Mulberry.' Of course then your readers response would be wtf kind of pedia is this and other editors would know what you were up to.

The misrepresentation of what is being presented, accepted summarization vs. what comes-off-the-top-of-your-head should be plain. When it comes off the top of your head, it is nothing but your bias, your pov. (How sources are different, should also be plain, to the extent reading sources is an observation, it is one we all, readers, everybody can have in-common (a common observation), and we all can repeat and reconfirm it endlessly. Alanscottwalker (talk) 17:25, 9 January 2023 (UTC)

(As for the rest, your initial rhetorical questions are so unresponsive to my comment, or beside the point, they are best ignored). Alanscottwalker (talk) 17:04, 9 January 2023 (UTC)

What made you assume that I invented the name of the town? WhatamIdoing (talk) 17:17, 9 January 2023 (UTC)

Because you said, your only means of knowledge was driving through it. If you have some other source (say, you read something), that is what the reader should know. Alanscottwalker (talk) 17:31, 9 January 2023 (UTC)

American towns tend to have signs at the city limits that contain the name of the town. Most are official, but some are unofficial. Merely driving through is almost always enough to discover the actual name of the town. In this case, the reason that I know the name of the town is because of it's semi-famous newspaper, The Mulberry Advance. What I would really discover from driving through is that it is a small town (and not e.g., a larger city). WhatamIdoing (talk) 00:02, 10 January 2023 (UTC)

So, you are basing it on the sign (not on driving through), the sign is a written, and it is something anyone can go read. You now have a source that is not you. (not a great source, not, alone what I would write any article on but that is your source.)Alanscottwalker (talk) 00:07, 10 January 2023 (UTC)

And what is my source for "small town" or "in Kansas"? WhatamIdoing (talk) 00:32, 11 January 2023 (UTC)

In the fuzzy wiki system trying to parse out categorical logic is not a good idea. Situations vary in how directly/clearly core policies clearly / unilaterally determine a result....most are not not slam-dunk clear and for those the policy merely influences the result. IMO the existence of that phrase is to emphasize the importance given to the concept of neutrality. Trying to read more into it that is not a useful exercise including trying to interpret difference between core policies based on whether or not they include that phrase. When it hits the fan, the more prescriptive/operative wording/ concrete rules in policies tends to have more effect than statements of general principles. NPOV is more lacking in usable prescriptive/operative wording/ concrete rules, so perhaps that is an argument to compensate by giving it an extra boost from uniquely having the "non-negotiable" wording. Sincerely, North8000 (talk) 23:38, 22 December 2022 (UTC)

What does "negotiable" mean in this context anyway? If a policy were negotiable, with whom would we be negotiating it? I would remove all instances of "is non-negotiable" from all policy pages as being meaningless rhetoric. If what we mean is that there are no exceptions, we should say that. Better yet, if what we mean is you can't IAR a core content policy, we should say that. Levivich (talk) 16:40, 30 December 2022 (UTC)

Good question, Levivich.

I understand the negotiability concept in terms of symmetry. If some issue, e.g. whether we choose the version "A" or "B", is negotiable, the choice is a subject of negotiation: if more arguments have been proposed in support of "A", this version is selected. If consensus is not achieved, a stable version (either "A" or "B") is restored. This symmetrical situation usually happens when both "A" and "B" comply with NPOV, V, or NOR. A simple example is a choice of an infobox photo: if all candidate are in PD, we are free to select any of them, and the candidate that got more voices wins.

However, when we speak about violation of NPOV, V, NOR, or BLP, no symmetry between "A" and "B" exists. Normally, in this type of discussion, some user claims that "A" (or "B) violates some policy. For example, it contains some unsourced claim, which is not obvious. This type dispute cannot be resolved by negotiations: the users who disagree that "A" violates the policy must present evidences (in this case, it should be some RS that directly supports "A"). If the evidences are insufficient, the users cannot refer to NOCON: it does not matter which version, "A" or "B", was stable.

In other words, if someone claims that "A" (or "B") violates V, or NPOV, etc, and there is no consensus if that claim is valid, the outcome of the discussion can be summarized in two symmetrical ways.

• "There is no consensus that "A" violates the policy", or
• "There is no consensus that "A" complies with the policy".

In my opinion, "non-negotiable" implies that we always chose the second interpretation.

And one more explanation: if some issue is negotiable, the proposer is supposed to seek consensus to implement the change, and if consensus has not been achieved, the edit is reverted. In contrast, when some user notes that some piece of text violates neutrality, verifiability, etc., the burden of proof is on those who want to keep that questionable text. Paul Siebert (talk) 07:14, 5 January 2023 (UTC)

I think of it as operating at a more basic or fundamental level. "Not optional" or "not a principle you can negotiate away/reject by local consensus" might be a fair substitute for "non-negotiable".

