Wikipedia talk:Verifiability/Archive 76

This is an archive of past discussions on Wikipedia:Verifiability. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

Archive 70

←

Archive 74

→

Verifiable ≠ cited

Latest comment: 2 years ago217 comments30 people in discussion

I feel like I've had too many conversations recently in which someone has selectively quoted BURDEN to claim that all uncited sentences are unable to be verified. The idea is that "other people" are entirely, hopelessly unable "to check that the information comes from a reliable source" unless a reliable source is presented to them on a silver platter.

I think it is probably time to add a sentence about the difference between cited and verifiable. NOR uses the line "verifiable, even if not verified", but "verified" could be understood as someone actually checking the source, rather than a source merely being listed in the article, so I'm inclined to suggest "verifiable, even if not cited" instead. A less stylish way might be to note that "if some material is not presently cited, but you can find a reliable source that directly supports it, then that material is already verifiable, and we would like you to edit the article so that the material is both verifiable and also cited".

What do you think? WhatamIdoing (talk) 07:38, 20 June 2022 (UTC)

I wonder if you are just trying to handle a situation where a room full of pedants still frequently use lazy shorthand, expecting others in the room to understand their shorthand, rather than to get diverted by someone dismantling their failure to be utterly precise. In other words, is this an issue to bother most editors with? The lead does already state everything must be verifiable but only "material whose verifiability has been challenged or is likely to be challenged" needs inline citations. So the message is already present in the lead that there is that there is text that is verifiable already but lacks and doesn't need a citation. Your proposed guidance would perhaps suggest editors should spend their time citing that Paris is the capital of France if they can find a source for it.

Perhaps the problem is with BURDEN:

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

is absolute. The two sentences combined make this bold text require one demonstrates verifiability with citations in all content, which is wrong. We could drop that first sentence that just repeats what the lead has already stated. Replace this short paragraph with

"If material has been challenged or is likely to be challenged, the burden to demonstrate verifiability lies with the editor who adds or restores that material. This is achieved by providing an inline citation to a reliable source that directly supports the contribution."

Remember this is the section on "Responsibility for providing citations" so a statement "All content must be verifiable." is not vital here. However, if editors wish to keep it, I think the caveat I've added would mean that first sentence could remain:

"All content must be verifiable. If material has been challenged or is likely to be challenged, the burden to demonstrate verifiability lies with the editor who adds or restores that material. This is achieved by providing an inline citation to a reliable source that directly supports the contribution."

-- Colin°^Talk 13:45, 20 June 2022 (UTC)

Or perhaps insert a sentence: All content must be verifiable. Content might be verifiable even while it is uncited, but sometimes it is necessary to demonstrate verifiability, by adding a citation. (No changes, just adding the second sentence.) WhatamIdoing (talk) 18:44, 20 June 2022 (UTC)

Way, way more than "sometimes". How much uncited text do you find in GAs and FAs? This isn't 2003 anymore where editors welcome people just plopping down whatever they believe to be true. The overwhelming majority of our text should be and is cited. Crossroads ^-talk- 01:24, 21 June 2022 (UTC)

Using GAs and FAs as a guide to what a binding minimum standard like WP:V requires doesn't seem well-founded. -- Visviva (talk) 02:04, 21 June 2022 (UTC)

There might even be a difference between when it is "necessary" to demonstrate verifiability and when it is "desirable" to do so. WhatamIdoing (talk) 04:19, 21 June 2022 (UTC)

An aspect to consider: say I add a paragraph of info to an article, but because I am baffled by WP's citation system, I include the URL to the source on the talk page or even as the edit summary. That is a completely valid verified piece of info with a source, just that the source isn't in an inline cite, and meets this guideline. Obviously, not including the citation will make it hard on other editors to track, so it would be expected experienced editors can help to include the cite, but just because the cite is absent doesn't make it unverified. --Masem (t) 19:34, 20 June 2022 (UTC)

Masem, I think the issue is about articles that end-up in the NPP queue or somehow make it into mainspace without any sources - totally unsourced articles - and the expectation is for NPP to find the sources that will (a) verify N, and (b) verify that it is not OR. NPP should not be burdened beyond identifying unsourced articles as being unsourced, and then send them to draft or PROD them for failing to meet GNG, and V. BURDEN states the onus is on the article creator to provide sources - NPP are not article creators and we have enough on our plates without having to finish the work that may be the result of a bot or UPE or an editor who just wants to take credit for article creation but doesn't want to do the work associated with V and GNG. NPP was not established to complete unfinished articles. Our job is to not allow them to be indexed, and to send them back to draft. We should not be encouraging the creation of unsourced articles by doing the work the creators failed to do, or were unable to do because sources don't exist, or the topic was not notable. Atsme 💬 📧 23:57, 20 June 2022 (UTC)

Sure, but that said, if I as a new editor created an article and still confused about citation style, dumped all my sources into the talk page, that's still "verified" for all purposes. Mind you, we absolutely accept bare URLs wrapped in ref tags as acceptable inline cites (they can be improved but they aren't failing to source their information) and a NPP patroller can inform the user how to do this quickly. Masem (t) 00:43, 21 June 2022 (UTC)

Even if you didn't dump any sources anywhere, the subject could be notable and the contents could be verifiable.

Have a look at Geriatric sexology, which I ran across yesterday. It's three sentences long. It's been unsourced since its creation about 17 years ago. But ask yourself: Do you reasonably expect someone to have written about this subject? Is there anything in those three generic sentences that sounds like the kind of thing that you'd never find in a reliable source? I don't see any problems here, and I suspect that the reason it remains unsourced is because nobody else thought it was all that important to have those three unsurprising, basic sentences followed by citations, either.

I'm not saying that the article benefits from being unsourced. I'm only saying that its completely uncited contents are also completely verifiable, even though it's obviously an article that made it into the mainspace without any sources. I'd expect a NPPer to glance at it long enough to see that it doesn't contain any inappropriate jokes, have enough general knowledge to realize that since old people have sex, someone's going to write about old people having sex, click the [Mark page as patrolled] button, and move on without giving it another thought. I would not expect NPP to spend time searching for sources (though if you want to, you can find several books on the subject as well as many journal articles), finishing the article (unless it just happened to be a subject area that interested the individual), or deciding whether it's worthy of being indexed by a search engine. WhatamIdoing (talk) 03:37, 21 June 2022 (UTC)

I would expect at minimum a competent NPP patroller to at least dump the term into Google and make sure its not something made up, because it is very easy to make a unsourced article with compelling enough language to seem legit. (I'm thinking of the hoaxes involving "dihydrogen oxide" being dangerous as a outside-of-Wikipedia example). The NPP doesn't have to read any more than 2-3 pages of results - a whole minute of work - to validate its a legit topic and then flag the article for "needs sources". But its like Schrodinger's cat - without sources present somewhere (talk page or bare url or the like), there's no way to know if the topic can be validated without at least identifying the potential for sourcing. Masem (t) 04:48, 21 June 2022 (UTC)

Masem, how about this idea - when NPP tags an unsourced article with a CSD or PROD, and an admin rejects the tag, they should go ahead and find the sources to satisfy Notability and Verifiability - it only takes a minute, right? And while they're at it, they can do some CE and make sure MOS is followed. I like that idea. Let the rejecting admins do the work and find the sources if they choose to not encourage article creators to make sure their articles are ready for mainspace, and are verifiable by adding RS since the ONUS is on them. Why won't that work? NPP has a 14k+/- backlog and that doesn't count AfC so it seems perfectly feasible for our admins to help out. After all, the article could be a hoax. j/s Atsme 💬 📧 12:37, 21 June 2022 (UTC)

This might be the only time you and I have agreed, Atsme. I don't know if you're being sarcastic or not but I do think that declining CSD/prods (by pretty much anyone) if it's largely unsourced/unreliably sourced, should absolutely put the onus on the decliner instead of allowing a vague notion toward "sources exist" which I know of at least 3 editors who mass de-prod, de-tag and de-csd articles but fail to even provide a single one of those "sources" that exist. PRAXIDICAE🌈 12:44, 21 June 2022 (UTC)

Not sarcasm, Prax – I've never faked a sarcasm in my life.^[FBDB] And we actually do agree on a lot more than you think, despite our occasional differences. We're both here to help build an encyclopedia. (PS: I love your new sig). Atsme 💬 📧 13:34, 21 June 2022 (UTC)

All I am saying is that asking for a NPP patrolled to do a 1-minute Google search for the topic where the new article lacks any type of source confirmation, just to make sure it doesn't appear to be a hoax, is a very reasonable check. They don't have to add sources, just be aware that potential sources exist. At worst, then.maint-tag the article for lack of sources before approving Masem (t) 13:21, 21 June 2022 (UTC)

It appears that the opposite is being proposed at Wikipedia talk:New pages patrol/Reviewers#Proposal: increase unreviewed new page search engine NOINDEX duration. We must prevent Google's users from seeing any content at all until we are satisfied that the subject is worthy of being shown to Google's users. WhatamIdoing (talk) 18:57, 29 June 2022 (UTC)

This sounds like a problem with NPP taking on a bit too much, frankly. The expectations at WP:NPPCHK are wildly excessive for the basic protections against spam, vandalism and other facially inappropriate material that we actually need patrollers to protect the project from. If reviewers are overloaded because they're taking on extra work checking cites, or because this voluntary project has been structured in a way that makes them think they have to do a bunch of extra work, that's unfortunate but not really something a core policy should be addressing. -- Visviva (talk) 02:04, 21 June 2022 (UTC)

I agree, @Visviva. NPP used to be our defense against CSD-worthy new pages. Now they're trying to stop anyone from getting an imperfect article into the mainspace. This is a hugely bigger scope than the original goal and a completely unfair expectation to put on them. WhatamIdoing (talk) 03:09, 21 June 2022 (UTC)

Your linked example of geriatric sexology is not, in fact, notable, and is a fitting example of why it's good that NPP enforces reasonable standards, and why we shouldn't be so forgiving of unsourced material. The phrase, in quotes, returns almost nothing on Google Scholar - there is no evidence that this is a distinct field of sexology. Rather, research on this topic belongs in the much-better article linked in the "see also" - Sexuality in older age. I intend to pursue deletion or redirection of this article once this discussion runs out. I would very much not want NPP to approve a lazy and misleading article like this. Crossroads ^-talk- 01:26, 22 June 2022 (UTC)

I'm a dedicated mergeist, so put me down as a support. Just please add a few sources while you're at it. ISBN 9789021905662 by Zwi Hoch and Harold Lief (quote: "The lifecycle subdivisions of sexology are: embryonal–fetal, infantile, child, pubertal, adolescent, adult, and geriatric sexology") is plausible, if old. You'll find a similar description in one of the books by John Money. But I point out that the question on this page isn't whether it should be handled as a Wikipedia:Separate, stand-alone article. The question on this page is whether someone is "able" to "verify" that this information comes from reliable sources. WhatamIdoing (talk) 03:51, 22 June 2022 (UTC)

I think it would be good if NPP could enforce those standards. The problem is that they can't. We don't have the volunteer capacity or the processes to quickly deal with unsourced material in the way that (some) NPP reviewers would like to, so they're getting frustrated and blaming the policies and processes that we do have. – Joe (talk) 09:13, 23 June 2022 (UTC)

The idea is that "other people" are entirely, hopelessly unable "to check that the information comes from a reliable source" unless a reliable source is presented to them on a silver platter. Doesn't matter. As soon as a good faith challenge occurs, it is the BURDEN of someone restoring a claim to demonstrate that the claim is verifiable. Why is the (frankly) bare minimum of "cite sources for your claims" so oddly controversial with some, getting derided as requiring sources "on a silver platter"? We should not water down BURDEN whatsoever with anything confusing about how uncited text is supposedly verifiable. Again, nowadays, we don't want people just lazily plopping down whatever they believe to be true. Crossroads ^-talk- 01:24, 21 June 2022 (UTC)

Whether an uncited fact is verifiable is really a separate question from whether an editor should do something.

Consider "The capital of France is Paris".

We

fully agree that if someone makes a good-faith challenge for that sentence, then editors must cite a reliable source for that uncited sentence.

The question we're addressing here, though, isn't whether WP:CHALLENGE rules apply to CHALLENGEd material. We all already agree to that and support those rules.

The question at hand is whether that sentence is actually verifiable right now, despite not having a citation after it right now, or if – according to some editors – the sentence is unverifiable unless and until someone adds a citation (...which nobody is "able" to do, because it's "unverifiable", right?).

Do you have an opinion on whether editors are able to verify that uncited sentence? WhatamIdoing (talk) 03:17, 21 June 2022 (UTC)

Responded below. Crossroads ^-talk- 01:41, 22 June 2022 (UTC)

Perhaps it is the nature of the discussions I've been in, but I think I've had more difficulty with editors trying to satisfy verifiability with vague and unspecific citations, for example citing an entire 10,000+ page book, instead of citing a specific chapter or page within that book, than I have with editors trying to do what WAID has said when opening this section. Sideswipe9th (talk) 01:33, 21 June 2022 (UTC)

FWIW I have also observed that this confusion seems to be increasingly common. I think the distinction is very important, and should be addressed in the policy, since the lack of clarity on this point only serves to encourage ever more aggressive gatekeeping that harms the project. -- Visviva (talk) 02:04, 21 June 2022 (UTC)

I mean, policy is already clear that citations are only required for things that are challenged or likely to be challenged. I'd be reluctant to tinker further in the way you suggest because I do think that when something is challenged, the citation needs to be clear - it is not sufficient, in the long term, to say "a citation exists on talk" or "there's a citation elsewhere in the article." I think that in situations where a citation can be easily produced WP:PRESERVE might lightly discourage massive wholesale removals, but ultimately we can't forbid them because the burden is on whoever wants to keep the text. So I don't think we want to encourage process-wonkery arguments about the difference between "verified" and "verifiable" - if you are at the point where you're actually arguing with someone, then the bottom line is that whoever wants to preserve the text is the one with the actual responsibility to affix an actual citation to the disputed text in question, end of story. Yes, on a proverbial silver platter, so to speak. If people are going around challenging massive swaths of text on the belief that absolutely all text requires citations even on very new articles, they're mistaken and policy is clear about that; but once that challenge is made I'd be very reluctant to do anything that could weaken WP:BURDEN. --Aquillion (talk) 03:32, 21 June 2022 (UTC)
What I want is for highly experienced editors to not say that an uncited sentence, such as The capital of France is Paris, is unverifiable or a violation of this policy.
It is very easy to follow this sloppy path, and I'm sure you've seen editors do it:
- Everything uncited is unverifiable.
- BURDEN says "All content must be verifiable".
- That means all uncited content is a violation of the Verifiability policy!
This conclusion is false, because it is simply not true that all uncited content is unverifiable. The capital of France is Paris is a verifiable statement even when it is uncited. WhatamIdoing (talk) 03:46, 21 June 2022 (UTC)
If there's an editor who's routinely challenging such WP:SKYISBLUE-style statements, then you might be able to make an argument that it's a conduct matter rather than an issue with the underlying content being challenged. However, I'm not sure that this issue is endemic enough that a change to policy is warranted...though I might support further clarification on the differences between "verifiable" and "cited". In my personal experience, when I challenge "obviously true" statements, I tend to end up in arguments with editors claiming that the information doesn't need to be sourced because it's "obvious"...nevermind that the most productive option would be to simply produce a source rather than bickering over the need for one (maybe that should be in the policy), and then frequently the first source they (finally) provide is a bad source. DonIago (talk) 03:54, 21 June 2022 (UTC)
I am looking only for a clarification on the differences between the words "verifiable" and "cited". I am not looking to change the policy on when or whether something should be cited. (Editors in those disputes should WP:Let the Wookiee win; it's faster to cite obvious information than to argue about it.) WhatamIdoing (talk) 04:25, 21 June 2022 (UTC)

I agree with the logic behind the concern. Editors can easily misinterpret BURDEN to mean that all material must have an inline citation. Instead of tinkering with the first paragraph, perhaps change the third paragraph to say something like: "Any material lacking an inline citation to a reliable source that directly supports the material may still be verifiable. If challenged, the material may be removed and should not be restored without an inline citation..." (bold is only to emphasize the suggested change). Maybe that's not front and center enough to cater to the concern, but just a thought. --GoneIn60 (talk) 04:17, 21 June 2022 (UTC)

What would that change even mean? The removal itself is a challenge. The policy already says, in bold, that only things that are challenged or likely to be challenged require a source; anyone who ignores such a flat statement is not going to be convinced by turning the entire section into tortured language that reiterates that point again and again. Additionally, I would strenuously oppose anything that would add additional bars to removing unsourced material. We are not required to do so, but unsourced material 100% can be removed, at any time, by anyone, solely based on the fact that it is unsourced, and when that happens anyone who wants to restore it should find a source. (If it really is WP:BLUE, good for you; finding a source should be trivial.) I would oppose any change to that in strongest possible terms. Indeed, my position is that entirely unsourced articles should be removable at any time for any reason. It is absolutely essential that we encourage people to provide sources for the things they add, and the best way to produce that pressure is to make it clear that if you don't add a source then anything you add can be instantly removed with no further discussion or explanation. Per WP:PRESERVE it is not always required to remove it; often a CN tag is better. But I would never support anything that would prevent anyone from doing so or add any barriers to doing so. Removing unsourced text is always, without exception, valid, fullstop. --Aquillion (talk) 09:04, 21 June 2022 (UTC)

"The policy already says, in bold, that only things that are challenged or likely to be challenged..." The "only" in your statement is assumed; it is not explicitly stated in policy, hence the potential for confusion. You and I get it, but novice editors can easily interpret this to mean that no citation always means not verifiable, and therefore any statement lacking an inline citation violates WP:V and should be fixed, tagged, or removed. BURDEN shouldn't be seen as an instruction to do so, that some action must be performed, because it is possible for verifiability to exist without a citation. If there's a way to add that clarification without being overly repetitive, then I'm all for taking that into consideration. --GoneIn60 (talk) 10:11, 21 June 2022 (UTC)

I agree with Gi60 here. The other false implication is that anybody can challenge any material for any reason, and that is absolutely not true and totally ridiculous. You must provide a valid reason in the edit summary. The thing we are not taking into consideration is that perhaps a challenging editor has mistakenly removed uncited material without realizing the material actually has been sourced elsewhere in the article. The confusion arises when editors think "uncited" means "unsourced", and this is due to a lot of our tools and templates such as Twinkle, Redwarn, and the Real-Time Recent Changes which propagate these misunderstandings by using such language as "added unsourced content" or whatever. If editors see that enough times they will think they are the same. Most, if not all, arguments tend to assume that a challenge, once issued, is infallible with the ultimate authority to force an inline citation to be produced without question, and this is simply unquestionably absurd reasoning that is in direct conflict with the spirit of collaboration and consensus. I think I have seen almost everyone agree that once a challenge has been issued the citation must be added, but why? Who says the challenger must always be right, or that their challenge is "indisputable"? I find it to be extremely odd, and even a little bit inappropriate that we are putting such an inordinate amount of good faith in challengers to the point that Doniago has even suggested it would be a matter of conduct issue if we had one making the mistake of routinely challenging uncited material that is clearly sourced, but any newish editor could easily make that same mistake and even veterans too. Huggums537 (talk) 04:20, 5 September 2022 (UTC)

@Huggums537, can you explain the distinction you're drawing between "uncited" and "unsourced"? Those two words are usually used interchangeably.

As a side note – and perhaps this is a key example of what you're thinking of – there has never been any clear agreement about whether a fact repeated in several places in an article must have a citation in every one of those places (assume someone who didn't read the whole article is complaining about it). WhatamIdoing (talk) 14:14, 5 September 2022 (UTC)

I know you're not generally asking, but to me it only makes sense that a repeated fact be cited each time (though that does somewhat beg the question of why the repetition is occurring). It's not an unreasonable burden to place upon editors to ask that they reiterate a source (indeed, it's easily accomplished), but it does seem unreasonable to put readers and other editors in a position of needing to confirm that a fact has in fact been repeated, and if so, whether it's sourced at any of its iterations. This would especially apply to newly-added information that may or may not be a repeat of information elsewhere in the article and that may or may not have been sourced when previously added. As a gnome, I have little interest in perusing the entirety of an article to confirm that yes, the information is repeated, and yes, it's sourced at least once. DonIago (talk) 15:06, 5 September 2022 (UTC)

It should be noted that this would also complicate the removal of sourced text, in that editors would then need to review the rest of an article to ensure that they weren't unintentionally removing a source for information elsewhere in the article. DonIago (talk) 15:12, 5 September 2022 (UTC)

Well, it's complicated, right? On the one hand, people blank "unsourced" content all the time, and then are told that WP:LEADCITE doesn't require citations in the lead, and they'll find the citation they're looking for in the body of the article. Or they'll be told that it's cited in the text and duplicated to the infobox (or vice versa). So this happens, and it's usually resolve amicably. In other cases, editors duplicate the citation to the contested spot. Since people don't read articles from top to bottom, sometimes a contested claim means that people are skipping to a section and not seeing the previous material, and editors decide that it's helpful (or just easier) to meet the apparent demand by duplicating the citation.

