Wikipedia talk:Verifiability

Update WP:V to require verifiability through article sources

Currently, WP:V permits editors to add unsourced content. This is a remnant of the Wikipedia of decades ago, not the Wikipedia of today, where editors who habitually add unsourced content will typically be shown the door, and where unsourced content is seen as a cleanup problem that needs to be addressed.

It's time to update V to require that all content not merely be verifiable to sources, but to be verifiable to sources in the article. I propose two changes to do this. In the lede, change Each fact or claim in an article must be verifiable. to Each fact or claim in an article must be verifiable to sources in the article. In WP:BURDEN, change A fact or claim is "verifiable" if a reliable source that supports it could be cited, even if there is no citation for it in the article at the moment. to A fact or claim is "verifiable" if a reliable source that supports it is cited.

This will reflect existing practice and will make it clearer for new editors what our expectations are, rather than expecting them to figure out that we say one thing but mean another. BilledMammal (talk) 03:40, 8 May 2026 (UTC)

Remember what I was saying on the other page about proposals to strengthen sourcing requirements fail due to overreach? This is an example of that.

You are proposing to change the very definition of what it means for a fact to be verifiable. That's a huge change. Policies, like community practice, evolve by baby steps. Try something like "Every article needs to contain at least one source". Copy the wording currently used for WP:BLPPROD. Don't lard any extra stuff on to the side of it, and don't tell anyone what your future hopes for maximizing citation density are. Just keep it short and simple, and you'll at least have a chance of getting a change to this policy. WhatamIdoing (talk) 04:03, 8 May 2026 (UTC)

It will change very little, as in practice we already require that all statements are verified in the article. It also won't lead to us deleting every unverified statement; the guidance about how to deal with unverified statements will remain as it currently is, encouraging the addition of sources or the addition of cleanup templates.

All it will do it update policy to reflect practice, and address one of the main issues that our policies have for new editors - that they say one thing, but mean another. BilledMammal (talk) 04:08, 8 May 2026 (UTC)

as in practice we already require that all statements are verified in the article Given that there are regularly sentences added to Wikipedia every day that aren't verified by sources in the article, how is this codifying existing practice? Katzrockso (talk) 04:30, 8 May 2026 (UTC)

The claim that in practice we already require that all statements are verified in the article is not true. The lead of the Wikipedia:No original research policy even gives an example of a statement that it explicitly says is not required to have an inline citation. WhatamIdoing (talk) 17:25, 8 May 2026 (UTC)

For what it's worth, I think I'd lean toward supporting this change...but WAID isn't wrong that it's a pretty significant paradigm shift, and I suspect you're going to face an uphill battle getting a consensus in favor of it. Good luck! DonIago (talk) 04:11, 8 May 2026 (UTC)

Non starter, because then we replace SKYISBLUE problems with a stranglehold on new article creation or expansion. Not kidding, I have been informed that my Wikipedia-style incessant citation habits are overkill for academic writing. Whoops. Jclemens (talk) 04:37, 8 May 2026 (UTC)

That's likely as many in academia expect citations to be used to build support for original research, rather than be the main focus. It seems highly unlikely there would be a SKYISBLUE stranglehold on new article creation or expansion, as those processes already expect good sourcing. Trying to think of common SKYISBLUE exceptions, one might be extrapolating geography (eg. attaching a country name to a state name, even if the country is not mentioned), but that usually comes with a citation for the local area so it isn't unsourced. In some cases, there is the usual lack of enforcement around pronunciations and non-English names. Perhaps some extrapolation of nationalities? CMD (talk) 04:47, 8 May 2026 (UTC)

As it stands any challenged content must be referenced with, and not restored without, an inline citation. So any editor with concerns about unverifiable content already has a strong policy backing, the question is do they need more. Personally I'd like to see Wikipedia moving to inline citation being the norm, it's what the readers expect and helps boost credibility (whether readers check them or not). Standardisation (not just in this matter) would also make the learning experience of new editors a lot easier. However I don't see this proposal being accepted. I also worry that the suggested language would exasperate the situation where some editors think something is unverifiable because the source doesn't match the article content word for word. -- LCU ActivelyDisinterested «@» °∆t° 13:07, 8 May 2026 (UTC)

Which makes this feel the better way to handle this is through behavior rather than policy. An editor that continues to add unsourced content without making attempts to source it after being cautioned about WP:V is being disruptive. It would feel that trying to change WP:V from what it is now would create a mess of enforcement problems. Masem (t) 13:33, 8 May 2026 (UTC)

The context for this is a discussion at Wikipedia talk:Notability#Proving notability in which the OP is trying to find a policy basis for overturning WP:NEXIST and WP:ARTN. WP:V apparently seemed like a hopeful candidate, especially if you apply an Infinite monkey theorem to WP:V's rule about content that is Wikipedia:Likely to be challenged.

Maybe we should consider adding a statement like "For reference, editors estimate that less than 10% of content actually gets challenged, so mathematically, most content is unlikely to get challenged"? WhatamIdoing (talk) 17:32, 8 May 2026 (UTC)

Deleting content for which proper sources exist but have not yet been added to the article is putting worship of policy over improvement of the encyclopedia. The more appropriate reaction to being upset about sources not having been added is to fix it, not to demand that other editors do the work for you on penalty of deletion. Requesting sources can be a reasonable thing to do, but mainly in cases where you have a reasonable suspicion that the sources might not exist. When you know they do, you also know what to do to fix it. —David Eppstein (talk) 17:49, 8 May 2026 (UTC)

Well, I agree with you in principle, but I also believe that there are a few editors in the community who have done the math and determined that they can get articles deleted faster than they can add sources to them, and so they see deletion, and the threat of deletion, as both a righteous end itself (readers will be protected from "bad" articles) and also as a force multiplier (because other editors will drop all the other work to prevent deletion of articles about notable subjects, so instead of "just me" adding sources, it's now Alice and Bob and Chris and David adding sources [but not me, because that was so effective that I'm going to spend my time threatening another set of articles, to get Eve and Frank and George busy adding sources to some other articles]). WhatamIdoing (talk) 17:53, 8 May 2026 (UTC)

Most V issue tend to come down to behavioural issues. -- LCU ActivelyDisinterested «@» °∆t° 19:35, 8 May 2026 (UTC)

This seems like a sweeping change proposed for insufficient reasons. I am far from convinced that new editors are actually confused to a serious extent about this, or that common practice is, in practice, out of line with the current phrasing. What people appear to need more help understanding is that not everything you can find on the Internet is reliable. New editors add "sources" that end up being press releases, SEO slop, etc. Stepwise Continuous Dysfunction (talk) 18:17, 8 May 2026 (UTC)

And also the other way around: A source will be perfectly reliable in context, but someone will remove it because press releases are never reliable, etc.

I don't think the community is ready to have a rule that says articles can be deleted if they're "insufficiently sourced", by which some editors mean that there isn't a citation after every sentence, and citations to certain kinds of sources don't count. We might be ready to expand the WP:BLPPROD rules to all articles.

If newcomers are confused, it's probably because immediatist–citation-maximalist editors assert that "policy requires" full citation, and when they look at other articles, they discover that this alleged policy is not enforced, and if they ask here, they are told that the alleged rule doesn't exist. IMO the solution is for the pro-citation editors to stop saying "policy requires" (because that appeal to the law is not true) and instead say "it's best" or "it's important" (because it is). WhatamIdoing (talk) 20:23, 8 May 2026 (UTC)

New editors are confused because Wikipedia doesn't have a way of doing things, instead it has many different ways. This leads to a learning curve that is a cliff. You can have a personal opinion on what the best way would be, but they're not confused by one particular way or another rather than all those ways are the 'right' way. -- LCU ActivelyDisinterested «@» °∆t° 09:20, 9 May 2026 (UTC)

That's very true. I've been Wikipediaing for years, and I've been continually surprised by how many things I thought were rules are actually just... suggestions that are written to strongly imply they are rules, without actually technically claiming to be so. I often wonder how Wikipedia conventions got settled at all, given that everybody seems to have a different idea for the reasons that underlie them. Dingolover6969 (talk) 06:32, 5 June 2026 (UTC)

• I've just written a (blissfully infobox-free) article. If this proposal had been the rule, then I wouldn't have bothered, as the sheer labour involved in sourcing and defending it would have prevented me. If anyone wants more sources, then it's up to them to specify the sentence or paragraph which they say is undersourced. And that's as it should be.
  But, I have the autopatrolled flag so I get to bypass NPP. That means my experience of writing articles is fundamentally unlike most other people's. If I didn't have that flag, then I would expect a NPP patroller or AFC reviewer to interrogate my sources fairly closely before they approved that article for mainspace. (They might have some trouble, considering that I've put together an article about a concept, from sources that are about procedures.)
  Counterproposal:

  With effect from %futuredate, for articles within the scope of CTOPs, or for all BLPs, any edit that introduces a new fact or claim into an article must have an inline citation.

  How's that?—S Marshall T/C 09:52, 9 May 2026 (UTC)

  No. Unenforceable and overbroad. And likely to be misinterpreted as requiring the addition of a new citation rather than allowing reuse of existing ones. —David Eppstein (talk) 17:37, 9 May 2026 (UTC)

  I don't think that's likely (i.e., someone might make that claim, but they'll get shouted down pretty quickly). I think that more likely problems are:

  • It's not actually a "new" fact or claim, but the reviewer didn't notice that the information being summarized is already cited elsewhere on the page.
  • "all BLPs" will be taken by some editors to mean "only biographies, narrowly defined, so that it includes Cher but not Cultural impact of Cher" and others will take it to mean "any material covered by the BLP policy, including obviously non-BLP articles like Volcanic eruption and Death of a Salesman and White cake, if they ever happen to mention any living people".

  I also think that these problems could be handled with a footnote.

  On that first bullet point, it might be a good idea to get the "one cite per fact/claim per article is enough" into this article, though it's not true for especially contentious BLP matter, and we should talk about whether duplicate citations are desirable if a direct quotation is repeated (fairly uncommon, especially outside of WP:PLOTSUM situations, such as the lines in a poem or the lyrics for a song). WhatamIdoing (talk) 18:22, 9 May 2026 (UTC)

  Where there's dispute, an article is a BLP if it's in Category:Living people.—S Marshall T/C 08:45, 10 May 2026 (UTC)

  So:

  With effect from %futuredate, for articles within the scope of CTOPs, or for any article in Category:Living people, any edit that introduces a new fact or claim into an article must have an inline citation.

  WhatamIdoing (talk) 20:18, 10 May 2026 (UTC)

  An excellent idea Marshall, in my opinion. But then, as you know, I am in the "over-protective" camp regarding quality and verifiability. I am sure some people will object, but we must push to strengthen the requirements for adding junk at will and worrying about it later. That was the motto 20 years ago and at the time "text bulimia" was not only overlooked but encouraged. But now we have an encyclopedia that mimics Gérard Depardieu and things must change. Yesterday, all my dreams... (talk) 11:59, 10 May 2026 (UTC)

• There are two sides to this: 1) When you are adding stuff to an article - if there is any doubt about whether to include a citation, do so. 2) When you are reading or reviewing an article - if there is any doubt, don’t be a dick about it. Consider tagging rather than removing. And if YOU can add a source, even better! Blueboar (talk) 18:55, 9 May 2026 (UTC)

  I think we've made progress over the years on your point #1, by losing ground on your point #2. WhatamIdoing (talk) 19:27, 9 May 2026 (UTC)

• I am opposed to this change. Verifiability has always meant that something is, well, verifiable. This is deliberately a low threshold, and only depends on whether reliable sourcing exists. Other policies and guidelines then address when citations are required, what kinds of sources are reliable, how citations should be placed, and what material belongs in the article. This proposal would make WP:V do too much work, and would risk turning a basic inclusion threshold into a general article-cleanup or citation-density rule. Sławomir Biały (talk) 08:58, 10 May 2026 (UTC)

Billed, I agree with your sentiment, but know that you will face opposition. I suggest a softer approach that would have a chance of success. In my view it is far too easy to add junk to articles, and not enough editors to monitor things. Yesterday, all my dreams... (talk) 12:08, 10 May 2026 (UTC)

I think the idea is good. I think the proposed phrasing is confusing. I also think we shouldn't change what "verifiable" means (unless that gets changed around all the time on Wikipedia unbeknownst to me), and should instead invent a new term for the new requirements and refer everything to that. I think the new requirement is a plausible use of the English word "verifiable" but Wikipedia has its own entrenched technical terms (or are they?). Dingolover6969 (talk) 06:42, 5 June 2026 (UTC)

We avoid changing the meaning of words, thought semantic drift affects communication on Wikipedia in the same way that it affects any other communication between humans. Some words don't officially have a definition (e.g., "reliable source"), and others have multiple official definitions (e.g., WP:SOURCE). The Wikipedia:Glossary might be interesting to look through some time (or Wikipedia:WikiSpeak, if you're looking for a more cynical version). WhatamIdoing (talk) 17:28, 5 June 2026 (UTC)

By the way, there is a discussion on Village pump idea lab about adding master theses, which may be related to verifiability. I made a few comments, then stopped. You people may want to comment there. Thanks Yesterday, all my dreams... (talk) 12:46, 10 May 2026 (UTC)

Clarification

There seems to be a lot of misunderstanding about this proposal, which is my fault. To clarify:

• This proposed change will not require inline citations, only general citations. For example, my work at Chile at the 1964 Summer Olympics.
• This proposed change isn't related to notability, though it came out of a discussion at WT:N; most non-notable articles are fully verifiable.
• This proposed change will not change how we handle content without citation; WP:BURDENWAIT will be unchanged.

In my opinion, editors should only add content when they have verified it - and since that means they should already have the supporting source, they should also provide it to save future editors the work of finding the source. However, if editors agree with the general principle but think this change is too dramatic, then I'm happy to consider lesser changes, if anyone has suggestions? BilledMammal (talk) 03:41, 13 May 2026 (UTC)

There is a big jump in reasoning in the first sentence of your last paragraph between "have verified it" and "have the supporting source", that completely disregards other ways of verifying content than through sources (WP:BLUESKY and WP:CALC). —David Eppstein (talk) 23:38, 13 May 2026 (UTC)

WP:BLUESKY is an essay saying "we don't need to verify because it's obvious", rather than "this is verified". It is disputed, rarely applies, and WP:IAR would allow its use in the few cases where it does.

WP:CALC is an exception to WP:SYNTH, and sources are still required for the figures used in the calculation. BilledMammal (talk) 08:09, 14 May 2026 (UTC)

• WP:BLUESKY should be ignored. It is neither a guideline nor a policy, just an essay with no formal Wiki impact. CALC still needs sources as you said and has a very narrow focus. So, please keep going. :::Yesterday, all my dreams... (talk) 10:51, 14 May 2026 (UTC)

  BLUESKY works if you interpret it as “this is obviously verifiable” rather than “this is obvious”. The Statement “Albany, New York is the capital of New York State” isn’t obvious to everyone, but it is obviously verifiable (just look at an atlas)…The statement is SO EASY to verify that we don’t have to provide a citation for the reader in our article. Blueboar (talk) 11:58, 5 June 2026 (UTC)

This proposal doesn't seem controversial to me. In most cases, if there's a content dispute about whether something is verifiable for inclusion, the reliable sources determine the outcome. It becomes very frustrating when someone keeps pressing an unverified claim (and I have been dragged through some very tedious discussions) only to end with no sources found and the claim rejected. The wording of this proposal may need to be tweaked so it doesn't overreach (e.g.: we don't need to cite something in the lead if there's ample sources in the body). But on the whole, this proposal is correct both as best practice and normal practice. Shooterwalker (talk) 18:53, 2 June 2026 (UTC)

@Shooterwalker, the FAQ at the top of this page has a question "Are sources required in all articles?" with links to some of the prior failed attempts to have WP:V include a requirement that each article should have any sources at all. Given that we can't even get a "Yup, even the most boring obvious sub-stub has to cite one somewhere on the page" adopted, what do you think the odds are that the community will adopt a rule that every single addition to an article has to be accompanied by a source? I put that chance somewhere around the proverbial chance of a snowball surviving in hell, and I therefore rate this proposal as controversial.

Yes, if there's a content dispute about whether something is verifiable, then WP:BURDEN applies. But this would apply to absolutely non-controversial content, too. WhatamIdoing (talk) 19:02, 2 June 2026 (UTC)

Strictly speaking, those two discussions were about creating a PROD category for unsourced articles, which would be a stronger action than merely having a guideline. Articles created via the Draft: route are routinely declined if they have no sources, and unsourced articles taken to AfD are routinely deleted if nobody comes up with sources (cf. WP:DILIGENCE). So there is some truth to the claim that our present culture wants sources, despite the large number of existing exceptions. Zero^talk 02:16, 3 June 2026 (UTC)

I agree that our practice, especially as enforced upon newcomers, is different from our written rules. But this is still a proposal to change the written rules, going far beyond more modest proposals that have been rejected (both the two linked in the FAQ and also others). WhatamIdoing (talk) 05:14, 3 June 2026 (UTC)

Catch 22 re ONUS vs NOCON

Wikilawyers rejoice! We have had multiple discussions and even some RFCs… and it is obvious to me that the community can not reach a consensus on ONUS vs NOCON.

That raises an interesting situation for wikilawyers - if we apply ONUS to ONUS itself, we can say that those who want to keep it have failed to achieve consensus, and thus we should remove it. However, if we apply NOCON to ONUS, we can say that the default of no consensus is to keep the long standing language and not remove it.