To build on your story, if someone claims that an article violates NPOV, then editors don't get to say "Oh, you're right, it really is non-neutral, but in this instance, it's okay for the article to be non-neutral, because writing a neutral article would have these other problems [legal? moral?], so we all agreed that it was okay to have a non-neutral article here." WhatamIdoing (talk) 06:34, 7 January 2023 (UTC)

Im cross-posting this from Help_talk:Citation_Style_1 as that page seems to be mostly for discussing technical difficulties.

What is template:cite thesis, if not a giant violation of Wikipedia's policy against original research? By definition a thesis is original research, as it is something submitted by a student to the head of a department at a university. Am I missing something? Are we now considering what is effectively a very large homework assignment to be on equal footing with peer-reviewed research? —Soap— 23:44, 3 February 2023 (UTC)

Per the Scholarship section of the policy at Wikipedia:Reliable sources;

Dissertations – Completed dissertations or theses written as part of the requirements for a doctorate, and which are publicly available (most via interlibrary loan or from Proquest), can be used but care should be exercised, as they are often, in part, primary sources. Some of them will have gone through a process of academic peer reviewing, of varying levels of rigor, but some will not. If possible, use theses that have been cited in the literature; supervised by recognized specialists in the field; or reviewed by independent parties. Dissertations in progress have not been vetted and are not regarded as published and are thus not reliable sources as a rule. Some theses are later published in the form of scholarly monographs or peer reviewed articles, and, if available, these are usually preferable to the original thesis as sources. Masters dissertations and theses are considered reliable only if they can be shown to have had significant scholarly influence.

PhD theses are not "original research" in the meaning of Wikipedia:No original research. Donald Albury 00:11, 4 February 2023 (UTC)

"Original research" in Wikipedia policy refers to research conducted by Wikipedia editors and cited by themselves. It does not refer to published research. Zero^talk 00:17, 4 February 2023 (UTC)

@Soap, the picture in the lead of NOR might help: "Outside Wikipedia, original research is a key part of scholarly work. However, Wikipedia editors must not base their contributions on their own original research." WhatamIdoing (talk) 06:43, 4 February 2023 (UTC)

Okay thanks. No, the essay that brought me here is not a doctorate or masters' degree course requirement, but is part of a bachelor's degree. I don't know if that matters, but yes, universities often require students to write a thesis to get a bachelor's degree as well, depending on the field. I don't know how to tell if it's been approved or just submitted, or if that matters either since the text above isn't clear.

It still seems to be a violation of our no original research policy to me regardless, as by definition a thesis submitted as part of a course requirement ... whether for a bachelor's degree, master's degree, or something else ... has not been peer reviewed or even published, unless we have looser definitions of peer review and publish that apply just to graduating university students. Thanks, —Soap— 08:20, 4 February 2023 (UTC)

It wouldn't be OR in the Wikipedia sense as that only applies to editors of Wikipedia. However the thesis of a bachelors degree wouldn't normally be considered a reliable source for referencing. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 10:11, 4 February 2023 (UTC)

Okay thank you. I will wait a bit on this so that there isnt a direct link between this conversation and my action. I will bring this up with the editor who added the PDF link. I should clarify this is certainly not a case of someone adding their own thesis ... it's an uninvolved third party. As for the distinction between primary sources and original research ... personally I think they're two sides of the same coin, but I understand that it's easier to keep the two policies separate. —Soap— 02:31, 5 February 2023 (UTC)

They can both fall into the bigger category of "stuff that shouldn't be in an article", but NOR is what Wikipedians must never do, and PRIMARY is about non-Wikipedians. The history of NOR is interesting; a lot of it was developed around a Usenet crank who believed he had disproved Einstein's theories of relativity, and he thought that Wikipedia would be a fine website for letting the world know about his crackpot ideas, since all those stuffy, out-of-touch physics journals just laughed at him. Looking back at it now, when almost everybody already knows that Wikipedia is solely for collecting information that came from other sources, it feels strange, but back then, Wikipedia really needed to write down a rule to tell people that, even if you were absolutely convinced that you knew more about physics than Einstein, you still had to get it published in a reliable source before you could put it in Wikipedia. That's the core point behind NOR: Wikipedia itself is not a primary source; Wikipedia itself is never the first place for Wikipedia's own editors to tell the world about their own discoveries. WhatamIdoing (talk) 04:09, 5 February 2023 (UTC)

Hi Zero0000

"Onus" is a very rare English word. Do you agree?

So it "needs" to link to an explanation. Agree?

Actually I wanted to simply make "onus" into a link. But Onus is a disambiguation page. (Besides the surnames) it offers 3 titles, none of them has the word "onus" in the beginning.

• The article Blame does not contain the word "onus" at all.
• Burden (disambiguation) even is another disambiguation page.
• I had the impression that Legal burden of proof (which redirects to Burden of proof (law)) fits best to understand the use in this article Wikipedia:Verifiability.