On the other hand, there are citations that ...don't make sense to repeat. Elephant (an FA) makes about a dozen direct and indirect claims that elephants are mammals. Should every instance of this claim be followed by an inline citation? Class: Mammalia^{[elephants are mammals]}... the largest terrestrial mammal^{[elephants are mammals]}... Unlike most mammals,^{[elephants are mammals]}... in most other mammals^{[elephants are mammals]}... the largest known among mammals^{[elephants are mammals]}... an unusual trait among mammals^{[elephants are mammals]}... as in most mammals^{[still mammals]}... While most mammals have^{[Have you figured this out yet?]}

Experienced editors mostly seem to know how this should work, but I'm not sure how you would write a rule for this that wouldn't invite wikilawyering. WhatamIdoing (talk) 15:41, 5 September 2022 (UTC)

I'm not sure it's all that complicated on a practical basis, most of the time, but I accept that opinions might vary on that. In my above text I was mostly referring to repeating information in prose. I think most editors understand that the lead is a special case, and that in theory an infobox is only supposed to be summarizing information that's repeated in the body of the article as well, so there should be sourcing in the article for anything in the infobox. I don't really have an issue with editors repeating cites in the infobox, though if they do so in the lead it tends to draw more scrutiny from me...but then, sometimes the lead isn't repeating information in the article body, which is a separate concern.

I'm assuming your elephant example as-presented is just hypothetical, because if the information is being repeated within the same section, much less within the same paragraph, then it's hard for me to imagine that any reasonable person would insist upon repeat citations.

I don't know whether a good rule is possible either, but based on my above statement, it might be that, "there's no need to repeat a citation within the same section of an article for repeated information" would address the most common scenarios. If we even needed this kind of rule, and I'm not really clear that we do. DonIago (talk) 16:42, 5 September 2022 (UTC)

The quotations from Elephant are scattered around the article. Few come from the same section. Except for the first (in the infobox, unsourced), the main point of the sentence is unrelated to the fact that elephants are mammals. For example, consider this sentence:

"Unlike most mammals, which grow baby teeth and then replace them with a single permanent set of adult teeth, elephants are polyphyodonts that have cycles of tooth rotation throughout their lives."

You want the citation for this to focus on cycles of tooth rotation, not on elephants' classification as mammals. WhatamIdoing (talk) 17:01, 5 September 2022 (UTC)

Hmm. But in that example there is a distinction being drawn between elephants and other mammals, so I would guess the source itself notes that contrast, though perhaps it only supports the final clause? If it does only support the final clause though, then that's a little eyebrow-raising as that would mean the sentence is synthesizing.

Otherwise, it being an FA and all, and not an article I'm especially invested in, I'd be very reluctant to try to edit myself. If the sourced claim about them being mammals is fairly early on in the article though, then I don't really see issues with it not being sourced elsewhere. I'm not sure that's what I'd consider optimal, but there's a world of difference and a world of relativism between 'acceptable' and 'optimal' in any case.

Anyway, I feel we're going a bit far afield here. I don't really have any proposals to make, nor do I necessarily feel there's the need to change anything anyway; I just think that editors shouldn't be expected to read an entire article before redacting unsourced material because the same information might be sourced elsewhere in the article, especially with regards to newly-added information. Mistakes may occur, but those shouldn't be laid at the feet of said editors, assuming anyone is even inclined to raise the argument, "But it's sourced three sections further down!" Rather in such cases it should be a calm, "Hey, that's sourced further down, but I can see how you might have missed that. I've reinserted the info and copied the source." DonIago (talk) 00:41, 6 September 2022 (UTC)

The problem isn't putting extra work on challenging editors to read the whole article before removing, the problem is that the only thing required to remove is that the material doesn't have an inline source. It then becomes a catch 22 for the uncited material that might actually be sourced if the material should not have been challenged in the first place. You say, "I don't want to do the due diligence", and I say, "I don't want my sourced material forced into the same category as quotations that require inline citations just because you refused to look at the fact it was possibly sourced elsewhere." Huggums537 (talk) 02:32, 6 September 2022 (UTC)

Depending on the size of an article, it can be a significant ask to require editors to read an entire article before they can remove any material that's apparently unsourced, especially if it's material that's been newly-added to the article.

Can you please establish how it's placing a significant burden on editors who wish to retain material lacking an inline citation to simply restore it with an appropriate citation if and when the material is challenged? Or, ideally, including inline citations at the time they're adding the material to begin with? DonIago (talk) 03:07, 6 September 2022 (UTC)

Well, that depends on what the editor values doesn't it? If the editor values their time, then they will simply add the citation and be done with it so your point would a good one that it would be asking more to check the article (you don't have to read it all) than to add the citation. However, if the editor values their principles enough to spend the extra time hashing it out on a talk page with several editors over a period of days or weeks about why the citation was not needed in the first place, then I would say that could turn into a much more significant task than simply checking an article for existing sourcing. Huggums537 (talk) 03:28, 6 September 2022 (UTC)

It just seems to me like we are expected to read the article and do due diligence to add material before writing, so why would it be any different for removing? We aren't supposed to just add willy nilly, so why remove that way? The problem is that most people don't think removing material is willy nilly if it doesn't have an inline citation, but they think it is if it isn't added with an inline citation, and that is a huge problem because no matter if it is easier or harder to do, it simply is not true. If I had any idea we were going to base rules on the easiest way to do things no mater what the truth is, then maybe I might feel differently. Huggums537 (talk) 03:45, 6 September 2022 (UTC)

You claim that we are expected to read the article and do due diligence to add material before writing. I don't think this is true. I think that the community's actual as-seen-in-practice policy is not to read the whole article unless you're planning to take the article to GA or FA.

For example, I edited Kings Dominion the other day. It's about 4200 words long. I checked (mostly with ⌘F) to see whether the content I wanted to add was already there, and I skimmed a couple of sections to figure out where it belonged. That's normal, expected, and acceptable.

I don't think that's a case of basing the rules on the easiest way to do things. I think that's a case of doing what's relevant to the task at hand. WhatamIdoing (talk) 04:02, 6 September 2022 (UTC)

I would like to resound those statements. Wikipedia has generally been very open and collaborative, and I support that approach to the maximum extent possible. Altanner1991 (talk) 04:37, 6 September 2022 (UTC)

Right, well I should have said "check" the article to keep in line with my previous statement, but your point is taken and I hope you see that the normal, expected, and acceptable amount of proofreading for adding material is not any more than I would be asking someone to do before they automatically yank some content just because it doesn't have an inline citation. Also, yanking content solely for that reason alone without even doing so much as a check to see if it could be sourced does nothing to support collaboration if the result is that it forces the content into a category of content that is now required to have inline citations such as quotes. If it only forced a mere discussion on the matter, but didn't force it into this category, then it would be collaborative. As it stands, I can essentially require any content (not just uncited) to have an inline citation no matter if it is already sourced or not even if I know it is sourced. This means that the choice of referencing or sourcing style is effectively taken away by rule. All one has to do to force this style is challenge for any reason, and it does not even have to be a valid challenge. In fact, it will be in favor of the challenger if they don't make a perfectly valid claim because once the challenge is made, others will say you can't put it back until you have argued both that the challenge was not valid, and for your reason to keep or that you simply give in to their demand to conform to their referencing style. Huggums537 (talk) 07:21, 6 September 2022 (UTC)

Another point I would like to make is that I would be perfectly fine with doing things the "easy way", if the easy way wasn't "inline citations only". IF users were offered more options for sourcing the uncited material other than strictly "inline citation only", then sure, I think it would be much easier and fair for everyone involved than asking someone to go through the trouble of going through the whole article to find sourcing, but we aren't asking editors to simply "verify the source" when we remove, or choose a way to reference the material. We are saying, "inline source or nothing!" It would be just as easy to ask to verify with a hidden comment as it would any other form of reference. I'm not saying hidden comments should take the place as references by any means, but it is a valid example for the sake of argument. Huggums537 (talk) 12:02, 6 September 2022 (UTC)

I guess the difference I am trying to draw between "uncited" and "unsourced" would be very similar to the same distinction you are drawing between Verifiable ≠ cited. It could more easily be explained by saying that by "uncited" I meant "without an inline citation", and by "unsourced" I meant that most editors editors erroneously think that if the inline citation is not present then the material is not otherwise sourced elsewhere in the article or anywhere in the world. So, they call it "unsourced" just because it lacks the inline citation, but that might not even be true since it could be sourced elsewhere in the article or somewhere in the world if that makes any sense. Huggums537 (talk) 03:05, 6 September 2022 (UTC)

I have also sought to distinguish "sourced" and "cited" in the policy page, and yes I would add "verifiable" to that list of often confused wording.

Unsourced should indeed be clarified: does it mean unverifiable or uncited? Altanner1991 (talk) 04:29, 6 September 2022 (UTC)

@Huggums537, is your list looking something like this?

verifiable: Someone could find a reliable source that says this, if they tried to.
cited: A very nice editor has made it really easy for others to find a reliable source that says this, by adding an inline citation.
sourced: Someone could find a reliable source that says this, and instead of starting with their favorite web search engine), they only need to look in the list of already-cited sources in this article to find at least one that Wikipedia:Directly supports this claim. (No very nice editor copied that citation over to this particular spot yet, though.)

The "un-" versions are the opposite: nobody could find it, no very nice editors have helped out, and the other sources already cited in the article for other material don't also cover this point.

If that sounds like what you're thinking of, then I think you need a different name for it. WhatamIdoing (talk) 05:06, 6 September 2022 (UTC)

Yes, that sounds right. Huggums537 (talk) 06:42, 6 September 2022 (UTC)

Editors generally don't make a distinction between the first and last state (you can find a source by asking Mr Google vs you can find a source by looking in the ==References== section), and they do use the words cited and sourced interchangeably.

Back in the day, we might have described this situation as using the ==References== section as (implicit) general references for the material that didn't have inline citations. WhatamIdoing (talk) 02:20, 7 September 2022 (UTC)

On rechecking I see that the policies were actually fixed sufficiently—no more need. Altanner1991 (talk) 07:47, 6 September 2022 (UTC)

As I've said before, the extreme overwhelming majority of our content does not consist of widely-known things like "Paris is the capital of France", so writing policy with that in mind is misguided. Rather, our content consists of things that most people don't know off the top of their heads. My concern is the 99.9999...% of content that tells most readers things they don't know, and that material should ideally be cited (and we definitely should not move toward saying or implying it's okay not to). And in that process, someone can quick cite the CIA World Factbook or something for where the capital of France is, rather than these lengthy debates about enabling failure to cite sources. Crossroads ^-talk- 01:41, 22 June 2022 (UTC)

I don't think that makes any difference. Consider:

Here is information you know, and you can find a reliable source if you try. You are "able" to "verify" it, so it's verifiable.
Here is information you don't know, and you can find a reliable source if you try. You are "able" to "verify" it, so it's verifiable.

Same end result, right?

I could agree that it's worse to have that second type of information unsourced, but I can't agree that it's unverifiable. The question here isn't whether uncited information is bad. The question here is whether uncited information is unverifiable. WhatamIdoing (talk) 04:19, 22 June 2022 (UTC)

As I've said at length at WT:NOR, I don't think that whether sources "exist" in some vague sense, but are never cited, really matters in practice. Trying to make the policy go on about that more is just confusing. Crossroads ^-talk- 23:44, 23 June 2022 (UTC)

To be honest, other than the two sentences at the top of the BURDEN section (which I think should be fixed because their juxtaposition suggests verifiable = cited), I keep seeing WP:V reminding me that "Verifiable ≠ cited". Read the sentences beginning "When tagging or removing..." Those would make no sense if uncited meant unverifiable. I've read a lot of completely made up and untrue things about polices in the last few months but I don't think adding more text to those policies always a solution. If NPP needs guidance is how best editors should handle utterly uncited articles, say, then perhaps there should be some guidance for NPP, rather than a policy text change. -- Colin°^Talk 09:01, 21 June 2022 (UTC)
To be frank, ^{because I'm always being honest,} the problem is not that NPP doesn't know what to do or cannot discern the difference between cited and verifiable. Verifiable means that the article, in its current state, should include RS that establish notability, verifiability and NOR. If it doesn't, then it fails and should not be allowed in main space until those issues are addressed. The policy needs to more clearly state that RS are required to establish both V, N and OR before an article or stub will be published in main space and indexed. Otherwise, V will conflict with all our other PAGs. What some may not be taking into consideration is whether or not the article includes any sources at all, it's not simply about inline citations. Stubs/articles that are sourced and have no inline citations are simply tagged by NPP with either ^{[citation needed]} or a big tag at the top of the page. NPP reviewers know the drill. My concerns are not about material that is added to an article that already contains RS. For all we know, some of those thousands of articles sitting in main space that are not sourced at all, and have slipped under the radar, could have been created by a UPE who has written a really nice promotional article that reads like an advertisement for a BLP or company they represent. How do those articles make it into main space? Could have been theft of a redirect, or an autopatrolled editor helped make it happen – I don't know, but we eventually find out if it's an admin or NPP reviewer with autopatrolled rights. There's a reason the decision was made to stop automatically giving all admins autopatrolled rights, and for having new reviewers either attend NPPSCHOOL or qualify via experience before they are given NPP user rights. Unsourced articles are much more serious than what some of the arguments have presented. It is not in the best interests of WP for any of us to turn a blind eye and simple accept unsourced articles under an assumption that they qualify. It's time to stop (a) putting bandaids on gaping wounds, and (b) encouraging editors to create articles in main space without RS or having first established V, OR and/or N. Perhaps some of our problems stem from a lack of qualified AfC reviewers, I don't know. I've read some of the comments by DGG who is inundated with new articles at AfC and carries more than his weight. He's beginning to burn out, and the same happens at NPP. Atsme 💬 📧 14:32, 21 June 2022 (UTC)
Atsme is absolutely right. An article without any RS has no right to exist here. Independent RS notability has not been established. There is no evidence it's not a hoax or OR. It doesn't have to be fully referenced, but there must be enough RS to first establish V, OR and/or N. Period. Without RS we should treat it as the opinion of the author, and we don't allow such content in any article here. Send it back to private userspace or draft space. The initial burden of establishing V, OR and/or N is on the article's creator. -- Valjean (talk) (PING me) 21:14, 21 June 2022 (UTC)
@Valjean, I disagree. @Atsme says:
"Verifiable means that the article, in its current state, should include RS that establish notability, verifiability and NOR."

The actual policy says:
"verifiability means other people using the encyclopedia can check that the information comes from a reliable source."

These definitions are not the same. Worse, one of them is wrong. If you read the next bit, you see her say 'The policy needs to more clearly state that "RS are required"...'. This indicates to me that Atsme's first sentence is a statement of what she wishes the definition of verifiability were, rather than what the definition actually is.

It is not necessary to say that uncited information is technically unverifiable or a violation of this particular policy to say that unsourced articles are a blight upon humanity. If you want to establish rules that say unsourced articles are bad or banned (i.e., even when the contents are technically verifiable because someone could add citations), then I suggest making a proposal at WP:NOT. Please insist that at least one of those sources be WP:INDY. WhatamIdoing (talk) 03:32, 22 June 2022 (UTC)
WhatamIdoing stated: "verifiability means other people using the encyclopedia can check that the information comes from a reliable source." That is exactly what it says, but it doesn't mean what you think it means. See Use-mention distinction which may be at issue here, so forgive me for using bold text and for highlighting the important aspects of the policy that are either being overlooked or misinterpreted. Based on what I've gleaned, the position in some of the arguments is that we should keep unsourced articles in main space and not worry that the article isn't sourced or is missing inline citations because it can be verified. The "how" of verification is what appears to be missing or misunderstood, as does the use of "material" which may include the entire stub/article. The following very important aspect of V supports my position: Any material that needs a source but does not have one may be removed. It goes on to say: Please immediately remove contentious material about living people that is unsourced or poorly sourced. Granted, not every sentence in an article requires a citation provided the article is properly sourced...only challenged material or material that is likely to be challenged are affected. If those sources are not provided the material can be removed and that comes straight from our core content policy as does the following: {{tq|All content must be verifiable. The burden to demonstrate verifiability lies with the editor who adds or restores material, and it is satisfied by providing an inline citation to a reliable source that directly supports[2] the contribution.[3] NOTE 2 is also extremely important: A source "directly supports" a given piece of material if the information is present explicitly in the source so that using this source to support the material is not a violation of Wikipedia:No original research. What part of the former and latter are causing editors to believe that unsourced articles are ok, and should not be removed from main space? But wait...there's more. Unsourced articles are automatically challenged as OR - and any experienced editor should already know that, especially if they've edited in a controversial topic area or BLPs. If a source is not cited, and/or there are no sources listed in the article at all, is it expected that our readers should leave the article to find those RS in order to verify the information in the article? If that's the case, why read WP? Here's a somewhat comparable comparison: a grad student turns in a master thesis without references or citing any sources, and then simply tells the prof not to worry about it – it's verifiable. What do you think happens? When an article is unsourced, it's more than just a matter of simply finding a source to verify that the topic is a valid one, and not a hoax. It involves finding multiple RS to satisfy GNG, V, and OR, as I've stated repeatedly, and am being criticized over. Far too many arguments are failing to acknowledge that the sources used must support what the author wrote, so if the author failed to provide the sources they used, and failed to provide inline citations for the material they added that is likely to be challenged, how are we supposed to find those same sources that were used to add that material? Cited sources and citations are how we separate the wheat from the chaffe. Bottomline: if the article creator failed to provide the sources they used and/or failed to include citations for material that is likely to be challenged so that our readers could easily verify a statement, opinion, or fact in that article it should not be allowed in main space. Atsme 💬 📧 12:39, 22 June 2022 (UTC)
I think you're missing the point. It's not about whether articles should be sourced. It's about whether you say that uncited material is impossible to verify. This is a use-the-correct-jargon problem. Whenever these steps can happen:
Alice adds uncited information

Bob slaps a fact tag on it

Alice adds an inline citation to a reliable source that Wikipedia:Directly supports the information.

then that information was never "unverifiable". That information was "verifiable, but uncited". In the third step, Alice now demonstrates verifiability, but the information already was verifiable from the very beginning. The policy says that all information has to "be" verifiable; it does not say that all information has to "be demonstrated to be verifiability".

The "how" of verification has never been defined or restricted. The policy does explicitly state that information does not need to be "easily" verifiable by anyone.

This is long, so perhaps numbers will help in case someone wants to pick out a bit for further exploration.
"The position in some of the arguments is that we should keep unsourced articles in main space and not worry that the article isn't sourced"
...is irrelevant. This discussion is about whether you declare something to be "uncited" vs "impossible to verify", not about whether it's a good idea to have completely unsourced pages in the mainspace.

We all agree that "Any material that needs a source but does not have one may be removed."
But there is no sentence in this whole policy that says all "uncited" information is "impossible to verify".

There is also nothing in here that says completely unsourced pages are banned, and you've produced no evidence that all unsourced pages always contain material that needs a source. Have a look at User:WhatamIdoing/Christmas candy. Do you see anything in there that qualifies as contentious matter about BLPs? And direct quotations? Any material that has been challenged? How about any material that you think is Wikipedia:Likely to be challenged (i.e., by anyone who isn't trying to get blocked for bad behavior)? My answers are nope, nope, nope, and nope, and I bet yours are, too. If you can't find some material in that page that falls into at least one of those four WP:MINREF categories, then the whole page does not contain "Any material that needs a source", and therefore the whole article does not (technically) need any sources, because the whole article already complies with WP:V even without them. BURDEN doesn't apply.

Granted, not every sentence in an article requires a citation provided the article is properly sourced...
The last phrase is wrong. Not every sentence in an article requires a citation even if the rest of the page is garbage. Whether this material is possible or impossible to verify does not depend on whether the rest of the article is properly sourced.

NOTE 2 is also extremely important: A source "directly supports" a given piece of material if the information is present explicitly in the source so that using this source to support the material is not a violation of Wikipedia:No original research. What part of the former and latter are causing editors to believe that unsourced articles are ok, and should not be removed from main space?
I think you might be reading more into Wikipedia:Directly supports and NOR than is really there.

I don't see anyone claiming that unsourced articles are "ok". I see them claiming that unsourced material is not always impossible to verify. We might be able to get an agreement that unsourced pages are Very Extremely Extra Bad – just not "unverifiable". Most of our unsourced pages contain verifiable information.

