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)

SAN@S Marshall ~2026-42517-72 (talk) 05:01, 31 July 2026 (UTC)

What's a SAN?—S Marshall T/C 07:51, 31 July 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)

Break re Catch 22 re ONUS vs NOCON

"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 agree that everyone has to work together, but if we interpret BURDEN as being an unending game of FETCH, then we don't have everyone working together. Instead, we have one self-appointed boss ordering around someone who actually does have to look constantly and continually over his shoulder to see what excuse the lazy boss is giving for rejecting the sources this week. WhatamIdoing (talk) 19:43, 28 June 2026 (UTC)

You are not interpreting it correctly, to the extent anyone needs to "fetch" (which is what, just a way to denigrate burden?), it is always supporters of the content (add, keep, or readd) who have the sufficient policy and capability requirements to do so (the multi-thought process I set out above) and it is true that, that class of persons is not necessarily only one person, it is made up of all such supporters of the content. (Indeed, a consensus implies more than one). But it is always fulfilled by an individual, with others agreeing. Which also means, it is never on the unsupporters, the no longer supporters, or those who don't have the sufficient policy and capability requirements to require or determine the rs. Alanscottwalker (talk) 12:26, 1 July 2026 (UTC)

It's not a way to denigrate any of this policy. It's a way to discourage lazy stonewalling, poor communication, and edit warring by POV-pushing blankers. At some point, "No, that's not a reliable source, so I blank it again" needs to be replaced by something that is functional for resolving the dispute instead of something that is only functional if your goal is to keep The Wrong™ Information out of the article. WhatamIdoing (talk) 16:45, 1 July 2026 (UTC)

There are several ways to resolve the dispute, they are laid out in Wikipedia:Dispute resolution. But once again, you assume bad faith on the second editor. You also assume that all editing occurs at once, when that's regularly not true. Alanscottwalker (talk) 22:18, 1 July 2026 (UTC)

No, I assume very good faith of the second editor. I assume that the second editor genuinely, deeply, truly believes that keeping the disputed content out of Wikipedia is what's best for Wikipedia. The first sentence of AGF says "Assuming good faith (AGF) means assuming that people are not deliberately trying to hurt Wikipedia, even when their actions are harmful", and I'm going well beyond that: I believe that the second editor believes (rightly or wrongly) that their actions are actively helping Wikipedia, especially in these three ways:

• Wikipedia is helped the most by keeping that content out of the article;
• Wikipedia is helped the most by minimizing the amount of their own time and energy spent in direct collaboration; and
• Wikipedia is helped the most by discouraging, blocking, or otherwise getting rid of any editor who would put that kind of content in this kind of article.

I believe this happens because I have been that second editor. However, I don't think that the second editor's behavior is a best practice. WhatamIdoing (talk) 01:14, 11 July 2026 (UTC)

You seem to have a skewed view of "collaboration". We are only here to collaborate on producing good content in line with policies and guidelines. We generally only write in either Wikipedia's voice or in the voice of an identified source, and we summarize reliable sources in a coherent policy compliant context. To do that collaboration, we edit and editing regularly and vitally involves also editing out. The good faith editor, who edits out is collaborating to the best of their ability. It's not best practice to support something that in their considered editorial judgement, they don't support. Alanscottwalker (talk) 11:36, 31 July 2026 (UTC)

Again, the problem I want to solve is not the removal of the information. The problem I want to solve is the incorrect reliance on BURDEN as a reason to revert the source and demand that the other editor go find another source.

Please go look at the desirable outcome I described above. How can we clarify BURDEN so that the removing editor stops saying "WP:FETCH me another source again" and instead says something like "Your source is unreliable, but don't waste your time looking for another, because AFAICT this doesn't belong in Wikipedia"? WhatamIdoing (talk) 16:32, 31 July 2026 (UTC)

It's not a problem to solve, certainly not in content policy, your scenario either relies on the bad faith or incompetence of the the second editor, or it seeks to force editors to do something they don't support. Alanscottwalker (talk) 17:26, 31 July 2026 (UTC)

No, my scenario only requires the reverting editor to be feeling very busy.