Do you mean that the article Legal burden of proof respectively Burden of proof (law)) does not fit AT ALL as an explanation for the use in this case?

The explanation and origin of a word does not say something about the current context it is used in. The English language uses many words in new meanings and contexts, some far away from it's original meaning and use. Agree?

A lack of an explanation is worse than an explanation which might give a slightly false impression. Agree?

To what would you link "onus"?

Ping welcome, Steue (talk) 15:25, 22 January 2023 (UTC)

'Onus' in a Wikipedia policy/guideline context (or any such context I can think of) doesn't remotely meet the standard required to meet a 'legal burden of proof'. So not 'slightly misleading', entirely misleading. An invitation to Wikilawyering. AndyTheGrump (talk) 15:35, 22 January 2023 (UTC)

The capacity for WP:ONUS to give rise to discussions without end never ceases to amaze me. Of course the sentence could be reworded so as to do without the word 'onus' - Inclusion of disputed content requires consensus. Selfstudier (talk) 15:40, 22 January 2023 (UTC)

Good suggestion. Zero^talk 22:10, 22 January 2023 (UTC)

I don't think it needs a link, but if others want one, we could link to wikt:onus. Crossroads ^-talk- 23:43, 22 January 2023 (UTC)

For a dictionary definition like this, linking to Wiktionary would make the most sense to me as well. Not convinced its necessary though. Sideswipe9th (talk) 23:47, 22 January 2023 (UTC)

• I disagree it is a terribly rare English word. I've known about it for as long as I can remember. I mean, it's not like "the" or "and", but it certainly isn't an unheard of word to me. --Jayron32 15:36, 30 January 2023 (UTC)

  OTOH, I checked a couple of lists of the most common words in English, and it was on none of them. WhatamIdoing (talk) 01:15, 31 January 2023 (UTC)

  The common English vocabulary size is usual reported as 20,000 words, how big are the lists you're checking? -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 11:08, 31 January 2023 (UTC)

  I didn't want to pay for https://www.wordfrequency.info/ so I looked for free lists. I found several that were 10K to 60K, none of which contained it. Tonight, I found it at wikt:en:Wiktionary:Frequency lists/PG/2005/08/80001-90000#82101 - 82200. If that position in the 80K range is at all accurate, and the reported size of a typical English vocabulary of 20K is approximately correct, then that probably explains why some people think it's rare. However, I'm a little doubtful about that list, though, because it puts it after words like jests and libel and membrane but before words like physic, pivoted, and pronoun, and none of those seem unusual to me. WhatamIdoing (talk) 06:21, 1 February 2023 (UTC)

  I also wonder if this is a VAR issue, as it's as I'm in the same camp as Jayron32. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 11:10, 31 January 2023 (UTC)

  Value-at-risk? (kidding)

  Agree with Jayron32 that it's a common word. Linking to Wiktionary would be fine, but I'm not sure that's necessary. DFlhb (talk) 12:28, 31 January 2023 (UTC)

  I agree. Onus doesn't seem unclear or uncommon to me, but a link to Wiktionary (noting that sense #5 there is intended) does no harm and would help those unfamiliar with the word. — Carter (Tcr25) (talk) 13:42, 31 January 2023 (UTC)

  One of my pet peeves: Editor A says "this is a problem" and Editor B replies "not for me." Well, Editor B, good for you. My hat is off to Editor C, who says "not for me, but here's something we can do for all of the Editor As out there." It looks like this conversation is coalescing around a consensus to add a link to Wiktionary. - Butwhatdoiknow (talk) 15:12, 31 January 2023 (UTC)

  Just as a counter-argument; every word is likely to be unknown to somebody on the planet. We don't need to link every word merely because one or two people haven't heard of it, lest we create a WP:SEAOFBLUE problem. I have no problem specifically with linking to Wikitionary if it is necessary; I am only arguing that it isn't necessary here. If others feel differently, have at it. --Jayron32 15:56, 31 January 2023 (UTC)

  Links would only be helpful to people who read English, though, and most words in a sentence are high-frequency words. People have written books with fewer than 100 words before, and we have a whole Wikipedia dedicated to the idea that we can write about almost anything with just a couple thousand words. I'd be sorry to have an WP:ALLWIKI model, but linking to words that some proficient English speakers say that find unfamiliar would not take us very far down that slippery slope. WhatamIdoing (talk) 06:25, 1 February 2023 (UTC)

  I worry about using "necessary" as the standard. How about "helpful" instead? While I agree with your reductio ad absurdum argument that one ignorant person does not create a problem, if a link would help a substantial number of folks then why not do it? - Butwhatdoiknow (talk) 07:01, 1 February 2023 (UTC)

  My question was whether your "substantial" standard had been met in this case. I've not known someone who was so unfamiliar with the word as to think it needs further definition, like my experience with grown adults unfamiliar with the word is currently at 1, being the OP of this thread. Whether or not you consider "1" to be substantial, I'll leave that up to you. --Jayron32 15:01, 1 February 2023 (UTC)

  Yeah, even "substantial" is problematic. But it - or maybe "significant" - seem better than "necessary" and "helpful." Unfortunately, they're both difficult to implement. Does the survey group consist of your acquaintances? Mine? Editors who choose to participate in a talk page discussion?