"Unsourced articles are automatically challenged as OR"
Um, nonsense? NOR is the policy that actually says Articles that currently name zero references of any type may be fully compliant with this policy. Anybody who is automatically challenging unsourced articles as a NOR violation needs to go read the policy. If you run across an unsourced article, please don't put Template:Original research on it. Please use Template:Unreferenced instead. You might find this table of problems and their matching templates handy.

(If you wanted the equivalent statement for this page, I think it would be "Articles that currently name zero references of any type may be fully compliant with this policy, assuming that they contain almost no information at all.")

"how are we supposed to find those same sources that were used to add that material?"
WP:V doesn't require that "those same sources" be cited, assuming anything has to be cited at all. If (and only if) we need to demonstrate verifiability, then we need to cite any reliable source that directly supports the material. It does not have to be the original source. In fact, there doesn't even have to be an original reliable source. The "original source" might be an editor's own memory, or something on Twitter. You should still cite a reliable source whose contents match the claims you're putting into the Wikipedia article. We do not insist on a strict interpretation of WP:SAYWHEREYOUGOTIT (so long as you actually did read the source that you put in the article).

The TLDR:
The absence of an inline citation doesn't actually mean that people aren't "able" to "verify" the material.

This policy permits (some) uncited material.

This policy permits (a few) completely unsourced pages.

Editors should call uncited material "uncited" or "unsourced" or "bad", but not "unverifiable".

Editors should use the word unverifiable to describe material that is impossible for anyone, no matter how determined, to find reliable sources that could support it.

WhatamIdoing (talk) 21:47, 22 June 2022 (UTC)
Why does this hair-splitting of terminology matter? When content is CHALLENGED and removed it makes no material difference whether, unbeknownst to everybody, a reliable source exists out there somewhere ("uncited") or not ("unverifiable"). Either way, the unsourced content must not be reinstated, and the editor(s) supporting it must cite RS that directly support the claim. Really, what we should do is update WP:V's clauses from like 20 years ago where Wikipedia was an obscure site that lacked content and welcomed people adding whatever they thought was true, and bring it in line with current practice by formally deprecating adding material without citation. Atsme's comparison with the student turning in a thesis with no citations is the perfect reductio ad absurdum of this outdated idea of 'verifiability is just when sources exist somewhere out there'. Crossroads ^-talk- 00:06, 24 June 2022 (UTC)
What you call "hair-splitting of terminology" is what other people call "clear communication". WP:Policy writing is hard, and it doesn't get any easier when people use the same words to mean significantly different things. WhatamIdoing (talk) 01:49, 24 June 2022 (UTC)

@Atsme, with regards to your highlighted points; please note that they are only referencing material that needs a citation or material that has been challenged. It does not apply to all unsourced material. This is per later in paragraph where WP:BURDEN tells us what needs to be sourced: Attribute all of the following types of material to reliable, published sources using inline citations: all quotations, all material whose verifiability has been challenged, all material that is likely to be challenged, and all contentious matter about living and recently deceased persons. Huggums537 (talk) 00:52, 1 September 2022 (UTC)
Citations do not verify; they merely facilitate. The process of verification is inherently difficult because, even when you have some evidence, analysis and understanding is required to confirm that the evidence corresponds with the assertion. The process of analysis itself requires verification and so you then get an infinite regress per What the Tortoise Said to Achilles. At some point, the reader just has to accept that something is true and that requires an act of faith. Often this faith is lacking and this generates much activity at WP:RSN.

As an example, consider the main page of Wikipedia and notice that none of the facts there are cited. But we mostly accept them on trust that someone has done the relevant checking to ensure that they all stand up. But this is a fallible process and so there is regular activity at WP:ERRORS.

And, even if everything has been done right, there will still be a disclaimer footnote which says emphatically that "WIKIPEDIA MAKES NO GUARANTEE OF VALIDITY".

So, as we are not claiming that any of our pages are exactly correct, we should just relax and save our energies for the most controversial and complex cases. It's like the dictum of Frederick the Great, "He who defends everything, defends nothing."

Andrew🐉(talk) 13:37, 21 June 2022 (UTC)

Andrew, I'm just going to jump in here and leave a comment. Tell me (ping) if I'm close to the point. You write: "Citations do not verify; they merely facilitate." I agree. We need to differentiate between "verifiability" and "verified". The initial content should be accompanied with a verifiable source. Hopefully, it also verifies the content. If it doesn't, then we have a case of "failed verification" and the "source" should be removed. The next question is what to do with the now unsourced content. If the content isn't seriously controversial or a possible BLP violation, it can be tagged with a "cn" tag. If controversial, then the content can be deleted. This is a matter of editorial discretion.

My point is that maybe we need to say something about the difference between "verifiability" and "verified". The existence of a citation/ref is not the end of the process, because some editors are careless or even sneaky vandals. Other editors should seek to "verify" the "verifiable" source. -- Valjean (talk) (PING me) 02:01, 10 July 2022 (UTC)

I can think of an example where it would reasonably be the end of the process, and that would be in the case where it is assumed that the subject of the article is the primary source like a book, and there is a plot section where it is obvious that the book (or the subject of the article) is the source, so a reference would not be needed for the plot section since anyone could verify from the book. See: Wikipedia:When to cite#When a source or citation may not be needed Huggums537 (talk) 23:46, 31 August 2022 (UTC)

Question - Something that needs clarification for this discussion to continue productively: are we discussing the situation where an article has NO citations (at all)… or are we discussing the situation where an individual fact (within a sentence or short paragraph) does not have a citation? I think these are two different issues. Blueboar (talk) 14:55, 21 June 2022 (UTC)
My understanding is that it's the latter; sentences or short paragraphs that do not have an inline citation. I agree that these are two totally separate issues, and judging from some of the responses so far, it looks like there may be some confusion. --GoneIn60 (talk) 21:26, 21 June 2022 (UTC)
Unless it's a 2 or 3 sentence stub like this one: Jagatipala. Or you have to deal with articles like 3rd Alley, originally deleted back in 2008. Right now there are 182 pages in the NPP queue with no citations/sources. From what I've gathered, some editors are of the mind that as long as an article is verifiable it meets V, but there's no guarantee that V aligns with N, or OR. And doesn't using V in that context also suggest that an article's notability is also verifiable if sources are automatically verifiable? Why even bother to add sources if it's ok to make these assumptions? We could just let WP run as an aggregator, I guess.^[stretch] + ¯\_(ツ)_/¯ Atsme 💬 📧 22:10, 21 June 2022 (UTC)
Information that (truly) meets V will always meet NOR. But there are no guarantees whatsoever that something complying with WP:V is a notable subject. If the only sources ever published about a business are the business's own self-published website, you could write a fully verifiable and even fully cited article, but it would not be a notable subject. WhatamIdoing (talk) 03:40, 22 June 2022 (UTC)

It doesn't really matter. Uncited information that could be cited to RS is verifiable. We are "able" to "verify" it; it is "verifiable". This is true whether we are talking about the only three sentences on the page, or the only uncited three sentences in an otherwise well-sourced article.

There might be some differences (e.g., whether it is a minor problem or a serious violation of all that is holy), but the presence or absence of other citations elsewhere on the page makes no difference in whether someone is actually able to verify that particular information. WhatamIdoing (talk) 03:13, 22 June 2022 (UTC)
Uncited information that could be cited to RS is verifiable. We are "able" to "verify" it; it is "verifiable". I like this. Effectively explains in a nutshell why verifiable ≠ cited. Sometimes it's the simple analogies that really hit home. Perhaps some variation of this is the answer. --GoneIn60 (talk) 05:49, 22 June 2022 (UTC)
Nope, it's misinformation - read the highlighted text above. Atsme 💬 📧 12:44, 22 June 2022 (UTC)
Atsme, do you really think that the statement: “Trenton is the capitol city of New Jersey” becomes misinformation if we say it without citing a source? There are thousands of sources that support this fact (atlases, almanacs, history books, tourist guides, news papers… etc, etc). It is an extremely verifiable statement, even if we don’t actually cite any of them. Blueboar (talk) 13:10, 22 June 2022 (UTC)
Blueboar, I've never even hinted at such nonsense – it defies common sense – but since you and I are usually on the same page in such matters, if you've seen something I've said that motivated your question, please point it out to me so I can fix it. My position throughout this discussion remains the same because I have not been convinced that it needs to change. I have made valid statements that are unambiguously supported by our core content policies, or at least I thought they were unambiguous. My highlighted text above remains as the crux of my position. Atsme 💬 📧 13:32, 22 June 2022 (UTC)

I think that there are two different important things being discussed here. The second raised one is the lobsided "prove/establish a negative" workload related to look for wp:GNG sources which is causing the in-progress collapse of NPP. It's not simply a google search. Even the easier ones (where a search of English language / Arabic character set sources is enough,) that still involves evaluating every hit for GNG sourcing criteria. And quadruple the person-hours required when the "source area" where we need to "prove a negative" has a non-english language with a different character set.

On the first topic raised, when there is a sincere concern about the verifiability/veracity of the material, I think that we need to strongly keep the burden for sourcing on the person seeking to include the material. But if you look at the problem cases, I think that a common theme is that there is no actual concern about the verifiability / veracity of the material, and none expressed, often resulting in deletion of uncontested sky-is-blue material. Probably the three most common cases of this are:

POV warriors wanting to knock out material as a way of pursuing that end. And using this rule in synergy with "RS" classifications.
Somebody that is having a pissing war with the editor involved
Somebody being obsessive, thinking that an overkill approach is "just enforcing the rules"

I've long advocated for a measure that I think would dramatically reduce these problems without upsetting the apple cart. And that is to require any removal on such a basis include a statement of concern about the verifiability/veracity of the material. That alone would constitute fulfillment of the challenge criteria. Sincerely North8000 (talk) 13:12, 22 June 2022 (UTC)

You're talking about a completely different problem. Compare these two:

Alice posts “Trenton is the capitol city of New Jersey”. Bob says "You put unverifiable information in Wikipedia! That's bad!"
Alice posts “Trenton is the capitol city of New Jersey”. Bob says "You put uncited information in Wikipedia! That's bad!"

Bob's first statement is wrong. Bob's second statement is correct. Are you able to understand why the first is wrong but the second is correct?

NB: The difference is not in whether it's bad. Also, it has nothing to do with NPP or notability. It is only about whether Bob is wrongly declaring that it's actually impossible for someone to find that information in a reliable source, or rightly declaring that this statement was not followed by a little blue clicky number. WhatamIdoing (talk) 20:19, 22 June 2022 (UTC)

I think that something is confused here. First, I noted the second topic infused into the discussion (NPP and verifiablility) mostly to identify it as separate but I did not post on that, and then I posted on the the main topic which is basic wp:ver. And, of course I understand that distinction, I understood it in my first 5 minutes of reading wp:ver over 12 years ago. Now, on applying my point to your examples....the case that I noted is neither of those. It is where Bob is a POV warrior that wants to minimize coverage of Trenton. So they just delete the material, saying "unsourced" without making either of those two statements that you described. Or if it is sourced but to an attack able source, they just delete it saying "not RS sourced" And what I advocate is requiring Bob to make one of those two statements that you describe (either the right one or the wrong one) which would cut down on what could be called bad faith deletions or even a weaker one like "I have a concern that the statement might not be verifiable" North8000 (talk) 13:39, 23 June 2022 (UTC)

Sure, and we've agreed on this for years. But that's outside the scope of this conversation. I really just want people to understand that a particular bit of content can be technically, according to this policy, actually totally verifiable even if there are no sources named on the page. That's it. I want your POV warrior to say "unsourced" instead of "unverifiable". Getting to the point that we CHALLENGE only material that we have an actual reason to doubt is for the advanced class.

Perhaps I should put this another way: I do not want any editor who reads this policy from top to bottom to be able to say that this policy says that everything is unverifiable unless it is cited. Right now, we have a few good(!!!) editors who are saying that. They genuinely believe, or at least fervently hope, that this policy bans unsourced articles (in particular) and that uncited material is a violation of this policy. I want good people to read this policy and discover that "All material must be verifiable" does not mean "All material must be cited", or even that "WP:V says that every article must name at least one source". WhatamIdoing (talk) 23:02, 23 June 2022 (UTC)

Thing is… even if Bob is acting in bad faith, the quickest and easiest way for Alice to tell Bob to shut the fuck up is to return the Trenton statement with a citation. Don’t worry about why Bob wants a citation. Give him one and move on. Blueboar (talk) 14:43, 23 June 2022 (UTC)

That happens after the bad-faith deletion and so is not related to my idea which is a simple soft requirement that would reduce bad faith deletions. BTW the most common case isn't for a totally unsourced item, it's when the POV wikilawyer uses this provision on an already-sourced item in tandem with a wikilawyer attack on the source or use of the source. North8000 (talk) 14:56, 23 June 2022 (UTC)

Exactly. That prevents this Bob this time, but does nothing proactive to prevent the next string of Bobs down the line... Huggums537 (talk) 06:21, 29 July 2022 (UTC)

There was a lack of sophistication in the initial guidelines, where it was assumed that any source on the internet would be satisfactory verification; considering that good sources accessible on the internet were many fewer than now, it produced generally poor quality verification by current standards. Most articles from that period that have not been extensively revised have inadequate sourcing by current standards. W :What is the operative definition of verifiable? I think the usable meaning is that we have many analogous articles, whicch we have sourced properly, and know that there will be similar sources available. But it can also mean an article translated from another WP, without copying their references (a great temptation, because the formatting usually needs to be reworked completely, and because some excellent WPs, such as deWP, accept extremely general references, such as "for references, see the entry in the national bibliography"). I think the safest course is to insist on at least one reference that verifies eexistence and the general notability -- which is the current standard at BLPPROD. DGG ( talk ) 18:27, 23 June 2022 (UTC)
@DGG, I think the operating definition of verifiability is "If I go to the trouble and expense of truly searching, I am able to find a reliable source for this specific claim".

This really isn't about notability standards. I'd love to see a rule that says all newly created articles must contain at least one source that is (a) reliable for at least one statement in the article and (b) financially independent of the subject matter. We don't have that rule; I suggest proposing it at WP:NOT. WP:NOT#UNSOURCED would be a good shortcut. In five or ten years, we can start apply it retroactively.

But really: All I want to achieve today is that people will stop saying "Aaaaarrgghhh, it's another stupid unverifiable sentence" and start saying "Aaaaarrgghhh, it's another stupid uncited sentence" whenever they have a reasonable expectation that someone could add a reliable source at the end of that sentence. That's it. Change one word. I want people to know that if someone is "able" to "verify" it, then it is "verifi-able". This really should not be a difficult concept. Maybe the simplicity of it is the actual problem? It's so patently obvious, so we need to drag something controversial or difficult into it, like NPP's challenges or how to establish notability or what to do about rules-lawyering POV pushers? But that's not what I want to accomplish. I want to tell people that "All material must be verifiable" means that it someone must be able to find a source, and not that someone must have already done so. WhatamIdoing (talk) 23:13, 23 June 2022 (UTC)
I have to quibble… I believe that it is the responsibility of those who wish to add material to actually have found a reliable source for it… before they add it. However, they don’t always need to cite that source (the “Paris is the capitol of France” exemption). Blueboar (talk) 11:54, 24 June 2022 (UTC)
As do I, and quibble I will because it appears to me that there is an attempt to lower the bar for verifiability in that material included in an article doesn't have to be corroborated if one RS has been cited. Granted, the primary purpose of V is to demonstrate that the material is not OR, and that content is determined by previously published information rather than the beliefs or experiences of editors. However, none of what has been proposed has taken into consideration the fact that there actually may be opposing views, and that we must maintain a NPOV. Furthermore, any effort attempt to dilute BURDEN requires community-wide consensus, especially the part that states Any material lacking an inline citation to a reliable source that directly supports[2] the material may be removed and should not be restored without an inline citation to a reliable source. That statement should not be altered, but I'm finding that this "no sources are ok" position to be rather disconcerting because it contradicts long standing policy and opens the door to contentious material and relentless debate. Atsme 💬 📧 01:40, 25 June 2022 (UTC)
@Atsme, you said "none of what has been proposed has taken into consideration"...

What, exactly, has been proposed? WhatamIdoing (talk) 18:06, 25 June 2022 (UTC)
WhatamIdoing, if you need "exactly", then may I suggest a keyword search using "propose", or some form of it? This discussion is now tl;dr for me, and I'm not inclined to keep repeating what I've already explained. I think we're pretty much on the same page on some issues with the exception that quite a few of us don't think the policy needs to be changed. The opening sentence is quite clear, which you've also quoted but I've added some bold underline: ...verifiability means other people using the encyclopedia can check that the information comes from a reliable source. It doesn't say search online for a RS if you want to verify it. The RS that the material came from has to be with the article, hopefully in the reflist. It's the only way we can check where the information came from, otherwise we must assume OR, or that it's possibly rumor, a hoax or taken out of context. If we have time, we can search for a RS to verify that the material is supported, but then it may require a rewrite as it may not exactly support what's stated in the article, not to mention that it could be taken out of context. The options we have include tagging it ^{[citation needed]}, or removal. If there are no sources with the article at all, we have options for deletion. It's pretty simple. This diff supports my position, and I also agree with what Aquillion stated in this diff. There's nothing more I can add. Atsme 💬 📧 01:39, 26 June 2022 (UTC)
Why do you think this policy requires editors to cite "the RS that the material came from", vs, "that the information comes from a reliable source" – "a" reliable source, as in "any reliable source in the whole world, even if it's not the specific RS that the material came from"? WhatamIdoing (talk) 04:50, 26 June 2022 (UTC)
P.S. "This diff supports my position" isn't a diff. It also doesn't say that we need to cite the exact RS that the material came from. WhatamIdoing (talk) 04:51, 26 June 2022 (UTC)

@Blueboar and @Atsme, I really do think that your "quibbles-n-bits" about a volunteer always being obligated to find a source before they add it is just a little bit of "Purina" bull stuffing because of what I said just a little earlier about how plot sections can easily be verified with the subject of the article (primary source) without a citation. The statement made by @Aquillion does not explain the fact that people like to argue that if something really is blue, then finding a source should be trivial, but in this case the source is already assumed, and easily verifiable so there isn't any need to cite it unless someone just wants to challenge it. Huggums537 (talk) 00:29, 1 September 2022 (UTC)

I don't feel we should use the current contested state of how we cite plot summaries as an example of anything. There have been a vocal number of people arguing that we don't need citations for plot summaries (or that "implicit" primary citations are sufficient); to my knowledge, discussions on that have never reached a clear consensus. For the record I strenuously disagree that plot summaries are in any way exempt from the clear requirements of WP:V or WP:PRIMARY - nor do I believe it would be possible for a consensus to override the most basic requirements of WP:V, since it is not subject to consensus; a challenged plot summary always requires a citation, fullstop. Primary sources can sometimes be used to answer that challenge, but only if the plot summary is not a "large passage" (or worse yet, the entire article), since that would violate the restrictions WP:PRIMARY places on the use of primary sources. But even the people who argue that plot summaries are somehow exempt from WP:V's requirement for citations when challenged or the WP:PRIMARY limitations on the use of primary sources have always held that they are a special case - I think that that's already a wrongheaded and deeply-mistaken argument, but I would strenuously oppose using that longstanding dispute as an argument to further erode our sourcing requirements elsewhere. If people are going to make that mistake, then I think we need to start being more aggressive about challenging uncited plot summaries that may not perfectly reflect the book, and to trim or reduce ones that have reached the point of violating the WP:PRIMARY restriction on using primary sources for large passages. --Aquillion (talk) 03:54, 1 September 2022 (UTC)

But, the whole point is not that we don't need the citations or that plots should be exempt. If some jerk wants to challenge a perfectly verifiable plot, then fine let the rules apply to the plot, but the point is that a simple straightforward plot is obviously verifiable through the primary source without even adding the citation. One could argue that a plot is already essentially sourced and cited by a general reference that isn't officially listed at the bottom in a reference section, but it does exist in the form of attribution to the title, author and so forth in the body of the article. By this logic, sources are not required if the appropriate sourcing is already provided elsewhere in the body of the article, and let's face it, if there is a full article on the topic, then both the primary and secondary sources likely already exist in the article for the plot, and whoever challenges, or argues it should be challenged is probably just being stubborn about their philosopyhy that every little "thee" or "thou" needs a cite. Huggums537 (talk) 20:58, 1 September 2022 (UTC)

@WhatamIdoing: OK, I'm starting to see your main point in this thread. (Although if I read your last post precisely, it's to get people to change a wording in a sentence that I don't see people say.) But I'm starting to see it as raising awareness of the fact that the general standard is verifiable, not verified. And I think a non-controversial small step in that direction would be adding a sentence to WP:ver that adds emphasis to that point. But going beyond that small step starts to get complicated. The most common removal verbiage isn't what you described, it's simply like "removed unsourced material" and the second most common is something like "not a WP:RS" with the WP:ver based removal material. And usually the person wants to remove the material, and merely used wp:ver as a justification. What would you propose regarding that? My idea is to require challenges to include an expression of concern such as the sentence in your post. North8000 (talk) 12:54, 24 June 2022 (UTC)