I want to change a string of "Rv unsourced per WP:BURDEN" – "Rv unreliable" – "Rv unreliable" – "Rv unreliable and I'm going to report you for edit warring because you've repeatedly tried to provide sources like I told you to" into "Rv unsourced per WP:BURDEN" and "Rv unreliable, and from my quick web search, this claim only appears in unreliable sources, so WP:ONUS applies". WhatamIdoing (talk) 18:36, 31 July 2026 (UTC)

Have a look at the most recent couple of edits to Neurodevelopmental disorder. The cited source has failed verification because it does not directly support the claim that NDDs (e.g., Intellectual disability and tics) are mental disorders.

Would you recommend that I just keep reverting the other editor, saying "Rv per WP:BURDEN; WP:FETCH me another source"? Is that the best practice encouraged by this policy? WhatamIdoing (talk) 18:43, 31 July 2026 (UTC)

I would recommend that you determine whether you are convinced that it is good content that belongs in the finished article. If you can't be convinced of that, there is nothing to do but edit it, up to editing it out. You as editor will like the first editor, have the BURDEN when citation is called for, only if you add or restore material that needs citation, in your editing. Alanscottwalker (talk) 10:53, 1 August 2026 (UTC)

I'm the removing editor in this case: another editor wants to have the MOS:FIRST sentence say that NDDs "are mental disorders" without citing a source that directly supports that. The editor wanting to include that information merely waves at the existence of the DSM-5 and the ICD-11, neither of which say this (I've checked), nor does the presently WP:CITED source.

IMO this editor is not meeting the WP:BURDEN. Should this policy recommend that I just keep removing it and saying "WP:FETCH me another source"? Is there really "nothing to do but" edit war out the unverifiable information? Or would you recommend a different approach for me? WhatamIdoing (talk) 18:07, 1 August 2026 (UTC)

If the editor came up with a source that actually did say that, would you support it being included? Nikkimaria (talk) 01:18, 2 August 2026 (UTC)

For the MOS:FIRST sentence, I think we need more than one such source, to comply with NPOV.

We actually do have one (1) reliable source that says this. I found it and posted it to the talk page myself. I also found one master's thesis, which is not MEDRS-quality. But that's apparently the end of the list. Everything else amounts to waving at sources that are known to not support the claim or at sources saying that ADHD is a mental disorder, so that "suggests" that all the non-ADHD NDDs (such as tics and cerebral palsy and intellectual disability) are mental disorders, too. WhatamIdoing (talk) 03:21, 2 August 2026 (UTC)

So then I think your concern is not so much BURDEN as BALANCE. Nikkimaria (talk) 03:26, 2 August 2026 (UTC)

Ultimately, it's WP:DUE (should a single MEDRS-quality source using a given term in its definition outweigh literally hundreds that don't?) but we actually do have a BURDEN problem here: "The burden to demonstrate verifiability lies with the editor who adds or restores material", and the editor who restored that material has not produced a source. That editor has failed to meet the BURDEN. Alan says "there is nothing to do but edit it, up to editing it out" in that situation. I don't think that's either true or the best practice. WhatamIdoing (talk) 03:36, 2 August 2026 (UTC)

So what do you think is best practice for that situation? Nikkimaria (talk) 03:38, 2 August 2026 (UTC)

I think the best practice usually looks like this:

• For the adding/restoring editor to produce one source (ideally, at a level of detail slightly more specific than waving at the Diagnostic and Statistical Manual of Mental Disorders in general). Technically, that editor needs to produce "one inline citation to a reliable source that directly supports the contribution", but it's the source that matters, rather than the inline formatting. If you dump a bare URL on a talk page or in an edit summary, any halfway competent editor can format it as an inline citation.
• For the removing editor(s) to look at the provided source, and either accept it or reject it.
  • If they accept it, then we're done.
  • If they reject it, then this policy imposes two requirements on them:
    1. The editors rejecting the proffered source need to explain why that source is insufficient (e.g., in this case, because the DSM does not actually contain the words that it was alleged to contain). This is in the policy: "Once an editor has provided any source they believe, in good faith, to be sufficient, then any editor who later removes the material must articulate specific problems that would justify its exclusion from Wikipedia".
    2. For everyone (adders, removers, and anyone else who is standing about) to proceed through the ordinary process of developing a consensus about what to put in the article. This, too, is in the policy: "If necessary, all editors are then expected to help achieve consensus". The process of achieving consensus includes, but is not limited to:
       • Starting a discussion on the talk page
       • Searching for sources
       • Checking how similar articles (if any) handle the situation
       • Suggesting alternative wording
       • Figuring out if there is any common ground
       • Seeking help from other editors (e.g., a Third opinion or asking for help at a noticeboard)
       • Forming a consensus through the Wikipedia:Requests for comment process
       • but not just insisting that since the other guy is trying to "restore" a word and I'm trying to "remove" it, that he has to do everything and all I have to do is keep reverting with "Nope, WP:FETCH me another source, because I say you've failed to meet the WP:BURDEN seventeen times in a row".

This policy also contains one relief valve for the removing editors: Once they've given the required explanation (which can usually be done with a simple link or half a sentence) and made a reasonable effort to achieve consensus (e.g., posting a Wikipedia:Fringe theories/Noticeboard to say that a newbie is peddling pseudoscience again), if the adder/restorer persists, they can step back per WP:ONUS and say that it's the adder/restorer's job to overturn the consensus that formed (e.g.,) when the FTN regulars landed on the talk page like a thousand of brick, or to produce evidence of a positive consensus for their proposal when none was previously found. WhatamIdoing (talk) 04:08, 2 August 2026 (UTC)

Editing is one of the ways to achieve consensus, indeed, it is the usual way to achieve consensus. And explaining your edit in the editing note is standard practice. And "up to removal" makes clear that removal is not the only option when you edit.

As you are not convinced it belongs in the article, make your position known. As your position is the edit needs at least two qualified sources that show that it is accepted in the field (and presumably not have readers read disinformation of misinformation), it's just incompetent or bad faith for you to tell someone else to just produce one qualified source. You still absolutely have no BURDEN to produce any qualified reliable source for material you are not and do not support adding or restoring to the article, by the plain language of the policy. (Above, you appear to try to make a joke about the 'burden on you not on me' but the truth is the burden is only on the person who adds or restores, regardless whether it is you or me. And no one has satisfied the BURDEN ever, if no one has added the qualified source.).

And nothing I said prevents or discourages also going to the talk page, or directing the other editor to look elsewhere, for substantive and additional editors support. If there is a dispute, engaging in WP:DISPUTE RESOLUTION is the thing to do, as I have previously made clear. And I have also previously made clear over and over, you are free to volunteer to satisfy the BURDEN -- if you are going to volunteer because you think it belongs in the finished article, just edit the article, and add the qualified source(s). There it is, "editing", again, for consensus. Alanscottwalker (talk) 09:18, 2 August 2026 (UTC)

It sounds like you're saying that BURDEN doesn't prohibit me from choosing a path that's different from edit-warring out the uncited content.

But I think the community's best practice goes a little beyond "Well, BURDEN doesn't say removal is mandatory".

I've got a scenario in which the other editor either believes that waving at a source that failed verification is sufficient.

1. Does BURDEN permit me to remove the content after the adding/restoring editor has named a source that I've deemed insufficient? Can I claim the BURDEN is still on the other guy, since I believe the supplied source doesn't "directly support" the content, or that it's not "reliable"? (What if I'm wrong, and it does directly support the content, and it actually is a reliable source?)
2. In your opinion, how many times do I get to remove the unsupported content before the BURDEN being on the other guy stops applying? Once? Three times each 24 hours? Seven times seventy?
3. Are there any limits articulated in BURDEN itself to me reverting this incorrect and poorly sourced word? (I'm asking about this being in BURDEN, because that's where a reverting editor is looking for a policy that justifies repeated reverting.)