  In this case, the word is crucial to the text and WhatamIdoing has given us an objective measurement of how common the word is. With that in mind, I thank you for your willingness to compromise on this issue. - Butwhatdoiknow (talk) 16:39, 1 February 2023 (UTC)

  That list checked had onus as more common than pronoun, I doubt its reliability. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 16:47, 1 February 2023 (UTC)

• One argument for linking: This article from the BBC states that people typically know 15,000 to 20,000 Lemmas (roots) in their native language, but second language learners (not immersive) typically learn only 2,000 or 3,000 lemmas in that second language, even after years of studying. As many of our users read English as a second language, we cannot expect them to know all of those lemmas in the 3,000 to 20,000 range. - Donald Albury 18:01, 1 February 2023 (UTC)

  Agreed; I now support linking. We should be mindnful of ESL people too. DFlhb (talk) 18:09, 1 February 2023 (UTC)

  That's a very good point. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 18:19, 1 February 2023 (UTC)

Link to Wiktionary? If we need a definition or synonyms, put in cites to RS dictionaries or thesauruses, not to a wiki (which pursuant to this policy are not RS). The major point of this policy is find, rely, and cite to RS. -- Alanscottwalker (talk) 22:39, 1 February 2023 (UTC)

That seems unnecessary. We're trying to help some people understand what the word means, not to prove that it could be found in a reliable source and wasn't made up by some editor. A Wiktionary link is adequate for that purpose. WhatamIdoing (talk) 06:42, 4 February 2023 (UTC)

That makes no sense, obviously it's just as easy to link to a real dictionaries [6], [7], including ELL, [8], which exist to help "understand what the word means". You are not helping them by linking to a wiki, you are indicating to them how to not rely on and use RS (here, real dictionaries are something that every user of English should use, and should know how to use), and you are actively harming the integrity of the project and this policy. -- Alanscottwalker (talk) 15:25, 5 February 2023 (UTC)

If we're talking about reliable sources then dictionaries would be tertiary sources, what we really need is a secondary source by a recognised expert discussing the meaning of the word onus. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 16:21, 5 February 2023 (UTC)

No. Tertiary sources from reliable publishers are general RS. Alanscottwalker (talk) 22:52, 5 February 2023 (UTC)

That makes no sense?

We want people to know what the word means. We can achieve that by:

1. telling them what the word means on this page, and/or
2. linking to a different page that tells them what the word means.

If we take the second option, then as long as the linked page tells them what the word means, who cares what page is linked to? It could be WP:Definition of onus. It could be wikt:onus. It could be a link to any of the dictionary websites. We don't need a "reliable source" for the definition, as we're not supporting article content. We're not trying to demonstrate that this definition is verifiable (see WP:NOTPART). We just need a link to a page – any page; even a tweet from Jimmy Wales would do – that contains the information that we need people to have. That's the sole firm requirement here: People have to end up knowing what the word means. So long as we get that information to the reader, then we are helping them and not harming anyone or anything.

I have a slight preference for linking internally, as we have better privacy practices than most of the external options. I don't feel a need to send someone to a page with advertisements and tracking cookies just to get this definition to them. WhatamIdoing (talk) 17:39, 9 February 2023 (UTC)

• Ok… time to think outside the box. If the word “onus” is so problematic, just swap in a synonym that gets our point across. Suggest “Responsibility”. Blueboar (talk) 17:12, 5 February 2023 (UTC)

  Or we could just add a note clarifying nthe meaning. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 19:38, 5 February 2023 (UTC)

  Typical WP… why use one word when an entire sentence would suffice? Blueboar (talk) 20:42, 5 February 2023 (UTC)

  Not convinced it is "so problematic"' It's supposedly derived from Latin, instead of Middle English (from which 'burden' comes), but many English words are from Latin, and it is usual to mix them in English writing. Alanscottwalker (talk) 23:07, 5 February 2023 (UTC)

  I'd be happy to replace the word onus with the word responsibility. WhatamIdoing (talk) 17:39, 9 February 2023 (UTC)

  I'm fine with that as well; synonyms are perfectly OK with me. Burden isn't good because it is being used for a different concept, but "responsibility" is fine so long as WP:ONUS remains as the redirect. --Jayron32 18:35, 9 February 2023 (UTC)

  WP:ONUS -> WP:RESPONSIBILTY, gotta be a plus, that:) Selfstudier (talk) 18:55, 9 February 2023 (UTC)

  WP:ONUS should remain, it's 3.25 times better than WP:RESPONSIBILTY (although that could be setup as well). -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 22:03, 9 February 2023 (UTC)

  Agreed on the shortcut, it's far too recognizable at this point.