Re WhatamIdoing's question on is the uncited sentence Paris is the capital of France verifiable, yes obviously. But if anybody challenges it then it still needs a citation. Even if you, or I, think that challenge is silly. Or even if made by a Nice supremacist who is POV pushing. Like Blueboar said above more poetically, the quickest way to dispatch that request is getting the requested citation, not argue over if it is technically verifiable or what the other editor's motives are. nableezy - 13:38, 24 June 2022 (UTC)

If there's a pattern of behavior in good faith editing that's tied to BURDEN, as observed by the OP, then it might be worth pursuing some additional clarity. However, it's a good point that the fix is often quicker than deciphering the motive. Maybe we need a clearer picture of the scale of the problem to help us weigh the benefit of action over inaction. --GoneIn60 (talk) 23:09, 27 June 2022 (UTC)

Also… as I have mentioned elsewhere in these threads… Wookies are MUCH more receptive to explanations about “why a source isn’t actually needed” AFTER they have “won” the challenge and a source has been provided. The conversation shifts tone from being confrontational “drama” to being friendly “education”. Blueboar (talk) 23:38, 27 June 2022 (UTC)

WE're talking bout the more reasonable common use of this which is on an unsourced item. But keep in mind that the most common disputed use by wikilawyer POV warriors is on a sourced item in tandem with nitpicking the source. North8000 (talk) 19:30, 29 June 2022 (UTC)

If a source does exist (or did exist and the editor is aware of it) but an editor is disputing it, then that isn't really the same thing as what we're talking about here - at that point you can point to and defend the source on talk, or take it to WP:RSN, or start an RFC over the specific and clear question of whether the source is sufficient for that specific statement. None of those things make sense if someone has removed a completely unsourced statement (ie. one that nobody has attempted to provide any source for at all at any point), because at that point there's nothing to constructively discuss - if you refuse to provide a source, all you can do is argue that the statement falls under WP:BLUE, which is overtly accusing the other editor of bad faith or incompetence (since BLUE is supposed to be immediately obvious.) That's not helpful, and if you do think one of those things are the problem you should take it to ANI rather than article talk. That said, even when there's already a source that someone found objectionable, unless the complaint seems glaringly absurd, I will usually first make an attempt to find an additional source - if I can find one, then doing that is faster and easier, and the article is ultimately better for having additional sourcing on a statement someone found objectionable, which will also avoid further rehashes of the discussion from others who might have the same objection. If you can't then sometimes you have to wade in and argue over the one source you have, but I don't think there's a harm in encouraging people to at least make an attempt to find a second source for contested statements first. --Aquillion (talk) 04:08, 1 September 2022 (UTC)

I'm a big Star Wars fan, and I think Let the Wookie Win is cute as hell, but I think letting someone know about a source you found for the challenged material or why it is verifiable on the talk page is just as fast, and probably ends up being just as productive as providing the citation. Just as N8000 suggested, you might be dealing with a POV warrior who could just as easily find some other excuse for removal after you add the citation, and you would just as likely find yourself ending up on the talk page anyway, so you might even save yourself some time by skipping adding the citation... Huggums537 (talk) 00:43, 1 September 2022 (UTC)

Verifiable, not verified

Wikipedia:Verifiable, not verified says this:

"Content is verifiable if, using all the means at the disposal of an educated and dedicated person, the content in question can be found in a source that Wikipedia editors deem to be reliable for that claim."

User:Altanner1991 and I discussed the problem below of editors not knowing that WP:V requires inline citations for most material (i.e., not for all material). I suspect that this knowledge gap affects our up-and-coming editors the most: they are experienced enough to know that the policies exist and want to help with the work, but they are not so old that they know all the complexities. In this section, we have been discussing the problem of being unclear about the difference between what's "able to be" verified and what's "already cited".

I wonder if we could clarify both of these points by adding this sentence as a footnote at the top of WP:BURDEN, and then adding a transitional sentence to lead in to the main point of the section. The result would be:

Proposed change
Yes Proposed change	No Old version
All content must be verifiable.^[1] Additionally, this policy requires that certain types of material have inline citations. The burden to demonstrate verifiability lies with the editor who adds or restores material, and it is satisfied by providing an inline citation to a reliable source that directly supports^[2] the contribution.^[3] ^ Content is verifiable if, using all the means at the disposal of an educated and dedicated person, the content in question can be found in a source that Wikipedia editors deem to be reliable for that claim. ^ A source "directly supports" a given piece of material if the information is present explicitly in the source so that using this source to support the material is not a violation of Wikipedia:No original research. The location of any citation—including whether one is present in the article at all—is unrelated to whether a source directly supports the material. For questions about where and how to place citations, see Wikipedia:Citing sources, Wikipedia:Manual of Style/Lead section § Citations, etc. ^ 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.	All content must be verifiable. The burden to demonstrate verifiability lies with the editor who adds or restores material, and it is satisfied by providing an inline citation to a reliable source that directly supports^[1] the contribution.^[2] ^ A source "directly supports" a given piece of material if the information is present explicitly in the source so that using this source to support the material is not a violation of Wikipedia:No original research. The location of any citation—including whether one is present in the article at all—is unrelated to whether a source directly supports the material. For questions about where and how to place citations, see Wikipedia:Citing sources, Wikipedia:Manual of Style/Lead section § Citations, etc. ^ 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.

What do you think? I want to hear from anyone interested, but I would especially value hearing from anyone who believes these two sentences are factually wrong (e.g., material is not verifiable if you can find a source, this policy does not require inline citations for certain kinds of material, etc.). WhatamIdoing (talk) 15:26, 5 September 2022 (UTC)

The factuality seems fine to me.
I think the footnote would add unnecessary fluff to the policy page, and it might be better to weigh for the encouragement of some kind of reference (say, bibliography), even if not cited.
I like the part about inline citations being required for only certain content, because indeed I had not known about that policy. But I am concerned about its affect on the flow of the already-choppy sentences. Perhaps something more like the following would be better: All content must be verifiable, whereby the burden to demonstrate verifiability lies with the editor who adds or restores material. This can be satisfied by providing an inline citation to a reliable source that directly supports the contribution. Additionally, certain types of material are required to have inline citations.

Altanner1991 (talk) 02:46, 6 September 2022 (UTC); edited 08:38, 6 September 2022 (UTC)

For future reference, the aforementioned preceding discussion was at Wikipedia talk:Verifiability#The policy page is, at this point, out-of-order. Altanner1991 (talk) 08:17, 6 September 2022 (UTC)

Does anyone object if I add "Additionally, this policy requires that certain types of material have inline citations." to BURDEN? WhatamIdoing (talk) 17:59, 11 September 2022 (UTC)

This sounds more than reasonable to me Moxy-

21:50, 11 September 2022 (UTC)

I think it would be a step in the right direction as far as distinguishing between required/non-required inline citations, but the community needs to eventually have a discussion about why we are including challenged or likely to be challenged material in this category because in my mind this type of material is not "challenged", but simply "questioned" or "debatable" meaning it could be discussed as to what kind of citations would be acceptable if any at all so I don't see why an inline citation would be a requirement the same way it would for a quotation or contentious material of a living person. There seems to be a huge disparity between these types of materials for us to be making it a requirement simply because anybody decided they would force a discussion on any material. If the material is worth debate, then I think quotations, and BLP libel covers most if not all of it, and if there is anything else not covered we can add it without effectively putting a blanket ban on all other types referencing by forcing inline citations on anything at all just because it got "challenged". Huggums537 (talk) 13:15, 12 September 2022 (UTC)

What exactly should be required to have citations feels like it should be a separate discussion. Would you like to start a new section to focus on that? WhatamIdoing (talk) 18:02, 12 September 2022 (UTC)

Somebody else can if they want to. I'm trying to ease myself out of policy discussions for personal stress related reasons. Huggums537 (talk) 18:08, 12 September 2022 (UTC)

I will also throw in my support for this. Altanner1991 (talk) 11:50, 14 September 2022 (UTC)

I have no objection to it and mildly support it. Two different meanings are implied in it. One is "not everything needs a cite" and the other is saying "a few things categorically need a cite". So it's not explicit on either, but that's sort of the Wikipedian way to write policies. North8000 (talk) 20:44, 12 September 2022 (UTC)

No, we should absolutely not add any text to WP:V that waters down the requirement to cite sources, makes it easier to refuse to provide sources or to revert unsourced material back in, or adds undue emphasis to sources supposedly being unnecessary when in fact this is very rarely the case. I feel like this same exact thing has been discussed for months and months on this page and those of us who don't want to read and respond to the same POV being argued over and over again are being worn down relentlessly. If anyone really wants to open up Wikipedia to "I think it's true so I'm gonna just put it in" lazy editing and make it harder to remove such because some source "exists" somewhere, then do an RfC at the Village Pump and end this already. If anyone tries to change WP:V in line with that based on these interminable discussions that everyone other than proponents has gotten sick of reading, they will be reverted. Crossroads ^-talk- 22:44, 12 September 2022 (UTC)
I don't see anything in the proposal that waters down the requirement to cite sources or makes it easier to refuse to provide sources. Just FYI. Newimpartial (talk) 22:47, 12 September 2022 (UTC)
I agree with Crossroads. The proposed amendment is a watering down, even if only by implication. The phrase "this policy requires that certain types of material have inline citations" implies in natural language (even if not in formal logic) that the material not covered by the "certain types" alluded to does not necessarily require inline citations. That is absolutely not something we should be implying. The current understanding is that a properly verifiable article is one that is basically fully cited. That's the standard we apply to main page (ITN/DYK/OTD/TFA) content, and we shouldn't write anything here that allows people to hide behind an assumption that they don't have to inline cite. Cheers — Amakuru (talk) 23:18, 12 September 2022 (UTC)
When I read this policy, I see this:

Attribute all of the following types of material to reliable, published sources using inline citations:
all quotations,

all material whose verifiability has been challenged,

all material that is likely to be challenged, and

all contentious matter about living and recently deceased persons.

I do not see this:

Attribute all of the following types of material to reliable, published sources using inline citations:
everything.

What do you see in the policy? Do you think it would be accurate to say that the long-standing text of this policy implies, in natural language, that the (small amount of) material not covered by these types does not necessarily require inline citations? If so, then it sounds to me like your objection amounts to noticing that this sentence accurately describes the policy as written, and the real problem is that you don't like the actual policy. WhatamIdoing (talk) 23:28, 12 September 2022 (UTC)

My point is that in any forum where the issue of verifiability is actually checked by editors other than the article authors (i.e. WP:FAC, WP:GAN, WP:DYKN, WP:ITN/C), the standard applied is that basically all facts mentioned must be cited. So that's the thrust of what we want the project to look like. That's the "big hairy audacious goal", and if anyone asks what the standard for verifiability is then that is what we should tell them. "It's fine to write an article with large swathes of uncited text but don't expect it to be permitted on the main page" is hardly the message I think we should be sending out; far better to just say "please cite everything".

Now where the current policy isn't wrong, and where I would probably be in agreement with you, is that the presence of uncited text doesn't mean we should automatically delete that text. In fact, going back to ITN/C again it is forbidden to simply remove text in a bid to get around the lack of citations. The goal is to salvage the text written and have someone actually check it, followed by either removing it or providing a citation, depending what they find. So, in summary, we shouldn't write anythign here that encourages uncited text at all, but if someone does happen to write some, then so be it. — Amakuru (talk) 23:40, 12 September 2022 (UTC)

As I recall, pushing for a Big Hairy Audacious Goal was a pretty reliable way to kill organizations in the 1990s. But even at the FAC level, not everything requires a citation. Wikipedia:When to cite is the FAC standard, and it has a whole section called "When a source or citation may not be needed", that lists four exceptions. It is therefore not actually true that "basically all facts mentioned must be cited". The goal is not "please cite everything". The goal is much closer "please cite stuff that isn't common knowledge to the people who know anything at all about this subject area" – not just "you don't have to cite that humans normally have two legs and two arms", but you also don't have to cite some undisputed definitional statements, such as "In a computer, the processor is the component that executes instructions."

Your belief about what's actually required does not seem, to me, to match what our long-standing policies and documentation says is actually required. This isn't really a question of what to do if material doesn't have an inline citation; this is about which material is required to have an inline citation and which material isn't. The community seems to believe that even FACs do not need to have "basically all facts mentioned" followed by an inline citation. You seem to think the standard is much higher than it actually is. WhatamIdoing (talk) 23:54, 12 September 2022 (UTC)

I haven't been around here long enough to know any better, but it sounds to me like Amakuru is saying the standard at FA, GA, DYK, is "must be cited", and "please cite everything", but I wonder if they have any evidence this actually means "inline citations" or if material could be otherwise cited or verified for these projects because there is a difference there and earlier when they were saying we should not imply that not everything requires inline citations, why not? Everything has to be verified, sure, but doesn't have to be with inline citations. Huggums537 (talk) 23:57, 12 September 2022 (UTC)

WP:When to cite is an essay page. A lot of this stuff is clearly a legacy from ye olden days of Wikipedia when the site had very little content and desperately needed users who would write stuff, and citations were just kind of a sometimes thing. I don't believe for a second that if we somehow had to write WP:V from scratch nowadays, in the modern internet awash in misinformation, tendentiousness, POV warriors, and the like, we'd at all bother to carve out exemptions for when citations are not needed. The overwhelming majority of our content is not "sky is blue" material, and writing the policies around that tiny, tiny minority of material will mislead.

Even without changing the technical meaning of the policy, we could change the emphasis the other way. We could rewrite it to say, All material must be cited to reliable, published sources using inline citations; the only exception is material that is not a quotation, does not concern living or recently deceased people, and has never been challenged and is unlikely to ever be challenged. Would this be an acceptable edit? I would be perfectly good with this. It properly lays out the reality that citations are the rule, not the exception. Crossroads ^-talk- 22:36, 13 September 2022 (UTC)

citations are the rule, not the exception +1 to that. Selfstudier (talk) 22:44, 13 September 2022 (UTC)

No, this would not be an acceptable edit - such a shift in policy text would be weaponized by editors to leverage the removal of text or terms with which they disagree, in instances where text at the same level of abstraction from the sources would be perfectly acceptable to the same editors if they happen to agree with it.

This already happens in cases where editors insist on the removal of paraphrase or content that is sourced elsewhere in the article, not out of any sincere concern for accuracy but simply because they DONTLIKE the content. (You did this to me recently with sourced content simply because you didn't like a watermark.) Let's not make the current situation even worse. Newimpartial (talk) 22:45, 13 September 2022 (UTC)

I don't think Crossroads proposal changes anything; the revision is just marketing. "You have to cite these four things" is not actually different from "You have to cite unless it's not one of the these four things". It just switches from a positive statement to a (grammatically permissible) double negative – more confusing, perhaps, but the same logic once you get past the extra words. WhatamIdoing (talk) 00:19, 14 September 2022 (UTC)

A proposal for this would look something like this:

Change the wording from:

"Attribute all of the following types of material to reliable, published sources using inline citations:

A
B
C
D"

to "Attribute all material to reliable, published sources using inline citations, unless all of the following conditions are met:

it is not A,
it is not B,
it is not C, and
it is not D."

Crossroads, do you want to make that proposal? WhatamIdoing (talk) 22:48, 15 September 2022 (UTC)

I mean, I am seriously good with it and would support it, but I'm not really feeling like trying to get the broad base of support to actually fix it and the whole ensuing discussion. If someone else wants to spearhead this, they have my encouragement. Crossroads ^-talk- 06:03, 16 September 2022 (UTC)

"Challengeable quotations" might be too narrow a requirement especially when others feel it should go the other way: that citations be more encouraged than discouraged. Altanner1991 (talk) 11:57, 16 September 2022 (UTC)

No sense in refusing sources based on this clarification; it only elucidates policy. This is useful if anything in later knowing how to improve policies. Editors would still be required to provide sources if challenged, per the rest of the policy. Policies should not be hidden to have an effect. If effects are desired, it should be done through explicit communication of policies, IAR notwithstanding. Altanner1991 (talk) 12:01, 14 September 2022 (UTC)

From the proposal: "Content is verifiable if, using all the means at the disposal of an educated and dedicated person, the content in question can be found in a source that Wikipedia editors deem to be reliable for that claim.". I think this is a terrible idea. I think that the person who adds the information must be able to provide an unambiguous indication of where the information can be verified. The rule should be "unverifiable until proven verifiable". Otherwise people can add stuff just on the recollection that they saw it somewhere. Zero^talk 06:56, 16 September 2022 (UTC)

Not sure what exactly "must be able" would mean, but I will throw my support for what @Zero0000 has boldly just said: it would challenge the framework, but a clean encyclopedia is the goal. Altanner1991 (talk) 11:43, 16 September 2022 (UTC)

I agree. The encyclopedia will never be completely sourced, but there is too much material in Wikipedia that is currently challengeable but unsourced. We need to encourage more citations, not less. - Donald Albury 13:27, 16 September 2022 (UTC)

It would be a good start to first clarify the policy as it stands. Too many ambiguities over what needs a citation leave too much room to feelings of the selective/biased application of policy. Altanner1991 (talk) 22:01, 16 September 2022 (UTC)

Agree that this is a terrible line. People concerned with whether a statement is verifiable shouldn't be required to pursue "all the means at their disposal" to do so, and editors certainly shouldn't have to do the same before they can tag or delete an unsourced statement they have doubts about. What does this phrasing even mean? Yes, I could embark on international travel to try to find a copy of a rare book...but I'm not going to, nor should anyone be expected to. I also feel there's something inherently snobbish in the phrasing "educated and dedicated"; it seems to be implying that anyone who would delete content that other editors believed to be verifiable simply wasn't "educated and dedicated" enough. DonIago (talk) 13:59, 16 September 2022 (UTC)

Except that it doesn't say that?

This is a definition of what it means for any statement to be "verifiable". A statement is "verifiABLE" if someone is "ABLE to verify" it in any reliable source (a cited source, or otherwise).

It is not a rule about whether anything requires inline citations.

Consider:

VerifiABLE: The capital of France is Paris.
VerifiABLE and cited: The capital of France is Paris.^{[number leading you to information about a reliable source]}
VerifiABLE and incorrectly cited: The capital of France is Paris.^{[ref leading you to a joke about Paris]}

That statement is always verifiABLE, regardless of whether it's uncited, correctly cited, or incorrectly cited. The only thing that could make this sentence become UNverifiABLE is if Paris stopped being the capital of France. There is nothing you can do, or fail to do, in an article that has the power to make this material impossible to verify in a reliable source. The most that you can do is to make it slightly inconvenient to verify.

Again:

VerifiABLE: When B cells internalize antigens for processing, membrane-anchored IgD is systematically excluded from the resulting vesicle.
VerifiABLE and cited: When B cells internalize antigens for processing, membrane-anchored IgD is systematically excluded from the resulting vesicle.^{[number leading you to the graduate-level textbook that says this]}
VerifiABLE and incorrectly cited: When B cells internalize antigens for processing, membrane-anchored IgD is systematically excluded from the resulting vesicle.^{[ref leading you to website that mentions B cells and vesicles but not IgD]}

Again, there is nothing you can do, or fail to do, in an article that has the power to make this material impossible to verify in a reliable source. The most that you can do is to make it very inconvenient for other people to verify (and this kind of claim is Wikipedia:Likely to be challenged, so you'd be violating this policy's requirements if you didn't cite a source when adding it).

On the other hand:

UNverifiABLE: Joe Film has been scientifically proven to be the best movie actor ever.
UNverifiABLE and cited: Joe Film has been scientifically proven to be the best movie actor ever.^{[numbers][leading you][to many][fansites and][other][unreliable][sources]}

There is nothing you can do in an article that will make this statement be verifiable: it is unverifiable with no sources cited in the article, and it is unverifiable with bad sources cited in the article, and there are no good sources in the real world, so those don't exist. This statement will always be impossible to verify in reliable sources; it is UNverifiABLE even when it's cited.

As the section heading says at the very top, verifiable ≠ cited.