Catch 22… the wording of our preferred P&G instruct us to do the opposite of what we prefer! Just noting the absurdity of it all. Carry on. Blueboar (talk) 16:24, 12 May 2026 (UTC)

Please update Wikipedia:Verifiability/Onus as you see more discussions. WhatamIdoing (talk) 18:27, 12 May 2026 (UTC)

The community outside a few regular editors here seem to have no problems with ONUS. -- LCU ActivelyDisinterested «@» °∆t° 18:41, 12 May 2026 (UTC)

No one knows what ONUS means. - Butwhatdoiknow (talk) 19:59, 12 May 2026 (UTC)

Activeletereste ~2026-34236-21 (talk) 14:33, 10 June 2026 (UTC)

It might be that 'the community having no problem with WP:Onus' is connected to their not being aware of it or not thinking through its implications. IMO, there should be a clearly delineated list of reasons for which it is acceptable to remove sourced information. Otherwise, WP:Onus makes it look as if there is a carte blanche for anyone to remove anything at whim - as long as there isn't a clear majority against the removal (which may be just because [almost] nobody is watching the page), you can just repeat 'there is no consensus for keeping the information'. There is no shortage of people willing to censor information they don't like.--Anonymous44 (talk) 07:19, 13 May 2026 (UTC)

As I have attempted in the past, one solution is to just add a note to ONUS indicating that it is disputed. We just have to come up with appropriate wording. That will, hopefully, both discourage people from citing it without fully understanding the implications, and encourage people who think they have a clear-cut understanding of it to weigh in here so we can eventually figure out what to do with it and / or how to refine it. --Aquillion (talk) 22:13, 13 May 2026 (UTC)

We could tag it with {{under discussion inline}}, and skip discussions about "appropriate wording". WhatamIdoing (talk) 00:58, 15 May 2026 (UTC)

The "list of reasons" is already sketched out in note C. Alanscottwalker (talk) 23:37, 13 May 2026 (UTC)

S Marshall's clear-cut understanding of ONUS

• ONUS gives us a process for dealing with a fact or claim that's verifiable, but in the wrong article.
• It's the rule that enables editors to move any fact or claim, including a fact or claim that's meticulously cited, to the article that's most appropriate for it.
• Because it involves moving material that's properly cited, WP:V is one of the most natural places where editors would tend to look for the rule.
• It has interactions with WP:PRESERVE. Looking at the two policies together, they mean that if the fact or claim belongs anywhere in the encyclopaedia, it must be moved rather than deleted.
• It has interactions with WP:UNDUE. Looking at the two policies together, they mean that if there's dispute about where to put the fact or claim, process means we tend to put it in the less prominent place. This can be good or bad.
• It's important for dealing with marketers and people with an agenda, who often won't want you to move their well-cited fact or claim to a place that gives it less prominence. I find it useful and well-thought-out in that situation.
• It's useful for those who want to whitewash an article. In that situation it's easily misused.
• There might be a perfect wording for WP:ONUS, but nobody knows what it is.
• In my experience it can and sometimes does directly conflict with WP:NOCON, specifically when a fact or claim that's been in the article for a long time gets disputed. This situation isn't all that common, though. This from five years ago is the most recent one in my RfC close log.
• When the dispute is between good faith editors who're displaying good judgment, I have personally tend to prefer WP:ONUS over WP:NOCON because ONUS is in a core content policy.

Hope this helps—S Marshall T/C 16:25, 14 May 2026 (UTC)

This is a way to cut the Gordian knot, though: ONUS isn't a blank check to remove whatever you want. Sourced info that is overweight or misplaced but otherwise solid should be relocated to a subtopic or related article. Andre🚐 04:48, 16 May 2026 (UTC)

If an editor is going around removing content and giving no good reason than ONUS, they should be taken to ANI. Nothing in ONUS states that you can remove content. If editors do give a good policy based reason to remove content, why should it be restored elsewhere without considering that challenge. -- LCU ActivelyDisinterested «@» °∆t° 10:17, 16 May 2026 (UTC)

That argument doesn't hold water on Wikipedia with the present state of the editing culture and the online dispute resolution mechanism state of the art such as it is. For example, I have seen admins and even an arb or other most trusted users "cite ONUS" to remove something despite not another substantive argument. Diffs can be furnished on request. The counterargument may be as Blueboar says (a substantive argument is any coherent argument, not an argument that has evidence and logic (either statutory so to speak, or precedence)) Or consider recent arbcom cases where the fundamental problem was an editor stonewalling while a cadre of likeminded fellows saw no problem at ANI for whatever reason. The remedy "ONUS-stonewalling just go to ANI/AE/RFAr" to me, is a forest-trees problem. The community doesn't agree whether ONUS is a blank check. Many people are arguing that it is not, which I agree with at this point, (re S Marshall below), but saying "ONUS abuse is sanctionable" today, isn't defensible on the existing data. Andre🚐 21:48, 17 May 2026 (UTC)

I'm sympathetic to your concern regarding "just go to ANI". I wish there was something like a referee to help with some of these disputes. Someone who wouldn't weigh in on content and wouldn't hand out penalties but would help when editors feel people aren't following the proper process for solving issues. Editor A adds something new. Editor B reverts it with a reasonable justification (UNUDE with at least a bit of reasoning). If Editor A restores then the referee could say, no, revert and go to the talk page. This might stop some of the edit warring where Editor A thinks B is wrong so they don't follow BRD/NOCON while Editor B gets frustrated that people aren't following the process. A RfC, if held, might show Editor B is wrong. Even a decently attended talk page discussion might do that. However, Editor A should not bypass the process and a ref could point that out. Springee (talk) 23:12, 17 May 2026 (UTC)

@AndreJustAndre, I wonder, if you check a few of those diffs, if any of those disputes had no explanation at all in prior edit summaries or discussions. This might be reasonable:

• Alice adds content
• Bob reverts with an edit summary of "WP:UNDUE"
• Alice disagrees and re-reverts.
• Bob reverts again, with an edit summary of "Rv per WP:ONUS"

I think going to ANI with a dispute about overreliance on ONUS is likely to be about as effective as going to ANI to complain that someone isn't using any edit summaries, or that their edit summaries are always unhelpfully vague ("Edit"). Sometimes, if several people at ANI say "Not going to block you over this, but you really should do better", then the editor will make an effort to improve. WhatamIdoing (talk) 16:57, 18 May 2026 (UTC)

• I agree with most of this… but a few thoughts: 1) sometimes there isn’t a “more appropriate article” to move the disputed material to. In this situation ONUS would indicate complete removal. 2) ONUS should be seen as process, not justification for removal. When there is a dispute, those wishing to move/remove the content do need to explain why they think the material should be moved/removed. Then ONUS kicks in and those wishing to keep it as is need to gain consensus to not move/remove. Blueboar (talk) 12:44, 16 May 2026 (UTC)

  This, pretty much. I remove stuff all the time but when I do I'm saying under what grounds in the edit summary and ONUS isn't a grounds for removal. Rambling Rambler (talk) 13:56, 16 May 2026 (UTC)

• Material in the article for a substantial time has "implied consensus". It doesn't lose that if someone removes it without giving a reason for its removal. It only becomes "disputed" if someone makes a case against it. The situation is different for new content. If something is added and immediately removed with "we need to discuss that first", it is enough to trigger ONUS. Zero^talk 13:52, 16 May 2026 (UTC)

  “Implied consensus" is extremely weak. It assumes that multiple editors have seen the “long standing” text and have no issues with it. Sometimes that assumption is correct (especially in high-traffic articles)… However, at other times the assumption is incorrect.

  In a low-traffic article, an editor might add something that sits quietly for a year (or more) before anyone notices it. Indeed the addition might be challenged by the next editor to visit the page - a year later! In such situations, the challenged addition would not have “implied consensus” simply because it was in the article for a long time. Blueboar (talk) 15:17, 16 May 2026 (UTC)

  There's also just the fact that a lot of times, editors simply can't be arsed to stir up a hornet's nest by challenging stuff. Rambling Rambler (talk) 15:27, 16 May 2026 (UTC)

  Yes, I think that it's extremely important to underline that implicit consensus is ultimately based on the number of people who at the very least saw something and accepted it (and ideally actually edited it and accepted it.) It doesn't apply to things that people have been objecting to for a while even if it has remained in the article - especially not stuff that has been constantly removed and then reverted back in, but also stuff that just has eg. a comment on the talk page saying "wait that's wrong, right?" Another thing I'd add is that all forms of consensus exist in a spectrum and implicit consensus has to be considered relative to the current state of discussion. If something is longstanding in a highly visible part of a reasonably high-traffic article, and discussion leans towards retaining it, that's a strong argument that it should remain in place. If the "discussion" consists of just two or four people who are evenly split, I would say that that's when implicit consensus becomes important, with the people arguing for removal needing to demonstrate something to overcome that; the thing to do is to seek wider opinions, but the most disruptive interpretation of ONUS, IMHO, is the one where people say "all right now I'm gonna take this out and you need an RFC to put it back" even when only a tiny number of people have actively weighed in. But if discussions lean even slightly against inclusion that's another story, and you'd need truly historically strong "implicit" consensus (to the point where it's basically explicit, eg. you can point to past discussions with lots of editors that obviously assumed it should stay) to argue that it should stay. --Aquillion (talk) 01:10, 17 May 2026 (UTC)

  Is this recommendation, that when just two or four people who are evenly split, I would say that that's when implicit consensus becomes important, with the people arguing for removal needing to demonstrate something to overcome that, supposed to apply only to removals?

  Many disputes aren't really about removal, or removal wouldn't solve the problem. For example, I've been told that the RFC years ago about the first sentence at Trans woman ended with no consensus, but we keep the no-consensus sentence in place because there must be a first sentence in an article, and there's no consensus for any sentence. So:

  • no consensus from the RFC years ago, so there's no real consensus;
  • it's been heavily contested for years, so there's no implicit consensus;
  • but: as a practical matter we can't actually have an article with no first sentence (even if you blanked the current first sentence, whatever's left would instantly become the new first sentence, and if you don't do a bit of fiddling with the former second sentence, it probably won't comply with MOS:FIRST), so we left it in.

  What should we do? Nothing, I think. Maybe have another RFC if someone comes up with a brilliant solution someday. But I wouldn't characterize a dispute like that as arguing for removal; I'd characterize it as arguing for a change. WhatamIdoing (talk) 05:53, 18 May 2026 (UTC)

  In addition to low traffic articles, there are times when material gets in to a high traffic/contentious article as part of a large series of changes. An editor makes a change to 3 parts of the article, two of those result in immediate reverts/revisions/talk discussions. The third simply may have been lost in the shuffle of all the other changes. Presumably, if something is part of a high traffic article we should be able to point to something showing consensus for inclusion. Conversely, if the content is obviously due reaching a consensus on a high traffic article would should be relatively easy. Springee (talk) 04:14, 17 May 2026 (UTC)

  I think that removal itself (under any circumstances that make us believe it's not vandalism or an error) is an actual objection to having the content on the page. Something as benign as copyediting is a challenge to the implicit consensus. What I think throws some people off is that you can lose an implicit/silent consensus and still have a real consensus. For example, a little while ago, an editor blanked an entire hot-button article during a discussion. We have a real consensus to keep that article, but the meaning was clear enough: that editor did not agree that the article should exist at all. ONUS could have been applied, and the "achieve consensus" step would have been achieved merely by reverting with an edit summary inviting them to look at the nine(!) prior trips to AFD. WhatamIdoing (talk) 22:37, 16 May 2026 (UTC)

  Zero says "Material in the article for a substantial time has 'implied consensus'." The same goes for material in the article for a nanosecond. In both cases, the implied consensus ends when an editor objects for a substantive reason (and "discuss first" is not substantive). - Butwhatdoiknow (talk) 00:53, 17 May 2026 (UTC)

Sure, but also remember that substantive does not necessarily mean policy based. Comments such as “I think this is too trivial to mention” or “The paragraph flows better without this” are substantive reasons for removal, even though they don’t point to any specific WP policy that is being violated. Blueboar (talk) 01:27, 17 May 2026 (UTC)

I 100% agree. Your examples tell the reverted editor what the reverting editor thinks needs to be discussed first if the reverted editor wants to press the matter. - Butwhatdoiknow (talk) 03:17, 17 May 2026 (UTC)

I also agree, and I think we need more official {{policy}} pages to contain phrases like "for reasons based in common sense, Wikipedia's rules, reliable sources, etc." Editors seem to forget that common sense and good judgment are valid reasons for taking (or not taking) an action. WhatamIdoing (talk) 05:38, 18 May 2026 (UTC)

Under editing consensus (the usual and most widespread form of consensus), what is in the article is evidence of consensus. The issue is 'how do you overcome that evidence of consensus' and in reality it varies in the circumstance, but the end result is the same, as long as it remains in the article, it is evidence of consensus.

Consensus of what? The consensus is, consensus that it substantially complies with policy/guideline and improves the article as an encyclopedia article. So, think about what it means when there are serious doubts (actual non consensus) about varifiability, pov, blp, cvio, etc, or even serious doubts over the ability of the general reader to understand and comprehend what is written in our sometimes prolix articles or tangential passages, that may be lost in the trees, missing the forest. Alanscottwalker (talk) 11:15, 17 May 2026 (UTC)

Wouldn't that same logic mean that content removed from an articles doesn't have consensus. As the evidence for it's consensus was that it was included. Also saying that content in an article automatically shows that the content complies with policy and improves the articles is more than a stretch. There is substantial amounts of long term content that isn't policy compliant, that noone has corrected it doesn't mean it's actually compliant. Many editors may have looked at it and decided they had better things to do then try to sort out the mess. -- LCU ActivelyDisinterested «@» °∆t° 12:14, 17 May 2026 (UTC)

No or perhaps, yes: a removal done appropriately is raising an argument that the passage does not have consensus (for consensus, see my consensus on what section). That it is in the article is evidence of consensus, that evidence can be and is regularly overcome, usually in editing consensus, or in more formal discussion. Alanscottwalker (talk) 12:19, 17 May 2026 (UTC)

Another problem with “implied consensus” is that “consensus can change”. If we remember this, then the question to ask is whether consensus has changed or not?

If adding something, and having it sit uncontested for a while, “implies” consensus… then removing it “implies” that consensus has changed (or at least, might have changed).

So… when text with an “implied consensus” is removed, we can not simply default to “No… this is consensus”. We need need to dig deeper… and either reaffirm the previous consensus, or reach a new consensus. Blueboar (talk) 13:21, 17 May 2026 (UTC)

Except implied consensus might also be actual and present. And reaffirming may be as simple as editing. -- Alanscottwalker (talk) 13:36, 17 May 2026 (UTC)

I think that what is in the article is sometimes evidence of consensus. Other times, it's evidence of one 'side' being more successful at edit warring than the other. Or an enthusiast filling an article with hundreds of details, and nobody has enough energy to oppose them. Some of these 'explanations' take months, and at the end, the person only hears "I've been told I can't do this" and never grasps "because a comprehensive catalog of every single Pokemon character ever seen, complete with detailed descriptions, is not encyclopedic content". WhatamIdoing (talk) 16:38, 18 May 2026 (UTC)

All that said, I wish we had a way of encouraging real blue pencil editors to comb through our articles as ruthlessly as needed. I have seen a few and its generally awsome. Alanscottwalker (talk) 15:14, 17 May 2026 (UTC)

Antique value? In my view, given the level of multidimensional suggestions above, this discussion will achieve no result. But I feel I need to make a couple of comments for clarity.

• "Time based implied consensus" is meaningless given the large number of articles and not so many editors per article. Once again, let me mention Jar'Edo Wens. Is 9 years enough for implied consensus? No, no, no.

• Justifying well sourced text is already handled by page watchers in most cases. Deletion of unsourced text should require no onus except a talk page mention. Look at the disaster called Unix security that I happened to play with a few days ago. Did the totally outdated material there have implied consensus? No way.

When we come across outdated, banal and unsourced items, we just have to get rid of them, regardless of their value as antiques. Yesterday, all my dreams... (talk) 16:39, 17 May 2026 (UTC)

Well, a time based consensus argument will only arise if someone makes that argument, and it will only likely be successful, if it is something like, 'this has been in the article a long time because, [good reason, good reason, good reason]'. Alanscottwalker (talk) 18:24, 17 May 2026 (UTC)

If they have several valid reasons for keeping the item (eg reliable sources) then antique value matters not, given that the essay WP:STATUS QUO is not even a guideline and has no impact. I do not antiques, specially in technical articles. Yesterday, all my dreams... (talk) 23:52, 17 May 2026 (UTC)

I don't think it's true that they'll only be successful if they have good reasons. QUO frequently gets invoked precisely because there aren't good reasons, beyond ILIKEIT. Invoking QUO as a justification for keeping an older version of an article in place is a bit like saying that your statement is protected by free speech: If you had a better argument than "Well, it's not actually illegal to say it", then you'd have used that better argument. Here, if you had a better argument than "Well, it's been like that for a long time", then you'd be using it. WhatamIdoing (talk) 16:42, 18 May 2026 (UTC)

Iteration

• ONUS gives us a process for dealing with a fact or claim that's verifiable, but in the wrong article.
• Only use ONUS when you give a reason for your change. ONUS is not a reason---it's a process.
• A key purpose of ONUS is to enable editors to move any fact or claim, including a fact or claim that's meticulously cited, to the article that's most appropriate for it.
• If there's no appropriate article, the options are either to start one, or else delete the disputed fact or claim.
• Because ONUS involves moving material that's properly cited, WP:V is a natural place where editors would tend to look for the rule.
• It interacts with WP:PRESERVE to mean that if the fact or claim belongs anywhere in the encyclopaedia, it must be moved rather than deleted.
• It interacts with WP:UNDUE to mean that if there's dispute about where to put the fact or claim, we tend to put it in the less prominent place. This can be good or bad.
• It's important for dealing with marketers and people with an agenda, who often won't want you to move their well-cited fact or claim to a place that gives it less prominence. It's helpful and well-thought-out in that situation.
• It's also useful for those who want to whitewash an article. This can lead to misuse.
• Editors should be mindful that longstanding versions of the article don't necessarily have consensus, particularly for low-traffic articles. The best way to tell whether a fact or claim has consensus is where it's been specifically discussed on the talk page. It probably also has consensus if it's been edited by a number of different editors over a reasonable period of time without challenge.
• (On the gripping hand:) We seem to be implying that where there's an apparent conflict between WP:ONUS and WP:NOCON, WP:NOCON prevails where there's some evidence that the fact or claim has consensus, and otherwise WP:ONUS prevails?