That aside, I favour keeping the policy worded as "onus" (due to its recognizability) and just put a footnote saying "responsibility", or bringing back the definition footnote. DFlhb (talk) 22:21, 9 February 2023 (UTC)

• Ok, based on what most of us are saying, I’ve been somewhat bold and changed the word “onus” to “responsibility” in the text. I have left the shortcut at WP:ONUS for now, pending further discussion. Blueboar (talk) 19:27, 9 February 2023 (UTC)

Last year the Chinese Wikipedia uncovered an incident in which an editor, Zhemao, had created a web of hoax articles. It was particularly insidious because it exploited a social gap in the community and hurt public trust in the Wikipedia project. Mainly, Zhemao used foreign, offline sources and the community assumed on faith (i.e., blind faith) that the invented sources both existed and were accurate. I haven't seen us reckon with how we're vulnerable or liable for the same loophole.

I often work from non-English language and out-of-print sources myself. I am very aware that adding a rare citation makes it likely that few people will ever undertake the effort to verify the citation. I also often work with articles that attract rare/non-English source text from editors who are not (or were never) active, which leaves reams of text that is, for practical purposes, unverifiable and susceptible to this Zhemao vulnerability. (For context, when I say practical, I mean that I have the means to verify rare citations and if it's inaccessible to me, it's definitely inaccessible to our readership.)

The easiest solution would be to fortify the existing WP:NONENG recommendation into a standard expectation, treating these sources as implicitly challenged rather than just challengeable:

As with sources in English, if a dispute arises involving a citation to a non-English source, editors may request a quotation of When working with highly inaccessible (e.g., rare or non-English) sources, editors should proactively provide a quotation of the relevant portions of the original source be provided, either in text, in a footnote, or on the article talk page.

...

If you quote a non-English reliable source (whether in the main text or in a footnote), a translation into English should accompany the quote. ... The original text is usually included with the translated text in articles when translated by Wikipedians, and the translating editor is usually not cited.

This would close a gap in our social practice where inaccessible citations meet WP:AGF. If we have the appetite for such a change, I think it would help forestall the inevitable fallout of an editor pulling a Zhemao-style exploitation of community trust on the English Wikipedia. Given our size, it's a matter of time, if not already underway and undetected. czar 08:10, 7 January 2023 (UTC)

If the Zhemao vulnerability lies with an editor who is inventing sources, how would this proposed change help protect against that same editor from inventing a quote and translation? It would be more work for them but a committed fabulist could still meet the proposed standard without necessarily increasing the possibility of detection. — Carter (Tcr25) (talk) 13:02, 7 January 2023 (UTC)

It lowers a natural barrier to verification. It's much easier for a non-conversant individual to request a page scan of an inaccessible offline volume and verify that a string of non-Latin text matches the one quoted in WP citation than it would be to find a literate third party who can read, interpret, and verify the citation. That's why our guideline already says to provide quotations upon being challenged—it aids in verification. My suggestion is that we hold inaccessible sources to be inherently challenged. We can't change the fact that a bad actor can make it all up but we can make it easier to verify. czar 15:33, 7 January 2023 (UTC)

Can you clarify what The original text is usually included with the translated text in articles when translated by Wikipedians, and the translating editor is usually not cited. means? -- asilvering (talk) 03:45, 20 January 2023 (UTC)

When quoting non-English text directly, whether in a footnote or the article, the current guidance is to include the non-English original text next to the Wikipedian-translated English text. And the person doing the translation is not credited. I think that can be put more clearly but more a secondary point to the larger question at hand. czar 05:42, 20 January 2023 (UTC)

Oof. I don't think this is "usually" the case at all, though I wish it were. I think your change is sensible, with the caveat that editors shouldn't be copy-pasting entire pages of text since copyright restrictions may limit how much can be reproduced. If a sentence is cited to "pp. 104-118" or something we wouldn't want all 14 pages of it. But, frankly, I think it would already be a huge improvement if current guidelines like "machine translation is worse than no translation" and "verify the sources when you translate" were more commonly known/enforced. -- asilvering (talk) 05:54, 20 January 2023 (UTC)

I don't think this will help. Your theory of change seems to be that:

1. People will read the directions.
2. The occasional bad-faith "committed fabulist" (a lovely phrase) will be at least as likely to follow the directions than the main good-faith contributors.
3. If someone cites such a source, another editor will ask them to deliver a "scan of an inaccessible offline volume".
4. It will actually be possible for this to happen, which means that both editors will have e-mail enabled (I'd estimate a 75% chance of this being true), because you can't upload pages of copyrighted sources to the wiki under our fair-use rules, so offline delivery of the scanned page(s) is likely necessary, and also the contributing editor will still have access to the sources at the time the reviewing editor makes the request (probably depends on how soon the request appears).