There is nothing in this sentence that says editors shouldn't cite their sources for absolutely everything. There is nothing in this sentence that says you can't tag, remove, or otherwise WP:CHALLENGE unsourced statements. This sentence clarifies the difference between "verifiABLE" (=what the policy requires for absolutely all material without exception) and "cited" (=what the policy requires for the four large classes of material listed in the bullet points in WP:BURDEN). In particular, this clarifies that "uncited" does not mean "automatically unverifiable": Statements that are ABLE to be verified in a source are verifiABLE even when the source isn't handed to you on a silver platter (or at least in a little blue clicky number). WhatamIdoing (talk) 15:20, 16 September 2022 (UTC)

As I've said before, this is a completely unimportant distinction and a legacy of early 2000s Wikipedia that we should be moving away from, not towards. Try submitting a paper to a professor or an academic journal with no citations because all the claims in it are "verifiABLE" and see how far it gets you. I agree with the criticism that the sentence completely reverses the burden of proof and comes off as snobbish and implying we must engage in absurd amounts of effort before removing uncited material. Crossroads ^-talk- 17:16, 16 September 2022 (UTC)

I think this is a very important distinction. We can't actually move away from a requirement that everything be verifiABLE towards a requirement that everything be CITED as long as we pretend that those words mean the same thing.

(Also, we're not writing for professors or academic journals. We need to follow our rules, not theirs.) WhatamIdoing (talk) 17:20, 16 September 2022 (UTC)

Crossroads, I cannot imagine that a draft paper that did what you are insisting on for Wikipedia - providing a citation for each and every element of every descriptive statement every time such a statement is made - would be accepted by any peer-reviewed publication in any field. That you are blind to this obvious fact leads me to suspect that you would use such a policy much more "selectively" than the plain (and totally impracticable) meaning of your comments would directly suggest. Newimpartial (talk) 17:23, 16 September 2022 (UTC)

Do other editors really think it's a problem to specify that verifiability is defined in terms of what educated person thinks is verifiable, rather than what children or other uneducated people think?

If that's an example of snobbery, do you also feel snubbed by the NOR policy, which has specified the "reasonable, educated person" standard since 2006? (The language there was adopted to solve a burgeoning edit war over whether to say "reasonable person" or "reasonable adult".) WhatamIdoing (talk) 19:35, 16 September 2022 (UTC)

I'd have less of an issue with "educated" by itself, but "dedicated" to me just opens the floodgates to accusations of editors removing uncited and/or unverifiable material because they're not "dedicated" enough to locate and insert citations for sources themselves. When you also factor in, "all the means at the disposal..."...as I said above, what does that even mean? Even saying "a reasonable effort" opens the door to arguments over whether people who feel information should be cited are simply not 'making a reasonable effort'. DonIago (talk) 03:10, 17 September 2022 (UTC)

I think "dedicated" is a pretty fair description of Wikipedia:Reliable sources/Cost (from the top of PAYWALL section): information is verifiable even if you have to learn another language and travel around the globe at your own expense to check the source yourself.

What word would use you to describe someone who spends a lot of time, effort, and money to check a source? WhatamIdoing (talk) 15:41, 17 September 2022 (UTC)

Do you really think we should imply even a little that learning another language and booking an expensive and time-consuming international flight is a reasonable prerequisite to removing unsourced text on the grounds that it is unverifiable? It sounds to me like in practice, nothing could ever be unverifiable, because for all I know a source in China in Chinese verifying that claim exists and I could never prove such a source doesn't exist. "Unverifiable" becomes an empty concept and meaningless. To be honest I'm shocked and baffled at what's being considered reasonable here.

Oh, and of course that's another essay. Rather than altering V to make it sound like people need to go on literal expeditions before deleting text, what we should actually do is clean house on these ancient essays that make this outdated distinction. Crossroads ^-talk- 21:04, 17 September 2022 (UTC)

Content is verifiABLE under these circumstances (i.e., that you have to learn another language and book a flight to check it out with your own eyeballs).

VerifiABLE content is correctly CHALLENGEd every hour of the day. Saying that ____ is verifiABLE does not mean that you can't CHALLENGE it. The only thing you can't CHALLENGE is something that is already CITED. (If it is cited but has problems [e.g., not mentioned in the cited source, UNDUE, SYNTH, etc.], you don't technically CHALLENGE it; instead, you WP:DISPUTE it.)

This is another reason why we need to stop using these words interchangeably. This is the kind of discussion we want to encourage:

A: I don't think this is verifiable. I challenge it.
B: Of course it's verifiable! There's a source in China, written in Chinese, that verifies this claim.
A: Great. Post that citation right now, or I'm removing the text.

This is not the kind of discussion we want to see more of:

A: I don't think this is verifiable. I challenge it.
B: Of course it's verifiable! There's a source in China, written in Chinese, that verifies the claim. I've just added the citation to the article.
A: I don't care if you added the citation. It's still not verifiable. No matter what NONENG and PAYWALL and RS/Cost say, I shouldn't have to learn Chinese and travel to China to read the source myself. It's not possible for anyone to verify this if the source isn't reasonably easy for me to check.
B: But I'm already in China, and I already know how to read Chinese, and I just checked the source myself...

WhatamIdoing (talk) 15:46, 19 September 2022 (UTC)

Chinese isn't a problem as much because of Google Translate and the ability for discussions to elucidate meaning if needed. The bigger issue would be paywalls or travel expenses, etc. Altanner1991 (talk) 22:17, 19 September 2022 (UTC)

Setting your strawman aside, Newimpartial, the fact we are not an academic journal is a reason to be more strict about requiring citations, not less. Anyone is able to edit, and that means the misinformed and the axe-grinders. Crossroads ^-talk- 23:17, 16 September 2022 (UTC)

Why not do both—clarify the current policy, and encourage more citations? In other words I do not see how the clarification would imply reversing the burden of proof or mandating "we must engage in absurd amounts of effort before removing uncited material", but I would consider it if otherwise... As it stands, burden/policy would/should be the same, albeit some "clarification". Altanner1991 (talk) 22:13, 16 September 2022 (UTC)

I still think that Sentence 1 is a bad idea. Of course it is true that we want all content to be verifiable but that is not all that we want. The criterion should be that all content is either (1) uncontentious general knowledge of the "Paris is the capital of France" variety, or (2) verified. Some editor in good standing must have actually verified that the content appears in a reliable source and be willing to share the identity of the source. That doesn't necessarily mean that the source is named in the article (though that is highly desirable). "I found it in some reliable book and you can too" and "search in Scholar like I did" are insufficient to make something verified even though the editor is attesting to verifiability. Also, if an editor inserts something with no source and leaves Wikipedia without giving one, the information should be considered unverified until someone else comes up with a source. Zero^talk 08:56, 17 September 2022 (UTC)

@Zero0000, Wikipedia:When to cite says that inline citations are generally not necessary in these four circumstances:

General common knowledge
Subject-specific common knowledge
Plot of the subject of the article
Cited elsewhere in the article

You have agreed with the first. Do you agree with the other three? WhatamIdoing (talk) 15:43, 17 September 2022 (UTC)

In my opinion, the second would be like the first in that context, and the fourth is about citing and not verifiability. Not sure about the third. Altanner1991 (talk) 16:32, 17 September 2022 (UTC)

Regardless, my thought was the same about Sentence 1: it's a bad idea and does not seem to enhance any policy process or procedure. Altanner1991 (talk) 17:47, 17 September 2022 (UTC)

Again, that's an essay page. "Subject-specific common knowledge" sounds like a green light for "Undid revision...I don't need to cite sources for this claim because anyone familiar with the topic knows it." Crossroads ^-talk- 21:04, 17 September 2022 (UTC)

I've had those fun discussions. "If you knew anything about this subject then you'd know this is obvious and consequently it doesn't need a source..." and such. DonIago (talk) 23:08, 17 September 2022 (UTC)

It was shocking to me at first, but then I found it was common Wikipedia procedure. Here's to hoping the policy can become more clear! Altanner1991 (talk) 23:13, 17 September 2022 (UTC)

Again, Crossroads, this is the exact standard used in FAC articles. FAs are our best work, and all of that, remember? If this is what FAC supports, then it is extremely likely that the community supports this, too. Policy comes primarily from actual practice, not from one guy saying that there should be an inline citation after every sentence, including basic definitions that nobody would ever dispute.

DonIago (every time I see your name, I think of a stunning performance of Othello I saw years ago), I've been in those conversations, too, and I tend to find that the editor asking for the citation is correct. But once someone has challenged it, then it's 100% Wikipedia:Likely to be challenged and requires an inline citation, even though the citation might turn out to be the first page of every textbook in the field. WhatamIdoing (talk) 15:57, 19 September 2022 (UTC)

Regarding subject-specific common knowledge, I have to say that it would be more ideal if citations could be given, if anything just for propriety's sake, even if the content in question is already "common". Altanner1991 (talk) 22:06, 19 September 2022 (UTC)

This part of the policy is talking about the absolute, rock-bottom minimum requirements, rather than the ideal. WhatamIdoing (talk) 15:10, 20 September 2022 (UTC)

@WhatamIdoing: To answer your question about the four points: the second can provide some opportunity for arguments, but it describes a situation where something has been verified; that's what common knowledge within the group of subject-savvy people means. (It should still be sourced if challenged, though.) The third one is where everyone understands what the source is, so it isn't an example of something being unsourced. The fourth is a case where a source has been provided already. Zero^talk 08:15, 18 September 2022 (UTC)

@Zero0000, I think most editors use the word "verified" to mean something different. Imagine that we write "A computer processor is an electronic circuit that performs operations". Here are the words I'd use to describe various situations:

verifiable: It's possible to find a source that supports this, regardless of whether this has been done.
cited: Someone added a little blue clicky number at the end of the sentence, that contains information about a source that (allegedly) supports this.
verified: Someone actually checked a source to make sure this matches a source, regardless of whether the source is mentioned in the article.

The article says: "A computer processor is an electronic circuit that performs operations". There is no blue clicky number. You have no idea whether anyone has checked any sources. How would you describe this? (For myself, I would say: it's verifiable, it's uncited, and it's unknown whether anyone verified it.) WhatamIdoing (talk) 16:04, 19 September 2022 (UTC)

@WhatamIdoing: I used "verified" exactly as you define it. See my edit of Sep 7. Your example here is one of common knowledge to a large fraction of people without narrowly specialised expertise, so I would call it verified. "Common knowledge" implies "commonly known from reliable sources". It isn't a precise concept and will often need discussion and consensus forming. Zero^talk 14:18, 20 September 2022 (UTC)

So if I (and lots of other editors) remember this definition from years ago, then you'd call that "verified", even if nobody actually checked (e.g., during the year that the content was added or reviewed) to make sure that our memories are right?

I'm willing to call that "correct" and "undisputed", but I'm not sure that I'm willing to call that verified/actually checked. WhatamIdoing (talk) 15:13, 20 September 2022 (UTC)

You are stretching the meaning of "common knowledge". Something remembered from years ago is not common knowledge. "Paris is the capital of France" is common knowledge because everyone sees it stated in reliable sources on a regular basis, not because we were taught it in the first grade. Zero^talk 02:16, 21 September 2022 (UTC)

The definition of a computer processor is common knowledge because everyone with any connection to computer science knows it, not because we see it stated directly in reliable sources on a regular basis.

For that matter, when was the last time you remember actually seeing a reliable source that directly stated that Paris is the capital of France? High school geography class, maybe? Common knowledge is known by lots of people, but not necessarily oft-repeated. WhatamIdoing (talk) 19:00, 22 September 2022 (UTC)

Your distinctions I believe are correct, however the "operations" of the processor are not specific enough; this is grammatically problematic because "operations" has many meanings, for example, medical surgery; "mathematical calculations" would be more correct than "operations". I would only support "operations" if it had a reliably sourced citation, and even then the grammar is not ideal, but otherwise the statement is subject-specific common knowledge. Altanner1991 (talk) 00:06, 20 September 2022 (UTC)

In terms of good style, you might want to link to Logical operation in such a definition, but the use of words that have multiple possible meanings does not mean that the definition is incorrect or is not generally known to people. Consider, e.g., how many English speakers recognize the equation

E=mc^{2}

but have no idea what E, m, or c represent. Einstein's most famous equation does not require a citation just because people don't know what the terms mean. WhatamIdoing (talk) 15:21, 20 September 2022 (UTC)

@WhatamIdoing: What you are saying is too general, and it is hardly relevant to this policy. You say:

verifiable: It's possible to find a source that supports this, regardless of whether this has been done.

but that is totally senseless. If we expect that a reader is supposed to go to a library and look for sources by themselves, then what is a need in Wikipedia? That would be against a spirit of our policy. WP:V requires that

A. Everything that we are writing (except most obvious and trivial things) must be supported by some published source (i.e. that we didn't take it out of thin air, or didn't reproduced from memory, possibly incorrectly), AND
B. Wikipedia provides a reader with all necessary means to make sure that the information they are reading meets the condition A. That can be achieved by provided a reference to some reliably published source, and "reliably published" in this context means "non-volatile": i.e. that source is not expected to disappear in few days, and there is a serious reason to expect that after several years it will still be available by the reference provided. If this condition is met, a reader can go to a library, find the source and independently make sure the Wikipedia text matches the source.

In that sense, yours "cited" means "verifiable".

With regard to your "verified", you are also wrong: we are dealing with a process, not a state. Wikipedia texts are constantly changing, because users are making small additions, copy-editing, regrouping, etc, so even if some original text was matching a source pretty well, that may not be the case after years of editing. I can give you examples if you want. As a result, verification should be considered as a constant process of checking if some piece of text is still matching the source.

I already proposed once, and I maintain: we must clearly separate the concept of "verifiability" (which means a reader must be provided with an opportunity to independently make sure that Wikipedia text matches some published source, and a reader is not supposed to guess what source we were using, so the reference must be added), and the concept of "reliability of sources". And there is one additional consideration in support of this concept: we are considered amateurs, so we are not supposed to be smarter than our readers. Our readers are pretty capable of deciding which source are reliable and which are not, and some readers may be even more critical to sources, than our policy does> For example, I, as a Wikipedia reader, apply more strict criteria to Wikipedia sources, and I am frequently seeing some statements in Wikipedia articles that are supported by really lousy (but acceptable, per WP:SOURCES) sources. But that is ok, as soon as a reference is provided.

In summary, our policy should say:

You shalt provide a reference, and that reference must be to some stably published source (that is the first part of the policy);
The sources that you are using should be of the best possible quality, although, depending of a topic and a concrete case, their quality may vary (that is the second part of the policy).

And the second part is much less strict than the first one. Paul Siebert (talk) 22:04, 22 September 2022 (UTC)

Nope. The policy doesn't say B. The policy has never said B.

The first sentence of the policy defines the word. It says "other people using the encyclopedia can check that the information comes from a reliable source."

Notice that the definition doesn't say people "can check that the information comes from the specific reliable source cited at the end of that material". It just says "can check". There are no limitations on the methods used for checking. If someone "can check" the information at their local library, or someone "can check" it by searching through Google News until they find a suitable article, then someone "can check" – and that's all that's required.

See also the FAQ at the top of this page. Look particularly for " It must be possible to provide a bibliographic citation to a published reliable source", and consider what that word possible means. WhatamIdoing (talk) 02:03, 23 September 2022 (UTC)

Alas, it does. It says Readers must be able to check that any of the information within Wikipedia articles is not just made up. This means all material must be attributable to reliable, published sources.

How should we interpret it?

Two interpretations are possible:

1. We may add some text that, as we believe, is supported by some reliable source, and a reader is free to google for that source if they wish, or
2. We add some text and, unless it is totally obvious, add a reference to some reliably published source that directly supports what the added text says.

I think #2 is closer to the policy's spirit. However, we may probably start an RfC, if you think otherwise. Paul Siebert (talk) 02:16, 23 September 2022 (UTC)

The words "must be able to check that any of the information within Wikipedia articles is not just made up" cannot be interpreted as "we can write whatever we want, and if a reader may check it if they want", because reader's behaviour is not regulated by this policy. Obviously, these words should be interpreted as "we must do something to make it easier for a reader to check that any of the information within Wikipedia articles is not just made up", and I see the only way to do that: to provide a reference to some source. Paul Siebert (talk) 02:22, 23 September 2022 (UTC)

But "attributable" does not mean "already attributed". The -able means only that something is possible, that something is able to be done. It does not mean that it has already been done.

Compare:

"Readers must be able to check that any of the information within Wikipedia articles is not just made up."
"Readers must be able to check that any of the information within Wikipedia articles is not just made up – and they can check the information by looking at the citation at the end of the sentence".
"Readers must be able to check that any of the information within Wikipedia articles is not just made up – and they can check the information by using a web search engine".
"Readers must be able to check that any of the information within Wikipedia articles is not just made up – and they can check the information by visiting a library and asking for help".

The requirement that people "must be ABLE to check the information" does not have any restrictions on the methods that people are allowed to use in the process of checking the information. If people (other than the editor who first posted the information) are ABLE to find that information in a reliable source through any means whatsoever, then the material is verifiABLE and attributABLE.

Put another way: incorrect citations don't make it impossible to verify good content. Paris is still verifiably the capital of France, even if someone cites Facebook's main page at the end of the sentence. WhatamIdoing (talk) 05:58, 23 September 2022 (UTC)

Another line to follow:

The policy requires that everything, without exception, including all the material you would describe as "totally obvious", be verifiABLE. There are no exceptions to this rule, not even for water-is-wet material. If it's in the mainspace, (some) people must be ABLE to find (through any and all means) a reliable source that says the same thing.

This policy – additionally and separately from the above – requires that we add inline citations to said sources for four classes of material.

The spirit of this policy, as well as the community's practice, also encourages editors to add inline citations generously, even when an(other) inline citation is not necessarily required.

But if an inline citation is clearly not required by the separate enumeration of the four classes of material, then the policy only requires that people be ABLE to find a source if they want to. WhatamIdoing (talk) 06:05, 23 September 2022 (UTC)

The problem is that the same policy says:

Any material that needs an inline citation but does not have one may be removed..

In connection to that, I am wondering how all what you say is consistent with this clause.

Does it mean that some verifiable information needs online citation, whereas other does not? And if some verifiable (but not absolutely obvious) information does not need citations, how can we determine that no citation is required? Paul Siebert (talk) 14:50, 23 September 2022 (UTC)

No. It means that 100% of material must be verifiABLE, and that merely being verifiABLE (i.e., but not already cited) does not protect the material from being challenged.

Being verifiable does not protect material from being challenged. Why should it? How could it? Just because you know that ____ is common knowledge within a field doesn't mean that the kid doing RecentChanges patrolling knows that.

The policy sets up this conversation:

A: "CPUs are a kind of computer processor." Hmm, everybody who knows anything about computers knows that. It's not a quotation, it's not about BLPs, it's not Wikipedia:Likely to be challenged, and I just added it, so it of course hasn't already been CHALLENGEd, so I won't bother citing it.
B: What? I've never heard of that. Granted, I don't know anything at all about computers, but I think that should be checked. I hereby CHALLENGE this possibly wrong information!
A: Here's the first chapter of the last 15 textbooks published for teenagers about computers. Are you satisfied now?
B: Thanks for adding the citations. I didn't know any of that.

The material was verifiABLE the entire time. As a result of the CHALLENGE, it is also now CITED. WhatamIdoing (talk) 15:09, 23 September 2022 (UTC)

You haven't convinced me. Your interpretation implies that all materials can be subdivided onto three categories.

Absolutely obvious statements (that are deemed a priori verified, such as "Paris is a capital of France);
Verifiable statements that need no inline citations;
Verifiable statements that need an inline citations.

That division follows from the clause cited by me ("Any material that needs an inline citation but does not have one may be removed"). This quote directly says that the absence of an inline citation is some abnormal situation that must be fixed (either by adding the citation or by removal). Moreover, the preceding sentence ("...any material whose verifiability has been challenged or is likely to be challenged, must include an inline citation to a reliable source that directly supports the material..") clearly outlines what materials need an inline citation: ANY material that is "likely to be challenged" needs it. Obviously, "likely to be challenged" means "non-obvious". "Paris is a capital of France" is an obvious statement, and it is unlikely to be challenged, so it beliongs to the first category. "Uncertainty principle was discovered by Heisenberg" is not absolutely obvious, and it is likely to be challenged, so it falls into the the third category (i.e. it needs an inline citation). I cannot imagine a non-obvious statement that is unlikely to be challenged. I would say "non-obvious statement that is unlikely to be challenged in an oxymoron".

Therefore, all information can be subdivided on TWO categories: obvious and non-obvious, so your second category ("Verifiable statements that need no inline citations") does not exist, and "verifiability" (according to the policy) implies that a reader must be provided with an online citation that directs them to the source that we were using for each particular statement.

If you disagree (and it is very likely), than this dispute may be of interest to a broader audience, so I propose to start an RfC. It would be interesting to know what our community thinks on that account. Paul Siebert (talk) 15:43, 23 September 2022 (UTC)

@Paul Siebert, I claim only two categories of permissible content for articles:

verifiable statements that need no inline citations, and
verifiable statements that need an inline citations.