Comments of any kind are welcome but I'm particularly interested in your thoughts on those last two points, please? I've extrapolated them from what I think we're saying.—S Marshall T/C 21:31, 17 May 2026 (UTC)

I'm broadly in favor of this. Andre🚐 21:42, 17 May 2026 (UTC)

I think this is a good summary of where we currently stand… I don’t agree with all of it (for example, I still maintain that WP:V is the wrong place to say all of this, although I do understand your rational for why it is here)… however, most of my concerns are more quibbles than objections. It is definitely overly wordy, so we will need to brainstorm better wording if we put this in the policy. Alternatively, perhaps this could be made into an explanatory essay or guideline page to supplement the current terse language. The one thing I do know is that I need to take a break. Blueboar (talk) 21:46, 17 May 2026 (UTC)

• Yes: I hope that we're headed for an explanatory essay about WP:ONUS which would be linked from (I suggest) ONUS, NOCON, UNDUE and PRESERVE.—S Marshall T/C 21:56, 17 May 2026 (UTC)

I don't think I agree with the first point (and by extentension some others), onus should generally also apply to disputes about verifiability, including whether the source supports the claim and whether the source is sufficient (reliable) for the claim. Alanscottwalker (talk) 22:19, 17 May 2026 (UTC)

Isn’t that handled by WP:BURDEN? Blueboar (talk) 22:25, 17 May 2026 (UTC)

A dispute about whether there is consensus that the burden is satisfied -- no. Alanscottwalker (talk) 22:27, 17 May 2026 (UTC)

It doesn't say ONUS doesn't apply to those situations. It begins with "ONUS gives us a process..." If it began with, "One of the things ONUS does is to give us a process...", would that be OK by you? I propose to table the question of whether ONUS does anything else for the moment.—S Marshall T/C 22:45, 17 May 2026 (UTC)

My friend, iteration often means an endless loop. Yesterday, all my dreams... (talk) 23:53, 17 May 2026 (UTC)

In my experience it's usually an endless loop, on Wikipedia policy pages! But the occasional exception makes iterating worthwhile.—S Marshall T/C 08:11, 18 May 2026 (UTC)

Alan, we simplified BURDEN years ago: If there's a little blue clicky number at the end, and the person who put it there claims to genuinely believe that FakeWebsite.com is a reliable source, then we're done with BURDEN. This has reduced our Wikipedia:Bring me a rock problems, and we no longer have as many disputes about whether there is consensus that BURDEN is fulfilled.

Instead, we have disputes about whether the cited source is reliable and/or verifies the claim made. One thing that's not so clear in this policy is what happens in {{failed verification}} or {{unreliable source inline}} situations. I think this is partly because we don't need it (editors mostly do the right thing, so telling them what to do is WP:CREEPY), but also because we don't want to hand weapons to the "if it's not copyvio, then it's OR" minority. WhatamIdoing (talk) 16:49, 18 May 2026 (UTC)

I know that's been written but surely it can't actually mean that a burden is satisfied by a subjective belief of one person, when what we need is agreement that the objective factors (direct support, and qualified source) have been satisfied. Alanscottwalker (talk) 13:43, 23 May 2026 (UTC)

The BURDEN is satisfied by the subjective belief of one person. WP:V in general, however, is not satisfied. The process we want don't want looks like this:

• A: Adds uncited (and possibly bad) content
• B: Issues a WP:CHALLENGE
• A: Adds a citation
• B: "That's a lousy source! Go fetch another source!"
• A: Adds another citation
• B: "That's a lousy source! Go fetch another source!"
• A: Adds another citation
• B: "That's a lousy source! Go fetch another source!"
• A: Adds another citation
• B: "That's a lousy source! Go fetch another source!"
• A: Gives up

The process we do want could look like this (i.e.,this is one of several different desirable things):

• A: Adds uncited (and possibly bad) content
• B: Issues a WP:CHALLENGE
• A: Adds a citation
• B: "That's a lousy source! I spent a few minutes in my favorite search engine, and I don't see any reliable sources that directly support this claim. I'm removing this from the article."

WhatamIdoing (talk) 18:51, 23 May 2026 (UTC)

Well, its called the, "burden to demonstrate verifiability". And, it remains, the burden is only, "satisfied by providing one inline citation to a reliable source that directly supports the contribution." So, burden is not satisfied by 'providing one inline citation to a [UN]reliable source[, or] that [DOES NOT] directly support[] the contribution.'

And note [c], which has the language about belief, makes clear that providing the individually believed in source is not the end. The satisfaction of the burden needs be confirmed in agreement with others, that 1) the source is contextually reliable, and 2) it directly supports the contribution. (In addition to other issues). Alanscottwalker (talk) 16:07, 5 June 2026 (UTC)

Yes, eventually we need to come up with an actually reliable source. But the burden on the OP is limited: They have to supply one source that they believe (genuinely but possibly incorrectly) is reliable for this statement. Usually, their source is reliable, so that's the end of that dispute.

Sometimes it's not reliable. For example, I remember someone copy/pasting AIDS denialist garbage from virusmyth.net; he genuinely believed it was correct. When the OP has supplied one (1) source that they genuinely but incorrectly believe is reliable, then the BURDEN is fulfilled. At that point, we move on to ordinary WP:V procedures, the relevant points of which I'd summarize as:

• Everything that needs an inline citation should get one before the WP:DEADLINE.
• Once we're done with BURDEN, supplying a citation to an actually reliable source is anyone's/everyone's problem.
• Unreliable sources can be removed.
• Uncited material (including material that is uncited because you removed a citation to a source that was deemed unreliable during the BURDEN phase) can be tagged or removed per WP:CHALLENGE.
• If the material is removed as a CHALLENGE, it can't be restored until an actually-reliable source is provided.

WhatamIdoing (talk) 17:54, 5 June 2026 (UTC)

The burden was not fulfilled under the plain language of the policy. It is expressly not "satisfied" by an "[un]reilible source" and it can't be satisfied by a source that does not "directly support". Alanscottwalker (talk) 20:31, 5 June 2026 (UTC)

We do not intend BURDEN to be an endless game of WP:FETCH. You only have to present one (1) source. We expect you to do so as a genuine effort to prove that the material can be supported by a reliable source, but if your genuine effort is inadequate, then we are not going to make you present another source, and another, and another. The BURDEN is fulfilled when you've provided one (1) source.

After your one (1) source is rejected, as explained in the footnote, the material can be WP:CHALLENGED (including removal). Restoring CHALLENGED material requires at least an implicit consensus that the provided source is reliable. But BURDEN is not CHALLENGE; BURDEN only requires one (1) fair attempt.

How do you think we could re-word BURDEN to make this clearer, with an emphasis on it being clear to people who will dislike this limitation on BURDEN? We have many editors who would rather order other editors to WP:FETCH sources than do any work themselves. The one-source limit in BURDEN was added specifically to stop that game and force deeper discussions and compromises. WhatamIdoing (talk) 21:23, 5 June 2026 (UTC)

What we intend BURDEN to be is to clearly state what is needed to satisfy it, and what is needed to satisfy it is "providing one inline citation to a reliable source that directly supports the contribution" If you don't do that, you have not satisfied the burden - plain as day. The limit is not to "one source", and never has been, its to one source that has the required qualities. Alanscottwalker (talk) 21:53, 5 June 2026 (UTC)

In the footnote, what do you suppose we meant when we wrote words like "any source they believe, in good faith, to be sufficient" and "all editors are then expected to help"? Does that sound like it authorizes an endless game of "fetch me a source – no, that's not a reliable source, so the burden is on you to fetch me another source – no, that's not a reliable source, so the burden is on you to fetch me a third source"?

If you think these ideas conflict, I would be happy to hear your suggestions for how we can make the goal clearer. WhatamIdoing (talk) 03:49, 6 June 2026 (UTC)

I have talked about the footnote, what it does not mean is everyone has to accept and not correct anyone else's erroneous belief. Not for our readers, sake. Correcting is help. And this is what flows from the bedrock principle at the top of the policy that 'Wikipedia is not based on your beliefs'. More directly, helping make the encyclopedia article better involves several different tasks, including finding agreement/explanation of what is/is not a reliable source for this contribution, and what is/is not direct support for this, and what is/is not an improving particular editing out or addition to a particular encyclopedic subject. There is not contradiction, and in no way can we say that, "the burden to demonstrate Varifiability" is not what it is, "inline citation to a reliable source that directly supports the contribution." Alanscottwalker (talk) 11:00, 6 June 2026 (UTC)

The problem being that your definitions of "reliable source" and "directly supports" might differ from mine.—S Marshall T/C 13:10, 6 June 2026 (UTC)

If that's a "problem", its either inherent in us both writing here or it is a matter of consensus. Presumably that's why we spend many pixels on outlining what is and is not, and in different situations, recognizing nonetheless, that the editing context always matters. Alanscottwalker (talk) 13:24, 6 June 2026 (UTC)

'Wikipedia is not based on your beliefs' – but perhaps you think BURDEN should be based on one editor's belief about whether a given source is reliable?

Here's what I'd like BURDEN to accomplish:

• If you wrote/want to retain something, it's your job to supply the first source.
• If your first source is accepted, then we're done with BURDEN.
• If your first source is rejected as being unreliable for that statement, then:
  1. The unreliable source can be (but is not required to be) removed from the article, which leaves the material uncited.
  2. If the unreliable source has been removed, then now-uncited material can be (but is not required to be) removed per WP:CHALLENGE; if it is removed, it's kept out of the article until an agreed-to-be-reliable source is provided by someone (←doesn't have to be you).
  3. If the material remains in the article, then everyone (←not just you) is responsible for finding a reliable source.

Does this sound like an appropriate goal to you? WhatamIdoing (talk) 18:58, 6 June 2026 (UTC)

Your first question makes no sense, since I have been talking about editors examining things together.

Your second set of points are formulaic, there are several different issues that arise in different ways on whether efforts should be expended.

At any rate, I don't think we are getting anywhere, so I'll reiterate and leave it there: The satisfaction of the burden needs be confirmed in agreement with others, that 1) the source is contextually reliable, and 2) it directly supports the contribution. (In addition to other issues). Alanscottwalker (talk) 21:36, 6 June 2026 (UTC)

Folks… why are we getting sidetracked by a discussion about BURDEN - when the thread is about ONUS? These two shortcuts point to different sections of the policy. Can we refocus and get back to ONUS. Blueboar (talk) 16:02, 10 June 2026 (UTC)

BURDEN is a way to sidestep ONUS. If Alan gets his way, ONUS becomes irrelevant, because all I have to do is say "Nope, that's unreliable, and I've CHALLENGED it, so out the material goes until you FETCH me another source". I don't have to be correct about it, and I don't have to care about "editors examining things together", because BURDEN is a one-on-one action: I tagged it, and either you cough up a source, or I can remove it. If that becomes "you cough up a source that I agree is reliable", then – surprise! – no source in the world will ever be reliable enough to suit me, and the BURDEN to prove verifiability remains on you forever, or at least until you drag the whole thing through every dispute resolution process. WhatamIdoing (talk) 20:58, 10 June 2026 (UTC)

This kind of seems like a bad-faith assumption? If someone adds unsourced information and I CHALLENGE it on the basis of BURDEN and they re-add it while citing, for instance, IMDb or a wikia, I'm well within my rights to remove it again and point to WP:IMDB or WP:SPS. But if I'm claiming without evidence that a source isn't reliable (or reliable enough), then that seems like a situation where the other editor would be well within their rights to escalate the matter via the dispute resolution processes or otherwise getting other editors involved. DonIago (talk) 14:13, 11 June 2026 (UTC)

See above, many of your arguments seem to be premised on the 2nd editor (or the remover) doing stuff in bad faith or being obtuse and obstinate. The first requires admin intervention, the second requires others requesting, explanation and engagement, and if that does not work admin intervention as it at least looks like bad faith. Also, "unsourced" goes to at least three overlapping, yet important issues, V, NPOV, and NOR. At, any rate BURDEN is no less a community issue than ONUS or any other part of policy. In no way is ONUS irrelevant in my view. In my view, they must work together, one is discussing what is needed, and the other is discussing reaching consensus on that and other things. Alanscottwalker (talk) 14:43, 11 June 2026 (UTC)

I assume good faith on the part of both hypothetical actors. That is, I believe that the first editor believes that the Wikipedia article is improved by containing _____, and the second believes that the Wikipedia article is harmed by containing _____. They are therefore both acting in ways that, to the best of their beliefs, will improve Wikipedia.

I assume that both hypothetical editors are behaving honestly: The first editor, having been requested by the second to provide one (1) source per BURDEN, has added a citation to one source that the first editor honestly, genuinely believes is reliable per WP:RSCONTEXT. The second editor honestly does not believe that this is a sufficient source for this particular statement.

We don't know why they disagree. Maybe the first editor is uninformed about our ordinary sourcing standards, or perhaps the second editor has confused unreliable with undue, or perhaps the second editor deems the material to be WP:ECREE, and is trapped by ECREE's desire for "multiple" (emphasis in the original) sources and BURDEN's "one inline citation" (emphasis added). All we know is the fact that they do disagree.

I don't think that admin intervention is needed. I think that dispute resolution is needed, and I think that dispute resolution needs to proceed under the belief that everyone needs to cooperate in solving the problem, and that this idea of "everyone" cooperating overrules the idea that only the first/including editor has any obligation to find out whether sources exist.

Here's what I want this interaction to sound like:

• Newbie: I put some content in the article!
• Me: It needs an inline citation to a reliable source.
• Newbie: Sure, I read about it on CovidBleach.com.
• Me: For this kind of content, we actually need a source that meets WP:MEDRS standards. Give me a minute while I go check PubMed and Google Books...

What do you want the interaction to sound like? WhatamIdoing (talk) 20:02, 11 June 2026 (UTC)

In any given situation, it might be best to teach newbie what a reliable source is and how to find one. As the saying goes, 'teach how to fish, instead of give a fish'. Or direct them to others who can id a proper source in a particular domain. Or tell them, eg.,'that's not a proper source and I can't see how this bit matters to the article anyway, indeed where you found it and what that source says convinces me it almost certainly does not belong.' -- Alanscottwalker (talk) 20:43, 11 June 2026 (UTC)

Yes, there are a wide variety of appropriate responses. In terms of BURDEN specifically, though 'I can't see how this bit matters to the article anyway, indeed where you found it and what that source says convinces me it almost certainly does not belong' means "thanks, we're done with BURDEN, and now we need to move on to DUE, which this ain't." WhatamIdoing (talk) 22:30, 11 June 2026 (UTC)

Sure it may be done, because it is unsatisfied. Alanscottwalker (talk) 23:57, 11 June 2026 (UTC)

Because what is unsatisfied? WhatamIdoing (talk) 23:58, 11 June 2026 (UTC)

Burden is either satisfied or unsatisfied. That's what we have been talking about. Alanscottwalker (talk) 00:16, 12 June 2026 (UTC)

Maybe this will make sense:

The specific burden on the specific individual who added/kept the material is satisfied, because the person supplied the one (1) source required they're required to supply, even if I disagree that the source is actually reliable.

The general burden on everyone to have only verifiable material in articles is not satisfied. WhatamIdoing (talk) 00:19, 12 June 2026 (UTC)

So, the specific individual who is one of everyone has not satisfied the burden, either individually or as part of the collective. Alanscottwalker (talk) 00:30, 12 June 2026 (UTC)

The specific individual has satisfied their individual duty under BURDEN, which is to provide "any source they believe, in good faith, to be sufficient", even though a second editor disagrees that it's reliable. (NB that we don't know whether, in this story, the second editor is correct to disagree, nor whether the material is verifiable in a reliable source that has not yet been provided. "CovidBleach.com" sounds unreliable to me, but if the material is "Drinking bleach is unsafe and does not cure COVID-19 infections", then the material is definitely verifiable in other reliable sources.)

What you call "the collective" has not yet supplied a reliable source that the second editor will accept as reliable.

I wonder whether we need a more general, non-BURDEN-specific section on "What to do about badly sourced material" section. WhatamIdoing (talk) 04:23, 12 June 2026 (UTC)

It's the consensus in line with our policies and guidelines that must be satisfied that the source is reliable and directly on point. It's not the 2nd editor. The 2nd editor raises the issue. And it will never be the collective consensus supplying a source -- only one person brings it, usually the writer/proposer. The BURDEN of demonstration is on the writer/proposer and is only satisfied by providing a reliable source, on point.

If your issue is, is it best practice for the 2nd editor to supply the source (thus becoming the sponsor of the content, the taker of the burden) under certain circumstance, than, yes. But they must believe multiple things simultaneously and in no particular order: 1) the information is of the kind needing an inline source; 2) the information is verifiable (reliable and can be found on point); 3) the information belongs in the "finished" article; 4) they are qualified to identify such a source for this particular matter; 5) they have access to such a source and can supply it. And then their source and assumed proposal is open for inspection by consensus. On the other hand, if they don't have all that and all they can do is, recognize that bleach.com is unreliable for this medical info and that this is information that needs a qualified reliable source, they have done good work both for the 1st editor and the project in raising the issue.Alanscottwalker (talk) 08:51, 12 June 2026 (UTC)

In my practical editing experience (which certainly may not represent the experience of all editors), what tends to happen is):

1. Well-meaning editor adds unsourced information.
2. Unsourced information is challenged.
3. Original editor does one of the following...
   1. Adds the information with a source that seems reliable enough that it's unchallenged (or nobody notices).
   2. Adds the information with a source that's unreliable. Return to Step 2.
   3. Doesn't try to add the information again.