...all so that one reviewer can check one page in one source.

IMO #1 is unlikely to happen often; #2 is probably backwards to a significant extent, and #3 is very unlikely to happen, and even if #3 happens, #4 has logistical challenges.

In practice, even this could be gamed. It'd be simpler, faster, and easier for your committed fabulist to create a sock account and post a fake request on the talk page, which they would pretend to fulfill, and then – because the problem here is that we trust editors – every editor after that would trust the sock account's public claim that it had received and checked the scanned page.

Then, having established that this is appropriate for "inaccessible" sources, we'll see more editors feeling entitled to demand free copies of anything that isn't already available online and free-as-in-beer, which will probably lead to worse sources. Why should I buy a book, pay for a paywalled article, go to the museum and read the sign on the wall, etc., if citing such a source will only lead to the belief that in doing so, I have WP:VOLUNTEERed to provide scans of all of these "inaccessible" sources to other editors? Once established as the principle, it will not be limited to non-English sources. You take the first step towards that in this proposal, by indicating that not only non-English, but also "rare" sources are included in this rule. WhatamIdoing (talk) 07:07, 20 January 2023 (UTC)

My theory of change is that editors more frequently challenge citations from inaccessible sources without quotations, and then more inaccessible sources have quotations. I.e., encourage enforcement rather than a recommendation. I'm more concerned with our general practices that protect hoaxes rather than magically rooting out hoaxes. This wouldn't involve scanned pages but text quoted in references, as our current guideline recommends. czar 08:49, 20 January 2023 (UTC)

I thought you were saying above that "it's much easier for a non-conversant individual to request a page scan of an inaccessible offline volume..." WhatamIdoing (talk) 21:24, 28 January 2023 (UTC)

Agree with WhatamIdoing on all the above. Here's my take:

1. This would be ineffective at reducing the problem.
2. A lot of the time, entire paragraphs are being paraphrased with a single sentence. This proposal would require copyvio for any long passages.
3. Past instances seem to have been caused by a handful of bad actors, rather than being a systemic issue that would require a policy change. It's extremely bad practice to change policies to address edge-cases; that always leads to a vicious cycle of ever-increasing bureaucracy. Non-systemic problems must be met with non-systemic solutions.
4. As for the general issue of Wikipedia reliability: Wikipedia is inaccurate, and it's caused by edior laziness, carelessness, or emotional bias. Policy can't fix that. Either increase the amount of eyeballs on each article by an order of magnitude (instead, we've been losing editors), or wait 5-10 years until AI gets good (not like the CNET AI), and then take humans completely out of the equation. The latter is inevitable anyway.

DFlhb (talk) 05:06, 26 January 2023 (UTC)

I'm not sure AI will inevitably take over here, but I agree that trying to write policy based upon the potential for a rare bad actor to act badly is likely to limit good contributions in a way that doesn't improve the project. — Carter (Tcr25) (talk) 14:33, 26 January 2023 (UTC)

No one is arguing that verification rules stop bad actors... and stopping bad actors isn't even the point. In my experience, WP:TSI mistakes are far more often accidental error than overt "laziness, carelessness, or emotional bias". The point is to reduce the potential for mistakes by providing ready proof. It's already a recommendation in our guidelines but has no enforcement. Lack of verification for non-English and rare refs is absolutely a systemic issue: Inaccessible refs clearly go unchecked. For any author (AI or human) to be trusted, they need accessible sources, not just citations, and the fastest way there is through partial quotation. czar 02:47, 27 January 2023 (UTC)

In re "Inaccessible refs clearly go unchecked": I wonder what percentage of accessible refs actually get checked. WhatamIdoing (talk) 21:46, 28 January 2023 (UTC)

I just know that many times when I try to look at an on-line source, I discover that the link has been dead for years. I suspect that on-line sources are most likely (but, how likely is that) to be checked very shortly after they are added, and very seldom after that. Donald Albury 23:43, 28 January 2023 (UTC)

Seconding czar here. I'm greatly surprised to see this problem described as "caused by a handful of bad actors, rather than being a systemic issue." -- asilvering (talk) 09:02, 29 January 2023 (UTC)

A related (perhaps more common) problem is that when quoting an obscure source that actually exists that it can be near impossible to review compliance with other the other aspects of verifiability (e.g whether or not the source truly supports what they wrote.). One "middle of the road" solution covering both would be that whoever is advocating retention of the source or the text which it purportedly supports needs to make reasonable efforts to respond to inquiries about it. Regarding existence of the source, where they know it is available etc. And regarding what is in the source. to provide details and page numbers etc....the proposal above would be a subset of this. If that is not done upon request (including that nobody that responds to the request) then such would be considered to failing verifiability and also a basis for removal of the source. BTW, structurally, "removal of the source" is a separate question from wp:verifiability and rightly so. North8000 (talk) 16:23, 20 January 2023 (UTC)

Retaining sources is a problem for editors who use libraries. We really can't require it.