The first category includes "absolutely obvious statements" and statements with strongly implied sources (e.g., simple plot summaries). It might not include much else, but whether you believe something is WP:LIKELY to be challenged is a judgement call, and different editors have different judgement. "The capital of France is Paris" certainly falls into this category. For example, I would say that a summary statement like "The 2020 US presidential elections were conducted in the context of the COVID-19 pandemic and concerns about misinformation" also falls into this category, but I would also not fault others if they disagreed with me.

The second category includes the four points named in this policy. An example of that would be: every single sentence presently in Immunoglobulin D, explicitly including the first sentence. The caption on the lone image might not need a citation, but I would also not be surprised if a good editor cited it anyway. (Nobody ever has challenged it, and it is not actually Wikipedia:Likely to be challenged, even in an FA-level review, but it is still the kind of semi-technical statement that we often cite anyway.)

The definition of verifiability is not dependent on the rule we added much later about being allowed to remove uncited information regardless of whether it is verifiable. You can remove uncited verifiable information; you can remove uncited unverifiable information. You do not need to know whether the information is verifiable when you decide to challenge it. If you want to challenge something, the only thing you need to know is whether it's cited. If it's cited, you can't WP:CHALLENGE it. If it's uncited, you can WP:CHALLENGE it – even if it's a claim that you think is perfectly obvious (though I hope you wouldn't, in that case) WhatamIdoing (talk) 15:26, 24 September 2022 (UTC)

If permissible content for articles fall into two categories, and one of them has a clear description, then there is absolutely no reason to describe the second category. Therefore, if you remove all redundancies from your statement, it will look like this:

"I claim only two categories of permissible content for articles:

statements that are absolutely obvious to a reader with average knowledge in the field (something similar to PHOSITA) and statements with strongly implied sources (e.g., simple plot summaries); these statements are deemed verifiable by default, so they need no inline citations;
all other statements need inline citations to be considered verifiable.

That's it. Paul Siebert (talk) 17:54, 24 September 2022 (UTC)

More precisely, "All other statements need inline citations to comply with this policy", but not technically "to be considered verifiable". Verifiability is a statement about the relationship between reliable sources in the real world and the statement in the article. If a reliable source says ______, then the statement that ______ is verifiABLE before I cite it.

The inline citation is not needed to make the statement become verifiable/check-able in a reliable source; it is needed to make the verifiable content comply with the parts of the policy that require some (i.e., most) verifiable content to be not merely verifiable but also cited. WhatamIdoing (talk) 21:23, 24 September 2022 (UTC)

See my comment in the next section. How can we make sure some statement is verifiable if no sources have been presented? Paul Siebert (talk) 23:33, 25 September 2022 (UTC)

For "common knowledge", see the first paragraph of Wikipedia:Common knowledge. What someone thinks is "common knowlege" may not be such. In one of the raging battles over "common knowledge" on Wikipedia many years ago, one editor made what he thought was a telling argument by asking me if I thought we needed to provide a citation for the fact that George Washington had wooden false teeth. I said no, that we needed to cite a source on what his false teeth were actually made from. So, IMHO, "common knowledge" must be narrowly defined, and is always subject to challenge. Donald Albury 18:53, 18 September 2022 (UTC)

Common knowledge can however generally be validated through consensus alone, since that is the idea, (assuming editors could represent a common range of the population) unlike verifiability, which is in principle rooted in external sourcing, and expert opinions. Altanner1991 (talk) 23:25, 19 September 2022 (UTC)

I agree in principle, but in practice it's usually faster and easier to dump a source in the article than to argue with someone about whether his personal unfamiliarity with something means that it's not known to most people.[1] WhatamIdoing (talk) 15:25, 20 September 2022 (UTC)

Indeed—any narrowing is probably a good idea. Altanner1991 (talk) 03:25, 24 September 2022 (UTC)

Who must be able to verify?

There is an unasked question in this thread, so let me ask it: Who, precisely, must be able to verify the material? - Me? You? Everyone? Someone else? Blueboar (talk) 12:39, 24 September 2022 (UTC)

In my opinion the answer obvious: "me." Others may disagree, but they are wrong. - Butwhatdoiknow (talk) 15:19, 24 September 2022 (UTC)

All statements must be verifiABLE in a reliable source (not necessarily the one, if any, cited in the article) by someone ("at least one human") who is not the editor that added that information to the article.

Per WP:PAYWALL and Wikipedia:Reliable sources/Cost, this someone must be willing to invest whatever time and effort is necessary. As a practical matter, we usually expect this someone to be someone who either has an extraordinarily high interest, or someone who has a lower than usual cost (e.g., student at a university that contains this source, and the student already knows how to read the language that the source is written in). WhatamIdoing (talk) 15:30, 24 September 2022 (UTC)

Sources should be verifiable by 'anyone' (or at least, any literate adult) with the time, motivation, and financial means (if necessary) to do so. We are writing for readers. In a source we ourselves don't consider 'reliable'. If readers cannot verify things, we shouldn't be including them. AndyTheGrump (talk) 15:42, 24 September 2022 (UTC)

@Blueboar: In patent law, there is a term Person having ordinary skill in the art. It seems we should apply the same approach: if some person expresses interest in the Uncertainty principle, it is reasonable to assume they have some minimal basic knowledge in physics, so some simple statements need no inline citations. If some person wants to read about Paris, it is highly likely that they already know that Paris is a capital of France (or, at least, they are familiar with such a term as "geography handbook"), so the statement "Paris is a capital of France" is obvious to a reader with average knowledge in geography. However, the information about population size of Paris is not obvious, so it should be supplemented with an online citation.

In other words, we should not assume that our reader is absolutely ignorant, we assume that different articles have different audience, therefore, the answer to your question should be: "a reader with average knowledge in this concrete field".

Indeed, consider the "Uncertainly principle" article as an example. The "Matrix mechanics interpretation" section IS verifiable, because it cites a source. However, this source is the Landau&Livshitz book (which to 99.9% of Wikipedia audience is like Greek). Are you sure that an average Wikipedia reader is capable of independently verifying, using that book as a source, that the section was not made up? Obviously, no. Only a person with an average knowledge in physics can do that, and that is a category of readers that article is intended for. In contrast, the Game of Thrones article has a totally different audience (at least, people with average knowledge in physics is hardly a majority of it), so the answer to the question "who must be able to verify?" is different in that case. Paul Siebert (talk) 18:27, 24 September 2022 (UTC)

I asked this question because I don’t think we address it clearly in the policy. A lot of editors seem to interpret the policy as: “I must be able to verify the material” (especially in challenge situations) … and they are not happy to be told “No, it can still be verifiable… even if you can’t do so, someone else can.” We should perhaps clarify this. Blueboar (talk) 20:22, 24 September 2022 (UTC)
That would probably require either re-writing the first sentence, or adding a ==Definitions== section (perhaps modeled on the WP:GNG).
If you the first appeals to you, then you are probably looking for something like this:
- Original: In the English Wikipedia, verifiability means other people using the encyclopedia can check that the information comes from a reliable source.
- Change: In the English Wikipedia, verifiability means it would be possible, with enough time, effort, and resources, for at least one person (not counting the Wikipedia editor who added that material) to verify that at least one reliable source also contains that claim.
If the second appeals, then it might look something like this:
"In the English Wikipedia, verifiability means other people using the encyclopedia can check that the information comes from a reliable source. This means:
- "Other people" means people other than the editor who added the material originally. If you are the only person in the world who knows something, or if no sources saying this are available to the general public, then the material is not verifiable. "Other people" does not mean everyone or every Wikipedia editor.
- "Can check" means that it is possible for some other people to compare the material in an article against reliable source(s). This is not limited to any cited sources. For example, if the cited source is difficult to access, or if no source is cited, but other reliable sources also make the same claims, then people still "can check" that the material is found in other reliable sources. (Not citing sources may be a violation of this policy, but it does not technically make the material be unverifiable. Uncited material can be tagged, removed, or otherwise CHALLENGED regardless of whether it is verifiable.)
- "Comes from a reliable source" means that reliable sources make the same claims.
I suppose that "comes from" is really the wrong wording anyway. The point of verifiability isn't to WP:SAYWHEREYOUGOTIT (even though that is a very good idea). If verifiability depended upon knowing the exact source that the material came from, then we'd have all kinds of problems, and should start by deleting Template:Obsolete source and Template:Better source needed. The point is to be able to find the material in any reliable source. It doesn't have to be the exact reliable source that the material originally came from. WhatamIdoing (talk) 21:43, 24 September 2022 (UTC)

In practice this would discourage people from CHALLENGEing material, which is really bad because lots of wrong and outdated information exists on Wikipedia. We want people to challenge material if it is wrong. And citation templates have quote sections if needed for a reason. The sort of claim that exists only in a source that can't be found online, on Google Scholar, or through our Resource Exchange should be considered UNDUE at best. Crossroads ^-talk- 22:38, 25 September 2022 (UTC)

It says " Uncited material can be tagged, removed, or otherwise CHALLENGED regardless of whether it is verifiable." How exactly does explicit permission to challenge uncited content result in discouraging people from challenging uncited content? WhatamIdoing (talk) 00:00, 26 September 2022 (UTC)

Oh, and we don't want people to challenge material "if it is wrong". We want people to challenge material "if it is uncited and they have any genuine doubt, no matter how small or weak, that it might not be verifiable in a reliable source".

We do not, frankly, want anti-science folks going around to articles on, e.g., evolution, climate change, or the pandemic and saying "I challenge this because science is wrong". Or imagine the mess we would have at Attempts to overturn the 2020 United States presidential election if "all those sources are wrong, because YouTube says Trump really won" were an acceptable battle cry. We want people to challenge uncited content when they believe that reliable sources don't say that, not because some gullible people believe liars. WhatamIdoing (talk) 00:06, 26 September 2022 (UTC)

Well, now I understand what you mean.

Consider this scenario:

I write, "Group II introns possess only a very few conserved nucleotides, and the nucleotides important for the catalytic function are spread over the complete intron structure", and I believe that this statement is verifiable, because I (or somebody else) can find some source that supports this statement. Am I right per our policy? Formally, yes: if I am 100% sure that some source directly supports that statement, that statement is considered verifiable. However, in what situation can I be 100% sure such a source exists? Only if I already found that source and compared what it says with what I wrote. (If I don't have that source before me, and I am writing from memory, I cannot be 100% sure that the text that I am adding is supported by sources.)

That means, to make sure some statement is verifiable, I must verify it.

Verifiability of any statement, except absolutely obvious statements or statements with strongly implied sources, can be demonstrated only after that statement have been verified (a needed source has been presented).

In other words, if a burden to prove verifiability lies not on the editor who wrote the text, then the editor who adds that text cannot be completely sure the text they added is verifiable. But that means the verifiability criterion is not met.

However, if a burden of proof rests with those who adds the text, their mission is consideered accomplished only after the editor has verified the statement (for, as I already demonstrated, it is not possible to prove that some source exists until such a source has been presented).

Therefore, the answer to you question is:

I must be able to verify the material” (especially in challenge situations). If I cannot do that, but I believe that other people may potentially do that, that means the statement I am adding is not "verifiable", but "potentially verifiable". However, the policy we are discussing does not say that Wikipedia content must be "potentially verifiable". It says "verifiable".

Paul Siebert (talk) 23:31, 25 September 2022 (UTC)

That's an interesting argument, but it elides the distinction in English between verifiable and verified. The criterion in policy is not that an editor be 100% sure such a source exists, by verifying it - it is simply that the source must exist. What is more, your argument seems to rest on an odd interpolation concerning the burden of proof, which completely defies the way policy currently works. Currently, and outside of special situations like BLPs and medical content, it is fine to add material which is verifiable even if it is not backed up by a citation. It is also fine to remove material that lacks citations, even if it is verifiable, so long as it has not actually been verified (through a citation). The idea that "verifiable" somehow means "already verified by a citation" defies both plain English and Wikipedia policy logic. Newimpartial (talk) 23:55, 25 September 2022 (UTC)

Paul, (accurate) statements about conservation in Group II introns would verifiable regardless of whether you cite them. Whether a citation is required is a completely separate consideration from whether the material is verifiable.

The questions work like this:

Is this verifiable?
- Did this material appear in any one (1) reliable source, in any language, anywhere in the world? Oh, of course it did: I remember seeing this in a book a few years ago, and an article in this morning's paper reminded me of it.
Okay, then it's verifiable. Do I need to cite it?
- Is this a direct quotation? No, so maybe I don't.
- Is this contentious matter about BLPs? No, so maybe I don't.
- Has anybody already challenged the material that I'm writing? No, so maybe I don't.
- Is this the kind of thing that editors tend to challenge? Ooooh, yeah. Technical, scientific, and statistical content is definitely likely to get challenged. So an inline citation is required, and while the policy does not technically require me to add the citation in the same edit as adding the information, I might as well add the citation now, instead of waiting around for someone to challenge it later, when I might have forgotten about it.

Would the material still be verifiable even if you didn't cite it right away? Of course it would. Verifiability is about the relationship between the material and the real-world sources, not about whether you cited it.

But not citing it would separately violate the requirement to cite all content that you personally judge to be WP:LIKELY to be challenged. It is a best practice to provide those citations before the content is challenged (preferably immediately, or with a delay of no more than a few minutes). WhatamIdoing (talk) 00:26, 26 September 2022 (UTC)

Again, per Webster's dictionary, "verifiable" means "capable of being verified".

Imagine I have a box, and I believe there is a cat in it. Is the statement "There is a cat in this box" verifiable? Yes, if we use the term "verifiable" in its colloquial meaning. However, wouldn't it be more correct to open the box and to check if the cat is there before you write in some Wikipedia article that there is a cat inside of it?

Wikipedia is not a collection of information that, according to our belief, is true. If you believe some information is verifiable, that your belief should rest on a more solid ground than your memory. Before writing that "Pi number is roughly equal to 3.14159265395...." it makes sense to open a source and check if that is correct. If you add some information to Wikipedia, you are supposed to be responsible for its accuracy and veracity.

I see two approaches to verifiability in Wikipedia:

1. I added some information that, as I believe, can be found in some reliable sources, but I am too busy (or lazy) to provide these sources. You are free to do that if you want. This approach is superficial and unsatisfactory. Remember, Wikipedia has a very bad reputation of fact checking and accuracy, and by pushing this approach we are making it even worse.
2. I added some information that does not seem absolutely obvious. I found this information in a source X, and I provide a reference to this source, so everybody who has access to that source will be able to check that my statement is correct. Therefore, the information that I added is verifiable. This approach is much more serious, and its consistent implementation will improve Wikipedia's credibility.

Formally speaking, any statement that is not false can be verified, the difference is in the amount of time and efforts needed for that. For example, when I was a graduate student, my supervisor told me a story about Alchemy that seemed very realistic, and many experts with whom I discussed it told me that this story may be true. I was trying to verify it for several years, however, extensive googling yielded no sources. Does it mean that that fact is not verifiable? No, it is highly unlikely that my supervisor invented that story, so it is quite likely that the source does exist, maybe, as a printed version (not in a digital form). Can that story be considered verifiable? Formally speaking, yes, I still have a hope to find the source. However, from the point of view of the current policy it should not be considered verifiable until some source has been found.

Therefore, if an average reader of some concrete article is capable of independently verifying some statement in a reasonably short time, and that requires a reasonable amount of efforts that statement is considered verifiable. Otherwise, it is not. The statement that Paris is a capital of France can be verified in one mouse click, which means it is verifiable. In contrast, the statement

Spliceosome is likely derived from Group II self-splicing ribozymes.

can hardly be verified equally easily by an average reader of the Ribozyme article, so it would be correct to conclude that statement does not meet verifiability criteria. However, the statement:

Spliceosome is likely derived from Group II self-splicing ribozymes.^[1]

directs a reader to some concrete source that may help them to independently verify it, so the second statement does meet verifiability criteria. Paul Siebert (talk) 02:38, 26 September 2022 (UTC)

"Capable of being verified" does not imply "capable of being verified by looking at a specific cited source".

It might make sense to double-check the digits of pi in a source before adding them to an article, but if you somehow manage to produce the correct numbers merely by guessing, or mashing your keyboard, then it is still verifiable.

Verifiability is not about what the editor does. Verifiability is about whether Wikipedia content = reliable sources.

Editor reads about pi on a website, finds that it starts with 3.14159, and writes that in the article. The website the editor was reading is not a reliable source, but the editor cites it anyway. However, this fact is "capable of being verified" in reliable (e.g., mathematical) sources, so it's verifiable information.
Editor remembers pi from geometry class and writes 3.14159 in an article. The editor doesn't cite any sources. However, it is "capable of being verified" in reliable sources, so it's verifiable.
Editor carefully researches and obtains copies of the six most popular university-level geometry textbooks. They all agree that pi begins with 3.14159, so the editor writes that in an article. Unfortunately, the editor has a heart attack before citing any of the sources and does not return to Wikipedia for months. However, the information is "capable of being verified" in reliable sources, so it's verifiable.
Editor reads about pi on a website, finds that The Bible says pi is exactly 3, and puts this in the article, complete with nicely formatted citations. However, this false claim is not "capable of being verified" in reliable sources, so it's not verifiable.

So long as the editor is writing accurate information that is present in some reliable sources, then the editor's behavior cannot make the material unverifiable.

The editor's behavior can make the article "violate this policy", but the editor's behavior cannot make that material "become unverifiable". There is nothing that an editor can do that would remove 3.14159 from reliable sources all over the world. Removing the numbers 3.14159 from all of the reliable sources all over the world is what would actually be necessary to make "pi = 3.14159" become unverifiable. Wikipedia editors can't do that. They cannot make verifiable information become unverifiable. WhatamIdoing (talk) 03:05, 26 September 2022 (UTC)

The problem is that my statement "Pi number is roughly equal to 3.14159265395...." is NOT verifiable. If you check sources, you see that the correct value is 3.14159265359...

Therefore, by typing "3.14159265395" from memory and by not having bothered to provide a source and verify it I added a wrong information, and you haven't noticed that. If you haven't noticed it, what is a reason other users or readers will? That is a demonstration of a difference between a statement that some editor believes to be verifiable, and a statement that IS verifiable.

There is a big difference between the statement

"If I remembered this fact correctly, this information is correct, and I am sure it may be possible to find a source that confirms it" , and

"I checked this information, it is correct, and it may be found in a source X and possibly in other sources."

The first statement expresses nothing but my belief that that information is verifiable. The second statement describes some information that IS verifiable.

Remember my example: the story about Alchemy is verifiable according to your criteria (because it is, at least theoretically, possible to find some source that confirms it), but it is not verifiable according to my criteria (because one day of googling yielded no sources). And here we come to the Russel's teapot analogy: you literally say that if someone believes that it may be possible to find a source that supports a statement X (for example, you remember you saw it is some book when you were a child), then "X" is verifiable. However, this your statement is non-falsifiable: if I claim "X" is not verifiable because I was unable to find a source after two days of googling, you may argue that the source could be found after 3 days. If I still was uncapable of finding any source after four days search, you may argue that the fifth day could be successful, and so on and so forth. That is nonsense. Paul Siebert (talk) 03:41, 26 September 2022 (UTC)

You may have noticed that I shortened my reply to 3.14159, because I was certain of those digits, and I didn't want to proofread yours. ;-)

Even if you do look it up, you could mistype it. Information that does not match what a reliable source says is not verifiable, no matter what pains the editor has taken to check in advance.

Verifiability isn't really about beliefs. If the material can be found in a reliable source, then it's verifiable. If it's not, then it's not. It doesn't matter what any editors believe. And, once again, material that is uncited can be CHALLENGEd even if it is verifiable. Here is a list of circumstances for when you can challenge material:

There is no inline citation for that material.

The complete list of circumstances under which you are not allowed to challenge material for which no inline citation has been provided are:

If you've personally been blocked or banned from editing (that article; that subject; anything).

That's it! That's the whole list.

If it is verifiable information with no inline citation to a reliable source that directly supports the material, and you're allowed to edit, then you can challenge it.
If it is unverifiable information with no inline citation to a reliable source that directly supports the material, and you're allowed to edit, then you can challenge it.
If you have no idea whether the information is verifiable or not, but you know that it has no inline citation to a reliable source that directly supports the material, and you're allowed to edit, then you can challenge it!

You don't have to spend even five seconds asking your favorite web search engine about it. If there is no inline citation to a reliable source that directly supports the material, then you can challenge it.