3-1 resolves the matter. 3-2, to my mind, may techically satisfy WP:BURDEN, but presumably leaves the original editor in violation of other P&G. 3-3 may or may not be a desirable outcome depending on the nature of the information the editor was trying to add.

This of course assumes both editors are operating in good faith. There can be tension over whether or not a source is reliable, but that's when the editors should seek a consensus (possibly via dispute resolution), and the information should be excluded from the article until a consensus favoring its inclusion has emerged.

I'll agree that in an ideal world best practice would be for the editors to collaborate to find a reliable source...but as noted, that's often not a realistic expectation, especially if the second editor doesn't especially care whether the information is added. We have no way to realistically mandate that editors work together to find sources, even if we concluded that it was a good idea to try. DonIago (talk) 14:32, 12 June 2026 (UTC)

Wrt BURDEN, I'm mainly concerned about the procedural aspect, rather than the citation aspect. (I think we probably need a non-BURDEN section that says the obvious, like "you can remove stuff that failed verification" and "if it's uncited, or if the cited source is unreliable, that doesn't automatically prove that the material is unverifiable".)

Specifically about the procedural process, BURDEN must not be written or wielded as authorization for the second/challenging editor to sit on the sidelines and say "No, not good enough, try again – of course I can't help; the burden is only on you. Try again, because that one won't do, either". It is a one-time BURDEN, not an unending string of requests or a sneaky way to block a change on procedural grounds.

If anyone's curious about the history, then this footnote started with this major discussion and RFC in 2012, and was later narrowed to specify "one" source. WhatamIdoing (talk) 19:04, 12 June 2026 (UTC)

Something like (off the top of my head) "BURDEN is satisfied once a citation has been applied in good faith. If there are questions regarding the specifics of the citation used (e.g. IMDb is cited in violation of WP:IMDb), BURDEN is not considered a valid reason to contest the edit. If editors disagree as to the appropriateness of the source being cited, they should discuss the matter at the article's Talk page and consider pursuing dispute resolution."? DonIago (talk) 19:12, 12 June 2026 (UTC)

I wonder whether Alan would agree specifically to the "BURDEN is not considered a valid reason to contest the edit" line. It's my impression that he believes that an unreliable source (e.g., IMDb being wrongly cited) is still specifically a BURDEN-ALWAYS-ON-YOU-AND-NEVER-ON-ME problem. WhatamIdoing (talk) 19:23, 12 June 2026 (UTC)

We'll see what Alan comes back with. If IMDB is cited then IMO (if it's not clear above), it's transitioned from a BURDEN problem to a WP:IMDB problem, similar to citing any other unreliable source (WP:RSP, etc.). DonIago (talk) 19:30, 12 June 2026 (UTC)

"The burden to demonstrate verifiability lies with the editor who adds or restores material, and it is satisfied by providing one inline citation to a reliable source that directly supports the contribution."

If you don't add or retore material, you don't have the burden. (You have a best practice to volunteer under certain conditions to assume (take up) the burden, set out above. And of course, you can volunteer to take on yourself anything that improves the article). And the burden begins immediately, not waiting for anyone else to show up and it is satisfied only per above. As I understand it IMBD, is a consensus of unreliability in most cases, so in most cases it does not satisfy the BURDEN (it is NOT "a reliable source that directly supports the contribution"). Editor 2 can challenge. Editor 3 can challenge. Editor 4 can challenge. As long as it is there or readded. Alanscottwalker (talk) 20:54, 12 June 2026 (UTC)

The question is whether there's ever a point at which offering a source stops being the duty of just one specific individual ("the editor who adds or restores material") and starts being a duty owed by the community generally.

Consider the case of IMDb: It's generally discouraged. In most cases, it is not accepted as reliable. It's also very common for newcomers to genuinely believe that it is reliable. BURDEN also has a statement about "an editor [who] has provided any source they believe, in good faith, to be sufficient". BURDEN also says that "all editors are then expected to help achieve consensus, and any problems with the text or sourcing should be fixed" – all editors, not just the one who added or restored it.

How do you reconcile your claim that the burden is always exclusively on the one editor, with BURDEN's statement about "all editors"? WhatamIdoing (talk) 21:42, 13 June 2026 (UTC)

As that footnote makes clear "(e.g., why the source is unreliable; the source does not support the claim; undue emphasis; unencyclopedic content; etc.)", there are multiple things to work to achieve consensus on, in broad terms, work to achieve consensus on the content of the article. That's the main reason why editors exist at all, and why our content reflects consensus: editors work to achieve consensus on whether the burden has been satisfied, does the addition have one inline citation to a reliable source that directly supports the contribution; they work to achieve consensus on whether the addition complies with all other policies and guidelines as suggested in that footnote and in common sense hashed out among themselves, or is it excluded. None of that does not make, "The burden to demonstrate verifiability lies with the editor who adds or restores material"-- it's there in bold. And as I indicated, another editor is free to advocate for that content addition and adopt the burden as their own but it's not their burden until they decide that's what they want to do, ie. they themselves want the addition or restoration, consistent with the policies and guidelines. Alanscottwalker (talk) 23:46, 13 June 2026 (UTC)

So BURDEN says that "the editor" must do this, and also that "all editors" must do this. Why do you think that the former overrules the latter? WhatamIdoing (talk) 00:36, 14 June 2026 (UTC)

I have explained how it goes together not "overrules", the "the" talks of burden, the "all" talks of consensus making (Although your "this" is vague, at best. Also, it's obvious with its bolding and being in the main, the first is central to the policy). The "all" certainly does not mean what you suggest, that they are to be forced into a false consensus. That's not how consensus is made or helped.-- Alanscottwalker (talk) 09:42, 14 June 2026 (UTC)

The problem is that we have a consensus, as documented in multiple discussions in the archives of this page, that say that an editor can't endlessly be forced to play a game of WP:FETCH. However, if we interpret the bold text as meaning that the burden is exclusively on that editor, and that it continues to be in force until the opponent agrees that the supplied source is reliable – No, not that one, bring me another! – then an endless game of FETCH is exactly what we'll get.

How can we clarify in the policy that once the editor has tried to provide a reliable source, the burden of making a second attempt actually isn't on that editor any longer? As in:

• You want to keep it, so you go find a source.
• I say your source isn't reliable (and I'm correct).
• The burden for finding a reliable source is not on you any longer. You did your part. Now it's up to "us" instead of "just you". "We" now have to find a reliable source, instead of "just you (again and again and again)", or come to an agreement that no such reliable source exists.

WhatamIdoing (talk) 04:23, 22 June 2026 (UTC)

No. They don't have to endlessly fetch, they can decide not to support at any time. Surely, that's the correct thing to do in many instances. If anyone else decides they want to support, they can provide the source. Indeed, their support makes it their burden. Surely, that's the correct thing to do in many instances. All never provides the source, only one does. Alanscottwalker (talk) 12:20, 23 June 2026 (UTC)

At what point in time does someone else have to help? Ever? Is this a case of "Welcome to Wikipedia. We're a collaborative project. Everyone has to work together – oh, except if you're asking for a source. That doesn't require you to lift a finger, because the burden is always on the other guy, and he has to choose between endlessly supplying sources or losing the material, and all you have to do is say 'Nope, that's not reliable either. Try again, or hold an RFC and try to make me stop rejecting all these sources'." WhatamIdoing (talk) 19:00, 23 June 2026 (UTC)

Of course it may be a long time, we don't require anyone to look constantly and continually over anyone's shoulder. Everyone has to work together but we don't have to support poor edits, indeed part of working together is rejecting poor edits, which benefits both other editors and the project. I sometimes think users who have trouble being edited should put up something like 'yes, I know I am being edited here' on their own page to remind themselves of what this project entails. Alanscottwalker (talk) 21:04, 23 June 2026 (UTC)

I guess I should add, that I don't know if I have ever said BURDEN to anyone in an editing dispute, I have said something like 'that's not a reliable source' and 'I don't support that addition', etc. Alanscottwalker (talk) 22:34, 12 June 2026 (UTC)

Blueboar, we are discussing the interplay of two sections of the policy, that they are both part of policy means they must work together and both apply. If it will help, add "See ONUS." to the end of the comment you responded to. Alanscottwalker (talk) 15:11, 11 June 2026 (UTC)

I think I only tend to invoke BURDEN when editors get mouthy at me after I've deleted their unsourced addition (e.g. "If that's such a problem for you then you add a source, leave me the fuck alone!"). DonIago (talk) 01:30, 13 June 2026 (UTC)

Iteration 2

• One of the things ONUS does is give a process for dealing with a fact or claim that's verifiable, but in the wrong place.
• Only use ONUS when you give a reason for your change. ONUS is not a reason---it's a process.
• ONUS enables editors to move any fact or claim, including a fact or claim that's meticulously cited, to the article that's most appropriate for it.
• If there's no appropriate article, options are to start one, or to delete the disputed fact or claim.
• Because ONUS involves moving material that's properly cited, WP:V is a natural place where editors would tend to look for the rule.
• It interacts with WP:PRESERVE to mean that if the fact or claim belongs anywhere in the encyclopaedia, it must be moved rather than deleted.
• It interacts with WP:UNDUE to mean that if there's dispute about where to put the fact or claim, we tend to put it in the less prominent place. This can be good or bad.
• It's important for dealing with marketers and people with an agenda, who often won't want you to move their well-cited fact or claim to a place that gives it less prominence. ONUS is helpful in that situation.
• It's also useful for those who want to whitewash an article. This can lead to misuse.
• Longstanding versions of the article don't necessarily have consensus---particularly for low-traffic articles. The best way to tell whether a fact or claim has consensus is if it's been specifically discussed on the talk page. It likely also has consensus if it's been substantially edited by a number of different editors over a reasonable period of time without challenge. (Examples of edits that aren't "substantial" include minor MOS-related edits, categorization, autowikibrowser, and typo fixes).
• Where there's an apparent conflict between WP:ONUS and WP:NOCON, WP:NOCON only prevails where there's good evidence that the fact or claim has consensus. Otherwise WP:ONUS prevails.

This essay needs a title. It's about "How ONUS interacts with other policies and guidelines" but that's too long. Suggestions?—S Marshall T/C 10:58, 21 May 2026 (UTC)

Maybe try something along these lines:

• Wikipedia:What is consensus?
• Wikipedia:What "no consensus" means
• Wikipedia:What "Ignore all rules" means
• Wikipedia:What "Ignore all rules" really means
• Wikipedia:What does "per" mean?

WhatamIdoing (talk) 21:02, 21 May 2026 (UTC)

Thank you. Wikipedia:What WP:ONUS means, then?—S Marshall T/C 21:52, 21 May 2026 (UTC)

Maybe Wikipedia:What ONUS means. The "WP:" bit looks odd to me (but your choice). WhatamIdoing (talk) 22:06, 21 May 2026 (UTC)

It's up. Further thoughts welcome.—S Marshall T/C 14:47, 24 May 2026 (UTC)

Grey area: user-generated demonstrations based on otherwise well sourced statements

It seems to me an omission of this Verifiability article, that it doesn't mention any possibility of creating original work that attempts to provide a better demonstration of some propositions which are otherwise well sourced.

Image use policy clearly does allow original demonstrations in "user-made images may be wholly original. In such cases, the image should primarily serve an educational purpose" [Image use policy > Diagrams and other images], but this problem applies not only to images, but to any media, including audio, where it has lead to questioning of the demonstrations of audible phenomenon (e.g. Just intonation's "citation needed" template, relevant Talk discussion, citing Verifiability as forbidding original demonstrations, but only ones published elsewhere).

I want to make clear that I am not looking for a total permissiveness for any user-made demonstrations - I do recognize that demonstrations should unconditionally have verifiability information for the basic claims that the work is trying to demonstrate and those that lack it are due for improvement in this regard, or even removal.

It seems however that correctness of demonstrations is a separate issue altogether. Admittedly, it is not without its own pitfalls. Some demonstrations' correctness is easier to verify just by just observing them (e.g. geometry's Pythagoreas theorem animation) whereas some works, in their present format, are complex enough to be able to stealthily propagate wrong knowledge (e.g. for acoustic phenomenon wrong sine waves problem) and definitely require addressing eventually (ideas of solutions).

But it seems to me that it would be very detrimental to Wikipedia if the only demonstrative works allowed were those almost identical to versions published elsewhere, especially that it pushes the platform away from providing user value through best explanations, fully utilizing Wikipedia's digital nature, and presses it towards self-inflicted copyright infringement hell.

Having this in mind, is there any chance that any hint of the distinction is drawn in the Wikipedia's verifiability rules? Zlamma (talk) 12:11, 21 May 2026 (UTC)

• The interesting aspects of just intonation and 12 equal temperament reduce to understanding ratios, which is mathematically trivial and falls within WP:CALC.
  We allow user-generated files. In biographies, it's quite normal for Wikipedians to upload their photographs of the person; in articles about places, it's quite normal for Wikipedians to upload their photographs of the place. And for example, I personally drew all the site plans in Ness of Brodgar.
  In my view you do not need a source for those files, and the user who says otherwise is mistaken.—S Marshall T/C 13:10, 21 May 2026 (UTC)

  Thank you for your opinion. The "WP:NOR > What is not original research" seems to indeed be tackling the very matter, including non-image media.

  What I am hoping in this topic is for WP:V to also help discover what the boundaries to the rule are, to avoid overzealous actions. Perhaps it should have a section akin to "What is not an infringement of Verifiability", even if it just links to the relevant sections in other rules? Zlamma (talk) 14:16, 21 May 2026 (UTC)

  WP:V and WP:NOR are basically the same. NOR says you can't put new, previously unpublished ideas in Wikipedia. WP:V says that things in Wikipedia must have been published in a reliable source (and that some things are required to be WP:CITED, though not necessarily by any particular WP:DEADLINE). If something is acceptable under NOR, it's probably also acceptable under WP:V.

  Glancing briefly at the discussion at Talk:Just intonation, it appears that what's needed is more attention on the "in such cases" phrase. Someone might be interpreting that as "in all cases". WhatamIdoing (talk) 19:41, 21 May 2026 (UTC)

  Thanks @WhatamIdoing. It is likely true that that user here likely missed the conditional premise in WP:IUP that was making the requirement apply just to cases of 'user-made version of a presentation that is copyrighted', but it was me who brought WP:IUP to the discussion, whereas earlier the user pointed at just WP:V as justification to the arguments (1, 2) and to edits that removed some contributions. I do feel that especially certain strong verbiage of WP:V suggests that there are no exception to the "provide reliable-source" rule, and it's not immediately clear from WP:V that "WP:V and WP:NOR are basically the same", especially to the exceptions in WP:NOR, so I feel this could use addressing. Zlamma (talk) 00:50, 22 May 2026 (UTC)

  Images (and related media) are supposed to be illustrating what's already cited in the text. Therefore they do not need separate sources as a general rule. WhatamIdoing (talk) 01:28, 22 May 2026 (UTC)

  Again. I don't see WP:V mentioning any guideline for this case, so I feel this is something that could be improved.

  (That is whichever way the matter is decided, as some users see the problem with demonstrations-complexity exhibiting full spectrum, including cases where verifiability is borderline-impossible with the current state of tooling. I would not mind there being at least encouragement of the users to prefer open-source demonstrations, original files in the editor-software format, or prose giving the recipe for the creation.) Zlamma (talk) 11:25, 22 May 2026 (UTC)

I had a discussion with Trumpetrep about this on my talk page after he came to me for advice; I invite interested parties to see our comments at User talk:UpTheOctave! § Audio files, since I will only summarise them here. While I seem to have reached broadly the same conclusion as the users here that sources are not required in the article, and that the maths behind the needed calculations is routine, I do think Trumpetrep's argument has some merit. This is genuinely a grey area: the image use policy is expressly for images, so while analogies can be drawn, I'd be wary at saying whether certain passages definitively apply to audio files.

I'd like to shift the focus slightly, as my concern is not the lack of a source for audio files, but the current, opaque nature of most file descriptions. WP:WHYCITE makes it clear that for an image or other media file, details of its origin and copyright status should appear on its file page. My reading of this is that we already require a provenance for each file in some level of detail, but it's just not enforced. Looking at examples like File:A Major Scale, Triads, and Fifths Just.ogg and File:A Major Scale, Triads, and Fifths Equal.ogg, most of these audio files simply give a description of what they are, and not information on how they were created (i.e. their origin). They may well be what they purport to be, but any mistakes are very hard to find without any indication of how they were produced. As such, I think it might be worth considering drafting some guidance on how to write good file descriptions for audio demonstrations. UpTheOctave! • 8^va? 01:44, 22 May 2026 (UTC)

Images are not audio files.

Most of the cited precedents here are from Wikipedia's policy on images. The DIY ethos of those guidelines is clear enough. Images and audio files might seem closely related, but audio files present different problems.

User:S Marshall cited the Routine calculations policy, which also might seem to solve the problem. However, the policy specifies much simpler calculations than ratios, "Basic arithmetic, such as adding numbers, converting units, or calculating a person's age, is almost always permissible."

Note in particular the closing admonition, "Comparisons of statistics present particular difficulties." Guess what has also presented particular difficulties over thousands of years of human civilization: the mathematics of just intonation.

The File:A Major Scale, Triads, and Fifths Just.ogg file is an example of the problem. Simply saying the file is "just" is akin to ordering a "sandwich" at McDonald's and leaving it at that. There are too many different formulations of just intonation to not go into more detail for our target audience, which is the general reader.

User:UpTheOctave! is on the more productive track here, which is how to ensure that any audio file used in an article is accurate. Requiring editors to precisely document how the files were created seems reasonable. Doing so invites the kind of peer review process that seems out of bounds for Wikipedia, but it's better than nothing.

I would also encourage guidelines about when user-generated audio demonstrations can be used, and Wikipedia would be a better encyclopedia if those guidelines were extremely narrow. Trumpetrep (talk) 03:13, 22 May 2026 (UTC)

Yes, it does say "Comparison of statistics present particular difficulties", because I wrote that. A marketer or POV pusher has an incentive to compare dissimilar statistics -- ones that derive from different areas, or different time periods, or from studies that used different methodologies -- in order to draw you towards the conclusion they desire you to reach.