Also, I find that in many cases, we don't actually need to double-check the source. I already know that smoking tobacco causes lung cancer; I don't need to check a cited source to make sure that's a verifiable statement. One wants to be able to check some source (not necessarily the specific cited source) for claims that seem unexpected, doubtful, contrary to prior knowledge, etc. WhatamIdoing (talk) 21:45, 28 January 2023 (UTC)

@WhatamIdoing: I agree. I was digressing a bit. In essence that WP:VER excludes text (based on it's sources or lack thereof), not sources. North8000 (talk) 15:10, 30 January 2023 (UTC)

Re "will probably lead to worse sources", it's worth remembering that more expensive sources are often less reliable: Prestigious Science Journals Struggle to Reach Even Average Reliability. Nemo 12:41, 29 January 2023 (UTC)

That article says that Publish or perish causes scientists to submit shoddy work to scientific journals. It bases this conclusion on comparing the number of retracted articles in the 0.5% of most prestigious scientific journals against the average, and concludes that scientific fraud happens in journals of all levels of prestige, not just the unpopular ones (though some very bad ones have an especially high rate of fraud). More relevantly for us, it also doesn't say that scientific journals are less reliable than social media posts, or that we're better off with random websites than with books from the library. WhatamIdoing (talk) 23:05, 29 January 2023 (UTC)

Retractions are not the main focus of the article. The findings of the article are about statistical findings in the section "The Other 99.95% of the Literature". And it's not a case of "it happens everywhere"; the higher the journal impact factor, the lower the quality of a journal. Nemo 06:35, 30 January 2023 (UTC)

The Conclusion section says things like this:

• Biomedicine – Image duplications: Higher ranking journals show a lower incidence of image duplications
• Neuroscience, psychology – Statistical power: Either no correlation of journal rank with statistical power or a negative correlation
• In vivo animal experimentation in disease models – Reporting of randomization and blinded assessment of outcome: Lower reporting of randomization in higher ranking journals and no correlation with reporting of blinded assessment of outcome
• Medicine – Criteria for evidence-based medicine: Two studies found that higher-ranking journals met more criteria, while two failed to detect such an effect
• Biomedicine – Reproducibility of experiments: Reproducibility is low, not even “top” journals stand out

These are the subject areas that happen to interest me, and I interpret this list like this:

• Good for them, but if they're comparing the 0.5% at the top against the whole 99.5%, which contains some pay-to-play predatory journals, then it's hardly surprising.
• "Top-ranking journals would rather publish scientifically interesting pilots than Phase III trials." (Phase III's raison d'être is statistical power, but it's often boring scientifically.)
• When you're talking to someone who can speak knowledgeably (and probably passionately) about whether a third-party telephone-based randomization service is better than a random number generator on the researcher's computer, or if they have developed views about whether you randomize per animal instead of per rack location, then this is doubtless a huge big deal. But for a lot of other purposes, describing this is just wasted space in a paper. Most people need to know that an attempt was made at randomization, but they don't actually care about the details. (If you get stuck talking to one of those people, then the correct answer for human trials is to hire a specialty service.)
• This seems to favor higher-ranking journals.
• This seems to say more about the replication crisis than about the journals.

What the paper doesn't say is: Why compare the 0.5% against the 99.5%? Do you really expect journal quality at 99.4% to be materially different from 99.6%? Why not break them up into quintiles? The research that compares the top quintile against the bottom quintile produces a more favorable impression. WhatamIdoing (talk) 01:14, 31 January 2023 (UTC)

I'm afraid this entire comment is the result of a misunderstanding. The article doesn't say anything about a "0.5% at the top", nor about a 0.5% of anything. Instead it points out that "retractions cover only about 0.05% of the literature", and therefore proceeds to look for data about the entirety of the literature (as opposed to past studies which focused on the narrow set of works which got extra scrutiny and got retracted). Nemo 23:54, 9 February 2023 (UTC)

You're right; I misunderstood that. I see no information at all about how many journals fall into their category of "highly selective tier of the most prestigious journals". WhatamIdoing (talk) 04:03, 15 February 2023 (UTC)

There’s a big difference between borrowing from a library, which is normally free (uk) and paying $6/month for 12 months for a paywall? Riskit 4 a biskit (talk) 08:41, 17 February 2023 (UTC)

There are many different ways to get details of a source, including Wikipedia Library and the Resource Exchange. It's also possible that Google will allow you to find a way to access the source. Either way such sources are accepted as per policy, see WP:PAYWALL. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 12:36, 17 February 2023 (UTC)

• Meh… sure, if you live near a free library it is cheaper to use that library… but consider places where you have to travel miles to get to a library? With the cost of transportation it might actually be cheaper and more convenient to pay for the Paywall.