The alchemy story may or may not be verifiable, but the minute someone challenges it, the material is assumed unverifiable unless and until proven otherwise. As a result of this rule, we regularly remove verifiable information. Perhaps years from now, you will run across a reliable source that tells the same story about alchemy. At that point, you will realize that the material really was verifiable all along, and we made a mistake by removing it. But that's okay, because it's also possible that no such source exists, and our rules are optimized for removing potentially unverifiable information, instead of keeping challenged information. WhatamIdoing (talk) 04:08, 26 September 2022 (UTC)

Formally speaking, any statement that is not false can be verified – This is not true in the real world, and it is also not true for Wikipedia's purposes, since we require verification in published reliable sources. I fixed a cheese sandwich for lunch last Thursday. This is not a false statement, but nobody can verify that.
if an average reader of some concrete article – This policy does not limit itself to average readers. PRIMARY limits the use of primary sources to "straightforward, descriptive statements of facts that can be verified by any educated person with access to the primary source but without further, specialized knowledge", which is similar to the concept of a Person having ordinary skill in the art, but information can still be verifiable even if it violates that policy, even if it is only understood by experts, etc.
in a reasonably short time, and that requires a reasonable amount of efforts – This policy has never supported this idea, unless you think reasonable efforts include things like buying expensive sources, hiring translators, and traveling to specific locations.

We do not require content to be easily verified by an average reader. We do not require content to be easily verified by anyone. But I think that the bigger disconnect is that the policy defines verifiability as something like an intrinsic property of the content. Content is verifiable if a reliable source said that. Content is unverifiable if a reliable source didn't say that.

We use citations as a sort of cheat sheet towards determining whether a statement is verifiable, but the presence of a citation doesn't make the material become verifiable, and the absence of a citation doesn't make the material become unverifiable.

Using your example above:

Spliceosome is likely derived from Group II self-splicing ribozymes. – verifiable, because the real world has sources saying this
Spliceosome is likely derived from Group II self-splicing ribozymes.^[1] – verifiable, because the real world has sources saying this
Spliceosome is likely derived from Group II self-splicing ribozymes.^{[bad ref]} – still verifiable, because the real world still has sources saying this, even though, unfortunately, none of them got cited here

You seem more concerned with the epistemology of verifiability instead of whether the material is verifiable. Your approach answers questions like "How can the reader be sure that verifiable material truly is verifiable?" instead of "Yes or no: Does the real world have a reliable source that says this thing?" WhatamIdoing (talk) 03:34, 26 September 2022 (UTC)

Do you realise that the core idea that you are pushing is that citations are optional? That turns Wikipedia into a collection of garbage. That is totally against the spirit of Wikipedia. Paul Siebert (talk) 03:48, 26 September 2022 (UTC)

No, I'm not. I'm arguing that the problem is not with unverifiable material, but with uncited material. I like to believe that editors are capable of using jargon with precision, and that being precise about where the problem really lies will help us solve the problem. Using the term unverifiable can lead to 'tis/'tisn't arguments. Using the term uncited does not have any ambiguity about it: The options in that discussion are "You're right; There's no citation at the end of that sentence" or "I'm an argumentative idiot who needs to be blocked". WhatamIdoing (talk) 04:12, 26 September 2022 (UTC)

Maybe this will help:

Three situations for the same statement
Statement	Is it verifiable?	Is it cited?	Verifiability demonstrated?
Spliceosome is likely derived from Group II self-splicing ribozymes.	Yes	No	No
Spliceosome is likely derived from Group II self-splicing ribozymes.^{[good source]}	Yes	Yes	Yes
Spliceosome is likely derived from Group II self-splicing ribozymes.^{[bad source]}	Yes	Yes	No

The statement is verifiable every single time, even though that's not demonstrated in two of the cases. It is verifiable every single time because it is possible (as demonstrated in the middle row) to find a reliable source that says this. The difference between these situations is not whether the statement is verifiable. The difference is whether the statement is cited, and whether those citations demonstrate verifiability. WhatamIdoing (talk) 03:48, 26 September 2022 (UTC)

Again, if you write some text based of what you remembered, there is a risk that you remembered it incorrectly (which means, you just believe your text is verifiable). To make sure you typed a correct information, it is necessary to have some source before you when you are typing.

If you are writing some Wikipedia text and you believe there is some source that directly supports it, you may be right or wrong, but to make sure such a source really exist, you must have it before you. Otherwise, the text that you are typing is not verifiable, but it is just believed to be verifiable.

In other words, when you are typing some information, which, as you believe is correct, and, as you believe is supported by some reliable sources that as you believe may be found by some other users, that is not sufficient for a conclusion that your text is verifiable: the only legitimate conclusion that we may make is that when you were typing that text, you believed (without any evidences) that that text was verifiable. But Wikipedia cannot be based on user's belief.

The only case when you can claim your text is really verifiable is when you have a source. But if you already have a source, why cannot you add a citation? Paul Siebert (talk) 04:05, 26 September 2022 (UTC)

An editor's beliefs are completely irrelevant. What an editor claims about the text is completely irrelevant. The only case when the material is verifiable is when the real world has a source that says the same thing. WP:V requires material to be verifiable in real-world reliable sources. It does not require editors to believe that the material is verifiable. It does not require editors to claim that the material is verifiable. The requirement is that 100% of material in 100% of articles be verifiable. WhatamIdoing (talk) 04:15, 26 September 2022 (UTC)

How can you make sure a real world has a source if you don't have it before you? If you are writing some piece of text and you don't have a source that directly supports it, than your text reflects your belief that such a source exist.

The only way to make sure a real world has a source that supports your claim is to have it before you. But if you already have it, why not add a citation? Paul Siebert (talk) 04:21, 26 September 2022 (UTC)

Why not add a citation? Maybe you should start by finding a time when I added content and didn't also add a citation, so that we can talk about specifics instead of hypotheticals. That said, whether I cite the material I add is actually irrelevant to the question of whether the material is verifiable. "Why not cite?" is a question about whether the material is cited, and whether it is demonstrated to be verifiable. It is not a relevant question for whether the material is verifiable. WhatamIdoing (talk) 04:29, 26 September 2022 (UTC)

And, if we define "verifiable" as "something that can be independently verified by at least one person using previously published sources", then we should define "likely to be challenged" as "something that may be challenged by at least one good faith person". That means EACH non-obvious statement is likely to be challenged, so each of them needs an inline citation. Paul Siebert (talk) 04:18, 26 September 2022 (UTC)

I basically agree with you, and so does Wikipedia:Likely to be challenged.

The two details I'd clarify are that "may be challenged" is not just a remote hypothetical, but a plausible likelihood, and that "good faith" is a condition that should not be inquired into too closely. Challenges from POV pushers should produce inline citations, too. WhatamIdoing (talk) 04:26, 26 September 2022 (UTC)

POV-pushers may request citations even if no citations are needed. If consensus does not support POV-pusher's demand, it may be ignored. If consensus supports POV-pusher's argument, that person should not be considered (in this concrete context a POV-pusher is acting as a good faith person and should be treated accordingly). --Paul Siebert (talk) 17:10, 26 September 2022 (UTC)

As soon as a citation is requested, then a citation is needed, per this policy. WP:V says that when uncited content is challenged (e.g., by adding {{citation needed}} to request a citation), then a citation is required. WP:V does not say that consensus has to support the request.

I have seen one widely agreed-upon rejection of a fact tag by consensus alone: it was at the end of a sentence saying that human hands normally have five digits. I have seen editors point out, e.g., that the whole paragraph is supported by the same citation at the end of that paragraph, or that WP:LEADCITE is being followed and the claim is cited in the body of the article, but I don't remember seeing other cases in which content that was not cited elsewhere was deemed unnecessary to cite merely because we say so. The rule is to WP:Let the Wookiee win. (Someday we'll make that a blue link, and editors will look back and wonder how it was linked in 2015 but the page creation appears to be from many years later.) WhatamIdoing (talk) 14:45, 27 September 2022 (UTC)

That "digits" example is a "The sky is blue" type of fact that requires no RS, and a local consensus (that develops in reaction to the POV pushing of fools) is enough to establish that the statement does not need a source. -- Valjean (talk) (PING me) 14:55, 27 September 2022 (UTC)

We have not traditionally followed that advice, because a Wikipedia:Local Consensus is frequently out of step with the policy. You don't want a WP:FACTION to overrule basic rules about citing sources. Instead, editors appear to be more comfortable with editors invoking IAR in incontestable cases. WhatamIdoing (talk) 15:13, 27 September 2022 (UTC)

Good points. My personal preference is "when in doubt, cite it". One too many citations is better than too few. One cannot satisfy everyone, and if a point is constantly being questioned, it's better to just cite it and end the disruptions. For example... even if that means we deviate from our typical practice of not including refs in the lead. Some articles (and thus their leads) are so contentious that the lead does need refs. While many editors will not bend at all, resulting in much wasted time at various drama boards, I'm the type of editor that would rather just defang the issue by providing a source. That ends the disruption. Individual cases are usually dealt with at the local level, and I'll still abide by the consensus. If that local consensus really does violate PAG (You're right about that!), I may pursue dispute resolution. -- Valjean (talk) (PING me) 16:11, 27 September 2022 (UTC)

References

^ Fica SM, Tuttle N, Novak T, Li NS, Lu J, Koodathingal P, et al. (November 2013). "RNA catalyses nuclear pre-mRNA splicing". Nature. 503 (7475): 229–234. Bibcode:2013Natur.503..229F. doi:10.1038/nature12734. PMC 4666680. PMID 24196718.

A modest proposal

Since some editors clearly feel there is not enough emphasis on actually citing sources not always being necessary, how about we rewrite WP:V's lead to make it even clearer? For example:

In the English Wikipedia, verifiability means other people using the encyclopedia can check that the information comes from a reliable source, although you don't have to provide that source. Wikipedia does not publish original research, meaning that any source anywhere in the world must have said it previously even if not currently cited. Its content is determined by previously published information rather than the beliefs or experiences of editors, as long as it has been published anywhere on Earth even if you don't cite it. Even if you are sure something is true, it must be verifiable before you can add it, meaning someone can go search for and find a source, not that you actually have to cite it. If reliable sources disagree, then maintain a neutral point of view and present what the various sources say (citing the sources is optional), giving each side its due weight.

All material in Wikipedia mainspace, including everything in articles, lists, and captions, must be verifiable, even if not cited to a source inline. It is only quotations, and any material whose verifiability has been challenged or is likely to be challenged, that must include an inline citation to a reliable source that directly supports[2] the material; anything else does not need a citation. Any material that needs an inline citation but does not have one may be removed, but if one is not needed, you do not need to cite a source. Please immediately remove contentious material about living people that is unsourced or poorly sourced, but for other topics that doesn't matter.

There we go. Now we'll finally be able to keep all that wonderful content from drive-by randos who think 'anyone can edit' means that they can just post whatever they think is true like Twitter, and stop those meanies who think we should cite sources like any respectable work. Crossroads ^-talk- 22:59, 12 September 2022 (UTC)

Amusing, I suppose, but I would oppose that. Cullen328 (talk) 23:04, 12 September 2022 (UTC)

Yeah. Funny, but no. Huggums537 (talk) 00:07, 13 September 2022 (UTC)

Yes although it might need a "humorous" tag. Altanner1991 (talk) 12:05, 14 September 2022 (UTC)

You must verify it in a reliable source before you add it

Latest comment: 2 years ago39 comments8 people in discussion

Proposal withdrawn as it may not be a real improvement.

The following discussion has been closed. Please do not modify it.

Current sentence:

"Even if you are sure something is true, it must be verifiable before you can add it."

Proposal:

"Even if you are sure something is true, you must verify it in a reliable source before you add it, and, if it isn't an obvious fact, usually add that source as an inline citation, so that it can be verified by others who may doubt your addition.

Valjean (talk) (PING me) 16:28, 26 September 2022 (UTC)

Note, the usual caveats and exceptions found in the policy still apply to my proposal, so I shouldn't need to spell out every single one above. Use a bit of common sense. Yes, I'm talking to Jc3s5h and Newimpartial

The point is that OR and adding personal opinions of what is "truth" is not allowed. The one adding non-obvious content has the burden of proof. They must be documenting existing info found in reliable sources they provide so others can verify they are not spouting their own ideas. "The sky is blue" is not the topic here. -- Valjean (talk) (PING me) 18:29, 26 September 2022 (UTC)

The existing language using the concept of "verifiability" already specifies that adding personal "truth" is not allowed and that editors must be documenting existing info found in reliable sources they provide so others can verify they are not spouting their own ideas. Giving additional ammunition for editors to remove verifiable content - which may well be supported by sources already in the article - because citations have not been plunked down in a particular sentence - isn't prudent and doesn't contribute to a healthy encyclopaedia, IMO. Newimpartial (talk) 18:37, 26 September 2022 (UTC)

This only makes explicit what is currently implied. Actual practice doesn't really change. -- Valjean (talk) (PING me) 18:41, 26 September 2022 (UTC)

I disagree. The current practice is that editors must know the content to be verifiable; they do not have to verify it (and usually add that source as an inline citation, a proposal that, if adopted, would rapidly lead to an absurd amount of cite overkill as editors demand citations for minute passages that other editors add, including paraphrases from existing sources and summaries of content elsewhere in the article). Newimpartial (talk) 18:45, 26 September 2022 (UTC)

"Giving additional ammunition for editors to remove verifiable content..." what do you mean? The only way to make sure some source is verifiable is to verify it: if you believe something is verifiable, but you cannot find a source, than your belief is probably wrong. If you believe some text is verifiable AND obvious, obtain a consensus on that matter. If you believe some text is verifiable, and you have a source, what prevents you from adding a ref? Paul Siebert (talk) 19:26, 26 September 2022 (UTC)

The proposed text would privilege, for example, the removal of material that is not supported by a reference given in the same sentence, and would mitigate against the current practice of discussing sourcing - and whether or not a citation need be added in a particular sentence, or whether e.g. it is covered by citations provided elsewhere in the article - on Talk. These matters should be determined by consensus based on a global discussion of what the RS say, not based on what markup text is present in any particular proposed edit.

To be clear, I expect that the proposed text would be weaponized by editors who allege that paraphrases and lead material based on existing sources are "unsupported by citations" and who would then selectively remove them - not based on any good faith concern for the material being "unverifiable" but because the editors in question disagree with the content. I think it is quite easy enough for editors to do this already, without handing them more ammunition. Newimpartial (talk) 19:52, 26 September 2022 (UTC)

No, my understanding of the proposed text is different. It does not describe any concrete procedure, it just explains how a good text should look like: any non-obvious statement is considered verifiable if it is supplemented with a reference to some reliable source.

The proposed text does not replace other parts of the policy, so the old procedure (placement of the "cn" tag, talk page discussions, etc ) is still in effect.

That means the proposed text literally says that, unless some statement is obvious, or a consensus is achieved that some particular statement needs no citation, or the statement is in a lede, or it is a summary of other (properly verified) statements, it must be supported by a reference to some RS.

What is wrong with that, and how can it be misused? Paul Siebert (talk) 22:16, 26 September 2022 (UTC)

The text literally invites editors to expect that a source be added to accompany any added content. This sentence will inevitably - as was already happening in the discussion thwt prompted the RfC - be decontextualized so that all the caveats and procedures you are pointing to would be ignored by editors who, feeling that anyone doubting an addition should find a citation already at hand, will believe the content to be inappropriately added and will remove it (particularly if they find it disagreeable). Newimpartial (talk) 23:29, 26 September 2022 (UTC)

Have you seen the words "if it isn't an obvious fact" in the proposed text? If something is obvious - no citation is needed. How can we decide that some statement is obvious? If consensus is achieved that an average reader can found that fact in a couple of mouse clicks, or it is obvious to any average reader even without googling - that fact is obvious. That is simple.

Using the @Jc3s5h:'s argument as an example, I can type capital of Connecticut, and I immediately get this. That means the statement "Hartford is a capital of the State of Connecticut" is obvious (even those who didn't know that can verify it in one mouse click).

At least, we can write "Hartford is the capital city of the U.S. state of Connecticut." (and the blue link directs to the list with all needed references. Paul Siebert (talk) 00:00, 27 September 2022 (UTC)

Yes I read it. I divide the proposal into the following clauses:

Even if you are sure something is true, you must verify it in a reliable source before you add it
and, if it isn't an obvious fact, usually add that source as an inline citation, so that it can be verified by others who may doubt your addition.

So "if it isn't an obvious fact" only appears to the part of the proposal that appears after that phrase. The obvious fact leniency only applies to adding an inline citation. Because the first clause is before "if it isn't an obvious fact" that leniency does not apply, and you must verify every single statement before adding it. Jc3s5h (talk) 00:07, 27 September 2022 (UTC)

As far as the obvious fact exception is concerned, my experience on Wikipedia tells me that editors will argue that something is not a fact, or is not obvious, when they wish to subject said "something" to more rigorous requirements for documentation (or simply disagree with it). I get that we have to produce sources in response to good faith queries or objections from other editors, but I do not see any reason why the playing field should be shifted further in favor of SPAs who demand sourcing each time a formulation with which they disagree appears in article text (even if the source is clearly present elsewhere in the same article). Newimpartial (talk) 02:14, 27 September 2022 (UTC)

"(even if the source is clearly present elsewhere in the same article)" covers it. No need to find a source again if it's already in the article. An editor who asks in good faith just needs to be pointed to that source (and maybe the exact quote in the source to prove verification). Often the refname should be copied to the spot to head off further queries. This is all standard practice, and this proposal doesn't change anything. -- Valjean (talk) (PING me) 02:26, 27 September 2022 (UTC)

No, this proposal says that editors should usually add this source as an in-line citation (if it isn't obvious). This is an invitation to recalcitrant editors to wikilawyer what counts as "obvious" (or "a fact"), and will in fact encourage editors to remove content that they believe is unverified rather than engaging in discussion on Talk to evaluate sourcing. It would also enhance "reference-bombing" of Wikipedia prose (prose that is often already overburdened by references), because it will encourage the placement of refnames (and, in principle, distinct quotes) to head off further queries about state capitals and unit conversions, and will also force editors to cite the same source repeatedly, sentence by sentence, to forestall borderline-good-faith queries about where various elements of a description are sourced. Neither of these changes would represent good encyclopaedic writing. Newimpartial (talk) 02:47, 27 September 2022 (UTC)

What you say is oxymoron. WP:WL defines wikilawyering as "Using the rules in a manner to achieve a goal other than compliance with the rule". Every rule can be used in a manner that is not consistent with that rule, but that does not make that rule bad.

Yes, some disruptive editors may use a new rule to disrupt Wikipedia by adding redundant references or by removing useful information. However, currently, disruptive editors are already using omissions in the policy to keep poorly sourced or non-verifiable information of Wikipedia. Therefore, a correct argument should be based on a comparison of possible negative and positive consequences of implementation of new policy.

Every good thing can be misused, that is not a reason to reject it. Paul Siebert (talk) 03:30, 27 September 2022 (UTC)

Re: disruptive editors are already using omissions in the policy to keep poorly sourced or non-verifiable information of Wikipedia - I do not believe this to be true, to any significant extent. While I have certainly seen unsourced (and potentially unverifiable) information added to articles, I have absolutely never seen disruptive editors use omissions in policy to keep it there. Newimpartial (talk) 03:53, 27 September 2022 (UTC)

Jc3s5h You write: "This still requires editors to look up obvious facts, such as every single conversion factor, before adding material." Exactly. Wikipedia is not your personal blog, you must be sure that what you are writing is factually correct. If you write some obvious thing, for example, that Pi equals to 3.14159265953... it makes sense to check it. Paul Siebert (talk) 20:44, 26 September 2022 (UTC)

Since Paul Siebert is unable to correctly report the value of π I will suspend consideration of his remarks until he corrects the error and cites a reliable source. I, of course, have two independent reliable sources that demonstrate the falsity of the value stated, but decline to cite them because I have been insulted. I am indeed capable of stating that 1 m = 1000 mm without having to consult a reference book every time. Jc3s5h (talk) 21:07, 26 September 2022 (UTC)

You noticed the error! Good job! I deliberately swapped "9" and "3" to check if you will be able to see the error. I am glad you are.

And now, please, answer two questions:

Did I understand it correctly that you used some reliable sources to verify my statement "Pi equals to 3.14159265953..."?
Do you think it would be correct to say that it would be my responsibility to make sure my statement is verifiable? In other words, even if I believe that Pi equals to 3.14159265953..., it would be correct if I verified that statement by looking at some RS, and, as soon as I already have these sources before me, it would seem logical to add references to these sources?