Audio files containing only pitch comparisons are fundamentally dissimilar. Nobody is trying to sell anything or win an election based on such a file.

We don't require editors to document how they generated a file and we have no plans to start.

What we don't want is a problem that looks like this:

Anne: This is a diagram of a right angle triangle.

Bob: Prove it!

Anne: Look, it's got a right angle at the bottom left corner. See? It's got three straight sides. It's a right angle triangle.

Bob: Only if you drew it on a flat surface. Did you?

Anne: Yes.

Bob: Prove it! Or I'm deleting your image.

Anne: Okay, whatever. (Gives up encyclopaedia writing and goes away forever.)

Bob: Yay, I win! The badwrong image can never be used to mislead people, and the badwrong person is gone!

In this scenario Anne is in the right and Bob is the problem.

NB: We use media files from Wikimedia Commons so this isn't even entirely up to us. It's also a Commons issue.—S Marshall T/C 07:33, 22 May 2026 (UTC)

The scenario you described played out recently with this image that claimed to be an E 9th chord. It was created for and used for years as the lead image at Nashville tuning. It was absolutely incorrect. Not only did it not depict an E 9th, but it was not an accurate notation of Nashville tuning. The picture has now been replaced by a user-generated image with two citations that readers and editors can check to verify its accuracy.

As for POV pushing not being a problem in music theory, the Talk page at the recently redirected "Regular temperament" article offers a window into just how much of a problem bias can be in the topic.

And again, audio files are not images. Wikipedia should have a specific policy for audio files, because they are an entirely different medium.

Returning to the Routine calculations policy, the section is immediately followed by a reminder of Wikipedia's first principle:

"Wikipedia's content is determined by previously published information rather than by the personal beliefs or experiences of its editors. Even if you're sure something is true, it must be verifiable before you can add it. The policy says that all challenged or likely to be challenged material and all quotations need a reliable source..."

Why would Wikipedia abandon this requirement when it comes to audio files? Trumpetrep (talk) 13:10, 22 May 2026 (UTC)

Not to pour any of the Anne & Bob babies with the bath, I still feel an improvement to the guidelines could help (I remind that the topic's thread is whether any improvement to the guideline is warranted. While I think both sides' concerns are valid, I would suggest to make it clear when there is no contention to coming up with some guideline improvement):

On the one hand, there could be a recognition and a reminder that some verifiability problems are not down to lack of citations, so citation-needed are the wrong templates to use (even when there's a citation, because of the need to avoid copyright-infringement like described in this case, the demonstration can still be wrong and unverifiable). Additionally, when the demonstration's correctness is doubted, the challenger should first consider whether the intent to provide a demonstration was helpful. If so, the course of action should not be outright edit reversals and mostly discouragement, but rather an encouragement to improve the demonstration, to allow verification for correctness.

On the other hand, there could readily be a guideline with encouragement of the users to prefer open-source demonstrations, original files in the editor-software format, or at least providing prose giving the recipe for the creation (a template could help, both in Wikipedia, but I feel that one for Commons could be even more effective, explaining that following the guidelines would be very valuable for the encyclopedic use of the work). Zlamma (talk) 13:37, 22 May 2026 (UTC)

@S Marshall: I get what you're saying, but I don't think that's an entirely fair representation of the argument. No-one's saying that file creators would be forced to prove they did what they say they did: as you imply with the Anne–Bob argument, that would be unreasonable as a violation of WP:AGF. All we're saying is that to avoid the issues with these black box-like file descriptions, which from Trumpetrep's examples are a real problem, we should consider encouraging better file descriptions locally. Since a lot of files on Commons are destined for the English Wikipedia anyways, and it would be here that they are actually embedded, I don't think we are overshooting our jurisdiction if we pursue local guidance. In reply to your comment, we don't require editors to document how they generated a file and we have no plans to start: with the Anne–Bob issue sidestepped, my question is, why not? Better file descriptions don't just serve verifiability in the abstract, they make errors findable and fixable, which is especially important since audio errors can fall below the threshold of epistemic detection (see just-noticeable difference). UpTheOctave! • 8^va? 15:29, 22 May 2026 (UTC)

User:UpTheOctave!'s admirable invocation of epistemic detection cuts to the heart of what I've been hinting at when I stressed that audio files present different issues from images. (I didn't want to get distracted with technicalities.)

S Marshall's diagrams at Ness of Brodgar are supported by a citation. Other editors can see where he got the information for the images he created. That practice comports with Wikipedia's core concept of verifiability. It seems reasonable to ask the same of user-generated audio.

There are only a few of us participating here, but three of us seem to agree that a guideline about audio demonstrations would be worthwhile. What's the process for doing that? Do we just hash out language here or take it to some broader forum? Forgive my ignorance about Wikipedia's internal processes. Trumpetrep (talk) 15:54, 22 May 2026 (UTC)

@Trumpetrep this would be something you'd want to float at the Village pub. (proposals) if you're getting input on a draft proposal, (policy) if you feel that that draft is ready for community consensus yea or nay. If it's at policy, you want to frame it as an RfC.----3family6 (Talk to me|See what I have done) 16:35, 22 May 2026 (UTC)

@Trumpetrep, in my mind this problem isn't specific to audio-files, but rather applies to other complex works where only the output of the black box is published. We all did seem to agree here in Talk:Just intonation that this dilemma may appear even for images, when they are complex enough (there is more that images can erroneously claim than just right angles :) ).

I feel that the just-noticeable-difference realm is not the essence of the content that is problematic, but a broader category of content form which is more complex to analyze cognitively than practically feasible. Even if one can't perceive the difference between just-intonation and TET, there are forms of providing an audio demonstration which one can verify and those that would be very hard.

This is the reason I saw it pertinent to raise the discussion here. I also feel that, by not trying to create a prescription for specific medium (though by all means audio can be listed as example), it will be easier to create something that mostly just shows correspondence with the already existing first-principles of Wikipedia. Zlamma (talk) 00:08, 23 May 2026 (UTC)

Yeah, I agree with everything you are saying. So, what's the logical next step? Trumpetrep (talk) 00:46, 23 May 2026 (UTC)

Regarding formalities, I myself am quite ignorant here, but I only found what 3family6 had in mind: Wikipedia:Village pump.

Judging by W:VP (proposals) being described there as 'non policy related', I suppose the right place for discussing a draft of changes would more likely be just W:VP (idea lab), and in case of approval, submitting to W:VP (policy).

Regarding producing some draft/scaffold of the text, if it is helpful for me to produce something, I think at some point in the week I may find a moment. If someone readily feels under the muse, my quick share of thought for how I would see it is that I wish the text at least reflected what I wrote above , especially about reminding both sides to take good faith efforts to avoid any of the two extremes.

Same regarding my point that "citation needed" is not always the right template to use. Here, my ideation about template(s) for this case is probably a bit wishful but, in either case, I feel WP:V does need some text giving discoverability of what the best "template of behavior" is rather warranted. Ifeel that WP:V is currently a bit blind to the facts that 'publication' and 'citations' for verifiability is neither always necessary (as per WP:NOR > What is not original research and W:IUP#Diagrams) nor sufficient (as per "even when there's a citation, because of the need to avoid copyright-infringement like described in this case, the demonstration can still be wrong and unverifiable").

To anyone else in this topic, you may have and idea for a different course of action or choice of words into words, and I'm also happy to just assist.

In either case, any feedback, both encouraging and principled-discouraging is appreciated! Zlamma (talk) 00:00, 24 May 2026 (UTC)

A late 2c: There needs to be space in policy for simple demonstrations of concepts such as the ones described here that are not merely copies of the same demonstrations from copyrighted material (which would likely be a copyright and fair use policy violation), because if we make our sourcing requirements so strict as to prevent those demonstrations we would likely be left with no demonstrations at all. I have in mind, for instance, the lead image of Featured Article Euclidean algorithm, which provides a graphical demonstration of a specific instance of the algorithm. If it merely copied the numbers and graphical appearance from a textbook, it would be a copyvio problem, but if our sourcing rules were tightened so much as to prevent us from using demonstrations like this unless they were copied wholesale from textbooks, it would also present a Catch 22 for article content preventing us from including any such demonstrations at all. Maybe in this specific case we are saved by WP:CALC but I am not sure that is true more generally for the types of demonstrations under discussion here, and I am also not convinced that even WP:CALC is safe from well-meaning but poorly thought out "let's improve our sourcing and verifiability requirements" proposals. —David Eppstein (talk) 00:17, 24 May 2026 (UTC)

I completely agree with regards to requiring a source for every demonstration; I think others have mostly moved on from that original proposal as well. Would you mind if I pick your brain on the alternative suggestion of pursuing better documentation for how these demonstrations are created? UpTheOctave! • 8^va? 01:04, 24 May 2026 (UTC)

Media files and verifiability

Wanting to unpack this and take a close look the principles that underlie verifiability as it applies to media files.

Per policy, verifiability applies to facts and claims. Media files can be used to support various different facts or claims.

Where verifiability comes in

Here is an example of some information.

This information isn't making a fact or claim, and therefore verifiability doesn't apply to it.

Here is the same information with a fact or claim. I've highlighted the part where verifiability comes in.

• This is Beethoven's 5th Symphony.

But let's imagine that I use the same information to support other claims.

• This is a musical phrase in C minor.
• This is information that's out of copyright.
• This is wikicode that automatically generates a media file.
• This shows how tuplets are written in musical notation.
• This is the mating call of a wild piano.

Each one of those needs to be separately evaluated for verifiability. Having a source for one claim doesn't verify any others.

Also note that if someone breaks or rewrites the score extension, the way that information looks or sounds will change.

Media files and Commons

Now let's take the case of a file that's hosted on Wikimedia Commons. These files are usable on Wikipedia per policy, but not within our control. (Over there on the right ---->)

The Eiffel Tower

Someone on Wikimedia Commons could replace this file. This affects the accuracy of the information, and is not within our control or the control of anyone on en.wiki. We don't get to make rules about that.

The claim that "This is the Eiffel Tower" is one you can check, but you have to use your eyes and your knowledge (or a personal visit to Paris). You can't check it with a citation.

"How did you produce that?"

Let's now imagine, hypothetically, that I produce media files demonstrating just intonation vs 12 equal temperament. You challenge me to explain how I produced it. Here are some examples of answers I could give you.

• It's raw sine waves, tuned and played using Audacity.
• It's a real guitar, tuned by ear and played into a microphone.
• It's a sample of a violin, pitch-shifted in Ableton.
• I generated them by typing the frequencies into onlinetonegenerator.com.

You can't check a single one of those claims with a citation. You can only test them with your ears and your knowledge.

This is quite separate from WP:V. Verifiability is about making sure anyone can check a fact or claim, i.e. without any specialist knowledge.—S Marshall T/C 09:44, 25 May 2026 (UTC)

Quibble… Verifiability is about making sure that someone (not necessarily you) can check without having specialist knowledge.

There are many things that I can not personally check (a book in a language I don’t speak, for example) - but since I can ask others to check on my behalf, they are verifiable. Blueboar (talk) 11:54, 25 May 2026 (UTC)

Okay, quibble acknowledged; but I feel it's, at best, tangential to this matter. Someone with the right equipment might be able to verify a claim like "this is a sine wave at 440 hertz". Nobody but the uploader can verify how a sine wave at 440 hertz was produced because there are so many ways to do it.—S Marshall T/C 12:23, 25 May 2026 (UTC)

@S Marshall. Thanks for your response. Could you clarify one thing missing for me? I'm unsure whether you argue against always requiring a form of reproduction steps (source code, editor file or prose), which I feel most here find too extreme, or whether you against even encouraging them for at least some cases, or whether you argue against any initiative to change the WP:V text, even to just remind of the notable distinction and record the consensus.

Second thing which I feel underpins a difference in which w look at the matter is, re:

"I generated them by typing the frequencies into onlinetonegenerator.com. You can't check a single one of those claims with a citation. ... Verifiability is about making sure anyone can check a fact or claim."

If a different generator was described that allows downloads, and whose files produced are deterministic (I think I found one on audiocheck.net), then to me, using simple file contents comparison "anyone can check a ... claim" about what the file contains (the claim that someone use that website is surely uninteresting for anyone). An even better version of the commons file would be one that provides source code. The discussion is about whether, in order not to limit how high quality Wikipedia can ever become, it shouldn't provide encouragements to provide these types of verifiability. Zlamma (talk) 11:59, 25 May 2026 (UTC)

Those are quite specific questions!

You asked: Do I argue against always requiring a form of reproduction steps? Yes I do. An editor might download a CC0 sound file from someone's blog, and then upload it to Wikimedia Commons. Then the blog goes poof and nobody knows how the sound file was produced, and the editor who uploaded it vanishes. But it's a legitimate sound file that should still be usable.

You didn't ask: Do I argue against encouraging editors to explain how they produced a file? If you'd asked me that, I would have no objection.

You asked: Do I argue against any initiative to change the WP:V text? At this stage, yes I do. If after consulting Wikipedia:Village pump/Policy, Wikipedia:WikiProject Media, Wikipedia:WikiProject Professional sound production it's clear that the consensus is against me, then I might withdraw that objection.

You didn't ask: Is WP:V the right rule to edit in this situation? If you'd asked me that, then I would say that the right rule to edit in the first instance is Help:Creation and usage of media files#Audio.—S Marshall T/C 12:49, 25 May 2026 (UTC)

Re source code: some of the mathematical images that I have uploaded to commons were generated by a script that I wrote, whose source code I have uploaded to the commons description page (example). Others were drawn by me by hand using a graphical user interface, most frequently in Adobe Illustrator (example) or, occasionally, they include a combination of programmatic generation and hand touchup. It might be possible for a knowledgeable reader to look at the source code and verify that it is generating what the image claims to be about, but I think in most cases it is far easier to look at the image. I don't think stating how an image is generated, in these cases, is very helpful for verifiability. It may be helpful for other reasons (for instance to generate variations of the image) but that is not about verifiability.

Incidentally, in the case of the programmatic example above, what needs to be verified is that the image has the same geometric features in roughly the same positions as the original image by Catalan that it copies. This is not something one can verify by looking at the code. And the positions are not exactly the same. In this case, the geometry is exact, but that is also not always needed. I have seen arguments that for some editions of Euclid's Elements the geometry of the illustrations is deliberately inexact, not so much because of the limitations of the person creating them but because an inexact illustration forces the reader to pay attention to the mathematical arguments and not to rely too much on their visual intuition. —David Eppstein (talk) 23:30, 25 May 2026 (UTC)

I feel as if mathematics is a special case, in terms of WP:V.

I feel as if a well-formulated mathematical equation, or a well-written computer program, is self-verifying (if it has a unique solution). Maybe not to Randy from Boise, but to a David Eppstein or a Charles Matthews, to i.e. to someone who "speaks maths", a valid mathematical proof is self-verifying.

This is really important in maths because the topic area benefits so much from worked examples, and because you can't just copy/paste a worked example from a textbook without violating copyright.—S Marshall T/C 00:00, 26 May 2026 (UTC)

Yes, mathematics may be special in a sense but there was a very long discussion about cartography a month or two ago on RSN. It is more complicated than math diagrams. User Geosage was the only one who understood the field there. So the special list may be longer than appears. As for "speaking math" the issue of dialects is serious. I am certain that there are topics that Eppstein knows better than Matthews and also the other way around. Do either of them know the recent issues in fluid dynamics? Anybody's guess. So if someone adds a diagram on that who checks it? Relativistic gravity? Quantum computing theory? Does David Deutch edit here? The list of topics can be quite long. Overall, I do not see a clear solution yet. Yesterday, all my dreams... (talk) 03:31, 26 May 2026 (UTC)

Twenty years ago when machine translations were ghastly unreliable, Wikipedians managed to agree that a foreign language source is still a potentially good and valid source, and if you can't verify a claim because you don't speak the language, then you just have to ask someone who does. We conspicuously failed to agree the same thing for mathematics.

This might be because such a high proportion of Wikipedians graduated in the arts and humanities.

Nowadays, when machine translations are confident and plausible, but sadly not necessarily accurate, we have editors "translating" articles from and to languages that the editor does not speak. I suggest not allowing people to self-certify that their mathematical proof is valid. We would need a panel of trusted reviewers.—S Marshall T/C 09:07, 26 May 2026 (UTC)

Yes, panels of trusted reviewers in all technical areas would be good, as would the end to world hunger. I am not sure which will happen first. The unspoken truth is that Wikipedia is seriusly understaffed in technical areas. So we will struggle along, as is. Yesterday, all my dreams... (talk) 11:50, 26 May 2026 (UTC)

Thanks for your thoughts @David Eppstein.

Regarding "reading the code" not always being the best way to verify, while this is true, I think it's also true that it is sometimes helpful, like in the case of audio from plain waveforms that is given in the initial comment of this topic.

Moreover, I'd like to emphasize that providing the code means also that the code can be executed, which vastly expands the group possible verifiers (think even simple human errors on any level - wrong program ran, wrong parameters, wrong constant value, etc), whereas reading+changing+executing loop gives tools for further expansion, as users can even upgrade their understanding to become verifiers.

Overall, because the original post is just claiming that there are some cases where this is a real problem and is just looking for a good-faith guideline that prevents deletion from WP:V on one side and stalling quality verification on the other, I hope that it's not going to be buried under comments about the cases for which this problem is likely never going to surface. Zlamma (talk) 11:54, 27 May 2026 (UTC)

Thanks for clarifying that you also, like most here so far, are merely looking to consider encouraging users to consider leaving more details on how they produced a file. The way I formed the questions was just reflecting my genuine unknown which matters you are in agreement with. Apologies if they sounded too specific.