WhatamIdoing (talk) 00:10, 3 August 2026 (UTC)

Second thought: "Editing is one of the ways to achieve consensus". But is there any place in this policy does it say that I have to do that? Remember, I'm the removing editor. BURDEN's on my side, and so is ONUS. Achieving consensus is the other guy's job, and AFAICT you don't want either of them to put any limits on my ability to remove the disputed content. WhatamIdoing (talk) 00:47, 3 August 2026 (UTC)

Policy does not generally give you a choice to edit war except perhaps in service of BLP (or a lenient administrator).

You should decide whether you support the content, but if you don't add or restore, you don't have the Burden. Only the adder or restorer has the Burden. Policy can't make anyone fulfill the Burden, it can but assign it. Similarly, it can't make anyone take-up the ONUS, it can but assign it. Its best practice to support good content and oppose bad content. You'll get plaudits for supporting good content and opposing mistake, misinformation or worse. Policy should support the good editing.

Policy can't make you edit, but if you decide to edit content, including removal, explain, usually in your edit summary. If you are asked to explain further, explain. If you think they got it or the misunderstanding can be ameliorated, edit again up to removal but I don't recommend edit warring. If the cycle of removal threatens to continue, go to more formal dispute resolution (and yes disputes cover content (such as the things in NOTE C) and conduct).

If you support the content, it at least behooves you to make sure it is no longer under threat of possible removal, but again you can't be forced to do any editing by policy. Also, unless you're editing in bad faith or incompetently, you should not generally remove content that you agree improves the article (by your own analysis, you are making the article worse).

It remains, the Burden lies where it lies per its terms, and the Onus lies where it lies, per its terms. And for good reason, the burden of persuasion is on the side that's: 'yes, let us tell the world this.' -- Alanscottwalker (talk) 15:28, 3 August 2026 (UTC)

Where in BURDEN does it say this? What words in BURDEN would give me, the remover, some idea that my policy-given right to remove content that (a) contains a citation, but (b) I personally, with no evidence of consensus behind me, have deemed that citation to be insufficient doesn't extend to repeated removal?

Where in BURDEN is there any suggestion that maybe, after someone has WP:CITED the material, just maybe we should move on from "Ha ha, BURDEN says I get to revert you, because I say that's not reliable, and BURDEN says you have to FETCH me a 'reliable' source" to some more sophisticated form of editing the article (e.g., to ordinary Wikipedia:Consensus work)? WhatamIdoing (talk) 17:02, 3 August 2026 (UTC)

We are not going to write the whole of policy in Burden. We are not going to list endlessly all the possible permutations of what may happen. See, KISS. Once again, you betray yourself, "Ha, Ha", really you should get away from trying to have content policy assume editors are in bad faith or incompetent Alanscottwalker (talk) 17:51, 3 August 2026 (UTC).

We are not going to write the whole of Wikipedia's ruleset in BURDEN. But perhaps we could put the footnote, which says:

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

in the main text, in the hope that a somewhat larger number of editors would notice that this policy says (and has for many years) that the next goal is explanations and finding consensus, rather than repeating the first step of BURDEN in a loop. WhatamIdoing (talk) 00:35, 4 August 2026 (UTC)

No. That's no, "next goal". Consensus on content is the goal from the start. So much so, that once an article has content provided by any good faith editor, it is presumed to have consensus until edited again. And policy lays out what consensus on content should look like, it should be NPOV, V, NOR, non-CVIO and BLP compliant content. There are many statements on Wikipedia that should have a qualified cite, that don't. Does that mean they are not verifiable, and in conformance with the other policies? No, not necessarily, it means no one has edited to do that. The BURDEN remains where it is, always has been, and always will be, with the adder or restorer of the content. -- Alanscottwalker (talk) 10:50, 4 August 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 re Catch 22 re ONUS vs NOCON

• 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)

IMO both from the original entry and the acceptance of it and the explanation, and the fact that it is in the verifiability policy IMO the original intent of Onus was to prevent using verifiability of something to force into it into an article. IMO then the usage drifted off course. To randomly put a finger on the scale toward exclusion of content. It conflicts with the wp:consensus policy, tampers with the decision processes under that policy. IMO the best thing to do would be to reword it to follow its IMO original intent. Sincerely, North8000 (talk) 02:21, 3 August 2026 (UTC)