Paul Siebert (talk) 21:17, 26 September 2022 (UTC)

I did check two reliable sources, because I do not remember more than 3 significant digits of π. But other well-known facts are easier to remember, and the decimal ratios of the metric system were intentionally chosen by the system creators to be easy to remember. Indeed, to be able to read or write at all in any particular discipline, it is necessary to have in one's mind, ready for instant recall, the bare essentials of the discipline. The proposed policy allows no exceptions at all. It would be so burdensome as to prevent thinking. Adopting such a policy would amount to a statement that Wikipedia hates and despises its editors. Jc3s5h (talk) 21:40, 26 September 2022 (UTC)

As I already wrote (see my 18:12, 26 September 2022 (UTC) post), I agree that Valjean's proposal should not be applicable to obvious statements (which are by definition verifiable), as well as lede and similar texts. Let's think how to make the policy clear with respect to that. Paul Siebert (talk) 21:47, 26 September 2022 (UTC)

Jc3s5h, you should now read my comment for you further down at your !vote. Above you write "The proposed policy allows no exceptions at all." That's ridiculous. My proposal allows (without stating it) all the exceptions found in the existing policy. My proposed addition fits in an existing spot at the beginning of the policy, and the rest of the policy contains many qualifications and exceptions on how to apply the policy. Those all apply to my proposition, so it makes no sense to list all of them here. You should know those exceptions and qualifications and apply them. Instead, you're proposing ridiculous ideas and objections. I added the qualification about non-obvious facts to meet concerns here. Be happy for that. -- Valjean (talk) (PING me) 22:05, 26 September 2022 (UTC)

I concede this my example may be an overkill, because there are many ways to independently verify the Pi value, so the information about Pi number may be considered obvious. However, this example demonstrates that it may be dangerous to put less obvious information that only seems verifiable if you haven't verified it by looking at sources. Paul Siebert (talk) 21:29, 26 September 2022 (UTC)

Valjean, how many of these sentences do you expect me to look up before adding them to an article?

Penicillin is a type of antibiotic.
A piano has 88 keys, with more white keys than black keys.
The worldwide web became available to the public in the early 1990s.
Vehicles that use gasoline and diesel fuel may be fitted with a catalytic converter to reduce air pollution.

None of these are obvious to everyone. All of these are verifiABLE at your nearest web search engine. Do you really think that I need to look them up before adding them? (NB that I'm not asking whether you think I should cite them; I'm asking whether you think the typical editor's memory is so poor that we'd easily confuse penicillin with Pepcid.) WhatamIdoing (talk) 15:48, 27 September 2022 (UTC)

Support

Paul Siebert (talk) 17:10, 26 September 2022 (UTC)

I support this proposal, but there is one reservation: it must be clear from the proposed text that it relates only to non-obvious statements (i.e. the statements that are unlikely to be challenged). Obviously, no citations are required for obvious statements (as well as for lede sections and, probably in some specific cases). That reservation partially addresses criticism expressed by Jc3s5h&Newimpartial. Paul Siebert (talk) 18:12, 26 September 2022 (UTC)

Donald Albury 17:31, 26 September 2022 (UTC)
Only makes sense. Anyone arrogant enough to add things for which they are unwilling to provide a citation are likely partisans or cranks. Anything added to Wikipedia should have a citation provided or there is really no point in adding it as we cannot know if the assertion is true. Chris Troutman (talk) 18:55, 26 September 2022 (UTC)

Oppose

Oppose. Requires editors to waste time looking up information that virtually all the readers of a somewhat advanced article already know. If the information is not readily available online, editors would be forced to buy paper books or travel to a library. As an example, every single use of the {{convert}} template would require the editor to look up the conversion in a reliable source, since the convert template is part of Wikipedia and thus not a reliable source. Jc3s5h (talk) 17:41, 26 September 2022 (UTC)
- Reaffirm and amplify opposition Valjean modified the proposal after several editors had responded to it, to read "Even if you are sure something is true, you must verify it in a reliable source before you add it, and, if it isn't an obvious fact..." [addition underlined]. This still requires editors to look up obvious facts, such as every single conversion factor, before adding material. Valjean would spare us the effort of writing a citation for 1000 mm = 1 m, but we still have to look it up before we add a conversion that depends on it. Jc3s5h (talk) 19:54, 26 September 2022 (UTC)
  - That's a second !vote, even though this isn't an RfC. I updated the proposition to meet the expressed concerns. That's proper in a discussion like this. We refine the proposition until it meets all concerns and is acceptable. You should be happy I have responded favorably to your concerns. Note the location in the policy of the content under discussion. The policy still contains all those exceptions and caveats, and they still apply to my proposition, so I shouldn't need to include them in the proposition. Even adding the part about "if it isn't an obvious fact" shouldn't be necessary here, as that concept is already in the policy. I assumed you would factor that in, but you didn't. Start doing it now and stop this line of objection.
  - You're really being ridiculous with your nitpicky and extreme "conversion factor" example. That's quite the rarity. Do we currently require a source for that situation? If not, then this does not change that. If we do, then this just reaffirms our common practice. Use common sense and stick to the big picture and stop trying to find some 1% exception you can use to overthrow a 99% improvement. We have lots of exceptions and caveats in this policy, and this does not remove them. I proposed this here and assumed I was talking to experienced editors who would use common sense. I see I assumed too much.
  - I'm not suggesting anything that changes how we already do things. I'm just making it explicit that the one who adds content should base it on existing reliable sources that they know of and can cite. What's wrong with that? Anything else is OR. That's not new. Whether they should "usually" (note I'm leaving that open for exceptions) add an inline citation for non-obvious facts is also not in question. We already "usually" require that. If they don't provide a RS, we often delete the content and/or require them to find their source. We are not talking about obvious facts like "The sky is blue", so stop commenting as if we are. Discuss like an experienced editor, not a newbie. I helped create this policy, so I know a bit about what lies behind it. "Verifiability, not truth" was our mantra back in 2003 when I started here, and it's still true today. -- Valjean (talk) (PING me) 21:53, 26 September 2022 (UTC)
    - "...the one who adds content should base it on existing reliable sources that they know of and can cite. What's wrong with that? Anything else is OR." No, that's false. In my grade school I learned that Hartford is the capital of Connecticut, and that has been reinforced may times by my many visits to that city. At the end of the school year my teacher collected my text book, so I can't cite it. The fact that I can't compose a citation from memory does not make it OR for me to state Hartford is the capital of Connecticut. No verifiable statement is OR, regardless of the state of mind of the editor who added it. Jc3s5h (talk) 22:25, 26 September 2022 (UTC)
      - Your example is an obvious fact that even a mere wikilink covers. We are not talking about such facts. -- Valjean (talk) (PING me) 23:01, 26 September 2022 (UTC)
        The proposed revision requires editors to look up every single fact, no matter how obvious, in a reliable source. Valjean's interpretation of the proposal is wrong. Jc3s5h (talk) 23:20, 26 September 2022 (UTC)
        Jc3s5h, what wording would be better to avoid any confusion? Provide a suggestion. -- Valjean (talk) (PING me) 16:13, 27 September 2022 (UTC)
        I believe any attempt to delve into this in the lead would make the lead too long and confusing. I believe the proposal should just not be adopted. Jc3s5h (talk) 16:16, 27 September 2022 (UTC)
        I'm beginning to think you're right. This may not be an improvement, since what I'm trying to get at is already in the policy. My tweak may just complicate matters. I think I'll just withdraw this now. Thanks for the discussion. -- Valjean (talk) (PING me) 16:39, 27 September 2022 (UTC)

Oppose - this would encourage the proliferation of unnecessary citations (especially in lead sections) by empowering editors to delete uncontroversial content unless the citation is presented in the specific sentence under discussion. That isn't where the onus should be. Newimpartial (talk) 17:56, 26 September 2022 (UTC)
Oppose - this would absolutely certainly be weaponized in content disputes, to the detriment of the encyclopedia and of the editing environment. All while achieving no positive goal. Thparkth (talk) 16:13, 27 September 2022 (UTC)

WP:SELFSERVING should link here...

Latest comment: 2 years ago3 comments2 people in discussion

... specifically to Wikipedia:Verifiability#Self-published or questionable sources as sources on themselves, which mentions this as a key element of that branch of the policy. Seems like it would be navigationally useful. I already requested this at WP:AfC, but it occurs to me that as it'd presumably involve a corresponding edit to this page, editors here might have a view either way. 109.255.211.6 (talk) 10:21, 3 October 2022 (UTC)

Oops, apologies! Wrong section. Fixed. 109.255.211.6 (talk) 13:54, 5 October 2022 (UTC)

I Will Create An Article With This Name SpyridisioAnnis (talk) 10:51, 6 October 2022 (UTC)

Semi-protected edit request on 8 October 2022

Latest comment: 2 years ago3 comments2 people in discussion

This edit request to Wikipedia:Verifiability has been answered. Set the |answered= or |ans= parameter to no to reactivate your request.

Amir k wurma (talk) 08:32, 8 October 2022 (UTC)

The best way Amir k wurma (talk) 08:33, 8 October 2022 (UTC)

Not done: it's not clear what changes you want to be made. Please mention the specific changes in a "change X to Y" format and provide a reliable source if appropriate. MadGuy7023 (talk) 08:37, 8 October 2022 (UTC)

vouching for accuracy

Latest comment: 2 years ago9 comments4 people in discussion

does wikipedia have a mechanism by which a user can declare that they have verified content against provided references?

In vanilla wikipedia there is no way to know if information is accurate without consulting references. This can be a time consuming process, for example if many sources have to be cross-referenced.

Like a crowdsourced publishing service.

As it stands, there is no reliable way to tell if a particular piece of information was entered by a dilettante or by a highly meticulous person. Or to even tell if there had been a sneaky act of vandalism some time in the last year.

If users vouch for accuracy, track records are generated that can be used to estimate the chance that a certain claim of accuracy is true.

if verified content on wikipedia is linked to by other wikipedia articles, the discovery of an error can be escalated downstream.

If such a mechanism was in place, users are probably going to be more motivated to ensure accuracy.

Effort spent verifying information can be used like a currency. Users could verify content and in return be awarded with an effort of verification of content of their choosing.

If a record of this activity existed, these efforts would not go unnoticed. The quality of content would increase. Nowakki (talk) 14:12, 17 October 2022 (UTC)

Presumably, the editor that adds a citation is vouching for that citation. Of course, some citations that currently exist were added more than 20 years ago, and the editor that added one may have stopped editing almost that long ago. Citations to web sites generally have an access date, and editors who have clicked through to the web site and found it alive and relevant are encouraged to update the access date. Editors who find problems with citations can try to fix the problem. They also have a number of templates they can add to the article to indicate a problem they have not fixed. In the end, however, Wikipedia is edited by volunteers who are, more or less, anonymous. The closest thing to a reward system on Wikipedia is the reputation that editors acquire through their contributions to the project. In general, the community has rejected proposals that complicate editing, or smack of bureaucracy. Donald Albury 15:59, 17 October 2022 (UTC)

> Presumably, the editor that adds a citation is vouching for that citation.

There is no way of telling if an editor is reliable. There is no way of knowing which iteration of an article an author deems reliable. wikipedia editors do not apply the same level of meticulousness to all their edits.

> The closest thing to a reward system on Wikipedia is the reputation that editors acquire through their contributions to the project.

nobody is stopping an editor to go into a quid-pro-quo with other editors.

> In general, the community has rejected proposals that complicate editing

vouching for a piece of content is optional and anyone can add content just as before.

The community has not rejected the general proposal to make a reliable encyclopedia. And i think the community would be willing to pay a price to achieve it.

In my opinion, a reliable wikipedia is possible, and as it stands now wikipedia by itself cannot provide any reliability, unless external sources are verified every time or you already know that something you read is true. Nowakki (talk) 17:26, 17 October 2022 (UTC)

There is equally no way of telling if the user who wants to declare that they have verified content against provided references is reliable. We assume that Wikipedia:You are not a reliable source – neither the person who adds the content originally, nor anyone who claims to have verified it.

Additionally, we don't necessarily want people to verify the content "against provided references"; it's often more important to verify them against sources that aren't cited in the article. What if someone cherry-picked the one source that said the Moon is made of green cheese? Or the one source with an unfortunate typo in a critical statement? Inline citations are convenient, but if you want to know whether the article is good, you often need to look beyond them. WhatamIdoing (talk) 05:09, 19 October 2022 (UTC)

> There is equally no way of telling if the user who wants to declare that they have verified content against provided references is reliable.

yes there are ways.

you may know the verifier for some time, because you are working on the same project and quacks like a good faith helper bee
the track record of a verifier can be used otherwise, a percentage of verifications that have until the present day not been falsified

there is no guarantee, but at this point you have to take into account that the reference is used to write another wiki article, and that new article will point to the original verified claim and the new article can be flagged as needing maintenance if the verification is falsified.

> it's often more important to verify them against sources that aren't cited in the article.

you can and should still do that. verification of proper citation to make wikipedia more dependable is not required to solve the problem of accuracy in every conceivable way. it is just another tool to be used to achieve the goal. Nowakki (talk) 10:19, 19 October 2022 (UTC)

sorry forgot one thing: the original editor can introduce an insignificant error deliberately. a verifier can never know whether they are being audited and would have to be always meticulous. under such conditions if a verifier has a track record of 98% after 200 verifications, that would be good enough to rely upon in matter less severe than life and death. Nowakki (talk) 10:31, 19 October 2022 (UTC)

Please don't use Wikipedia for matters of life and death. While Wikipedia is relatively reliable, and in my opinion just as if not significantly more reliable than other encyclopedias that aren't Wikis (per the famous Nature study, more reliable than Brittanica), we don't allow citations to Wikipedia because it's not itself reliable enough as a source. So always follow the citations and pay attention to the edit history and content tags if you care about veracity to a high standard. Andre🚐 21:23, 19 October 2022 (UTC)

i am making a proposal for a change to wikipedia.

why do you explain the current state of wikipedia as if i am living under a rock?

if a table made from 10 crappy looking pdfs can be used to sum one column of it, without first having to go through 10 crappy pdfs, that would be an improvement.

wikipedia 1.0 will look different from wikipedia 0.4 we have now. Nowakki (talk) 21:55, 19 October 2022 (UTC)

i made a mockup

User:Nowakki/sandbox/Reliability

what do you think? if you were writing an article about the shipbuilding history of california in the 20th century on wikipedia, would you dare to link to this section with a <ref>?

your article will get notified when it changes.

if you instead copy over 5 references to your article, if one of the references is later proven to have been errorneous, you can only hope that the person who discovers that did so while browsing your article. Nowakki (talk) 11:31, 21 October 2022 (UTC)

Document hosting websites

Latest comment: 2 years ago3 comments3 people in discussion

Not sure if this is the right place to discuss this or not… but I have noticed that recently we are getting more discussions at RSN about the reliability of document hosting websites (ie websites that compile and republish documents). It is obvious that many newer editors do not understand HOW to cite the documents hosted on such websites (ie cite the original document, but also include a note that the website is a “courtesy link”). Do we need to clarify this at WP:V or WP:RS? (Or someplace else?) Blueboar (talk) 16:28, 22 October 2022 (UTC)

It should be in WP:CITE even if it is elsewhere too. Zero^talk 22:09, 22 October 2022 (UTC)

I don't think that it's normal to explicitly note that a URL is a Wikipedia:Convenience link. I think we've always assumed that editors were smart enough to figure that out. WhatamIdoing (talk) 16:46, 27 October 2022 (UTC)

RfC [Proposal] about tags, and content removal being dubious valid challenges.

Latest comment: 2 years ago6 comments3 people in discussion

Per the directive made here I am requesting comments for the proposal being made at: Wikipedia talk:Inline citation#Unverified statements about valid challenges Huggums537 (talk) 21:23, 29 October 2022 (UTC)

I have commented there… but I think one flaw in your RFC is that you start off by defining certain forms of challenges as “valid” (and thus, by implication, other forms of challenge as “invalid”). First, I doubt there is consensus on what constitutes a “valid” challenge (and, thus, by implication what would constitute an “invalid” challenge), and second… I am not sure that making such a distinction is appropriate to begin with. As I see it, there are multiple ways of challenging - all of them are “valid”. I suppose you could say that some are weaker/stronger than others (a “cn” tag would be a weaker challenge than blunt removal)… some are more/less helpful than others (a removal accompanied by a full explanation on the talk page is certainly more helpful than an unexplained removal)… but all are “valid”. Blueboar (talk) 15:15, 4 November 2022 (UTC)

(Also, it's not an RFC.) WhatamIdoing (talk) 21:24, 4 November 2022 (UTC)

Dang, I just realized that. What a goof! I will replace the "RfC" in the title with "Proposal" instead... Huggums537 (talk) 00:38, 5 November 2022 (UTC)

Thank you for commenting, @Blueboar. Why don't you enlighten us with one of your history lessons about where this idea about "valid challenges" originated from? 😁 Huggums537 (talk) 01:13, 5 November 2022 (UTC)

Also, I responded to your thoughts about defining "valid" and making such distinctions at the proposal... Huggums537 (talk) 01:49, 7 November 2022 (UTC)

RfC for clarification at MINREF

Latest comment: 2 years ago1 comment1 person in discussion

There is a request for comments at Wikipedia_talk:Inline_citation#Alternate_proposal_for_valid_challenges on a proposal to clarify the differences between content removal related to missing citations, and the ordinary removal of content not related to missing citations. Your voting decision is appreciated. Huggums537 (talk) 20:44, 7 November 2022 (UTC)

The "new" Twitter

Latest comment: 2 years ago3 comments3 people in discussion

Given all of the concerns that have been recently raised related to Twitter's new way of operating, I'm wondering how this would affect things such as WP:TWITTER and WP:BLPSELFPUB. It's seems that up to now, verifiable tweets from account with a blue check mark were assumed to come from the subject of the account. However, there've been quite a number of reports in the news how this can no longer be necessarily assumed since a fee was put into place to get accounts verified. There also appears to be quite a bit of upheaval going on at Twitter since new ownership took over which means whatever problems it had before per WP:RSPTWITTER might've increased, particularly with respect to citing verified accounts or subject-matter experts. -- Marchjuly (talk) 21:41, 13 November 2022 (UTC)

I think we can probably wait a couple of weeks before updating policies. Editors are generally pretty smart about this kind of thing. WhatamIdoing (talk) 22:10, 13 November 2022 (UTC)

Agree that this is concerning enough that we should pay attention to what is happening, but also agree that it is too soon to change the policy. “Wait and see” seems like the best approach. Blueboar (talk) 22:20, 13 November 2022 (UTC)

RFC to change "verifiable" to "verified"

Latest comment: 2 years ago3 comments2 people in discussion

At the pump: Wikipedia:Village pump (policy)#RFC: change "verifiable" to "verified" Levivich (talk) 16:53, 19 October 2022 (UTC)

@Levivich, to follow up on that short-lived discussion, would you like to talk about a mini Wikipedia:Glossary for this policy? I think we could start with something like this:

verifiable: Someone (other than the person who originally added the material) is able to find this claim in at least one published, reliable source. There are no limits on the amount of expense, effort, or prior knowledge required to verify content. A claim can be verifiable even if it is not cited.
cited: There is an Wikipedia:Inline citation to a source associated with this claim. The cited source may or may not be reliable for this claim. A claim can be unverifiable even if it is cited (e.g., due to errors introduced by subsequent editors); similarly, a claim may be verifiable even if the cited source does not verify it (e.g., due to accidentally citing the wrong source).
verified: A Wikipedia editor (other than the person who originally added the material) compared the contents of the cited source(s) against the contents of the Wikipedia article and determined that they have the same meaning. A claim that is not verifiable in the cited source(s) may or may not be verifiable in other, uncited sources.

(Alternatively, we could just copy from Wikipedia:Glossary#Citation, Wikipedia:Glossary#Source, Wikipedia:Glossary#verifiable, and Wikipedia:Glossary#uncited.)

I think that one major source of opposition to your proposal is that people didn't like the word you chose to describe it. Standardizing the jargon could help if you want to revive that idea in the future. WhatamIdoing (talk) 20:54, 14 November 2022 (UTC)

Thanks for the invite but I'm gonna pass. Levivich (talk) 21:31, 14 November 2022 (UTC)

[1] Content is verifiable if, using all the means at the disposal of an educated and dedicated person, the content in question can be found in a source that Wikipedia editors deem to be reliable for that claim.

[directly_supports-2] A source "directly supports" a given piece of material if the information is present explicitly in the source so that using this source to support the material is not a violation of Wikipedia:No original research. The location of any citation—including whether one is present in the article at all—is unrelated to whether a source directly supports the material. For questions about where and how to place citations, see Wikipedia:Citing sources, Wikipedia:Manual of Style/Lead section § Citations, etc.

[3] 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.

[directly_supports-4] A source "directly supports" a given piece of material if the information is present explicitly in the source so that using this source to support the material is not a violation of Wikipedia:No original research. The location of any citation—including whether one is present in the article at all—is unrelated to whether a source directly supports the material. For questions about where and how to place citations, see Wikipedia:Citing sources, Wikipedia:Manual of Style/Lead section § Citations, etc.

[5] 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.

[pmid24196718-6] Fica SM, Tuttle N, Novak T, Li NS, Lu J, Koodathingal P, et al. (November 2013). "RNA catalyses nuclear pre-mRNA splicing". Nature. 503 (7475): 229–234. Bibcode:2013Natur.503..229F. doi:10.1038/nature12734. PMC 4666680. PMID 24196718.

[1]

[2]

[3]

[1]

[2]

[1]