Re:

> "You asked: Do I argue against any initiative to change the WP:V text? At this stage, yes I do. If after consulting Wikipedia:Village pump/Policy, Wikipedia:WikiProject Media, Wikipedia:WikiProject Professional sound production it's clear that the consensus is against me, then I might withdraw that objection."

Wikipedia:Village pump/Policy actually encourages using talk page like this to "for drafting with a more focused group", so we may as well address any remaining arguments here before they repeat in high-visibility forums.

Re:

> "You didn't ask: Is WP:V the right rule to edit in this situation? If you'd asked me that, then I would say that the right rule to edit in the first instance is Help:Creation and usage of media files#Audio

If you can, please elaborate why the earlier explicit implorations to not make an audio-files specific solution should be rejected, by responding in our latest thread. Zlamma (talk) 00:37, 28 May 2026 (UTC)

@S Marshall, another quibble from me. Your example with the Beethoven excerpt is inconsistent with the practice used in the field: at FAC and other venues, excerpts are still expected to be sourced to the score wherever they appear per WP:V, as they are reproductions – i.e. direct quotations. I agree with the rest of your essay though, which is why I suggested encouraging some form of documentation for user-created demonstrations earlier. I'm glad to hear above you have no objections to this route, as I think it is the path most people have now coalesced on. Thanks, UpTheOctave! • 8^va? 14:08, 25 May 2026 (UTC)

The excerpt isn't what needs to be sourced! It's the fact or claim that it's making. In that case, what needs to be sourced to the score is the fact or claim that "this is Beethoven's 5th".—S Marshall T/C 14:42, 25 May 2026 (UTC)

I'm not quite getting your point. The claim that "this [the excerpt] is Beethoven's 5th" cannot exist without the excerpt it refers to. Although in a purely technical sense, the citation is attached to the prose, the text only exists to provide attribution for the excerpt, as is required for any other direct quotation: in practice, the excerpt would otherwise likely be removed for a lack of relevance. In my mind, it's a similar situation to quotations of text, where most will have some kind of introductory clause describing the speaker. WP:V requires a citation for direct quotations regardless, and does not consider these clauses: why would it not apply to musical quotations? UpTheOctave! • 8^va? 16:08, 25 May 2026 (UTC)

The confusion may be about an assumption that I've made. So let me state it explicitly.

The proposal we're considering is to encourage editors to explain how they produced a media file, in this case a sound file.

I have assumed that this explanation would be attached to the file, rather than to the claim. This allows for one file to be reused to support several different claims in several different articles. ("This is the sound of an electric guitar"; "This is the sound of a Fender Stratocaster being played through a Marshall amplifier"; "This is a piece in E pentatonic minor"; "This is an instrument being played at 120 bpm".)

But this policy is about claims, not files. What WP:V relates to is individual claims in context, so to comply with WP:V, each one of the above claims, if challenged, might need a separate citation.

Therefore the proposed rule isn't very compatible with how WP:V works.

I also want to add that my method statement, hypothetically "I generated this file myself by playing my Fender Stratocaster through my Marshall amplifier into a microphone", would I think, be remarkably difficult to prove to someone who challenges it.—S Marshall T/C 22:25, 25 May 2026 (UTC)

Ok, I will produce a file for Beethoven's 12th symphony. Let us see what happens. Yesterday, all my dreams... (talk) 22:42, 25 May 2026 (UTC)

Ah, I get you now, thank you. You're right about crossed wires: I thought you had pivoted to talking about the use of musical quotations in articles (see examples like Saxophone Sonata (Creston), Short Symphony, etc.), where common practice is as I described. I only invoked WP:V due to this misunderstanding – my own proposal is to instead pursue better standards of documentation for user-generated files – so I agree with what you're saying with regards to claims and the policy's irrelevance to file descriptions.

Since I think we've all come to a similar conclusion on the correct course of action, I wonder whether it would be worth porting this discussion over to Help talk:Creation and usage of media files? Pinging substantial participants to gauge opinions on that: @S Marshall @Zlamma @Trumpetrep. UpTheOctave! • 8^va? 22:58, 25 May 2026 (UTC)

Sounds good to me. I totally understand User:S Marshall's point. Even if I do wonder if this is true, "you can't just copy/paste a worked example from a textbook without violating copyright", it's neither here nor there.

The verifiability of audio files became an issue with me when I was tackling egregiously out of bounds music theory articles where people were just saying whatever they wanted and uploading random files. It all blurred together into a highly suspicious melange that did more harm than good. It confuses editors and readers alike. In going line-by-line to verify facts, it was impossible not to wonder, "Hang on, is this audio file actually the 'pythagorean major third' it claims to be?" How does one even verify that? What are the rules for such demonstrations?

That's why I started asking the question all over the joint. I'm glad to read other editors' perspectives, and whatever moves us forward is welcome! Trumpetrep (talk) 00:28, 26 May 2026 (UTC)

I've found in such cases that declaring the image/recording/whatever to be a poor illustration is usually effective. You don't have to prove that it's technically wrong; you just have convince people that it's unclear. This is easier to do if you can provide a better option. It changes the discussion from "This is factually wrong and a policy violation!!!" to "It's a bit hard to see/hear the thing that matters in that version". See WP:PERTINENCE for some ideas. WhatamIdoing (talk) 20:46, 26 May 2026 (UTC)

I'd like to go back to these "examples like Saxophone Sonata (Creston), Short Symphony, etc.":

These are both FAs. Short Symphony#Structure and Saxophone Sonata (Creston)#Movements both use <score> to provide brief "quotations". They don't use a File:, which what we've been discussing here.

The <score> feature doesn't support ref tags, so there are no little blue clicky numbers associated with these scores. IMO this is correct because this is WP:PLOTCITE territory: Anyone who genuinely can't guess, without having it spoon-fed to them in a ref, that one way to double-check the score snippet would be to find a copy of the musical score is, to put it bluntly, too stupid to be allowed to use the internet without supervision. WhatamIdoing (talk) 20:04, 28 May 2026 (UTC)

I realised my original comment was mistaken, which is why I made this reply to S Marshall. UpTheOctave! • 8^va? 20:15, 28 May 2026 (UTC)

That said, I've had a thought about your point and still disagree. This is not WP:PLOTCITE territory, as the MOS guidance is is based on (WP:PLOTREF) explicitly notes that if the summary includes a direct quotation from the work, then Wikipedia:Verifiability requires it to have an inline citation, just like any other direct quotation. Unless you're claiming that musical quotations don't constitute direct quotations, I don't see how this applies. UpTheOctave! • 8^va? 01:03, 29 May 2026 (UTC)

Maybe just a quibble too, but in your response to my message, you didn't respond to my counter-example to your earlier claim that statements about how a file is generated, things that can't be proven by a citation, fall outside Verifiability. Using your definition of verifiability, would you really argue that providing an instruction to a deterministic file-generator does not "make sure anyone can check a fact or claim, i.e. without any specialist knowledge".

Perhaps you want to update your definition, but I wonder why not just adopt a notion that citation is just one way of helping verifiability (and sometimes an insufficient one, as I noted here by saying "even when there's a citation, because of the need to avoid copyright-infringement like described in this case, the demonstration can still be wrong and unverifiable"). Zlamma (talk) 01:09, 28 May 2026 (UTC)

Break

@UpTheOctave! @S Marshall , am I understanding correctly that your current stance is that there is nothing in WP:V to change due to the file descriptions somehow not being related to WP:V?

If so, I feel this conclusion is misguided and sidetracking. Commons may be a separate project, but that's a technicality. Wikipedia has to use Commons and there is no other option, and so there must be a way of looking at the content placed there just like it was placed in Wikipedia, and bringing non-Wikipedia uses is irrelevant.

Another way to put it is that, for a Wikipedia article to continue to use a file in a specific article, that file needs to have the same claim as Wikipedia. If an article on Marshall amplifiers or Fender Statocasters is giving an example of "Sound of a Fender Stratocaster being played through a Marshall amplifier", the file should contain the same claim. If its only description/metadata was "an instrument being played at 120 bpm", it would be very incorrect to use it in the article on electric guitar equipment, because the metadata would not discourage future wikipedians to improve the file with a version played on the 2060 first-and-last edition of the CosmoTheremin™, the only acceptable instrument of the era.

As such, I really don't see much distinction in where the discussions questioning the correctness of the demonstration file should take place and that one place makes it not subject to WP:V. And here, the Commons project is just as good a place, and maybe the more practical, than in each of the Wikipedia uses of the file.

If I misunderstood that this is your current stance, then apologies. Feel from to share thoughts. Zlamma (talk) 12:32, 27 May 2026 (UTC)

Not speaking for S Marshall, but my reading is that WP:V applies to each fact or claim in an article. Until it a file is used in an article, therefore, the file and its description are not subject to this requirement.

It could be argued that a file's inclusion is itself a claim that the file is what it purports to be, regardless of any textual claim, but this implies that every media file must be able to be traced back to a reliable secondary source. I considered this stance myself, but I now think it's an unreasonable reading, since Wikipedia:Citing sources states that a citation is not needed for descriptions such as alt text that are verifiable directly from the image itself, or for text that merely identifies a source. The alternative option is what I think S Marshall is getting at, that verifiability only comes into play when a textual claim is made – it's not the file that needs sourcing, as the file effectively is the source.

This doesn't mean I don't think there's an issue: I believe that incomplete, black-box-like, detail-lacking documentation is still a problem worth fixing, but is a separate matter that should be fixed through separate means (i.e. building off of Help:Creation and usage of media files or pursuing an audio-specific policy/guideline similar to Wikipedia:Image use policy). UpTheOctave! • 8^va? 18:40, 27 May 2026 (UTC)

UpTheOctave has understood me.

I feel that the next move is to draft the exact wording for the guidance. It needs to be clear, specific, and unambiguous. I suggest that the wording should be worked out in someone's userspace in the first instance.

I feel that the new guidance should be in a place that's specific to sound files. I rather like UpTheOctave's suggestion of a Wikipedia:Sound file guideline. We would likely link to that guidance from WP:V.

Then I feel we should ensure that people interested in and affected by this proposed guidance are invited to speak up; do this by notifying relevant talk pages. Off the top of my head, Wikimedia commons, WikiProject Media, WikiProject Music, WikiProject Professional sound production, and the Village Pump (Policy). There may be others that I haven't thought of.

The last step is to hold a RfC to confirm that the community accepts this guidance as a formal guideline.—S Marshall T/C 21:39, 27 May 2026 (UTC)

Re @UpTheOctave!'s:

The alternative option is what I think S Marshall is getting at, that verifiability only comes into play when a textual claim is made – it's not the file that needs sourcing, as the file effectively is the source.

I still see this as artificial. Files always have names/title which is textual. The also have metadata which is textual. So, all files make some textual claims. The Wikipedia article must be repeating the same claim as if there wasn't a match, this would requires resolving or else there'd be nothing stopping people from introducing a new version under the very same Commons resource identifier, which would start making the Wikipedia articles incorrect. If the claim of the file name/title/metadata is detected to be wrong, the discussion should naturally first take place in Commons to attempt to make the file match the claims, so that correctness of Wikipedia articles or any other incoming link does not require changing any of these resources.

Re @UpTheOctave!'s:

> It could be argued that a file's inclusion is itself a claim that the file is what it purports to be, regardless of any textual claim, but this implies that every media file must be able to be traced back to a reliable secondary source.

If by secondary source you mean a citation, then for the types of files that I raised this issue for, this is not required. Image use policy > Diagrams and other images already punches a hole in WP:V by saying "user-made images may be wholly original. In such cases, the image should primarily serve an educational purpose". WP:NOR > What is not original research strangely doesn't mention the possibility of having wholly original educational images, though it punches another hole, especially through the CALC rule.

Re @S Marshall's:

> I feel that the new guidance should be in a place that's specific to sound files

From the original message, to my message preceding yours by 2 days, with others agreeing, I have been making a case that this problem is not specific to audio files, nor even media files. It seem the talk of going this direction results from not talking into account the points.

To me, the fact that original-image punch-hole is not mentioned in the WP:V or WP:NOR, both of which pertain to rules for satisfying Verifiability, is a problem which brings danger that certain demonstrations will be deleted based on reading just WP:V. And there isn't anything special about images that doesn't apply to some audio files (like the example of sine waves demonstrating just intonation, or the phenomenon of beating). Both are just demonstration of an exact science fact/phenomenon. I even frankly don't see that a file that allows to output two sine waves is an audio file by essence. Distant future could bring support for files like the mentioned list of sines (unlikely that one, but some score-like format, who knows), which would make many of "audio file" aspects irrelevant, just like <score> is arguably at least equally valued for just being a way to write in the musical-staff script.

That's why, in my mind, if we create yet another place that punches the same hole, it's not addressing the real issue. While "likely linking to the guidance from WP:V" sounds like a light in the tunnel, but the "likely" seems ominous. I was really hoping that the overarching problems of WP:V will be tackled: it suggests that the only way to verify are citations on the one hand, and on the other there's is no encouragement to consider whether verifiability of correctness of demonstrations isn't needed in some cases. For these cases it should best give guidelines how to achieve verifiability of correctness where citations cannot provide it (open-source demonstrations, original files in the editor-software format, or prose giving the recipe for the creation). I feel that if the problem was addressed using non-medium specific vocabulary, it could be that there's no real need for guidelines for every type of file. Zlamma (talk) 00:06, 28 May 2026 (UTC)

Ok, that's a lot to discuss; a few pointers here, numbered by reply:

1. I think you misunderstand me. It's true that files make textual claims, but these claims are not by themselves in an article, so WP:V does not apply. A filename is just a label and, per WP:CS, a citation is not needed for descriptions [in image captions]. The logic of your next sentence doesn't quite follow either: WP:V isn't a tool for preventing file tampering on a separate project, and besides, files are easily reverted, and Commons:COM:OVERWRITE already grants protections against substantial changes. Furthermore, if you think the discussion should take place on Commons, where WP:V doesn't apply, that's almost shooting your argument in the foot.
2. I misspoke slightly. What I meant was a citation to a reliable source, as is the wording in WP:V (there's no need for it to be secondary). To return to my reductio ad absurdum: the crux of this argument is that if we say that including a file in an article is a claim, we have to apply verifiability somewhere (presumably to the file description and name). The consequence of this is in direct conflict with established norms: per WP:SPS, user-made file descriptions are not reliable, so no wholly user-made files could be used in articles since the files' creators aren't reliable sources. Regarding your hole-punching, as we've mentioned before, the image use policy is explicitly only for images, so does not apply in this instance. I'm also not sure what you're getting at with invoking WP:CALC, as it is irrelevant in this instance. A meaningful reflection of an unreliable source is still built on a rotten foundation. Due to this, I think such logic is unworkable.

In short, no-one is denying that there are problems here; it's just that modifying WP:V is the equivalent of using a nuclear bomb to kill an ant: wrong tool for the right job. The gaps you identify are valid concerns, but should be solved through guidance for documentation. Categorically, WP:V does not, and should not, apply to the files themselves. Therefore, I don't see the need to include any warnings, as I don't perceive the "overarching problems" with WP:V that you suggest. Even if verifiability did apply, deleting a file based on WP:V would not be valid reasoning under WP:DP: if it were to be misused in this way, any reasonably policy-literate editor would defend the file's inclusion. Including anything more than a tangential link or signpost to more relevant guidance risks WP:CREEP.

On your last point, for what it's worth, I really don't mind whether we pursue an "audio file guideline" or a "demonstrative file guideline" (although the latter would be much harder to produce) – I just think something should be done, and at this point we are letting perfect be the enemy of good. UpTheOctave! • 8^va? 02:28, 28 May 2026 (UTC)

Re "no-one is denying that there are problems here": Ok, I'll step up.

There may be verifiability problems on rare occasions with some media and image files.

Nevertheless they are uncommon and minor in most cases.

On the other hand, introducing demands that all media and image files be backed by reliable published sources would introduce severe content problems: it would mean we could have almost none of the non-textual content that we currently have. Most of our technical articles would be unable to have any technical illustrations, for instance, because we could not copy them from other sources (copyvio) and we could not create them ourselves (under new much stricter interpretions of what is original research.

This well-meaning but badly thought out proto-proposal needs to stop in its tracks before it causes actual damage. The cost is far out of proportion to the benefit. —David Eppstein (talk) 06:05, 28 May 2026 (UTC)

I think that the proposal on the table (with Zlamma dissenting) is to create guidance encouraging people to document how they made their files.—S Marshall T/C 07:50, 28 May 2026 (UTC)

@David Eppstein and Sławomir Biały, S Marshall gets me here. I'm actively opposed to modifying WP:V for this purpose as I think it's out of scope, and strongly prefer the creation of new specific documentation guidance for audio/demonstrative files. The unfortunately nebulous "problems" I referred to are the lack of documentation and how that manifests further up the editing chain. I completely agree with both of your points here. UpTheOctave! • 8^va? 14:05, 28 May 2026 (UTC)

We could put a note in Wikipedia:Image use policy#Required information that it's nice to include citations to reliable sources in the description on File: pages. WhatamIdoing (talk) 19:48, 28 May 2026 (UTC)

That would be a good step forward, but I don't think it fully addresses the underlying issue here. UpTheOctave! • 8^va? 20:02, 28 May 2026 (UTC)

Can you name "the underlying issue" in 20 words or less?

Specifically, it would be helpful if you could decide whether "the underlying issue" is

• missing documentation ("This audio recording is correct, but the File: page should cite a source"),
• inaccurate descriptions ("This audio recording is not correct, no matter what that cited source says!"), or
• the application of non-existent rules ("I get to blank all the audio recordings until you say Mother, May I? by adding a citation").

WhatamIdoing (talk) 20:08, 28 May 2026 (UTC)

Yes, of course. My succinct summary would be "Insufficient documentation gives no indication of the correctness of a user-made demonstrative file, which lets potential mistakes go unnoticed". UpTheOctave! • 8^va? 20:11, 28 May 2026 (UTC)

Whereas if I put incorrect or misleading documentation in the description on the File: page, that mistake somehow will get noticed?