Of course the cheapest option is to find someone who does have easy access to the source (either because they live near a free library or because they have payed the paywall price) to volunteer to check the source on your behalf. Nothing says you have to be one to check the source. Blueboar (talk) 16:02, 17 February 2023 (UTC)

Where I live, we all have a local library based on our place of residence, but it might cost money or be impossible to obtain a library card for another local library with a larger collection or a more convenient location. So even physical libraries may have a proxy paywall attached. Imzadi 1979 → 17:24, 25 February 2023 (UTC)

Yes, the library is free, but many of the resources in my local library that are useful for WP are reference materials, and cannot be checked out. So, I sit in the library taking notes, and then find after I return home that I have left out something (sometimes a page number). On-line, I can find many scholarly sources that are now free. I also use the Wikipedia Library, and have a JSTOR account. More and more, I am finding sources (such as journal articles that are still behind a paywall) and chapters or entire books that have also been posted to a free site by the author(s) of the sources. Donald Albury 16:06, 17 February 2023 (UTC)

I began editing the article on The Washington Post and found that correcting and improving the sources, alone, has turned out to be an overwhelming undertaking. Please forgive me if I have come to the wrong place and direct me to the proper place. My concern with The Washington Post article is that The Washington Post, itself, is used as a source no fewer than 68 times! I understand the difficulty and temptation to use the Post as its own source, but I think that when we are creating articles about mainstream media we must resist the temptation to use that media as the main source. Even if the facts are accurate, our quest at neutrality and avoiding conflicts of interest compels us to find other sources for the information, or in the absence of outside sources that the information be removed unless it is critical to the reader's understanding of the entity, in this case, The Washington Post. What is the Wikipedia policy on such a thing? I have searched and searched and cannot find the policy that applies here. I covet your opinions and advice on how to address the ridiculous 68 instances of self-sourcing. I also invite you all to jump into that article to reduce its length and improve it overall. All the best. MarydaleEd (talk) 02:34, 20 February 2023 (UTC)

The policy on this is WP:ABOUTSELF, in the article about the Washington Post it's acceptable to use the Washington Post as long as the details being referenced are uncontroversial. It's always better to have reliable secondary sources, but not strictly necessary. -- LCU ActivelyDisinterested ∆transmissions∆ °co-ords° 09:11, 20 February 2023 (UTC)

@MarydaleEd, an article should be WP:Based upon sources that are Independent of the subject. It's not always a problem for a third of the sources cited in an article to be associated with the subject, but in the case of a famous subject like The Washington Post, I'm sure you could do better than the bare minimum. I appreciate your efforts to improve the quality of sources in that article. WhatamIdoing (talk) 03:13, 24 February 2023 (UTC)

• ABOUTSELF does say that it's important that the article is not based primarily on such sources, which is at least a reason to try and add secondary sources when available. My experience is that even for uncontroversial stuff, when an article has a huge amount of things cited to primary sources for which no secondary source exists, it's often trivia or even promotional in nature. Glancing at the Post's article, the big problem that leaps out at me is the "political endorsements" section, which is cited almost solely to the Post as a primary source - I don't think entire sections should rely almost entirely on one source in general, especially not a primary / WP:ABOUTSELF one. (It's also worth pointing out, though, that the raw number of citations may not tell us anything, because people often cite a primary source + secondary sources discussing it, with the primary cite being more of a convenience link rather than what satisfies WP:V.) --Aquillion (talk) 09:02, 24 February 2023 (UTC)

  Also, people often cite a good source (e.g., a history book) many times, and a news article once, with the result that 10 uses of a book + 10 newspaper articles actually is "half" the article, but it looks like it's 91% news articles. WhatamIdoing (talk) 18:21, 24 February 2023 (UTC)

  Looking at that section, most of the refs are for the actual endorsement, which doesn't seem problematic (there's nothing inherently better about a third-party source saying the Post endorsed Obama vs. linking to the endorsement). That said, Editor & Publisher usually tracks newspaper endorsements (at least for presidential races), so something like that could be substituted. The one non-endorsement citation there, used multiple times, is an analysis of the Post's endorsement history by the paper's ombudsman. It could be argued that despite being published in the Post, the ombudsman is actually independent of the paper. It would be good if there were third-party sources included in that section (perhaps this study would be relevant). While that section could be improved, it doesn't strike me as particularly problematic in its current form, at least from a sourcing perspective. —Carter (Tcr25) (talk) 19:26, 24 February 2023 (UTC)

For stuff without a specific challenge (e.g regarding veracity, POV), I think that we should not even discourage such self sources. For example, secondary sources usually don't cover the "boring" encyclopedic information about an organization and so such is needed to have an informative encyclopedic article. Also, when the organization topic gets larger and more diverse / less centralized, the concerns behind "self" mean less. An extreme case can illustrate. 100% of the sources for the Human article are humans so it is 100% self sourced. North8000 (talk) 13:49, 1 March 2023 (UTC)