Imagine that I upload two similar audio samples, but when I do, I accidentally copy/paste my prepared descriptions to the wrong files. Imagine that the prepared descriptions are everything you could wish: I tell you which version of Finale I used, what settings I used, which instruments, exactly how the two files differ, etc. The only problem is the mistake of putting the wrong description on the wrong file, so the one with "piano and violin" ends up on the "piano and viola" file, and vice versa. Do you think that my extremely sufficient documentation will result in the mistake being noticed? WhatamIdoing (talk) 20:24, 28 May 2026 (UTC)

First off, this is conflating two separate failure modes: I did not argue that detailed documentation will always be correct, I only said that with vague documentation, we cannot even begin to evaluate correctness. Perfection is not the threshold for usefulness. Let's try applying this logic to another separate Wikipedia process: citations can be wrong, as an editor can cite the wrong paper or a page range, the source itself can misstate a fact, or a citation could be mistakenly attached to the wrong claim. All are real problems, but it doesn't follow that they are reasons to avoid requiring citations on the whole.

For the sake of argument, I'll answer this particular question at face value. Yes, I do think that the act of including detailed documentation will make the mistake more likely to be noticed, precisely because the documentation is there. Any editor with basic musical training should be able to differentiate the sounds of a viola and a violin, at which point they will notice the mistake. Moreover, in this case, the fix is trivial (a round-robin file move), as the documentation even provides the correct state should be. When documentation is not included or lacking, when someone notices something seems off, they have no reference point to investigate, so the mistake is harder to fix as well. UpTheOctave! • 8^va? 21:05, 28 May 2026 (UTC)

When the documentation is not included, and someone notices that something seems 'off', then what's 'off' is probably in the Wikipedia article. It's probably not in 'the documentation', which doesn't exist (or is unhelpful). WhatamIdoing (talk) 03:47, 29 May 2026 (UTC)

Of course not, but my argument is that the documentation is what allows us to find the exact error, not that including it is what detects that something is off. UpTheOctave! • 8^va? 07:14, 29 May 2026 (UTC)

Re: proposal on the table (with Zlamma dissenting) is to create guidance encouraging people to document how they made their files

Just to clarify, my dissension is NOT opposition to creating the guidance encouraging to document. It is merely:

• To (at least, or additionally) create guidance curbing use of WP:V to ask for citations on legal original demonstrations (e.g. 1), or pitching for their removal (e.g. 1, 2).
• To make sure such guidance is discoverable from WP:V (e.g. 'Not to be confused with...' on top, or some subsection - other ideas welcome). Without it, the outcome doesn't address the reason the topic was created - multiple users involved in the conversations were fooled by WP:V's pretense to apply to all Wikipedia material, so they stopped looking.

Zlamma (talk) 23:55, 29 May 2026 (UTC)

I agree with Eppstein, but after looking at all this I think "stop in its tracks" is too mild. Bury it under concrete would be more suitable. Are you people not tired of this? I am. No further comments from me. Yesterday, all my dreams... (talk) 00:59, 29 May 2026 (UTC)

Re:

> introducing demands that all media and image files be backed by reliable published sources would introduce severe content problems

Does even one person here actually profess the above demand this under this topic? If anything, this discussion seems to be about making it more explicit that there's no such requirement. I hoped to have just clarified this in last paragraph of this message of mine. If your impression stands, if you could please point at a statement/response where anyone actually pitches for this, maybe it would be helpful. Zlamma (talk) 01:12, 29 May 2026 (UTC)

I think there may be a legitimate documentation problem for some media, but any change to policies and guidelines should be very narrow (if any). I worry about possible guidance being misinterpreted, either as a suggestion that code should be released for technical images (not always desirable or even possible), as a mandate to remove technical images that do not tick some formal WP:V-checkbox or delist articles for the same reason. So, while there is probably a real issue here, I would recommend against including it in policy, writing an essay instead, and an essay that is very narrowly tailored to the specific problems identified. Sławomir Biały (talk) 08:34, 28 May 2026 (UTC)

^ This. WhatamIdoing (talk) 19:45, 28 May 2026 (UTC)

@UpTheOctave! re:

Re:

> A filename is just a label and, per WP:CS, a citation is not needed for descriptions [in image captions]

If you included the rest of the sentence, "... that are verifiable directly from the image itself, or for text that merely identifies a source (e.g., the caption "Belshazzar's Feast (1635)" for File:Rembrandt-Belsazar.jpg)" you would see that this is an exemption only for claims in the filename/description that are self evident, or claims identifying the source (which are verifiable only by just comparing to the trusted representation of the source). Doesn't this explicit enumeration hint that there may be other claims made and the policy does not exempt these claims from providing means of verifiability?

Re:

> "It's true that files make textual claims, but these claims are not by themselves in an article, so WP:V does not apply."

> "WP:V isn't a tool for preventing file tampering on a separate project ...

> "...if you think the discussion should take place on Commons, where WP:V doesn't apply, that's almost shooting your argument in the foot."

The opening and the further sentences hint to me a misunderstanding rabbit hole. Regarding the first sentence, I actually did talk of matching arguments between the file and the article. To extra-clarify, a file may make many claims, and the article may only have few of those - it should only make claims that are pertinent to the article, e.g. when Fender article links a "120bpm piece played on a fender" file. It will just repeat the "played on fender" claim.

The second sentence must be another misunderstanding as I don't see myself claiming such thing. All I was trying to say was that in the event that the underlying file payload turns out not to fulfill some claim that both the Wiki article and the commons file are making, then the community will likely try the easiest way to fix this - fix the payload - because the Commons claim has already caused many wikipedia articles to refer to the file.

Finally, engaging the last statement, note that the discussions I am referring to are just discussions pointing at any identified incorrectness of payloads, not discussions about policies. And even then, I don't mind them taking place on Wikipedia articles. I was just pointing at the fact that since the Commons files will have the matching claim, it will be pretty practical to raise the issue there.

Re:

> What I meant was a citation to a reliable source, as is the wording in WP:V

Maybe the key to a large portion of our misunderstanding is not noticing whether the other party means " 'verifiability as in the wording in WP:V', there being a published reliable source" or "verifiability that the content entered into Wikipedia/commons is correct".

I was hoping to be understood which one I mean, because in creating this topic I thought I am pointing at the distinction. I was trying trying to point at shortcomings of WP:V wording in solving the latter problem, yet it appearing to many users as pertaining to that problem too (even in with its over-arching name 'verifiability', but also further by making sweeping prescriptions for all of Wikipedia content like "All material in Wikipedia articles must be attributable to a reliable published source" and similar), which is why users refer to it when trying to address correctness problems. The discussion had more than two participants, no users having the knowledge that this is not the applicable policy/guideline. The result is requiring citations (still there as we speak) as supposedly the only way to prove correctness. Again no one knowing this may be objected - users acquiesced or went quiet because they couldn't discover the correct policy/guideline. That's the reason I was hoping for an improvement in WP:V. WP:V, with its 'all encompassing' prescriptions makes users think that there is nothing more to look for, since all is already addressed. For some reason this is being strongly objected here, while I am giving proofs that deterioration and waste is being generated in Wikipedia due to this situation, and another such incident, with lengthy discussions may happen yet again, as users' state of knowledge of policy generates this pattern.

Apologies I have to end here this late today, but I hope these thoughts will be helpful for alignment. Zlamma (talk) 01:02, 29 May 2026 (UTC)

Zlamma, could you please try and be slightly less verbose? I will try and answer this, but I would really appreciate it if you could cut down on the amount to read through.

• Doesn't this explicit enumeration hint that there may be other claims made and the policy does not exempt these claims from providing means of verifiability? – yes, but these are not claims being made by the file. This refers to claims made in captions that are not directly verifiable from the file, as the sentence before makes clear.
• Regarding the first sentence, I actually did talk of matching arguments between the file and the article... – I do understand, but I still see this is inherently a problem with documentation: without adequate documentation, we wouldn't know when if a file was being played at 120bpm anyway. So yes, a file should only be used to support what it claims to represent, but that's fixed with just common sense and a bit of MOS:PERTINENCE, not a change to WP:V.
• The second sentence must be another misunderstanding as I don't see myself claiming such thing... and note that the discussions I am referring to are just discussions pointing at any identified incorrectness of payloads, not discussions about policies – granted; I only read them this way due to your repeated references to WP:V in your other comments.
• Maybe the key to a large portion of our misunderstanding is not noticing whether the other party means – yes, agreed. I have consistently understood any references to verifiability to be a reference to WP:V. Thank you for the clarification.

My issue with your position is that changing the wording to argue for "correctness" is a fundamental redefinition of what verifiability means that just isn't needed here. As a core content policy, you're right that it applies to all article content. Therefore, if we were to redefine it, the provision for "correctness" would apply to all prose, which is antithetical to WP:5P1 – we are an encyclopedia, so we are concerned with what other sources say, not universal truth, which the rather nebulous term "correctness" implies. I'm sympathetic to misunderstanding a policy, but that is not a reason to change it either. We should not be making a fundamental change to WP:V when the consensus here is that it doesn't actually conflict with demonstrative files. What we should be doing is trying to produce guidance for this particular edge case, which is exactly what guidelines are for. UpTheOctave! • 8^va? 01:57, 29 May 2026 (UTC)

• re: these are not claims being made by the file..." I am lost tracing statements to suggestions in this file-related part of the thread, so I must leave it. (given the misunderstanding about the citations vs correctness, I feel the discussion is but a child of this misunderstanding, or possibly another one, about whether by 'file', a party means just a payload, or also every metadata in Commons).
• re your position ... [of] changing the wording to argue for "correctness" ... if we were to redefine it .... Perhaps another misunderstanding. I don't intend to redefine verifiability or argue for correctness to take its place. I just argue for whatever action that solves the actual problem that made me create this topic, which is that users generate waste/deterioration and back it with WP:V. As in the topic opening message, I try to argue for an improvement to WP:V, the article, not 'verifiability' the principle. Some ideas:
  • Rather than trying to redefine things, I earlier suggested just having a section 'What is NOT an infringement of Verifiability'
  • or perhaps a '(Not to be confused with ...)' on top, or really anywhere, to keep the 'verifiability' concept untouched and point to the right concept. The problem I see here is that there doesn't even seem to be any WP page, any WP: concept that deals solely with correctness that isn't achieved via citations. I went through WP:Reliability and all of its resources/templates, despite having over-arching names relating to fact and truth, ultimately point to quotations as the solution. No wonder it's so hard to discuss this grey-area that original-demonstrations are.
  • Any other ideas welcome

Zlamma (talk) 23:26, 29 May 2026 (UTC)

Taken with your comment at 01:12, I think I understand your position fully now – thank you for clarifying where you stand on documentation guidance as well. Your idea of a pointer does have merit, and as S Marshall suggested, any guidance produced could be linked from WP:V to help direct towards the appropriate edge case. I think we should focus on that first task just now, as there's no point discussing the exact method of presenting the guidance when it doesn't even exist yet. UpTheOctave! • 8^va? 00:06, 30 May 2026 (UTC)

Image files contain information, which are statements. This can be content in a user generated graphic, or a statement about what the image is, amongst other things. IMO practical necessity dictates that wp:ver (and wp:NOR) can't be fully applied to these. An in-between route (and possibly what is currently common) is that when the veracity of the "statement" is challenged then it will need to at least buttressed and if not, be removed. But applying the full weight of wp:ver would create a giant mess. For example, that this can be triggered by a mere claim that it is unsourced (without questioning the veracity), or that buttressing the statement requires full wp:ver grade sourcing. For example, if editor John Smith took a photo of Elvis an put it in article, captioned as such. If someone doubts that it's a photo of Elvis, Smith would probably have to do something to buttress the claim. And if not, the editors could decide to leave it out. But it would create a giant mess to start requiring a statement by a WP:RS that says that John Smith took that picture of Elvis or that the photo is actually of Elvis.North8000 (talk) 15:40, 28 May 2026 (UTC)

I agree. Also, it's our long-standing rule that images don't have to "be" the thing; they only have to "look like" the thing they're illustrating. This allows us to have images without having debates about Is It Cake? If it looks like a rubber ducky, then that's good enough for any article that needs an illustration of a rubber ducky. We don't need a third-party certification that it's not a trompe-l'œil cupcake that's been decorated to look like a rubber ducky. WhatamIdoing (talk) 19:52, 28 May 2026 (UTC)

Agreed, the question isn't verification but whether an image is appropriate for an article. If it looks like a duck then it could be used it Duck. If someone challenges it because they thinks it's actual a chicken the question is whether it's appropriate to be used in the article, not whether there's a source confirming if it's a duck or chicken. -- LCU ActivelyDisinterested «@» °∆t° 20:29, 28 May 2026 (UTC)

Focused page for this initiative

Seems like the initiative by other users' to start a draft from over 3 week ago found no one with the time to do it, so, being the user who raised this whole topic, I'm following on my suggestion to start a draft, after a request from @Trumpetrep and @UpTheOctave!:

Please see the page User:Zlamma/Rfc:WP_docs_improvement_re_demonstrations* and help improve the outcomes there.
Move conversations to its Talk page as separate subjects, which will allow them to be more focused than this thread.

*) Side-note about that page's approach: After trying really hard, I struggle to internalize as my own the idea to add new format-specific page, so all I can do is present multiple positions I found in this discussion (including ones I object to) and ask other users who see merit in them, to present what text they are looking to add and address my concerns 'Contra' to the idea. My own suggestions are presented merely as alternatives. I also caught some good suggestions from others in this discussion and made them easier to find under the appropriate sections of my draft. This hopefully will also save time by bringing structure and focus to the discussions. Zlamma (talk) 09:20, 18 June 2026 (UTC)

Verifying domains of shadow libraries

Shadow libraries, such as Anna's Archive, Library Genesis, Sci-Hub, and Z-Library all face the same problem: they have to regularly pop up on new domains when their previous ones were taken down. Editors are disagreeing on how these domains should be cited (see the {{cn}} tags and talk page discussions). Some cited the site itself, but there has been a history of scam sites appearing. Some support citing non-reliable sites like SLUM, or sites of questionable reliability like Verts-Luisants. Some recommended citing reliable sources (such as TorrentFreak), but they may be slow to report on domain changes, or they might not report on them at all. Considering these, some have proposed not including URLs at all.

It must be noted that some of these sites regularly refer people to their Wikipedia articles for the currently active domains (LibGen AA), while others list the currently active domains on their subreddits' wiki (Sci-Hub ZLib). Kovcszaln6 (talk) 17:29, 5 June 2026 (UTC)

@Kovcszaln6, is this discussion about verifying article content (e.g., using citation templates and ref tags), or is this about whether to put the domain name in the infobox (which is a question for WP:ELOFFICIAL)? WhatamIdoing (talk) 17:58, 5 June 2026 (UTC)

I believe there is consensus not to exclude the domains on the grounds of copyright violation. The main question is what source should we cite for the domains (or is it better to not include them at all). Kovcszaln6 (talk) 18:11, 5 June 2026 (UTC)

That doesn't answer my question. My question is whether you're discussing:

1. whether there should be {{infobox website |website=https://example.com}} at the top of the article, or
2. whether the middle of the article should have a little blue number that looks like and which, if you click on it, should say something like "Anna's Archive" in the ==References== section, or
3. both.

WhatamIdoing (talk) 18:17, 5 June 2026 (UTC)

The question is should the infobox include the domains, and if yes, should there be a citation verifying them, and if yes, what source should we cite. Kovcszaln6 (talk) 18:45, 5 June 2026 (UTC)

In general we do not cite sources for external links, whether in an external links section or in the home page line of an infobox. —David Eppstein (talk) 18:18, 5 June 2026 (UTC)

So are the citation needed tags in the infoboxes invalid or would this be an exception where sourcing is expected for external links? Kovcszaln6 (talk) 18:43, 5 June 2026 (UTC)

Citations are never expected for Wikipedia:External links. This might be a case for which doing the unexpected is warranted. WhatamIdoing (talk) 19:06, 5 June 2026 (UTC)

Doing the unexpected might be warranted, and it doesn't seem costly. I think that an endnote, which explains how and why the domain names are not stable, would be more informative than an ordinary plain blue clicky number. Stepwise Continuous Dysfunction (talk) 22:45, 5 June 2026 (UTC)

ABOUTSELF - discussion at WP:RS

I posted a question regarding self-published sources at the talk page for WP:RS here: . Since this page uses the exact same wording at WP:ABOUTSELF, this is relevant here as well. I suggest to keep the discussion there, however. NisJørgensen (talk) 13:22, 14 June 2026 (UTC)

All quotations or "direct quotations"

WP:V had "all quotations" but in 2024 WhatamIdoing changed to "direct quotations" in this edit. The edit summary says "per talk" but I don't see in the talk around that time (archives 79-81) a thread saying that was consensus. Does everybody agree that everybody agreed? I think it's bad, if I claim that WP:V's This page in a nutshell says quotations, that's substantially a claim that WP:V's "This page in a nutshell" says "quotations". Peter Gulutzan (talk) 17:45, 15 June 2026 (UTC)

Can you give an example of "a quotation" that is not "a direct quotation" that you think requires an inline citation exclusively because it's a quotation (and which would not require an inline citation if it were not a quotation)? WhatamIdoing (talk) 18:01, 15 June 2026 (UTC)

Can you tell me where the thread is? If it's already been discussed and agreed as you claimed, then there's no need for argument here. Peter Gulutzan (talk) 18:05, 15 June 2026 (UTC)

The edit summary says "Listing the four types per talk page", and the edit is much bigger than just adding the word "direct". I therefore assume that the discussion is not specifically or exclusively about whether to add the word direct.

I suggest that instead of seeking a past discussion, we instead discuss whether it's what we want now. Therefore: Can you think of a situation in which changing "quotations" to (the IMO clearer) "direct quotations" would exclude anything that you want to include? If not, then the change is harmless. WhatamIdoing (talk) 18:19, 15 June 2026 (UTC)

I'm not sure it matters either way, but since you ask :). This is the example from our quotations article: He said: 'I am leaving now' (direct) versus 'He said (that) he was leaving immediately' (indirect). So why cite only the first and not the second? Alanscottwalker (talk) 19:24, 15 June 2026 (UTC)

If you can't back up your claim that your policy change was per talk, then I don't see why you act as if the onus is on me. But, as Churchill said, "we will never surrender". I trust we accept that a direct quotation is exact word-for-word, verbatim, and thefore inside quotes. Thus this doesn't apply to paraphrasing. Here's an example then, from Durham special counsel report: Law professor Jonathan Turley[148] and Republican senator Marsha Blackburn asserted the Russian collusion theory was a hoax invented by Clinton and her operatives. Neither quite said that, but since nothing's in quotes there's no need for a directly-supporting cite (And no, [148] doesn't quite say they did.) Sure, that's objectionable for other reasons, but it was formerly a violation of this part of WP:V because it was an indirect quotation, but due to your change it's not. Now, I'm throwing this open to the audience. If you ddon't have consensus, and if nobody else recalls that this was agreed when you said it was, that's enough to erase your bad change. (By the way, my quote of Churchill shouldn't have been direct, I wonder if everyone will catch that.) Peter Gulutzan (talk) 20:11, 15 June 2026 (UTC)

Isn't that the kind of content that would currently be required to have a citation due to being Wikipedia:Likely to be challenged if it were left uncited? WhatamIdoing (talk) 23:07, 15 June 2026 (UTC)

That's no answer. I'll come back in a week and see who supports what. Peter Gulutzan (talk) 01:58, 16 June 2026 (UTC)

I think, what we need is the case for excluding indirect quotations from the cite requirement. I can't see any reason to do that. So can someone explain, why? Alanscottwalker (talk) 14:57, 16 June 2026 (UTC)

Because it's either unnecessary to have a citation at all, or it's redundant with WP:LIKELY. We want direct quotations to be cited so we have a way to show that the exact words are correct. But an indirect quotation doesn't necessarily use the exact wording.

"Lao Tzu said that the journey of a thousand miles begins with one step" contains an indirect quotation. "Polonius' advice to his son about personal finance, especially his exhortation to neither a borrower nor a lender be, is famous" also does. Ditto for "Richard Nixon said that he was not a crook".

Do we need an inline citation in such cases for the quotation (i.e., because we want to be able to show that the exact words are correct), or do we mainly need an inline citation for some other, non-exact-wording reason (e.g., to show that these things were said by Lao Tzu/Polonius/Nixon instead of some other person or character)? I think it's the latter, and that's why I think that indirect quotations should be handled under LIKELY instead of as quotations. WhatamIdoing (talk) 22:30, 16 June 2026 (UTC)

So are indirect quotes without citation a form of plagiarism? You are too close paraphrasing? Or for either quotation types, we are literally or substantively putting words/ideas in someone else's mouth. And what we want to do with citation of either type is show we are representing the other person/people correctly (he, she, they said.) They literally or substantively said that -- you reader, have the proof in the citation. Alanscottwalker (talk) 10:32, 17 June 2026 (UTC)

I would agree that both direct quotations and paraphrases (indirect quotation) require a citation. We are just nit-picking over why the citation is required. Blueboar (talk) 11:26, 17 June 2026 (UTC)

Thanks. Yes. But I asked for the "why" we went from all quotes to only direct quotes because I did not understand why it was done. Why we would we exclude indirect quotes. Now that a "why" argument has been presented. I find it unpersuasive. And I agree we need cites for both direct and indirect, all quotes. Alanscottwalker (talk) 14:13, 22 June 2026 (UTC)

It's not plagiarism, because the information has WP:INTEXT attribution.

Attributing ideas to someone falls under WP:LIKELY. If you say "I saw Karp in the elevator, and he said it was probably np-complete", then that requires an inline citation because of the substance, regardless of whether Karp's actual words were "That's probably np-complete" (in which case, it's a fully attributed, non-plagiarized indirect quotation) or if you ask him "Do you think this is probably np-complete?" and he replied "Yeah, intuitively I think so, but intuition isn't a proof" (which is fully attributed and non-plagiarized, but not a quotation of any kind). WhatamIdoing (talk) 20:41, 17 June 2026 (UTC)

I knew you were going to say intext. So, we need a citation to show we have done that correctly. And if something is not a quote at all, it's not covered, by "quotations". Alanscottwalker (talk) 14:17, 22 June 2026 (UTC)

Imagine this scenario:

• The words that Karp actually said: "Yeah, intuitively I think so, but intuition isn't a proof".
• What the reliable source said: "I saw Karp in the elevator, and he said it was probably np-complete".
• What the Wikipedia article says: "In 2022, Karp said that it was probably np-complete".

My questions for you:

1. Is this required to have an inline citation because it is a quotation (as opposed to being required for some other reason, such as LIKELY)?
2. How would an editor who hasn't seen the source be able to tell if it's a quotation?

WhatamIdoing (talk) 16:32, 22 June 2026 (UTC)

I would have put in something like, 'According to X, Karp said, . . . (and yes, with cite). Alanscottwalker (talk) 17:04, 22 June 2026 (UTC)

You did not answer my questions. WhatamIdoing (talk) 20:33, 22 June 2026 (UTC)

I thought I did answer the questions, I would change the article to conform to source and I would cite my change as it is a quote.

But let me add, where I think now, you appear to have some rigid thinking: there is no reason to want to, or think, each of the four reasons to cite are always going to be mutually exclusive of each other in every circumstance. You only need one reason, but if you have say, 2 reasons for putting in the cite, that's fine. Alanscottwalker (talk) 21:08, 22 June 2026 (UTC)

I think that overlapping reasons happen (e.g., a surprising or dubious quotation about a BLP is at least three of them).

You have not answered my question about whether "In 2022, Karp said that it was probably np-complete" constitutes a quotation. You have said that a different sentence does, but you have not said whether this sentence constitutes a quotation.

This matters, because if "In 2022, Karp said that it was probably np-complete" is an actual, bona fide quotation, then so is "The studio announced plans to release the film on DVD" and "A 2022 review found that cancer rates were rising", and I think that if we tell editors that when we Wikipedia:Use our own words to report facts, we're actually quoting sources, they will be very confused. WhatamIdoing (talk) 01:07, 23 June 2026 (UTC)

Any time, you in substance say someone said, it is presented as a quotation. In this case, an indirect quotation. I would oppose your proposed quotation of Karp because it is not accurate to the source. From the source we only know what X thinks Karp said, so it should be presented in substance as, X said [that] Karp said.

Can editors be confused about too close paraphrasing, yes, but that goes beyond quotations, and it is analyzed under the rules of plagiarism, which have some give for common expressions. The indirect quotation (He, she, they said), like "the direct quotation", obviates any concern of plagiarism because you are not presenting them as your thoughts, you are fully presenting them as someone else's. Alanscottwalker (talk) 10:45, 23 June 2026 (UTC)

I don't agree.

First, we don't take a reliable source and say "Oh, well, that's only what the journalist thought the person said. I mean, Jo Journalist asked the CEO whether it was true, and the CEO said 'Yes', but maybe they misheard, and the CEO actually recited a poem or something." If it's a reliable source, we have to accept what the source said. And if the source said "I saw Karp in the elevator, and he said it was probably np-complete", then "Karp said that it was probably np-complete" is what we should be writing. We should not be writing "Jo Journalist thinks that Karp said that it was probably np-complete". WP:INTEXT warns against the "ways in-text attribution can mislead", and claiming that it's only what the author "said" that Karp said is a way of casting doubt on whether Karp said it. This is not, and IMO should not be, our standard. Imagine if someone tried that with a politician who suffers from chronic foot-in-mouth disease.

Second, I wonder if you are interpreting quotation as encompassing any type of attribution. Quotation is about using the same words (e.g., "When asked whether it was np-complete, Karp said 'Yeah, intuitively I think so'."). Attribution is saying where the facts came from (e.g., "Jo Journalist said that Karp once claimed that it was probably np-complete"). These are different things, but IMO this policy should require both of them to have an inline citation. WhatamIdoing (talk) 19:17, 23 June 2026 (UTC)

I am certain you are wrong. (It's your proposed edit that decided it's important that Karp said it). So, we can stop now. We have both had our say. The edit in question, you made to this policy has been undone and that's rather all that matters now. Alanscottwalker (talk) 20:51, 23 June 2026 (UTC)

I've delayed reverting but observe that nobody has explicitly supported WhatamIdoing's policy change or remembered supporting it before.

Re Lao Tzu and Polonius and Richard Nixon: the Laozi article doesn't have the quote (which is okay since it's disputed that he wrote the Tao Te Ching and that that translation of 千里之行始于足下 is right), the Polonius and Richard Nixon articles contain direct quotations (in the latter it's "Well, I'm not a crook"), so WhatamIdoing's examples, unlike mine, aren't easily shown to represent something real in Wikipedia.

Re WP:LIKELY: the problem is WhatamIdoing rewrote a policy to say "direct quoation" and her solution is to instead use an essay which she wrote to say "direct quotation", I said that's not an answer and don't see why anyone else would think it is.

There are 3 guidelines which say that this bit of WP:V says "all quotations" (emphasis mine)

WP:RS "The policy on sourcing is Wikipedia:Verifiability, which requires inline citations for any material challenged or likely to be challenged, and for ALL QUOTATIONS."

WP:CITE "Wikipedia's verifiability policy requires inline citations for any material challenged or likely to be challenged, and for ALL QUOTATIONS, anywhere in article space."

MOS:LEAD "The verifiability policy states that ALL QUOTATIONS, and any material whose verifiability has been challenged or is likely to be challenged, must include an inline citation to a reliable source that directly supports it."

... those guidelines are now false due to WhatamIdoing's change.

There is one policy which says "all auotations" (emphasis mine)

WP:BLP says "ALL QUOTATIONS and any material challenged or likely to be challenged must be supported by an inline citation to a reliable, published source."

... this policy maybe now contradicts WP:V due to WhatamIdoing's change, as it doesn't clearly say that it only appplies to quotations inside BLP articles.

Alanscottwalker, Blueboar: are you supporting or opposing or neutral about WhatamiDoing's change?

Peter Gulutzan (talk) 13:46, 22 June 2026 (UTC)

I oppose WAID's change, we don't need arguments about whether a quote is direct or indirect. Will that leave arguments about whether something is a quote at all? Perhaps, but then WAID says it falls under likely, if you are having that argument. Either way, let's get rid of the direct or indirect arguments and just go back to all quotes. Blueboar also says "I would agree that both direct quotations and paraphrases (indirect quotation) require a citation." Alanscottwalker (talk) 14:00, 22 June 2026 (UTC)

It does seem pointlessly complicated to impose different expectations on different quotes. CMD (talk) 16:04, 22 June 2026 (UTC)

Thanks for clarity. I changed with edit summary = Undid Revision as of 16:57, 21 May 2024 by WhatamIdoing. See talk page = All quotations or "direct quotations". Peter Gulutzan (talk) 16:14, 22 June 2026 (UTC)

Peter, do you think you could reinstate all the other changes, instead of reverting everything in that diff? It'd have been very easy to take out the one word "direct" if you wanted to have the original word, or even to put in the new word "all" if you wanted to emphasize it. You did not need to revert everything, including formatting changes. WhatamIdoing (talk) 16:28, 22 June 2026 (UTC)

Yes, I was not talking about any other change than that one. Alanscottwalker (talk) 17:00, 22 June 2026 (UTC)

The word "all" is not new. I was thinking that the paraphrasing in the guidelines becomes confusing if the WP:V statement isn't exactly what it was before, so I didn't reinstate. Certainly if there is consens that the formatting changes improve, I'll not think it's a big deal. The non-formatting matters caused by a full revert, I think, were: elimination of a wikilink to WP:CHALLENGE, removal of a mention of your WP:LIKELY essay, different emphasis about WP:BLP. So let's put it this way: should the revert have been of all the changes (my preference), or only a change to "all" rather than "direct" (perhaps your preference and perhaps Alanscottwalker's preference)? Peter Gulutzan (talk) 17:30, 22 June 2026 (UTC)

I very strongly believe that the lead of this policy should say that "contentious material about living people" is required to have an inline citation. It no longer says that. WhatamIdoing (talk) 19:47, 22 June 2026 (UTC)

It says "Please immediately remove contentious material about living people that is unsourced or poorly sourced." I noted that's a different emphasis about BLP. You strongly want to say "lacks an inline citation" instead, or as well? Peter Gulutzan (talk) 20:03, 22 June 2026 (UTC)

That's a true statement, to the extent that "different" means "less". WP:V should have a positive statement that says such an inline citation is required.

This policy currently says many things about BLPs:

• Please immediately remove contentious material about living people that is unsourced or poorly sourced.
• Do not leave unsourced or poorly sourced material in an article if it might damage the reputation of living people or existing groups, and do not move it to the talk page, per Wikipedia:Biographies of living persons policy (common shortcut: WP:BLP). The latter policy also applies to some groups.
• Be especially careful when sourcing content related to living people or medicine.
• Never use a self-published source as a third-party source about living people, even if the author is an expert, well-known professional researcher, or writer.
• Editors should not rely upon machine translations of non-English sources in contentious articles or biographies of living people.
• Take special care with contentious material about living and recently deceased people.

but nowhere does it give the basic instruction that contentious matter about BLPs requires an inline citation. WhatamIdoing (talk) 20:32, 22 June 2026 (UTC)

About "the paraphrasing in the guidelines becomes confusing if the WP:V statement isn't exactly what it was before": WP:RS also says "In the event of a contradiction between this guideline and our policies regarding sourcing and attribution, the policies take priority". In other words, if the difference between WP:V and WP:RS bothered you, you should have changed WP:RS to match WP:V. WhatamIdoing (talk) 20:28, 22 June 2026 (UTC)

You caused a difference, I removed it, you're (if I understand you) thinking it should go back in and then I should change the guidelines to match your preference. Well, that's one of the reasons I'd prefer to keep your changes reverted. I think the sole question at this stage should be, as I suggested earlier: should the revert have been of all the changes or only a change to "all" rather than "direct". Peter Gulutzan (talk) 22:09, 22 June 2026 (UTC)

I think a wikilink to direct quotation, as in the version of WP:V that existed before this discussion, will be ignored by many readers, because "Quotation" is merely a Wikipedia article and carries no authority. If we look at The Chicago Manual of Style 18th ed, § 12.1 to 12.5, we see it discussing quotations and paraphrases. Direct quotations are only mentioned in passing, and not distinguished from quotations. So I think saying "all quotations" in the policy will not reliably communicate to readers that paraphrases are a form of quotation, and will not put readers on notice that paraphrases need inline citations.

Chicago § 12.3 states "Whether quoting, paraphrasing, or using other's words or ideas to advance their own arguments, authors should give explicit credit to the source of those words or ideas." This does not say that the credit must take the form of an inline citation. I don't think it's clear that Wikipedia requires an inline citation for a paraphrase, although some form of credit is required. Jc3s5h (talk) 21:40, 22 June 2026 (UTC)

Back to square one, eh? You support WhatamIdoing's change? Peter Gulutzan (talk) 22:09, 22 June 2026 (UTC)

I support making a bullet list. I don't support the phrase "direct quotation". After writing my comment, I noticed this in the "Copyright and plagiarism" section:

Do not plagiarize or breach copyright when using sources. Summarize source material in your own words as much as possible; when quoting or closely paraphrasing a source, use an inline citation, and in-text attribution where appropriate.

So I suggest changing "direct quotation" to "quotation or close paraphrase". Jc3s5h (talk) 22:24, 22 June 2026 (UTC)

You're not talking about the matter that the thread started with i.e. WhatamIdoing's policy change, or the thing that we were ending with i.e. whether I reverted too much. I'm not merely against the suggestion, I'm against talking about it here. Why not let this thread end, then you can start a new one, which I predict will fail?

I'm switching to non-reply mode for a week. Peter Gulutzan (talk) 01:08, 23 June 2026 (UTC)

If we can't agree on what counts as a quotation, how will editors know what we want cited?

@Jc3s5h, I'd like you to look at the example I outline above. Is that either a "quotation or close paraphrase" in your opinion? WhatamIdoing (talk) 01:11, 23 June 2026 (UTC)

I would call the hypothetical Wikipedia passage a close paraphrase of the reliable source. Jc3s5h (talk) 03:35, 23 June 2026 (UTC)

But not, I assume, a close paraphrase of Karp. WhatamIdoing (talk) 06:17, 23 June 2026 (UTC)

It's a close paraphrase of X who paraphrased Karp. Alanscottwalker (talk) 11:01, 23 June 2026 (UTC)

It might be a close paraphrase of the reliable source (though it wouldn't violate our Wikipedia:Close paraphrasing rules), but it's not a close paraphrase of Karp, so it's not a quotation of Karp. WhatamIdoing (talk) 19:19, 23 June 2026 (UTC)

I agree, WhatamIdoing, that the hypothetical Wikipedia passage would not be a close paraphrase of Karp. It would be a paraphrase or summary. Jc3s5h (talk) 12:24, 24 June 2026 (UTC)

WAID said its not a paraphrase of Karp, I said it was; although I am more than willing to move on. Alanscottwalker (talk) 12:41, 24 June 2026 (UTC)

I would not add "too close paraphrasing" there. In a proper indirect quotation, too close paraphrasing is fine because you tell the reader that is what you are trying to do. Too close paraphrasing is not always otherwise, fine. Alanscottwalker (talk) 11:54, 23 June 2026 (UTC)

Proposal for a new WikiProject Intellectual Diversity

Please see proposal at Wikipedia talk:WikiProject Council#Proposing a new WikiProject Intellectual Diversity. TarnishedPath^talk 06:29, 19 June 2026 (UTC)